Psychological therapies including dialectical behaviour therapy for borderline personality disorder
Transcript of Psychological therapies including dialectical behaviour therapy for borderline personality disorder
Health Technology Assessm
ent 2006;Vol. 10: No. 35
Psychological therapies for borderline personality disorder
Psychological therapies includingdialectical behaviour therapy forborderline personality disorder: a systematic review and preliminaryeconomic evaluation
J Brazier, I Tumur, M Holmes, M Ferriter, G Parry, K Dent-Brown and S Paisley
Health Technology Assessment 2006; Vol. 10: No. 35
HTAHealth Technology AssessmentNHS R&D HTA Programme
The National Coordinating Centre for Health Technology Assessment,Mailpoint 728, Boldrewood,University of Southampton,Southampton, SO16 7PX, UK.Fax: +44 (0) 23 8059 5639 Email: [email protected]://www.hta.ac.uk ISSN 1366-5278
FeedbackThe HTA Programme and the authors would like to know
your views about this report.
The Correspondence Page on the HTA website(http://www.hta.ac.uk) is a convenient way to publish
your comments. If you prefer, you can send your comments to the address below, telling us whether you would like
us to transfer them to the website.
We look forward to hearing from you.
September 2006
How to obtain copies of this and other HTA Programme reports.An electronic version of this publication, in Adobe Acrobat format, is available for downloading free ofcharge for personal use from the HTA website (http://www.hta.ac.uk). A fully searchable CD-ROM isalso available (see below).
Printed copies of HTA monographs cost £20 each (post and packing free in the UK) to both public andprivate sector purchasers from our Despatch Agents.
Non-UK purchasers will have to pay a small fee for post and packing. For European countries the cost is£2 per monograph and for the rest of the world £3 per monograph.
You can order HTA monographs from our Despatch Agents:
– fax (with credit card or official purchase order) – post (with credit card or official purchase order or cheque)– phone during office hours (credit card only).
Additionally the HTA website allows you either to pay securely by credit card or to print out yourorder and then post or fax it.
Contact details are as follows:HTA Despatch Email: [email protected]/o Direct Mail Works Ltd Tel: 02392 492 0004 Oakwood Business Centre Fax: 02392 478 555Downley, HAVANT PO9 2NP, UK Fax from outside the UK: +44 2392 478 555
NHS libraries can subscribe free of charge. Public libraries can subscribe at a very reduced cost of £100 for each volume (normally comprising 30–40 titles). The commercial subscription rate is £300 per volume. Please see our website for details. Subscriptions can only be purchased for the current orforthcoming volume.
Payment methods
Paying by chequeIf you pay by cheque, the cheque must be in pounds sterling, made payable to Direct Mail Works Ltdand drawn on a bank with a UK address.
Paying by credit cardThe following cards are accepted by phone, fax, post or via the website ordering pages: Delta, Eurocard,Mastercard, Solo, Switch and Visa. We advise against sending credit card details in a plain email.
Paying by official purchase orderYou can post or fax these, but they must be from public bodies (i.e. NHS or universities) within the UK.We cannot at present accept purchase orders from commercial companies or from outside the UK.
How do I get a copy of HTA on CD?
Please use the form on the HTA website (www.hta.ac.uk/htacd.htm). Or contact Direct Mail Works (seecontact details above) by email, post, fax or phone. HTA on CD is currently free of charge worldwide.
The website also provides information about the HTA Programme and lists the membership of the variouscommittees.
HTA
Psychological therapies includingdialectical behaviour therapy forborderline personality disorder: a systematic review and preliminaryeconomic evaluation
J Brazier,1* I Tumur,1 M Holmes,1 M Ferriter,2
G Parry,1 K Dent-Brown1 and S Paisley1
1 School of Health and Related Research (ScHARR), University of Sheffield,UK
2 Department of Research and Development, Nottinghamshire HealthcareNHS Trust, Retford, UK
* Corresponding author
Declared competing interests of authors: none
Published September 2006
This report should be referenced as follows:
Brazier J, Tumur I, Holmes M, Ferriter M, Parry G, Dent-Brown K, et al. Psychologicaltherapies including dialectical behaviour therapy for borderline personality disorder: a systematic review and preliminary economic evaluation. Health Technol Assess2006;10(35).
Health Technology Assessment is indexed and abstracted in Index Medicus/MEDLINE,Excerpta Medica/EMBASE and Science Citation Index Expanded (SciSearch®) and Current Contents®/Clinical Medicine.
NHS R&D HTA Programme
The research findings from the NHS R&D Health Technology Assessment (HTA) Programme directlyinfluence key decision-making bodies such as the National Institute for Health and Clinical
Excellence (NICE) and the National Screening Committee (NSC) who rely on HTA outputs to help raisestandards of care. HTA findings also help to improve the quality of the service in the NHS indirectly inthat they form a key component of the ‘National Knowledge Service’ that is being developed to improvethe evidence of clinical practice throughout the NHS.The HTA Programme was set up in 1993. Its role is to ensure that high-quality research information onthe costs, effectiveness and broader impact of health technologies is produced in the most efficient wayfor those who use, manage and provide care in the NHS. ‘Health technologies’ are broadly defined toinclude all interventions used to promote health, prevent and treat disease, and improve rehabilitationand long-term care, rather than settings of care.The HTA Programme commissions research only on topics where it has identified key gaps in theevidence needed by the NHS. Suggestions for topics are actively sought from people working in theNHS, the public, service-users groups and professional bodies such as Royal Colleges and NHS Trusts. Research suggestions are carefully considered by panels of independent experts (including service users)whose advice results in a ranked list of recommended research priorities. The HTA Programme thencommissions the research team best suited to undertake the work, in the manner most appropriate to findthe relevant answers. Some projects may take only months, others need several years to answer theresearch questions adequately. They may involve synthesising existing evidence or conducting a trial toproduce new evidence where none currently exists.Additionally, through its Technology Assessment Report (TAR) call-off contract, the HTA Programme isable to commission bespoke reports, principally for NICE, but also for other policy customers, such as aNational Clinical Director. TARs bring together evidence on key aspects of the use of specifictechnologies and usually have to be completed within a short time period.
Criteria for inclusion in the HTA monograph seriesReports are published in the HTA monograph series if (1) they have resulted from work commissionedfor the HTA Programme, and (2) they are of a sufficiently high scientific quality as assessed by the refereesand editors.
Reviews in Health Technology Assessment are termed ‘systematic’ when the account of the search,appraisal and synthesis methods (to minimise biases and random errors) would, in theory, permit thereplication of the review by others.
The research reported in this monograph was commissioned by the HTA Programme as project number04/52/01. The contractual start date was in January 2005. The draft report began editorial review inSeptember 2005 and was accepted for publication in February 2006. As the funder, by devising acommissioning brief, the HTA Programme specified the research question and study design. The authorshave been wholly responsible for all data collection, analysis and interpretation, and for writing up theirwork. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and wouldlike to thank the referees for their constructive comments on the draft document. However, they do notaccept liability for damages or losses arising from material published in this report.
The views expressed in this publication are those of the authors and not necessarily those of the HTA Programme or the Department of Health.
Editor-in-Chief: Professor Tom WalleySeries Editors: Dr Aileen Clarke, Dr Peter Davidson, Dr Chris Hyde,
Dr John Powell, Dr Rob Riemsma and Dr Ken SteinManaging Editors: Sally Bailey and Sarah Llewellyn Lloyd
ISSN 1366-5278
© Queen’s Printer and Controller of HMSO 2006
This monograph may be freely reproduced for the purposes of private research and study and may be included in professional journals providedthat suitable acknowledgement is made and the reproduction is not associated with any form of advertising.
Applications for commercial reproduction should be addressed to NCCHTA, Mailpoint 728, Boldrewood, University of Southampton, Southampton, SO16 7PX, UK.
Published by Gray Publishing, Tunbridge Wells, Kent, on behalf of NCCHTA.Printed on acid-free paper in the UK by St Edmundsbury Press Ltd, Bury St Edmunds, Suffolk. G
Objectives: To summarise the available evidence onthe clinical effectiveness and cost-effectiveness ofpsychological therapies including dialectical behaviourtherapy (DBT) for borderline personality disorder(BPD).Data sources: Electronic databases were searched upto March 2005.Review methods: Relevant studies were assessedusing standard checklists and data were abstracted bytwo reviewers using standardised forms. Separateeconomic evaluations were undertaken for six selectedrandomised controlled trials (RCTs). Cost-effectivenesswas assessed in terms of cost per parasuicide eventavoided in all six trials and cost per quality-adjusted life-year (QALY) in four of them. All results are at2003–4 prices and for 12 months follow-up.Results: Nine RCTs and one non-RCT of moderate topoor quality were identified in the clinical effectivenessreview. They provided some evidence that DBT ismore effective than treatment as usual (TAU) for thetreatment of chronically parasuicidal and drug-dependent borderline women; that DBT-orientatedtherapy is more effective than client-centred therapy(CCT) for the treatment of BPD; and that DBT is aseffective as comprehensive validation therapy plus 12-Step for the treatment of opioid-dependent borderlinewomen. There was also some evidence that partialhospitalisation is more effective than TAU in thetreatment of BPD, good evidence that manual-assistedcognitive behavioural therapy (MACT) is no moreeffective than TAU in the treatment of BPD and someevidence that interpersonal group therapy is no more
effective than individual mentalisation-based partialhospitalisation (MBT) for the treatment of BPD.However, these results should be interpreted withcaution as not all studies were primarily targeted toborderline symptoms and there were considerabledifferences between the studies. The assessment ofcost-effectiveness found a mix of results in the fourtrials of DBT, along with the high levels of uncertaintyand the limitations in the analyses. The findings do notsupport the cost-effectiveness of DBT though theysuggest it has the potential to be cost-effective. Theresults for MBT are promising, though againsurrounded by a high degree of uncertainty and forMACT, the analysis suggests that the intervention isunlikely to be cost-effective.Conclusions: The overall efficacy of psychologicaltherapies is promising; however, at this stage theevidence is inconclusive. The cost-effectiveness of theintervention in six RCTs examined, however, does notsupport the cost-effectiveness of DBT althoughpotential is suggested. There is a need for considerableresearch in this area. This research should involveappropriately powered head-to-head RCTs ofpsychological therapies; a survey of current practiceand the use of the full range of services by people withBPD to inform future economic analyses; full resource-use data collected in the context of pragmatic clinicaltrials; psychometric assessment of the validity of theEQ-5D or other generic and condition-specificpreference-based measures in BPD, and thedevelopment of a more formal cost-effectivenessmodel using the above data.
Health Technology Assessment 2006; Vol. 10: No. 35
iii
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Abstract
Psychological therapies including dialectical behaviour therapyfor borderline personality disorder: a systematic review andpreliminary economic evaluation
J Brazier,1* I Tumur,1 M Holmes,1 M Ferriter,2 G Parry,1 K Dent-Brown1 and S Paisley1
1 School of Health and Related Research (ScHARR), University of Sheffield, UK2 Department of Research and Development, Nottinghamshire Healthcare NHS Trust, Retford, UK* Corresponding author
Health Technology Assessment 2006; Vol. 10: No. 35
v
Glossary and list of abbreviations ............. vii
Executive summary .................................... ix
1 Aim of the review ...................................... 1
2 Background ................................................ 3Description of underlying health problem ...................................................... 3Description of new intervention ................ 3Current evidence ........................................ 5
3 Effectiveness ............................................... 7Methods for reviewing effectiveness .......... 7Results ........................................................ 8
4 Cost-effectiveness ...................................... 23Systematic review of existing economicliterature ..................................................... 23Cost-effectiveness and cost–utility analysis ....................................................... 25Methods by study ....................................... 30Cost-effectiveness results ............................ 32Univariate sensitivity analysis .................... 47
5 Discussion ................................................... 49Main results: clinical effectiveness ............. 49Main results: cost-effectiveness .................. 49Assumptions, limitations and uncertainties ............................................... 50Need for further research .......................... 51
6 Conclusions ................................................ 53
Acknowledgements .................................... 55
References .................................................. 57
Appendix 1 Identification of studies ......... 63
Appendix 2 Database keyword strategies ... 65
Appendix 3 Evidence tables for BPD studies ......................................................... 73
Appendix 4 Excluded studies .................... 107
Appendix 5 Consensus trial quality ratingsaccording to Lackner’s quality checklist .... 109
Appendix 6 British Medical Journalchecklist for economic evaluations ............ 111
Appendix 7 Case studies ........................... 115
Appendix 8 Mapping BDI to EQ-5D ........ 117
Health Technology Assessment reportspublished to date ....................................... 119
Health Technology Assessment Programme ................................................ 133
Contents
Health Technology Assessment 2006; Vol. 10: No. 35
vii
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Glossary and list of abbreviations
GlossaryClient-centred therapy Model of supportivetherapy based on Carkuff ’s model whichemphasises empathic understanding of thepatient’s sense of aloneness and provides asupportive attitude on an individual basis.
Cognitive behaviour therapy The pragmaticcombination of concepts and techniques fromcognitive and behaviour therapies common inclinical practice.
Comprehensive validation therapy with 12-Step A manualised approach that providesthe major acceptance-based strategies such astherapeutic warmth, responsiveness andempathy in combination with the 12-Stepprogramme.
Dialectical behaviour therapy Combinationof standard cognitive behavioural techniqueswith acceptance-based strategies and strategiesdesigned to keep the therapy balanced betweenchange and acceptance (dialectical strategies).
Manual-assisted cognitive behaviour therapyA 70-page manual that consists of a brief formof cognitive behaviour therapy combined withdialectical behaviour therapy techniquespotentially suitable for widespread use inroutine healthcare settings.
Mentalisation-based partial hospitalisationIntegrates individual and group psychoanalyticpsychotherapy within a limit-setting,structured, flexible and reliable partialhospitalisation. The mentalisation-based partialhospitalisation reflects both the therapeuticand management difficulties, with an emphasison the relational aspects of the disorder.
Psychodynamic therapy Emphasisespersonality structure and development andaims to provide insight for people, allowingthem to understand their feelings and to findbetter coping mechanisms.
List of abbreviationsA&E accident and emergency
AUC area under the curve
BAI Beck Anxiety Inventory
BDI Beck Depression Inventory
BHS Beck Hopelessness Scale
BPD borderline personality disorder
BPDSI Borderline Personality DisorderSeverity Index
BPRS Brief Psychiatric Rating Scale
BSI Brief Symptom Inventory
CASP Critical Appraisal Skills Programme
CBT cognitive behavioural therapy
CCDAN Cochrane Collaboration Depressionand Anxiety Neurosis Review Group
CCT client-centred therapy
CEAC cost-effectiveness acceptability curve
continued
Technical terms and abbreviations are used throughout this report. The meaning is usually clear fromthe context, but a glossary is provided for the non-specialist reader. In some cases, usage differs in the
literature, but the term has a constant meaning throughout this review.
List of abbreviations continued
CI confidence interval
CRD Centre for Reviews andDissemination
CSRI Client Service Receipt Inventory
CVT+12S comprehensive validation therapywith 12-Step
DBT dialectical behaviour therapy
DES Dissociative Experiences Scale
df degree of freedom
DIB Diagnostic Interview forBorderlines
DSH deliberate self-harm
DSM Diagnostic and Statistical Manualof Mental Disorders
EQ-5D EuroQol 5 Dimensions
EuropASI European Addiction Severity Index
GAF Global Assessment of Functioning
GAS Global Adjustment Scale
GSA Global Social Adjustment
GSI Global Symptom Index
HADS Hospital Anxiety and DepressionScale
HAM-D Hamilton Depression Rating Scale
HARS Hamilton Anxiety Rating Scale
HSC-90 Hopkins Symptom Checklist
HSRS Health Sickness Rating Scale
ICD-10 International Classification ofDiseases 10
IGP interpersonal group psychotherapy
IQR interquartile range
IRT interpersonal reconstructivetherapy
ITT intention-to-treat
LAAM levo-alpha acetyl methadol
LOS length of stay
LPC Lifetime Parasuicide Count
MACT manual-assisted cognitivebehavioural therapy
MBT mentalisation-based partialhospitalisation
NA not applicable
NICE National Institute for Health andClinical Excellence
NR not reported
ns not significant
OBI Objective Behaviours Index
ONS Office for National Statistics
PD personality disorder
PDE Personality Disorders Exam
PDQ Personality Diagnostic Questionnaire
PH partial hospitalisation
PHI Parasuicide History Interview
POPMACT Prevention of Parasuicide byManual-Assisted CognitiveBehaviour Therapy
PS parasuicide
PSA probabilistic sensitivity analysis
QALY quality-adjusted life-year
RCT randomised controlled trial
SAS Social Adjustment Scale
SCID Structured Clinical Interview forDSM
SCL-90-R Symptom Checklist 90–Revised
SD standard deviation
SE standard error
SFQ Social Functioning Questionnaire
SHI Social History Interview
SSRI selective serotonin reuptakeinhibitor
TAU treatment as usual
TFT transference-focused therapy
TLFB Timeline Follow-Back
Glossary and list of abbreviations
All abbreviations that have been used in this report are listed here unless the abbreviation is well known (e.g. NHS), or it has been used only once, or it is a non-standard abbreviation used only in figures/tables/appendices in which case the abbreviation is defined in the figure legend or at the end of the table.
viii
Health Technology Assessment 2006; Vol. 10: No. 35
ix
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
BackgroundBorderline personality disorder (BPD) is a severeand complex mental disorder characterised bypervasive instability in moods, interpersonalrelationships, self-image and behaviour. In theDSM-IV system, the criterion for a diagnosis ofBPD is five of nine presenting symptoms. TheOffice for National Statistics 2000 survey ofpsychiatric morbidity in private householdsidentified seven people per 1000, which indicatesthat for a primary care trust of 500,000, therewould be 3500 individuals meeting the criteria for BPD.
Psychological therapies for BPD have many factorsin common, such as a high level of structure,consistency, theoretical coherence, taking accountof the relationship problems (including thedifficulty in engaging positively with thetherapist), and taking a flexible and individualisedapproach to care. Within these general principles,several specific therapies have been applied to,and developed for use with, patients with BPD.Mental health practitioners specifically trained inthe methods described deliver these treatments.The practitioners may have a qualification inpsychiatry, mental health nursing, clinicalpsychology or another mental health profession(e.g. occupational therapy or mental health socialwork).
ObjectiveThe aim of this project was to summarise theavailable evidence on the clinical effectiveness andcost-effectiveness of psychological therapiesincluding dialectical behaviour therapy (DBT) forBPD.
More specifically, the review aimed to:
● evaluate clinical effectiveness in terms ofreductions in self-harm and suicide
● evaluate effectiveness in terms of improvedpsychological functioning (e.g. in terms ofdissociation and mood)
● evaluate effectiveness in terms of interpersonaland social functioning
● evaluate effectiveness in terms of quality of life
● evaluate effectiveness in terms of presentationto mental health and other services (includingaccident and emergency attendance andpsychiatric hospital admission)
● evaluate the cost-effectiveness of the therapiescompared with treatment as usual
● identify the important areas of ignorance oruncertainty.
MethodsClinical effectivenessA systematic review of the literature aimed toidentify all references related to the clinical andcost-effectiveness of psychological therapiesincluding DBT for BPD.
Twenty electronic bibliographic databases weresearched, covering biomedical, health-related,science and social science literature. In addition,attempts were made to identify ‘grey’ literature bysearching appropriate databases (e.g. HealthManagement Information Consortium, Index toTheses, Dissertation Abstracts), current researchregisters (e.g. National Research Register, CurrentControlled Trials) and the Internet (e.g. bysearching Google and relevant websites, such asthe British Association for Behavioural andCognitive Psychotherapies, British PsychologicalSociety and Royal College of Psychiatry). Citationsearches of included studies were undertakenusing the Science Citation Index and SocialSciences Citation Index citation search facility, andthe reference lists of included studies and relevantreview articles were also checked.
The study quality of relevant studies was assessed using standard checklists and data wereabstracted by two reviewers using standardisedforms.
Cost-effectivenessThe cost-effectiveness assessment was in two parts.The first was a review of the literature. The secondwas an original assessment undertaken by thereview team using evidence from the clinical trialsand other sources.
Executive summary
x
It was not possible to apply a formal decisionmodelling approach given the complex carepathways for patients with BPD and the lack ofevidence. It was decided instead to undertakeseparate economic evaluations for the sixrandomised controlled trials (RCTs) that hadsufficient data using a combination of datareported in published papers, trial data sets sentby the investigators, a cost model using data fromthe POPMACT study and a utility mappingexercise. Cost-effectiveness was assessed in termsof cost per parasuicide event avoided in all sixtrials and cost per quality-adjusted life-year(QALY) in four of them, (which was done bymapping BDI results onto the EQ-5D for three.All results are at 2003–4 prices and for 12 monthsfollow-up.
ResultsNumber and quality of studies, anddirection of evidenceClinical effectivenessTen studies met the inclusion criteria of DBT,mentalisation-based partial hospitalisation (MBT),manual-assisted cognitive behavioural therapy(MACT), comprehensive validation therapy (CVT)and client-centred therapy (CCT), along withtreatment as usual (TAU). Of these, nine wereRCTs and one was a non-randomised comparativestudy. The quality of the studies ranged frommoderate to poor.
Cost-effectivenessThe review of published studies identified onecost-effectiveness analysis of psychological therapyfor BPD. This was based on data from an RCTcomparing DBT with TAU for the treatment ofBPD. Participants were women who were clinicallyreferred to a psychotherapy outcome study. Thereview of published studies also identified aneconomic evaluation of psychological therapies ofpartial relevance to BPD. This was a cost-effectiveness analysis of data from an RCTcomparing MACT with TAU for the treatment ofpeople with recurrent episodes of deliberate self-harm. A subgroup analysis was published but thisdid not present a full economic evaluation(although one was undertaken by the review team).
Evidence of effectivenessClinical effectivenessNine RCTs and one non-RCT of moderate to poorquality were identified in the clinical effectivenessreview. There is some evidence to support theeffectiveness of psychological therapies for BPD:
● There is some evidence that DBT is moreeffective than TAU for the treatment ofchronically parasuicidal and drug-dependentborderline women.
● There is some evidence that DBT-orientatedtherapy is more effective than CCT for thetreatment of BPD.
● There is some evidence that DBT is as effectiveas CVT with 12-Step (CVT+12S) for thetreatment of opioid-dependent borderlinewomen.
● There is some evidence that partialhospitalisation is more effective than TAU inthe treatment of BPD.
● There is good evidence that MACT is no more effective than TAU in the treatment of BPD.
● There is some evidence that interpersonalgroup therapy is no more effective thanindividual MBT for the treatment of BPD.
However, these results should be interpreted withcaution as not all studies were primarily targetedto borderline symptoms and there wereconsiderable differences in patient characteristics,comparison groups and outcomes between thestudies.
Cost-effectivenessReviewOne cost-effectiveness analysis used data from anRCT that compared DBT with TAU for thetreatment of BPD. The participants were womenwho were clinically referred to a psychotherapyoutcome study. Those receiving DBT (n = 22)incurred significantly higher psychotherapy costs,lower psychiatric inpatient costs and loweremergency room costs compared with TAU(n = 22). The two treatment groups did not differsignificantly with respect to median medical ortotal healthcare costs. The cost-effectivenessmeasures used were cost per week employed andcost per point of global adjustment, and nosignificant difference was found in either of thesemeasures for DBT compared with TAU. This studyhad limitations concerning the lack of importantcost data and the fact that it was undertaken usingdata from a small, underpowered trial with a highdropout rate.
The cost-effectiveness analysis comparing MACTwith TAU for the treatment of people withrecurrent episodes of deliberate self-harm found no significant differences between thegroups in the total costs across all patients oramong those with BPD (n = 62). The cost per 1%reduction in the proportion of patients with a
Executive summary
repeat self-harm episode was £120, with morethan a 90% chance of being cost-effective, but thisanalysis was not undertaken for the BPDsubgroup. The incremental mean effect asmeasured by EQ-5D was negative for MACT(–0.01118). The incremental cost per QALYgained from TAU was therefore £66,000, but theauthors argued that this was probably a chancefinding given that the difference in EQ-5D was notsignificant.
AssessmentIn three of the four DBT trials, the interventiondominated the control groups in terms ofparasuicide events or achieved a cost per eventavoided below £50. However, in a fourth DBT trialthe estimated cost per event avoided was £43,124.Although these studies seem favourable to DBT interms of mean incremental cost-effectiveness, theprobability of being cost-effective at £5000 perparasuicide event avoided was around just 60% ineach case. Only two DBT trials could be subjectedto a cost per QALY analysis, and for one theintervention again dominated and the other had acost per QALY of £273,801. The probabilisticsensitivity analysis showed substantial uncertaintysurrounding these results; the most favourablestudy had a probability of DBT being cost-effectiveof around 85%.
The MBT study group achieved a low cost perparasuicide event avoided, with a probability ofbeing cost-effective at £5000 per parasuicide eventavoided of 80%. While the cost per QALY wasmodest at £7242, there was substantial uncertainty,with a probability of being cost-effective at£20,000 per QALY of less than 60%. For thePOPMACT, the BPD subgroup analysis found thatthe intervention was dominated in terms of costper parasuicide event avoided. There was aninsignificant incremental QALY gain in BPD, withan associated cost per QALY of £84,032. Theseassessments of MACT were both associated with ahigh degree of uncertainty, where the probabilityof being cost-effective was less than 50% in eachcase.
These assessments must be viewed with great care.The trials on which they were based were often ofpoor quality, using a mixture of methods forcosting and assessing outcome (including QALYs)and of doubtful generalisability to the NHS formany of the studies. This mixture of results, highlevels of uncertainty and the limitations inmethods provides very limited support for thecost-effectiveness of DBT, but the results suggestthat DBT could be cost-effective.
ConclusionsThe overall efficacy of psychological therapies ispromising; however, at this stage the evidence isinconclusive.
This study attempted to examine the cost-effectiveness of the intervention in six RCTs. Themixture of results for the four trials of DBT, plusthe high levels of uncertainty and the limitationsof the analyses, do not support the cost-effectiveness of DBT, although they suggest that it could have the potential to be cost-effective.The results for MBT are promising, althoughagain surrounded by a high degree of uncertainty, and for MACT, the analysis suggeststhat the intervention is unlikely to be cost-effective. There is a need for considerable research in this area.
Recommendations for researchThe results from existing studies in this field haveproduced a body of evidence that has been largelyinconclusive. BPD is an important condition with anumber of resource-intensive therapies availableand it should be a priority area for future research.Suggestions for further research in terms ofpragmatic trials and studies to inform economicevaluation are presented below.
Pragmatic controlled trialsAppropriately powered head-to-head RCTs ofpsychological therapies are needed. The keyfeatures of these trials include:
● Where possible, a trial should have more thanone psychological therapy being compared.
● Studies must be designed with adequatestatistical power taking into account expecteddropouts.
● Patients from a variety of ethnic and socio-economic backgrounds must be included, withan age and gender mix comparable to thosereceiving treatment on the NHS.
● The level of severity and dysfunction must bewell defined.
● The definition of ‘dropout’ must bestandardised and reduced where possible in theRCTs examining psychological therapies forBPD. Where patients drop out of therapyconsiderable effort must still be undertaken tocollect data on them.
● The different therapies need to be properlydescribed, including a TAU arm (e.g.medication must be taken into account).
Health Technology Assessment 2006; Vol. 10: No. 35
xi
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
● The longest follow-up has been for 18 months,and 6 months was more common. Given thehigh cost of the interventions, longer termfollow-ups should be undertaken.
● Data should be collected on outcomes,including recognised generic measures ofhealth-related quality of life, includingpreference-based measures to permitcomparisons across programmes (see below).
● Data should be collected on resource-useservices (see below).
● Research teams should include independentresearchers.
Studies to inform future economicanalyses● A survey of current practice and the use of the
full range of services (including number of
sessions attended and type of therapist) bypeople with BPD is needed to inform futureeconomic analyses.
● Full resource-use data must be collected in thecontext of pragmatic clinical trials.
● A psychometric assessment is needed of thevalidity of the EQ-5D and other genericpreference-based measures in BPD.
● If the generic measures are found wanting, thena more condition-specific preference-basedmeasure that captures the impact of BPD onpeople’s lives should be developed.
● A more formal cost-effectiveness model needs tobe developed using the above data.
Executive summary
xii
Borderline personality disorder (BPD) is asevere and complex mental disorder
characterised by pervasive instability in moods,interpersonal relationships, self-image andbehaviour. There are several psychologicaltherapies, including dialectical behaviour therapy(DBT), and there is an emerging evidence base ontheir efficacy. The aim of this project is tosummarise the available evidence on theeffectiveness and cost-effectiveness ofpsychological therapies including DBT for BPD.
More specifically, the review aims to:
● evaluate clinical effectiveness in terms ofreductions in self-harm and suicide
● evaluate effectiveness in terms of improvedpsychological functioning (e.g. in terms ofdissociation or mood)
● evaluate effectiveness in terms of interpersonaland social functioning
● evaluate effectiveness in terms of quality of life● evaluate effectiveness in terms of presentation
to mental health and other services includingaccident and emergency (A&E) attendance andpsychiatric hospital admission)
● evaluate cost-effectiveness of the therapiescompared with treatment as usual (TAU)
● estimate the possible overall cost in Englandand Wales.
● identify the important areas of ignorance oruncertainty.
In undertaking to achieve the above aims thereview will consider factors such as the setting andprocess of therapy, including the professionalbackground of therapists involved, impact on theuse of other services, co-morbidity, and co-medication and patient characteristics includingchronicity and severity of the condition. Therapieswill be compared against any control orcomparator.
Health Technology Assessment 2006; Vol. 10: No. 35
1
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Chapter 1
Aim of the review
Description of underlying healthproblemBPD, one of nine or ten personality disorderdiagnoses [according to International Classificationof Disease-10 (ICD-10) or Diagnostic andStatistical Manual of Mental Disorders IV (DSM-IV)], is characterised by instability of self-image,interpersonal relationships and mood. The personhas an uncertain sense of identity, feelings of inneremptiness and often a fear of being alone.Interpersonal relationships tend to be intense butstormy, and there may be an intense fear ofabandonment and strenuous efforts to avoidabandonment, real or imagined. The person hasdifficulty regulating their emotions, with extremeand sudden shifts of mood to intense depressionor anxiety, often lasting for a few hours. Theindividual may be prone to respond to somesituations with intense anger, or with impulsive,often self-harming, behaviour. Under stress, theperson may dissociate or become paranoid. Self-harm can include attempted suicide, overdosingor self-mutilation (e.g. through cutting or burning).
EpidemiologyThe classification of personality disorders remainscontroversial and the borderline diagnosis inparticular has been criticised on scientificgrounds.1 There are high levels of co-morbiditybetween different personality disorder diagnosesand difficulties in reliable assessment, factors thatshould be taken into account when consideringepidemiological estimates.
Prevalence data for the UK are available from theOffice for National Statistics (ONS) 2000 survey ofpsychiatric morbidity in private households. Thissurvey suggests that the prevalence of borderlinedisorder was 7 per 1000, which indicates that for aprimary care trust serving a population of 175,000there would be 1250 individuals meeting thecriteria for BPD.
The ONS survey can be compared withinternational estimates of prevalence. Thesesuggest a greater prevalence of BPD in women,but the ONS survey found lower rates for women(4 per 1000 women compared with 10 per 1000men). This discrepancy may reflect differences in
sampling and instrumentation, and it is possiblethat the excess prevalence in women has beenoverestimated in some studies. An implication forresearch is that findings from intervention studieswith all-women samples may not generalise to thefull population.
The overall prevalence in the UK study is at thelower end of the range of 0.7–4.6% reported inother studies.2–7
AetiologyThe cause of BPD is complex, with adverseexperiences in childhood, such as neglect andabuse, including sexual abuse,8–11 interacting witha genetic predisposition for emotionaldysregulation.12,13
PrognosisEarly studies suggested that over 5-year follow-up,symptomatic patterns change little,14–16 but that bymiddle age, many people will no longer meet BPDcriteria, mainly through a reduction in impulsivity,self-harm and aggression, although otherborderline features, such as inner emptiness andaffective instability, may continue.17–19
More recent studies suggest that this may be toonegative a picture, with 75–80% losing the diagnosisover 4–10 years of follow-up.20–22 There has beenreplication of the finding that the change is in thebehavioural and reactive criteria such as self-harm,rather than in negative affectivity and anger.23
Significance in terms of ill-healthBPD represents a significant burden of ill-health,with reduced levels of functioning, difficultymaintaining relationships, difficulty maintainingemployment, high levels of service use, includingattendance at A&E departments and admission topsychiatric hospitals, and rates of suicide morethan 50 times higher than in the generalpopulation.13,21,24,25
Description of new interventionIdentification of patients and subgroupsThere are wide variations in presentation and inseverity. In the DSM-IV system, the criterion for a
Health Technology Assessment 2006; Vol. 10: No. 35
3
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Chapter 2
Background
diagnosis of BPD is five of the following ninepresenting symptoms:
● inappropriate intense anger or difficultycontrolling anger
● chronic feelings of emptiness● affective instability● transient stress-related paranoid ideation or
severe dissociative symptoms ● identity disturbance: striking and persistent
unstable self-image or sense of self● recurrent suicidal behaviour, gestures or threats;
or self-mutilating behaviour● impulsivity in at least two areas that are self-
damaging that do not include suicidal or self-mutilating behaviour
● frantic efforts to avoid real or imaginedabandonment
● a pattern of unstable and intense interpersonalrelationships characterised by alternatingbetween extremes of idealisation anddevaluation.
This means that two individuals may meet criteriafor the disorder with only one of five symptoms incommon. The implications of this for research areimportant, as some treatments [e.g. DBT, manual-assisted cognitive behavioural therapy (MACT)]were primarily developed for deliberate self-harm(DSH), which although commonly found, isneither a necessary nor a sufficient criterion forthe diagnosis of BPD. In addition, the wide rangeof possible presentations means that interventionsfor BPD are complex, which generates problemsin designing research and interpreting findings.
Criteria for treatmentIn addition to meeting criteria as above, allpsychological therapies rely on a minimumwillingness to attend regular sessions and form aworking relationship with the therapist. For peoplewho have severely problematic drug or alcoholdependence, coexisting psychotic symptoms, severeand intense suicidal behaviour or intellectualimpairment, this may not be possible.
InterventionPsychological therapies for BPD have many factors in common, such as a high level ofstructure, consistency, theoretical coherence, andtaking account of the relationship problems(including the difficulty in engaging positivelywith the therapist), and taking a flexible andindividualised approach to care. Within thesegeneral principles, a number of specific therapieshas been applied to, and developed for use with,patients with BPD.
Psychodynamic therapies used with BPD includetransference-focused therapy (TFT26) andmentalisation-based therapy.27 Cognitive andbehavioural approaches include DBT,28 schema-focused therapy,29 and MACT30 [an adapted formof cognitive behavioural therapy CBT)].31
Integrative (relational) approaches includecognitive analytic therapy32 and interpersonalreconstructive therapy (IRT). Evidence fromrandomised trials has not yet been published forall of these modalities, and the interventionsdescribed here are those for which trial data arereported, namely DBT, MACT and mentalisation-based partial hospitalisation (MBT).
DBTDBT33 was developed for women who self-harm.Five stages of treatment are outlined, but mostliterature and all research focuses on stage 1, whichaims to help the patient to develop motivation tostay in treatment and achieve behavioural controlover urges to self-harm. Weekly individual therapyand a weekly psychoeducational and skills traininggroup are offered concurrently over 1 year. Theaim is to achieve behavioural control, stability andconnection with the care provider. Patients move tothe second stage (emotional experience andreprocessing of past trauma) when behaviouralcontrol has been achieved. The key principles oftreatment include moving flexibly betweenacceptance and validation and behavioural changestrategies; this includes behavioural analysis,solution analysis and strategies, skills training,contingency management, exposure, cognitivemodification and psychoeducation. The DBTpackage also includes weekly supervision andconsultation meetings for the therapists, who workas a team, and telephone consultation, wheretherapists are available to patients outside officehours for coaching.
MACTMACT therapy30 was developed as a public healthintervention for the large numbers of people whorepeatedly attempt suicide (parasuicide) ratherthan for BPD per se. However, a high proportionof people in this population meet criteria for BPD,and this subpopulation is therefore similar to thatfor which DBT was developed. The intervention isa brief, cognitively orientated and problem-focused therapy comprising up to five sessionswithin 3 months of an episode of self-harm, withthe option of a further two booster sessions within6 months. Bibliotherapy, in the form of a 70-pagebooklet, is used to structure the treatment sessionsand to act as an aide-mémoire between sessions. Themanual covers an evaluation of the self-harm
Background
4
attempt, crisis skills, problem solving, basiccognitive techniques to manage emotions andnegative thinking, and relapse prevention strategies.
MBTMBT27 also termed psychoanalytically orientatedpartial hospitalisation, is based on anunderstanding of BPD as a disorder of the selfresulting from a failure in mentalisation, withintervention aimed at increasing the self-reflectivecapacity of the patient. Treatment is in the contextof a day hospital and comprises many elements,including weekly individual therapy, thrice-weeklygroup analytical therapy, weekly expressivetherapy with psychodrama and a weeklycommunity meeting, for a maximum of18 months.
Personnel involvedThese treatments are delivered by mental healthpractitioners specifically trained in the methodsdescribed. The practitioners may have aqualification in psychiatry, mental health nursing,clinical psychology or other mental healthprofession (e.g. occupational therapy, mentalhealth social work). Training routes vary, but afterthe core professional qualification, usually involvea 1- or 2-year part-time course followed bysupervised practice.
Current evidenceTwelve systematic reviews reporting at least someinformation regarding BPD, personality disorder,DSH and suicide attempters as populations andpsychological treatments as interventions wereidentified (Table 1). They were aimed either atbroad strategies such as dissemination ofguidelines34–37 or at particular target groups andproblem areas related to personality disorders.38–45
Most primary studies were included in more thanone review. There was a lack of common approachaccepted between the reviews and the inclusionand exclusion criteria varied considerably.Interventions were classified differently indifferent systematic reviews. The characterisationof BPD was also complicated. Because the BPD isa subcategory of personality disorder withDSH/suicide attempts as a main feature, thereviews on personality disorders and DSH ofteninclude borderline patients and there are very fewreviews conducting BPD subgroup analyses orreviews on BPD on its own.
No systematic reviews published before 1997 wereidentified. Four reviews conducted meta-analyses
of the results of the studies identified.36,38,39,40
Because of the broad inclusion criteria andheterogeneity of the studies included in thereviews the appropriateness of meta-analyses isuncertain. Table 1 presents the overlap ininformation of identified systematic reviews.Among the reviews only one high-qualitysystematic review38 was designed to look at studieson BPD. This review was completed by theCochrane Collaboration on behalf of the NHSNational R&D Programme on Forensic MentalHealth, UK. The authors were contacted and gavetheir permission for the review to be used in thisreport before its official publication. The reviewerssearched large number of electronic databasessupplemented by citation tracking of includedarticles and keywords. Only published data wereincluded in this review.
To assess the effectiveness the authors limited thetype of study to randomised controlled trials(RCTs). The quality of studies was assessedaccording to the Cochrane CollaborationHandbook46 and only trials in category A and Bwere included. The outcomes were data fromassessment scales such as global state, behaviourand mental state. The engagement with services,satisfaction with treatment, acceptance oftreatment and quality of life were also assessed.Meta-analyses were performed on relevantoutcomes using a random effects model. Theauthors described in detail the methods of testingheterogeneity and sensitivity analyses. Although awide range of psychological therapies is used totreat BPD, many were omitted from the reviewowing to a lack of RCT evidence.
The review by Adams and colleagues38 suggeststhat some problems of BPD patients may betreated by behavioural therapies. However, theauthors note that all reviewed therapies arecurrently at the experimental stage and thenumber and size of the trials are too small to cometo clear conclusions.
The scope of Adams’ review38 overlapped with thecurrent review in all aspects of inclusion criteria.Trials identified by Adams and colleagues38 arealso included in the current review and describedin the section ‘Results’ (p. 8). The populations,interventions and outcomes of the Adams review38
were compared with the scope of the present HTAreview to assess the degree of overlap and identifyareas not covered. The current review differs inthat non-RCT evidence is also included.
Health Technology Assessment 2006; Vol. 10: No. 35
5
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Background
6
TABLE 1 Summary of systematic reviews of BPD
Study Type of Condition Psychological BPD Main outcomeevidence therapy reported
intervention
RCT Non- PD DSH BPDRCT
Binks et al., 2005 Yes – – – Yes Yes Yes Improvement of (NHS National R&D symptoms Programme on Forensic Suicide and repetition of Mental Health)38 DSH
Treatment retentionTreatment duration
Bateman and Fonagy, Yes Yes Yes Yes Yes Effectiveness of the 200045 treatment
Boyce et al., 2003, Yes Yes – Yes – Yes Yes Improvement of (Australian and New symptomsZealand clinical practice Suicide and repetition of guideline)34 DSH
Cornah et al., 1997 – Yes – – Yes Therapeutic Yes Improvement of (NHS R&D Directorate community symptoms Report)37 Effectiveness of
therapeutic community
Hawton et al., 199839 Yes – – Yes – Yes Yes Improvement ofsymptomsSuicide and repetition ofDSH
Lees et al., 1999 Yes Yes Yes – – Therapeutic Yes Improvement of (NHS, CRD Report)36 community symptoms
Effectiveness oftherapeutic community
Leichsenring, 200242 Yes Yes Yes – – Psychodynamic Yes Effectiveness of therapy psychodynamic therapy
Perry et al., 199941 Yes Yes Yes – – Yes Yes Improvement ofsymptomsTreatment retention Treatment duration
Sanislow and McGlashan, – Yes Yes – – Yes Yes Effectiveness of the 199843 treatment
Van der Sande et al., 199740 Yes – – Yes – Yes Yes Suicide and repetition ofDSHEffectiveness of therapy
Warren et al., 2003 Yes Yes Yes – – Yes Yes Improvement of (Home Office Report)35 symptoms
Suicide and repetition ofDSHTreatment retention Treatment duration
Woods and Richards, 200344 Yes Yes Yes – – Nursing Yes Effectiveness of nursing interventions
CRD, Centre for Reviews and Discrimination; PD, personality disorder.
Methods for reviewingeffectivenessIdentification of studiesAim of the search strategyOne set of searches was undertaken to inform thereview of clinical effectiveness. The aim of thesesearches was to identify all studies relating topsychological therapies for BPD. A second set ofsearches was undertaken to inform the twoeconomic aspects of the assessment. The scope ofthese searches was broader, to satisfy the inclusion criteria of the cost-effectiveness reviewand to inform the broader requirements of theeconomic assessment. The aim of these searchestherefore was to identify all economic studiesrelating to BPD (i.e. not restricted byintervention).
Sources searchedTwenty electronic bibliographic databases weresearched, providing coverage of the biomedical,health-related, science, social science and greyliterature (including theses and conferenceabstracts). The publications lists and currentresearch registers of health services research, social care and mental health organisations were consulted via the World Wide Web. Keyword searching of the World Wide Web wasundertaken using the Google search engine. The reference lists of included studies andrelevant review articles were also handsearched. A list of the sources searched is provided inAppendix 1.
Search termsSensitive keyword strategies using free-text and,where available, thesaurus terms were developedto search the electronic databases. The selection ofkeywords was informed by DSM-III, DSM-IV, ICD-10 and clinical members of the assessment team.For the review of clinical effectiveness synonymsrelating to the intervention (e.g. psychologicaltherapies, dialectical behaviour therapy) werecombined with synonyms relating to thepopulation (e.g. borderline personality disorder,Axis II, Cluster B). For the economic searchessynonyms relating to the population only wereused. Keyword strategies for all electronicdatabases are provided in Appendix 2.
Search restrictionsThe clinical effectiveness searches were notrestricted by terms relating to study design. Theeconomic searches were restricted by termsrelating to cost and economics. The search ofPubMed was restricted to the last 180 days tocapture recent and unindexed MEDLINE records.Date limits were not used on any other database.Language restrictions were not used on anydatabase. Searches were undertaken in March 2005.
Inclusion and exclusion criteriaEligibility for this review was determined by thefollowing criteria:
● participants: adults with BPD (diagnosedaccording to DSM-III/DSM-III-R, DSM-IV orICD-10 criteria for BPD), with or without co-morbidity; studies on people with anypersonality disorder and DSH were alsoincluded, where subgroup analysis of BPD wasavailable
● intervention: psychological therapies, includingDBT
● comparators: any psychiatric or psychologicaltreatment, or no treatment
● outcomes: self-harm, suicide, interpersonal andsocial functioning, crisis presentations to mentalhealth services, quality of life, patient preference,satisfaction, acceptability of treatment and cost
● study type: published papers were assessedaccording to the accepted hierarchy of evidence,whereby systematic reviews of RCTs are taken tobe the most authoritative forms of evidence,with uncontrolled observational studies the leastauthoritative
● exclusion criteria: papers on personalitydisorder and DSH without separate BPDsubgroup analyses.
Figure 1 shows a summary of study selection andexclusion.
Quality assessment strategySystematic reviews were assessed according to theUsers’ guides to evidence-based practice.47 Thequality of RCTs was assessed using the CriticalAppraisal Skills Programme (CASP) checklist forappraising RCTs (http://www.phru.nhs.uk/casp/rcts.htm). The ten questions in CASP tool are
Health Technology Assessment 2006; Vol. 10: No. 35
7
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Chapter 3
Effectiveness
adapted from the users’ guides to the medicalliterature.48,49
The non-randomised study using quantitative datawas assessed with respect to validity using theCASP checklist for cohort studies(http://www.phru.nhs.uk/casp/cohort_studies.htm).It contains 12 questions for assessing the types ofstudy in terms of their validity, results andapplicability. The quality of the economicliterature was assessed according to the Guidelinesfor authors and peer reviewers of economicsubmissions to the BMJ.50
Key components of quality assessment are listed inAppendix 3 (Tables 18 and 19).
Data extraction strategyData were extracted by one researcher and checkedby another using the Reference Manager database.Any disagreements were resolved by discussion.The authors aimed to cover as many as possible
types of psychotherapy and find the best availableevidence on each treatment. Where there were noRCTs available on a certain type of intervention,lower level (comparative) studies were considered.
Data synthesisThe suitability of pooling data across studies wasassessed by examining study populations,comparators, outcomes and study type. Studieswere found to be too heterogeneous in theserespects for meta-analysis to be appropriate, and itwas not undertaken. The results are thereforepresented in tabulated format with narrativesynthesis of the results.
ResultsQuantity and quality of researchavailableTen trials were identified, of which nine wereRCTs and one was a non-randomised trial. Table 2
Effectiveness
8
Potentially relevant papersidentified and screened forretrievalRCT n = 1216Non-RCT n = 2455
Total abstracts screenedRCT n = 226Non-RCT n = 763
Papers rejected at titleRCT n = 990Non-RCT n = 1692
Papers rejected at abstractRCT n = 145Non-RCT n = 724
Rejected full paperRCT n = 72Non-RCT n = 38
Total full papers screened (plusfrom other sources)RCT n = 81Non-RCT n = 39
Studies included in this reviewRCT n = 9Non-RCT n = 1Total n = 10
FIGURE 1 Summary of study selection and exclusion
summarises the studies included in this review.(Excluded studies are listed in Appendix 4.)
Appendix 3 contains the evidence tables with dataextracted from the ten studies included in thisreport. RCTs and non-randomised trials arepresented in separate tables.
Study characteristicsStudy characteristics for the ten studies aredescribed in Appendix 3 (Tables 18 and 19).
Description of psychotherapiesRCTs The studies report problem-focusedpsychotherapies administered in the outpatient
Health Technology Assessment 2006; Vol. 10: No. 35
9
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TABLE 2 Studies included in the review
Study Sample Intervention Comparator Population Primary outcome size
RCTsBateman and 38 PH TAU Patients with severe Effectiveness of psychoanalytically Fonagy, 199951a parasuicidal BPD orientated PH for the BPD
symptoms
Koons et al., 20 DBT TAU Women Veterans Effectiveness of DBT for the 200157 with BPD treatment of BPD
Linehan et al., 44 DBT TAU Chronically Effectiveness of DBT for the 199153a parasuicidal women treatment of BPD women with
with BPD BPD
Linehan et al., 28 DBT TAU Substance abusing Effectiveness of DBT for the199954 women with BPD treatment of drug-dependent
women with BPD
Linehan et al., 23 DBT CVT+12S Heroin-dependent Effectiveness of DBT for the 200255 women with BPD treatment of heroin-dependent
women with BPD
Munroe-Blum 48 Time –limited Individual Patients with BPD Effectiveness of IGP for the and Marziali, (psychodynamic) psychotherapy treatment of BPD199556 IGP
Turner, 200057 24 DBT-orientated CCT Patients with BPD Effectiveness of a DBT-orientated therapy therapy for BPD
Tyrer et al., 480 MACT TAU Patients with Effectiveness of MACT in 200358a (PD recurrent DSH reduction of depressive
n = 391 (including BPD) symptoms and the rate of BPD parasuicide eventsn = 67)
van den Bosch 34 DBT TAU Female borderline Effectiveness of the DBT for the et al., 200259a patients with or BPD and substance use problems
without co-morbid substance abuse
Non-RCTWilberg et al., 43 Day hospital Day hospital Patients with BPD Effectiveness of the day treatment 199860 treatment (TAU) treatment with subsequent outpatient
plus post- (TAU) group therapy for the BPDdischarge group analytical therapy
a The follow-up and other relevant studies were assessed as part of the main studies, and only the main paper is cited in thecurrent report. All other relevant references can be found from the reference list. Linehan et al. (1993)61 was the 12-month post-treatment follow-up report of Linehan et al., (1991)53 and Bateman and Fonagy (2001)62 was the 18-monthpost-treatment follow-up report of Bateman and Fonagy (1999).51 Tyrer et al. (2004)63 was part of Tyrer et al. (2003)58
and Verheul et al. (2003)64 was included as part of van den Bosch et al. (2003).59
CVT+12S, comprehensive variation therapy with 12-Step; IGP, interpersonal group psychotherapy; PH, partialhospitalisation.
setting using various methods. Five RCTs, byLinehan,53–55 Koons52 and van den Bosch andcolleagues,59 assessed the effectiveness of DBT, thetherapy that balances problem-orientatedtechniques with supportive techniques such asreflection, empathy and acceptance of patient’sinherent ability. One RCT57 used a form ofintegrative therapy based on DBT, wherepsychodynamic techniques were incorporated toconceptualise patients’ behavioural, emotional andcognitive relationship schema and the skillsgroups were omitted. Instead, group sessions inboth the treatment and control groups emphasisedinterpersonal relationships. One trial51 used thepsychoanalytically orientated partialhospitalisation method, where the partiallyhospitalised group received a combination ofindividual and group MBT along with communitymeetings. One trial56 looked at the time-limitedgroup therapy approach, which addresses theinterpersonal problems and focuses on observingand processing the meaning of within-therapyenactment of interpersonal communication andbehaviour, among patients and between patientsand co-therapists. One RCT58 on DSH usedMACT, where patients were given a bookletcovering problem solving, basic cognitivetechniques to manage emotions, negative thinkingand relapse prevention strategies.
Non-RCT The study by Wilberg and colleagues60
used day hospital treatment followed bypostdischarge, group analytical therapy.
Study qualityRCTs The quality of RCTs was assessed using theCASP checklist for appraising RCTs(http://www.phru.nhs.uk/casp/rcts.htm), whichcovers five main categories: randomisationmethod (description of the randomisationmethod), blinded assessment, power calculation,reason for loss for follow-up and intention-to-treat(ITT) analysis (Table 3).
Because people with BPD have difficulty formingand sustaining collaborative interpersonalrelationships, psychological therapy is oftendifficult to deliver. A high dropout rate, failure to meet power calculation estimates andimpossibility of blinding are typical for the studies examining psychological interventions, and assessment of their quality with the standardcriteria seems inadequate. Therefore, the authors used an additional modified Lackner’schecklist based on the Cochrane CollaborationDepression and Anxiety Neurosis Review Group (CCDAN) scale, which covers a wider areaof design issues specifically for psychologicalstudies.
Lackner and colleagues65 adapted the CCDAN 23-item coding scheme (CCDAN criteria are onwww.iop.kcol.ac.uk/IoP/ccdan/index.htm)specifically to rate the methodological quality ofpsychological treatment trials. They incorporatedsix additional items into the measure, so that thefinal 29-item coding scheme reflected both
Effectiveness
10
TABLE 3 Study qualities assessed using the CASP checklist
Study Description of Blinded Power Reason for ITTrandomisation assessment calculation loss to
method follow-up
Bateman and Fonagy, 199951 N NR N N N
Koons et al., 200152 N NR Y Y N
Linehan et al., 199153 Y Blinded independent assessor N Y N
Linehan et al., 199954 Y Blinded independent assessor N N Y
Linehan et al., 200255 Y Blinded independent assessor N N Y
Munroe-Blum and Marziali, Y NR Y N N199556
Turner, 200057 N Blinded independent assessor N N Y
Tyrer et al., 200358 Y NR Y Y Y
van den Bosch et al., 200259 Y Independent assessora N Y Y
a Independent assessors were not informed about the treatment condition of the interviewees; however, patients mighthave given the information about treatment.
N, no; NR, not reported; Y, yes.
general, evidence-based medicine guidelines forrating the quality of clinical trials (e.g. post-treatment follow-up for all groups, declaredallegiance or therapy, avoidance or equality of co-interventions) and recommendations.
Appendix 3 (Tables 18 and 19) and Appendix 5show quality scores for each of the trials. All trialswere described as randomised with treatmentlasting from 6 months52 to over 18 months,51 andall trials except for Turner57 and Munroe-Blumand Marziali56 gave a description of therandomisation method. No studies were doubleblind. It is impossible to blind patients andtherapists in psychological treatments as both areaware that therapy is taking place. Therefore,assessments were conducted by independentassessors who were not aware of patients/treatmentconditions. Only three trials52,56,58 reported thatthey did a power calculation; however, the samplesizes were too small in most of them to detect thetrue effect of the treatment. All reported thenumber of withdrawals, but only five trialsreported the reasons for dropouts.52–54,58,59
Dropout rates were high, up to one-third of thetotal sample, often with no apparent reason, withpatients starting to dropout immediately afterrandomisation and pretreatment assessment andthroughout the active treatment period and post-treatment follow-up. ITT analysis is a strategy forthe analysis of RCTs that compares participants inthe groups to which they were originally assignedincluding all patients regardless of whether theyreceived the treatment or withdrew from the trial.Clinical effectiveness may be overestimated if anITT analysis is not done.66 Three trials51–53
reported data only for completers and assumedthat patients who left the trial had poor outcomes.The rest of the studies reported the use of an ITTanalysis. Most studies had a follow-up period of atleast 4 months, except for three trials.52,57,58 Twotrials56,58 had more than 50 participants in eacharm. Most participants were women. Most of thetrials gave full demographic details. All trialsclearly reported details of the inclusion/exclusioncriteria, diagnostic criteria and therapy conditions.However, important aspects such as allocationconcealment, therapy credibility and expectancywere omitted in all studies. No trial gave detailson side-effects. All described outcome measuresclearly and/or used validated instruments, andgave sufficient information on comparability ofgroups. The presentation of results was inadequate for later data synthesis in twotrials.56,58 All trials were considered to have‘mainly appropriate’ statistics, with someexceptions, such as the inappropriate use of one-
tailed tests. Conclusions were partially justifiedand all but two51,57 acknowledged support and/orfunding sources.
Non-RCT The Wilberg study60 was a comparativestudy with an active treatment period lasting forup to 1 year and had over 50 participants in eacharm. The study60 reported some information onsocio-economic background, but unclearinformation regarding participants’ gender. Thestudy had a small sample size, different baselinecomparability between two groups andretrospective assessment of diagnoses and somemeasurements [i.e. Health Sickness Racing Scale(HSRS)], which may lead to a serious bias inmeasuring outcomes.
Co-therapy or medicationRCTs The use of co-therapies was described inmost of the studies. Participants from the trialsconducted by the Linehan group53–55 aimed toterminate, taper off or replace psychotropicmedications. One trial51 reported thatpolypharmacy was discouraged, three trials52,57,59
reported the type of pharmacotherapy used byparticipants, and two trials56,58 did not provideany information about co-therapy.
Non-RCT Wilberg and colleagues60 did notreport co-therapy administration.
ComparatorsRCTs Six trials had TAU as a comparator,although the TAU was not always clearlydescribed. Koons and colleagues52 described TAUas 60 minutes of weekly individual therapy. Allpatients were additionally offered one or more ofseveral supportive and psychoeducational groups.Linehan and colleagues53 report that patients inthe TAU group were given alternative referrals.Nine patients received individual psychotherapyfor an average of 34.87 hours. The TAU conditionwas naturalistic and allowed participation in anytype(s) of therapy available in the community.Linehan and colleagues67 referred TAU patients toalternative substance abuse and mental healthcounsellors and programmes in the community, orthey continued their existing treatment. Van denBosch and colleagues59 report that TAU consistedof principal management from the originalreferral source: addiction treatment centre,psychiatric services with no more than two sessionsper month with a psychologist, psychiatrist orsocial worker. Tyrer and colleagues,63 report thatTAU patients were offered the standard treatmentin the area concerned or the continuation ofexisting treatment. This varied from problem-
Health Technology Assessment 2006; Vol. 10: No. 35
11
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
solving approaches, through dynamicpsychotherapy, GP or voluntary group referral, toshort-term counselling.
According to Bateman and Fonagy51 TAU consistsof general psychiatric services with regularpsychiatric review, hospitalisation as appropriatewith treatment focused on problem solving andoutpatient community follow-up. One trial55 onheroin-dependent women with BPD hadCVT+12S as a control therapy. CVT+12S isdesigned to control for provision of support,validation and general therapeutic acceptance,and other components of treatment not coveredby DBT. One trial57 used client-centred therapy(CCT). CCT is the model of supportive therapybased on Carkhuff ’s model,114 which emphasisesempathic understanding of the patient’s sense ofaloneness and provides a supportive attitude onan individual basis. One study56 used an individualdynamic psychotherapy as a comparator, which isopen-ended, individual therapy.
Non-RCT The Wilberg60 study design was TAUplus postdischarge group analytical therapy versusTAU, with TAU being day hospital treatment.Some of the control group participants wereuntreated postdischarge. Thus, in reported trialsmembers of the control group were not receivingthe same care.
Sample sizeRCTs Sample sizes were generally small (lessthan 50). Only two trials56,58 included more than100 patients. Three studies51,53,57 included from30 to 60 patients, and four trials52,54,55,57 hadfewer than 30 participants. Although threestudies52,56,58 reported a power calculation beingused to determine sample size, only two ofthese56,58 reached adequate power.
Non-RCT In the Wilberg trial60 the sample sizewas small (less than 50 in each arm).
Therapy detailsAppendix 3 describes the details of therapy(Tables 20 and 21).
RecruitmentRCTs Most trials recruited patients frompsychiatric hospitals or mental health centres, orpatients were referred by independent clinicians.The Linehan group55 also included participantswith substance abuse from methadonemaintenance clinics and HIV/AIDS preventionorganisations treating underserved minoritypopulations. One study57 recruited participants
from emergency services and community healthoutpatient clinics.
Non-RCT In the Wilberg study60 recruitment wasfrom patients at a day hospital.
Number and length of sessionsRCTs The number of sessions of DBT rangedfrom 48 to 56 (weekly for 1 year) according tostandard BPD treatment.53 In one study52 DBTwas administered for 6 months. A maximum of84 weekly sessions were administered for DBT-orientated therapy.57 Partial hospitalisation51 wasalso administered weekly over an average of1.45 years. Patients in IGP56 had 30 sessions(25 weekly sessions, followed by five twice-weeklysessions). The study on DSH58 reported thatpatients received six sessions according to sixchapters of the manual, but there is no indicationof the duration of the sessions.
Non-RCT In the Wilberg study60 the sessions were once a week for 1.5 hours over 1–33 months (mean length of treatment12 months).
Therapist contact between sessions andprofessional background of therapistRCTs Most studies reported that patients wereallowed to make a telephone call52–55,59 or have aface-to face57 consultation with therapists in acrisis situation. It was unclear whether there wassimilar access to therapists for comparator groups.Patients who used MACT were given the firstchapter of the manual by the therapist and in thelater sessions sought the therapist’s help only forspecific problems. More details are provided inTable 4 and Appendix 3 (Table 20). There is noinformation regarding the frequency and durationof the crisis situations and additional between-session consultations did not seem to be includedin the therapist time.
Non-RCT Therapist time was not reported.
Therapist’s professional background RCTs The therapist’s professional backgroundvaried from a psychotherapist with an average of22 years of experience57 to experienced graduatepsychology students53 and psychiatric nurseswithout a formal psychotherapy qualification.51 Allstudies report that therapists were trained in theintervention therapies. The details of thetherapists are given in Appendix 3 (Table 20).
Non-RCT In the Wilberg study60 the majority oftherapists (six out of eight) had a background in
Effectiveness
12
group analytical training. The details of thetherapists are given in Appendix 3 (Table 21).
Study site, follow-up and inclusion/exclusioncriteriaAppendix 3 (Tables 22 and 23) describes the studysite, follow-up and inclusion/exclusion criteria ofall included studies.
Study site and settingRCTs Five studies were conducted in theUSA,52–55,57 two in the UK,51,58 one in Canada,56
and one in The Netherlands.59 One of the trials58
in the UK was a multicentre RCT. In terms ofsetting, three studies were conducted in university-based research clinics,53–55 one59 wasbased in an addiction treatment centre and fivestudies took place in outpatient psychiatric units,either a hospital or a psychiatric centre.51,52,56–58
Patients in one study51 were partially hospitalised.
Non-RCT The Wilberg study60 took place inNorway and was carried out in a day psychiatricunit of a university hospital.
Follow-upRCTs The length of follow-up ranged from12 months57,58 up to 36 months51 (Table 5). Three studies52,57,58 did not follow up the patientsafter the active treatment period. Although the number of patients lost to follow-up wasreported in all trials, the reasons were not always given.55–57 Where reported, loss to follow-up was mainly caused by patients droppingout before, during or after treatment in bothgroups. Other reasons included death,54,48
distance and problems with transportation,52
refusal of treatment allocation52 or assessment.58
The non-acceptance of the treatment and/ordropout rate was relatively high, with an averageof one in three patients not completing theirtreatment.
Health Technology Assessment 2006; Vol. 10: No. 35
13
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TABLE 4 Therapist time
Study Psychological therapy Control therapy
RCTsBateman and Fonagy, 199951 Individual group session 1 hour per week NR
Thrice-weekly group 1 hour per weekOnce-a-week expressive therapy 1 hour per weekWeekly community meeting 1 hour per week Meeting with case administrator 1 hour per month
Koons et al., 200152 1.5 hours per week 1 hour individual sessionper week
Linehan et al., 199153 Individual session 1 hour per week Nine patients received Group session 2.5 hours per week individual psychotherapy
for an average of34.87 hours
Linehan et al., 199954 Individual session 1 hour per weekGroup session 2.5 hours per weekplus 15-minute wind-down NR
Linehan et al., 200255 Individual session 40–90 minutes per week Individual CVT+12S Group session 150 minutes per week 40–90 minutes per week
‘12-and-12’ NarcoticsAnonymous group120 minutes per week
Munroe-Blum and Marziali, 199556 30 sessions Individual session once or 25 weekly sessions for 1.5 hours twice per week. Total 5 biweekly sessions for 1.5 hours 210 hours for 30 sessions
Turner et al., 200057 NR NR
Tyrer et al., 200358 NR NR
van den Bosch et al., 200259 2–2.5 hours per week NR
Non-RCTWilberg et al., 199860 Once a week for 1.5 hours over a period of NR
1–33 months (mean 12 months)
Non-RCT Wilberg and colleagues60 reported thatthe follow-up information was available for 92% ofthe participants, of whom two were dead at follow-up, and the final results are based on 88% of theparticipants.
Inclusion and exclusion criteriaRCTs All studies clearly stated the inclusion andexclusion criteria using standardised criteria orscales for BPD. Exclusion criteria were similar inall trials and included schizophrenia, bipolardisorder, and organic mental or physicalimpairment. A primary diagnosis of alcohol anddrug addiction was also an exclusion criterion insome studies.53,56,58
Non-RCT Wilberg and colleagues60 used DSM-III/DSM-III-R criteria for BPD as inclusioncriteria.
Patient characteristicsPatient characteristics are described in Appendix 3(Tables 24 and 25).
Diagnosis of disorderAll studies reported the methods used to diagnoseBPD. The most common criteria were:
● DSM-III/DSM-III-R ● DSM-IV● Structured Clinical Interview for DSM (SCID)-I ● SCID-II● Personality Disorder Exam (PDE)● Diagnostic Interview for Borderlines (DIB)● ICD-10
● European Version of the Addiction SeverityIndex (EuroASI)
● Personality Diagnostic Questionnaire (PDQ-R).
Age, gender, ethnicity, background and patienthistoryRCTs Five studies were done only onwomen.52–55,59 The other studies also includedconsiderably more women than men. The age ofparticipants ranged between 16 and 70 years;however, the mean age was not always reported.Six trials52,54,55,57–59 reported the ethnicity,although some of them reported only thenationality of the participants. At least someinformation on education and socio-economicbackground, including level of education,employment, income, marital status and livingconditions, was reported in all but one trial.54
All studies reported explicitly the history ofpatients. The common elements in patient historyincluded history of parasuicide, psychiatrichospitalisation and alcohol or drug abuse. Twotrials51,52 reported history of childhood sexual andphysical abuse. Three studies52,54,59 reported thatpatients had a lifetime history of parasuicide. Fourtrials52,53,56,59 reported that patients had a historyof therapy or medication for BPD or otherpsychiatric problems.
Overall, the participants were mainly white,unemployed and single, and had graduated fromhigh school or college. There were, however, somedistinct differences between studies. For example,Linehan and colleagues’ 199954 study included
Effectiveness
14
TABLE 5 Study follow-up
Study Active treatment period Post-treatment follow-up Total follow-up(months) (months) (months)
RCTsBateman and Fonagy, 199951 18 18 36
Koons et al., 200152 6 No 6
Linehan et al., 199153 12 12 24
Linehan et al., 199954 12 4 16
Linehan et al., 200255 12 4 16
Munroe-Blum and Marziali, 199556 12 12 24
Turner, 200057 12 No 12
Tyrer et al., 200358 12 No 12
van den Bosch et al., 200259 12 6 18
Non-RCTWilberg et al., 199860 12 34 46
only women, who were relatively young andwealthy, whereas Bateman and Fonagy’s51
participants included men and women, who wereolder and mainly unemployed.
Non-RCT The information presented in theWilberg paper60 was unclear. However, it providedsome information on participants’ education,marital and occupational background, and data ontheir clinical history.
Baseline comparabilityRCTs All studies reported that there were nostatistically significant differences for importantvariables between two groups51,52,55–59 or thatparticipants were matched on age, severity of thesymptoms and number of psychiatrichospitalisations.53,54
Non-RCT Wilberg and colleagues60 reported thatthe control group were significantly younger thanthe treatment group and had spent less time inthe day hospital. Participants in the treatmentgroup had also been married less often than thosein the control group.
Outcomes and resultsTreatment outcome measures and instruments arepresented in Appendix 3 (Tables 26 and 27).
Outcomes to be reported in this review are:
● clinical effectiveness in terms of improvement inpsychological symptoms (parasuicide, suicidalideation, mood and emotional dysregulation)
● effectiveness in terms of interpersonal andsocial functioning (impulsive behaviour)
● effectiveness in terms of preference, satisfactionand acceptability of treatment
● effectiveness in terms of quality of life● cost (see Chapter 4).
Instruments and measurement periodsRCTs Outcomes were measured by a variety ofvalidated instruments for BPD, depression andanxiety, suicide ideation, treatment history andsocial functioning, at baseline, during and at theend of treatment. Most studies also reported theresults of post-treatment follow-up. The list of theinstruments is presented in Table 6.
Non-RCTs Outcomes were measured by a varietyof validated instruments. The list of instruments ispresented in Table 6.
Descriptions of the most common scales are asfollows.
● The Borderline Syndrome Index is a 52-itemforced choice questionnaire that measuresborderline psychopathology associated withborderline states and borderline personalityorganisation.
● The Parasuicide History Interview (PHI) is a semi-structured interview that is used to collectdetails regarding the time, circumstances,motivations and treatment of each parasuicideevent that a participant can recollect.
● The Beck Depression Inventory (BDI) is a 21-itemself-report scale used to determine depressionseverity. Items are scored on a 0–3 scale giving atotal range of 0–63. Total scores within the 1–9range indicate minimal depression, 10–18 milddepression, 19–29 moderate and 30–63 severedepression.
● The Beck Anxiety Inventory (BAI) is also a 21-itemself-report scale. Patients rate symptoms from 0to 3 according to severity. A score of 0–9 reflectsnormal levels of anxiety, 10–18 mild tomoderate anxiety, 19–29 moderate to severeanxiety and 30–63 severe anxiety.
● The Beck Hopelessness Scale (BHS) is a 20-itemtrue/false test that examines three aspects ofhopelessness: feelings about the future, loss ofmotivation and expectations. It is designed foruse with people aged from 17 to 80 years, andtakes 5–10 minutes to administer.
● The Beck Scale for Suicide Ideation is a 21-itemscale that assesses any potential suicidal intentand the severity of suicidal ideation.
● The Borderline Personality Disorder Severity Index(BPDSI) is a semi-structured interview assessingthe frequency and severity of manifestations ofBPD during a circumscribed period.
● The Hamilton Rating Scale for Depression (HAM-D, HRSD) is designed to be used on patientsalready diagnosed as suffering from an affectivedisorder of depressive type. There are 17variables measured on either a five-point or athree-point rating scale.
● The Hospital Anxiety and Depression Scale (HADS)is a self-assessment instrument for measuringdepression and anxiety independently. It wasdeveloped for use with physically ill patients. Itis limited to 14 items and scored on a four-point scale from 0 to 3.
● The Social Adjustment Scale (SAS) is a 54-itemself-report measure, using a five-point Likertscale. It is a comprehensive scale available forassessing detailed role performance within thefamily, as a parent and at work. Six major areasof functioning are covered: work (paid orunpaid), social and leisure activities,relationships with extended family, role asmarital partner, parental role and role within
Health Technology Assessment 2006; Vol. 10: No. 35
15
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Effectiveness
16
TABLE 6 Instruments and scales used as outcome measures
Instrument Abbreviation Studies
RCTsBeck Anxiety Inventory BAI Turner, 200057
Beck Depression Inventory BDI Linehan et al., 1991;53 Koons et al., 2001;52
Bateman and Fonagy, 1999;51 Turner, 2000;57
Munroe-Blum and Marziali, 199556
Beck Hopelessness Scale Linehan et al., 1991;53 Koons et al., 200152
Beck Scale for Suicide Ideation Linehan et al., 1991;53 Koons et al., 2001;52
Turner, 200057
Borderline Personality Disorder Severity Index BPDSI van den Bosch et al., 200259
Brief Psychiatric Rating Scale BPRS Turner, 200057
Brief Symptom Inventory BSI Linehan et al., 200255
Client Service Receipt Inventory CSRI Tyrer et al., 200358
Dissociative Experiences Scale DES
Global Adjustment Scale GAS Linehan et al., 1999;54 Linehan et al., 200255
Global Assessment of Functioning GAF Linehan et al., 2002;55 Tyrer et al., 200358
Global Social Adjustment GSA Linehan et al., 1999;54 Linehan et al., 200255
Hamilton Anxiety Rating Scale HARS Koons et al., 200152
Hamilton Depression Rating Scale HAM-D Koons et al., 2001;52 Turner, 200057
Hopkins Symptom Checklist HSC-90 Munroe-Blum and Marziali, 199556
Hospital Anxiety and Depression Scale HADS Tyrer et al., 200358
Inventory of Interpersonal Problems Bateman and Fonagy, 199951
Lifetime Parasuicide Count LPC van den Bosch et al., 200259
Longitudinal Interview Follow-Up Evaluation Linehan et al., 1999;54 Linehan et al., 200255
Base Schedule
Objective Behaviours Index OBI Munroe-Blum and Marziali, 199556
Parasuicide History Interview PHI Linehan et al., 1991;53 Linehan et al., 1999;54
Linehan et al., 2002;55 Koons et al., 2001;52
Tyrer et al., 200358
Generic Health Related Qualities of Life EQ5D Tyrer et al., 200358
(EuroQol 5 Dimensions)
Social Adjustment Scale SAS Linehan et al., 1999;54 Linehan et al., 2002;55
Bateman and Fonagy, 1999;51 Munroe-Blum andMarziali, 199556
Social Functioning Questionnaire SFQ Tyrer et al., 200358
Social History Interview SHI Linehan et al., 1999;54 Linehan et al., 200254
Spielberg State–Trait Anger Expression Inventory Linehan et al., 1999;54 Koons et al., 200152
Spielberg State–Trait Anxiety Inventory Bateman and Fonagy, 199951
Structured Clinical Interviews Linehan et al., 199954
Suicide and Self-Harm Inventory Bateman and Fonagy, 199951
Survival and Coping Scale Linehan, 199153
Symptom Checklist SCL-90-R Bateman and Fonagy, 199951
Target Behaviour Ratings Turner, 200057
The Reasons for Living Inventory Linehan et al., 199153
Timeline Follow-Back TLFB Linehan et al., 200255
Treatment History Interview THI Linehan et al., 199954
Non-RCTHealth and Sickness Rating Scale HSRS Wilberg et al., 199860
Global Symptom Index GSI Wilberg et al., 199860
family unit (including perceptions abouteconomic functioning).
● The Global Assessment of Functioning (GAF) is a100-point tool rating overall psychological,social and occupational functioning of peopleaged 18 years and older. It excludes physicaland environmental impairment.
● The Global Adjustment Scale (GAS) is a 100-pointscale for measuring the overall level ofimpairment.
● The Global Social Adjustment (GSA) is a five-pointscale that is more specifically related to socialfunctioning.
● The Global Symptom Index is measured using theSCL-90-R.
● The Symptom Checklist 90–Revised (SCL-90-R) is a90-item self-report inventory assessing currentlevels of mental symptoms patterns. Each itemis a description of a mental symptom rated on afive-point scale, and rates the degree of‘distress/discomfort’ during the week before itsadministration.
● The Ad hoc semi-structured interview for DSM-IIIcriteria for BPD is a semi-structured interviewbased on a 27-point scale made up of all theitems in each criterion category for BPD inDSM-III.
● The Health and Sickness Rating Scale (HSRS)includes a structured diagnostic interview,SCID-I and SCID-II, assessment ofemployment, social contact, suicide attemptsand treatment.
Results for behaviour (self-harm, alcohol and drug abuse), affect scales (BDI), therapy maintenance and hospitalisationoutcomesThe results for improvement in psychologicalsymptoms and interpersonal and socialfunctioning outcomes are presented inAppendix 3, (Tables 28 and 29). The results for theincluded studies are described below bycomparator.
RCTs Of the included nine RCTs seven were ofCBT: five of them were DBT,52–55,59 one was DBTorientated57 and one was MACT.58 Thecomparators were TAU,52–54,58,59 CVT+12S,55 andCCT.57 One study51 compared psychoanalyticallyorientated partial hospitalisation therapy withindividual therapy and one study56 comparedpsychodynamic structured IGP with individualtherapy (Table 2).
Cognitive behaviour therapy DBT versus TAU Of the four studies comparing DBT with TAU, three studies52–54
reported significantly greater improvement ofborderline symptoms such as parasuicide and/orsuicide attempts and drug abuse. However, onestudy59 did not find significant differences between DBT and TAU groups. This may haveoccurred because three studies52–54 were targetedto specific populations such as chronicparasuicidal or drug-abusing women, whereas thefourth study59 was a mixed group of parasuicidaland non-parasuicidal women with severe or lesssevere disorder. The study by van den Bosch andcolleagues59 was larger and conducted in Europe,in comparison with other smaller studies from theUSA. Thus, a different setting59 and a largersample size may have influenced the lessfavourable outcome. All trials reported thatmaintenance in the DBT group was greater thanin the TAU group. Koons and colleagues52
analysed the data of 20 (ten in each group) out of the original 28 randomised participants.Linehan and colleagues53 looked at the data on 44 (22 in each group) out of 63 randomisedpatients. In the second Linehan study,54 of 28 randomised participants seven in the DBTgroup and five in the TAU group were lost tofollow-up. Van den Bosch and colleagues59
reported that they lost almost 50% of the studyparticipants during the treatment period (in theDBT group ten out of 27 and in the TAU group24 out of 34). Three studies52–54 reported thatDBT patients received more therapy hours perweek. One study53 reported that control subjectshad significantly more psychiatric days per personhospitalised than patients who were receivingDBT. Two studies52,54 found no difference between DBT and TAU groups in terms ofhospital admission and one study59 did not report data on therapy contact and hospitaladmission. The details of these studies arepresented below.
Study: Linehan and colleagues (1991)53
Sample size: randomised n = 63; analysed ascompleters n = 44 (22 in each group); follow-upparasuicide assessment n = 39 (DBT n = 19,TAU n =20); all other follow-up assessmentsn = 20 (DBT n = 9, TAU n = 11).
Efficacy: the Linehan study53 looked at theparasuicide rate of chronically parasuicidalwomen, and found that the likelihood of anyparasuicide (DBT 63.6%, TAU 95.5%; p < 0.005)and medical risk scores (DBT mean 9.21, SD 8.22, n = 14; TAU mean 17.86, SD 20.94, n = 21; t = 1.70, df = 28.01, p < 0.05) were significantly higher for the TAU group. In thefollow-up year the suicide repeat rate (p < 0.01)
Health Technology Assessment 2006; Vol. 10: No. 35
17
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
and the likelihood of psychiatric hospitalisation(p < 0.07) were lower in the participantscompleting DBT.
Maintenance in therapy: 83.3% and 42% of patientscompleted the entire treatment year in the DBTand TAU groups, respectively.
Therapy contact: DBT patients had more group(z = 5.51, p < 0.001) and individual (z = 2.00,p < 0.01) therapy hours per week, and the controlgroup reported more day treatment hours perweek (z=1.83, p < 0.05). No significantrelationship was found between the number ofindividual and group therapy hours andparasuicidal behaviour, independent of treatment.
Hospital admission: TAU participants tended tohave more hospital admissions per person (DBT median 0, interquartile range (IQR) 1; TAU median 1, IQR 4, z = 1.47, p < 0.07).Control subjects also had significantly morepsychiatric days per person hospitalised thanpatients who were receiving DBT. It is unclearwhether these hospital admissions were voluntaryor involuntary.
Study: Linehan and colleagues (1999)54
Sample size: n = 28 (DBT n = 12; TAU n = 16).
Efficacy: the second study by Linehan andcolleagues54 was conducted on borderline drug-dependent women and their primary outcome wasreduction of drug abuse. The proportion of drugabstinence was significantly higher for DBTparticipants during the 4–8-month period andduring the 12–16-month period (DBT mean 0.94,SD 0.17; TAU mean 0.58, SD 0.36, F = 4.04,p < 0.05). However, there were no between groupdifferences on other psychopathology outcomemeasures (e.g. parasuicide episodes, GSA, GAS oranger). At the 16-month follow-up DBTparticipants showed better social and globaladjustment, with significantly lower (better) scoreson the GSA (DBT mean 2.25, SD 0.75; TAU mean2.92, SD 0.71, F1,12 = 3.98, p < 0.05 for bestscores) and higher scores on the GAS (DBT mean69, SD 12; TAU mean 49, SD 10, F1,12 =22.24,p < 0.001 for best scores).
Maintenance in therapy: 64% in the DBT group and27% in the TAU group remained in treatment(Fisher’s exact test, p = 0.1). In DBT, a subject wasconsidered a dropout if 4 consecutive weeks ofscheduled individual sessions were missed for anyreason. In TAU, a subject was considered adropout from therapy if the participant either
never went to therapy or dropped out of therapyany time following a first session.
Therapy contact: DBT participants receivedsignificantly more psychological therapy (thehours are not reported) than did TAU participants(DBT mean 43.14, SD 10.67; TAU mean 21.88,SD 32.32; F1,15 = 2.07, p < 0.05). TAU wasanalysed by summing hours of psychotherapy andsessions spent with a case manager that wereprovided to TAU participants. This total was thencompared with DBT individual psychotherapysessions.
Hospital admission: no between-group differenceswere found in types and number of medical andinpatient psychiatric treatments received.
Study: Koons and colleagues (2001)52
Sample size: randomised n = 28 (DBT n = 13, TAUn = 15); analysed as completers n = 20 (10 ineach group).
Efficacy: Koons and colleagues52 found that theproportion of patients who reported anyintentional self-harm (including suicide attempt)during the previous 3 months dropped from 50%at pretreatment to 10% post-treatment in DBT,and from 30% to 20% in TAU (p = 0.07). Thedifferences between groups at pre-treatment arenot reported. Participants in DBT changedsignificantly more than did patients in TAU withregard to suicidal ideation (p = 0.008),hopelessness (p = 0.004), Beck Depression(p = 0.012) and Spielberg Anger Expression Scaleanger out (p = 0.005).
Maintenance in therapy: 20 out of 28 patientscompleted the treatment. Three patients in theTAU group dropped out either before treatmentor after the first appointment. Three participantsin the DBT group and two in the TAU group werelost to follow-up during the treatment period.
Therapy contact: DBT patients received more hoursof group therapy than did TAU patients (DBTmean 32.1, SD 9.6; TAU mean 11.8, SD 11.2; t18 = 4.35, p < 0.001), while the TAU participantsattended more hours of 30-minute medication-management visits (DBT mean 2.7, SD 2.2; TAU mean 7.6, SD 4.2; t18 = 3.27, p < 0.01).There were no differences in individual therapyhours.
Hospital admission: the proportion of patients withany admission during the prior 3 months wasrelatively low at pretreatment. Neither group
Effectiveness
18
showed a significant change in this proportion bythe end of treatment (DBT group had 30% ofpretreatment hospital admissions and 10% post-treatment admissions; TAU group had 20% ofpretreatment admissions, which reduced to 10%after the treatment).
Study: van den Bosch and colleagues (2002)59
Sample size: randomised n = 58 (DBT n = 27, TAUn = 31); analysed as completers and followed up:n = 34 (DBT n = 10, TAU n = 24).
Efficacy: van den Bosch and colleagues59 reportedthat the frequency and course of suicidalbehaviours were not significantly different acrosstreatment conditions: neither treatment condition(p = 0.866) nor the interaction between time andtreatment condition (p = 0.639) reached statisticalsignificance. Fewer patients in DBT (7%) than inthe control group (26%) attempted suicide. Thisdifference was not statistically significant(p = 0.064) A significant effect was observed forthe interaction term-time × treatment condition(p = 0.003), but not for treatment condition alone(p = 0.055). In terms of self-mutilating impulsivebehaviour, participants in the DBT groupsignificantly improved over time (interaction termtime × treatment condition), but not for treatmentcondition alone (p = 0.315).
Maintenance in therapy: significantly more patientsassigned to DBT (63%) were retained in therapythan patients in the control group (23%) for theentire treatment year (p = 0.002).
Therapy contact: not reported.
Hospital admission: not reported.
DBT versus CVT+12SOne study compared DBT with CVT+12S.
Study: Linehan and colleagues (2002)55
Sample size: n = 23 (DBT n = 11, CVT+12S n = 12).
Efficacy: Linehan and colleagues55 found thatparticipants in both treatment groups showedsignificant improvements. BSI (pretreatment mean1.78, SD 71; post-treatment mean 1.17, SD 0.60;z = 3.17, p < 0.002) and GAS (pretreatment mean37.6, SD 5.6; post-treatment mean 47.4, SD 10.7;z = 3.59, p < 0.001) scores were statisticallysignificant in both groups and maintained at12 months. At the 16-month follow-up point, BSIscores continued to improve but were not reliablydifferent from the 12-month point (mean 0.98, SD0.74, z = 1.76, p < 0.08) in both treatments. No
difference appeared between treatment groups onGSA rating. The parasuicidal behaviour duringthe treatment year was low (17.4% of patients), butdid not significantly differ by treatment.
Maintenance in therapy: there were three dropouts(36%). in the DBT group. There were no dropoutsin the CVT+12S group.
Therapy contact: there was no statistical differencein the mean number of individual sessionsreceived across the treatment year betweentreatments.
Hospital admission: the incidence of psychiatric anddrug-related visits to emergency rooms andinpatient units was low over the year, but notsignificantly different between groups.
DBT-orientated therapy versus CCTOne study compared DBT-orientated therapy withCCT.
Study: Turner (2000)57
Sample size: n = 24 (DBT-oriented therapy n = 12,TAU n = 12).
Efficacy: Turner57 found that suicide/self-harmbehaviour (rate of parasuicide, BSI, number ofsuicide and self-harm attempts) significantlyimproved in patients for both treatments(F6,84 = 26.8, p = 0.001, R2 = 0.657, repeatedmeasures multivariate analysis of variance).However, the DBT-orientated therapy patients’gains were greater than those receiving CCT atboth 6 months and 12 months (F6,84 = 5.1,p = 0.001, R2 = 0.268). In addition, the rating ofparasuicide [95% confidence interval (CI) 0.559 to2.83], the BSI (95% CI 3.1, 11.3) and the numberof suicide/self-harm attempts (95% CI 0.24 to 5.7)favoured DBT at 6 months and 12 months. Bothtreatments also improved patients’ emotionalfunctioning, but patients receiving DBT-orientatedtherapy had significantly lower scores than thosereceiving CCT on impulsiveness, anger anddepression at 12 months. Global mental healthfunctioning was also statistically significantlyimproved for both treatments (p = 0.005). Therewere no significant differences between two groupson anxiety.
Maintenance in therapy: 15 out of 24 patients were still in treatment at 12 months. Four DBTand six CCT participants withdrew fromtreatment. Of these patients one participant in the DBT group returned to DBT treatment after a5-week break.
Health Technology Assessment 2006; Vol. 10: No. 35
19
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Therapy contact: there were no significantdifferences between groups regarding the averagenumber of treatment sessions.
Hospital admission: number of days of psychiatrichospitalisation was not significant (p = 0.08).
MACT versus TAUOne study compared MACT with TAU.
Study: Tyrer and colleagues (2003)58
Sample size: randomised n = 480 (MACT n = 239,TAU n = 241). Participants with personalitydisorder n = 391, participants with BPD n = 67.
Efficacy: the Tyrer study,58 which was a large trial,found no significant improvement between thegroups in terms of parasuicide events. There wereno statistically significant treatment effects ofMACT according to centre (p = 0.48), baselineparasuicide risk score (p = 0.64) or personalitystatus (p = 0.66). The trial included participantswith different types of personality disorder and the severity and type of disorder influencedthe repetition of the self-harm. Many patients had more than one personality disorder; however, the authors managed to examine eachdisorder separately. Of the 67 borderline patients, 44.8% made no parasuicide attemptduring the follow-up. Of the 55.2% who had made at least one parasuicide attempt, the 25th percentile time to parasuicide event (in days) was 89 and the frequency of self-harm(rate per year) was 3.15. It was evident thatparticipants with personality disorder, especiallyBPD, had a greater incidence of repetition of self-harm, compared with participants without a personality disorder; however, therewere no differences between treatment conditions.
Maintenance in therapy: 40% of patients in MACTwere lost to follow-up during the treatment period.The results of the TAU group are not reported.
Therapy contact: in many cases the amount oftherapeutic time given in TAU exceeded that ofMACT considerably.
Hospital admission: there was no difference in theproportion of self-harm in the year afterrandomisation, between MACT and TAU.
Mentalisation-based partial hospitalisationMBT versus TAU One study compared MBT withTAU.
Study: Bateman and Fonagy (1999)51
Sample size: randomised n = 44 (n = 22 in eachgroup); analysed n = 38 (n = 19 in each group).
Efficacy: Bateman and Fonagy51 found a highly significant reduction in self-mutilatingbehaviour in the MBT group (Kendall’s W = 0.21,�2 =11.9, df = 3, p < 0.008) compared with theTAU group (Kendall’s W = 0.05, �2 = 2.4, df = 3,p = ns) and the number of participants who wereno longer parasuicidal was significantly greater by18 months in the MBT group than in the controlgroup (�2 = 7.0, df = 1, p < 0.08). Anxiety scoresdecreased significantly in the MBT group(p < 0.005) while remaining unchanged in theTAU group. Beck depression scores alsosignificantly decreased in the MBT group(p < 0.0001). The SAS score was significantlylower in the MBT group (mean = 2.8) than in theTAU group (mean = 3.3, p < 0.006).
Maintenance in therapy: there was a 12% dropoutrate in the MBT group. Three patients in the TAUgroup crossed over to the MBT group and werenot included in the analyses.
Therapy contact: patients in the control groupreceived considerably more staff time duringfollow-up than did patients in the MBT arm.
Hospital admission: no patient who completed theMBT programme was admitted to hospital within6 months after discharge. Within 1 year after theend of the trial one patient from the MBT groupand 14 patients from the TAU group had beenadmitted to hospital.
Psychodynamic therapyIGP versus individual therapy One study comparedIGP with individual therapy.
Study: Munroe-Blum and Marziali (1995)56
Sample size: randomised n = 79 (IGP n = 38,individual psychotherapy n = 41); analysed n = 48 (IGP n = 22, individual psychotherapyn = 26).
Efficacy: Munroe-Blum and Marziali56 reported nostatistically significant differences in outcomesbetween the two treatment groups. However, bothtreatment groups experienced significantimprovements over time on outcomes such asbehaviour, social adjustment, global symptoms anddepression from baseline.
Maintenance in therapy: 31 participants withdrewfrom the study at the point of randomisation.
Effectiveness
20
Therapy contact: patient–therapist contact time wasconsiderably lower in the IGP group (90 hours)than in the individual group (120 hours).
Hospital admission: the use of mental health andsocial services decreased significantly by the end of12 months of follow-up (p < 0.038) for the totalcohort.
Non-RCT Appendix 3 (Table 29) presents theresults of improvements in psychological symptomsand interpersonal and social functioning. The trialreviewed in the current report56 was day hospitaltreatment (TAU) plus postdischarge groupanalytical therapy versus TAU.
Day hospital treatment (TAU) plus postdischarge group analytical therapy versus day hospital treatment (TAU)One non-RCT compared TAU plus postdischargegroup analytical therapy with TAU alone.
Study: Wilberg and colleagues (1998)60
Sample size: n = 43 (day treatment plus subsequentoutpatient group therapy n = 12, day treatmentonly group n = 31).
Efficacy: Wilberg and colleagues60 foundsignificantly (p < 0.05, two-tailed) higher (better)HSRS scores at discharge and follow-up for thetreatment group compared with the control group.They also found a significantly lower (better) GSIscore at follow-up for the treatment groupcompared with control (p < 0.05), although not atdischarge. Eight per cent (n = 1) of participants inthe treatment group attempted suicide comparedwith 18% (n = 5) in the control group (p = ns); 75%(n = 6) of the participants in the treatment groupshowed remission from substance use disordercompared with 41% (n = 7) in the control group.
Maintenance in therapy: two participants died, onefrom suicide and one from natural causes. Fourpatients refused to participate in the follow-up.The results were available from the data of 43patients (88%).
Therapy contact: between-group differences werenot reported.
Hospital admission: one participant (8%) in thetreatment group was rehospitalised, comparedwith 12 participants (43%) in the control group.The difference did not reach significance(p = 0.06, Fisher’s exact test, two-tailed).
Patient preference, satisfaction and acceptabilityInformation on patient preference, satisfactionand acceptability of treatment is presented inAppendix 3 (Tables 30 and 31).
RCTs One trial of DBT53 reported that treatmentsuccess did not improve general satisfaction,despite significant improvements in angerreduction and social adjustment.
One DBT study59 reported that all participantswho continued in therapy viewed the programmeas helpful and judged the treatment as veryimportant.
Non-RCT These aspects were not reported.
Quality of lifeOnly one study58 comparing MACT with TAUreported utilisation of a specific quality of life scale (EQ-5D). The EQ-5D scores were morefavourable at 6 and 12 months than at baseline(for both groups); however, they showed nodifferences between MACT and TAU (Table 7).
Summary of the assessment ofeffectivenessTable 8 presents a brief summary of the clinicaleffectiveness results. Ten studies were included inthis review, nine of which were RCTs and one wasa non-RCT. Active treatments included DBT, MBT, MACT and IGP. Comparators were TAU,CCT, CVT+12S, individual and day hospitaltherapies.
RCTs The results of the included studies aresummarised as follows. Three studies52–54
compared DBT with TAU and one study57
compared psychodynamically modified DBT with CCT. Three showed significant improvementsin the DBT group compared with the TAU and CCT groups. One study59 found improvement in both DBT and TAU groups;however, DBT was not more efficacious than theTAU group and both groups were equallyeffective.
Comparison of DBT with another activetreatment, CVT+12S,55 showed that both groupswere significantly effective; however, there were nodifferences between the two therapies.Comparison of MACT with TAU58 found MACTto be no more effective than TAU. One study51
found MBT to be more effective than TAU, andone study56 found IGP to be no more effectivethan individual therapy.
Non-RCT The study by Wilberg and colleagues60
found day hospital plus postdischarge groupanalytical therapy to be more effective than dayhospital on its own.
Health Technology Assessment 2006; Vol. 10: No. 35
21
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Effectiveness
22
TABLE 7 EuroQol scoresa at 6 and 12 months in MACT and TAU groups (by Tyrer et al., 2003)58
Baseline 6-month scores Difference (MACT – TAU)
n MACT (SD) TAU (SD) n MACT TAU Unadjusted (SE) Adjusted (SE)476 0.5 (0.3) 0.5 (0.3) 390 0.7 0.7 –0.01 (0.03) –0.01 (0.03)
Baseline 12-month scores Difference (MACT-TAU)
n MACT (SD) TAU (SD) n MACT TAU Unadjusted (SE) Adjusted (SE)476 0.5 (0.3) 0.5 (0.3) 400 0.7 0.7 0.00 (0.03) 0.00 (0.03)
a Higher scores represent better states
TABLE 8 Summary of clinical effectiveness
Study Study Study size Intervention Comparisons Evidence of clinical quality (total analysed) effectiveness for BPD
symptoms
RCTsBateman and Fonagy, Moderate to 38 MBT TAU Improvement in both groups, 199951 poor but MBT more effective than
TAU
Koons et al., 200152 Moderate 20 DBT TAU Improvement in both groups,but DBT more effective thanTAU
Linehan et al., 199153 Moderate to 44 DBT TAU Improvement in both groups, poor but DBT more effective than
TAU
Linehan et al., 199954 Moderate to 28 DBT TAU Improvement in both groups, poor but DBT more effective than
TAU
Linehan et al., 200255 Moderate to 23 DBT CVT+12S Improvement in both groups, poor but DBT group maintained
the efficacy longer thanCVT+12S
Munroe-Blum and Moderate to 48 IGP Individual Improvement in both groups, Marziali, 199556 poor psychotherapy but IGP no more effective
than individual psychotherapy
Turner, 200057 Moderate to 24 DBT- CCT Improvement in both groups, poor orientated but DBT more effective than
CCT
Tyrer et al., 200358 Moderate 480 MACT TAU Improvement in both groups, (PD n = 391, but MACT no more BPD n = 67) effective than TAU
van den Bosch et al., Moderate 34 DBT TAU Improvement in both groups, 200259 but DBT no more effective
than TAU
Non-RCTWilberg et al., 199860 Poor 43 Day hospital Day hospital Treatment plus TAU more
treatment (TAU) treatment effective than TAUplus postdischarge (TAU)
group analytical therapy
This chapter is in two parts. The first part is areview of the literature on the cost-
effectiveness of psychological therapies for BPD.The second part presents original cost-effectiveness analyses of the psychologicaltherapies based on data collected in the RCTsdescribed in the last chapter. The results arepresented in terms of cost per unit of effect, wherethe unit of effect is determined by the outcomesmeasured in the trials. The results are presentedas a series of incremental cost-effectivenessanalyses and the uncertainty is summarisedgraphically using cost-effectiveness acceptabilitycurves (CEACs).
Systematic review of existingeconomic literatureSearchStudies were identified using the search methodsreported in the section ‘Identification of studies’(p. 7) and Appendix 2.
Economic evaluations assessing the cost-effectiveness of any psychological therapy for BPDwere selected for inclusion. Studies were includedif both costs and benefits were reported and eithera cost-effectiveness ratio was reported or sufficientdetail was reported to enable a cost-effectivenessratio to be calculated. Studies where BPD subjectswere only partially represented were excludedunless a subgroup analysis was performed on theBPD subjects. The references retrieved by theeconomic searches (n = 1748) and the referencestagged as being RCTs from the clinicaleffectiveness searches (n = 1216) were assessed forinclusion in the review of cost-effectiveness. Aninitial title sift was carried out by one reviewer and2458 references were excluded. Two reviewers readabstracts of the remaining 506 references.Although some studies did contain relevantinformation, only one met the inclusion criteria asbeing suitable for review.68
One other study has been reviewed. The study byByford and colleagues69 was an economicevaluation alongside a clinical trial of MACT thatcollected costs and outcome data, but it wasconcerned with self-harming patients. Although
BPD patients were a subgroup in this study, fulleconomic evaluation was not undertaken for thissubgroup and so it should have been excludedfrom this review. However, an exception has beenmade for several reasons. First, it was the onlycost-effectiveness study identified that waspublished in a peer-reviewed journal. Secondly,the trial data from this study form an integral partof the cost-effectiveness analyses presented in thischapter. The study is therefore considered ofsufficient interest to be included in the review.
ReviewThe quality of the studies identified by the searchwas assessed using the British Medical Journalchecklist50 for economic evaluations (Appendix 6).
Cost-effectiveness studiesHeard (2000). Cost-effectiveness of dialecticalbehavior therapy in the treatment of borderlinepersonality disorder68
This dissertation by Heard presents the cost-effectiveness results of two psychotherapy outcometrials involving DBT. The first study is included inthe effectiveness review;53 however, the secondstudy is unpublished. In the first study subjectswere randomly assigned to receive 1 year of DBT(n = 22) or to receive TAU in the community(n = 22). Analyses suggested that at the end of1 year of treatment, subjects receiving DBT hadincurred significantly higher psychotherapy costs,lower psychiatric inpatient costs and loweremergency room costs compared with TAU. Thetwo groups did not differ significantly with respectto mean medical costs (DBT US$1161, 95% CI$589 to 1733; TAU $1799, 95% CI $710 to 2888)or total healthcare costs (DBT $9856, 95% CI$7292 to 12,420 (1988 prices); TAU $19,745, 95%CI $11,144 to 28,345). Analyses also revealedimportant differences in variance between the twoconditions, with the TAU arm having significantlygreater variance in terms of psychotherapy,inpatient, emergency room and total healthcarecosts. The two arms also did not differ in terms ofemployment or global functioning cost-effectiveness ratios.
In the second study the DBT subjects from thefirst study were compared with subjects whoseexisting therapists in the community had agreed to
Health Technology Assessment 2006; Vol. 10: No. 35
23
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Chapter 4
Cost-effectiveness
provide them with 1 year of stable psychotherapy(n = 16). At the end of 1 year of treatment,subjects receiving DBT had incurred significantlylower psychiatric inpatient costs and had a trendtowards better global functioning cost-effectivenessratios. The mean costs and cost-effectiveness ratiosdid not differ significantly between the two groupson any of the remaining variables. Finally, theTAU again had significantly greater variance interms of psychotherapy, inpatient and emergencyroom costs.
Comment This was a good-quality study thatscored highly on the BMJ checklist for economicevaluations. The cost-effectiveness analysis did notsupport the hypothesis that DBT is a more cost-effective treatment than TAU, though it sufferedfrom small numbers. Another shortcoming of thisstudy is the choice of economic outcome measures.Cost-effectiveness measured in terms of cost pernumber of weeks worked and cost per one-pointimprovement in global functioning is notcomparable to other disease areas and is of littleuse to decision-makers. A limitation of the study isthe lack of societal cost data and the cost ofconsultation that therapists receive for workingwith these patients. Societal costs are relevant forthis population and team consultation is stronglyemphasised in DBT. Excluding these costs mayhave given the DBT group an unfair advantage.
Byford and colleagues (2003). Cost-effectivenessof brief cognitive behaviour therapy versustreatment as usual in recurrent deliberate self-harm: a decision-making approach69
The Byford study is a cost-effectiveness analysis ofdata from an RCT [Prevention of Parasuicide byManual-Assisted Cognitive Behaviour Therapy(POPMACT)] comparing MACT with TAU for thetreatment of people with recurrent episodes ofDSH. The trial was conducted by Tyrer andcolleagues,58 in five centres in Glasgow,Edinburgh, Nottingham, West London and SouthLondon. The economic outcomes were cost per1% reduction in the proportion of patients with aself-harm episode and cost per quality-adjustedlife year (QALY) gained. The study took a broadsocietal perspective and included costs of hospital,community, voluntary and social services,community accommodation, the criminal justicesystem and productivity losses due to time takenoff work. The time horizon was the length of thetrial (12 months) and costs and benefits (self-harmepisodes and EQ-5D) were therefore notdiscounted. An interim cost analysis at 6 monthswas also conducted. Unit costs for hospital serviceswere based on local costs; all other unit costs were
based on national sources. The analysis wasperformed on an ITT basis. Mean costs in the twogroups were compared using standard t-tests withordinary least squares regression for adjustedanalysis. Bootstrapping techniques were used toconfirm the validity of the results.
Univariate sensitivity analysis was conducted byreplacing local unit costs with national ones,adjusting productivity costs, excluding communityaccommodation and including a mean cost forcourt appearances. Probabilistic sensitivity analysis(PSA) was conducted (based on repeated samplingusing bootstrapping techniques) and reported bymeans of CEACs.
Mean cost of treatment The overall mean cost perpatient was £13,450 and £14,288 (mean difference–£834, 95% CI –£2142 to 466) for the MACT andTAU groups, respectively. No statisticallysignificant differences between the groups werefound in the total costs or in the individualresource-use categories. No statistically significantcost difference was found between the MACT andTAU groups in any of the univariate sensitivityanalyses.
Cost-effectiveness analysis The cost per 1%reduction in the proportion of patients with arepeat self-harm episode was £120. However, theincremental mean effect as measured by the EQ-5D instrument, was negative for MACT(–0.01118). MACT was therefore cheaper but lesseffective than TAU using EQ-5D, although notsignificantly so. The incremental cost per QALYgained from TAU was £66,000, but this canpossibly be dismissed as a chance finding as theconfidence intervals for the mean difference incost are large.
PSA showed that the probability of cost-effectiveness of MACT compared with TAU usingthe percentage reduction in repeat self-harm wasover 90%, whatever the willingness to pay. UsingQALYs based on the EQ-5D, MACT was morecost-effective than TAU up to a willingness to payof £60,00 per QALY.
Comment This was a good-quality study thatscored highly on the BMJ checklist for economicevaluations. For the primary outcome measure ofcost per 1% reduction in the proportion ofpatients with a self-harm episode, MACT wascheaper and more effective than TAU. However,this unit of measurement is difficult to interpretsince it is a relative measure and so cannot becompared between studies. Furthermore, there are
Cost-effectiveness
24
questions about its appropriateness as the primaryoutcome for patients with BPD. Another limitationis that this outcome cannot be compared withother disease areas and so is not helpful todecision-makers.
The cost per QALY gained of £66,000 of TAU ishigher than is generally accepted by the NationalInstitute for Health and Clinical Excellence(NICE), but tells us little about the potential cost-effectiveness of MACT since it is based on a smalland non-significant difference in EQ-5D scores (–0.0118). This small difference, and one that wasin the opposite direction to the primary outcome,may have been the result of chance or the lack ofsensitivity of the EQ-5D to changes in this patientgroup.
An important limitation of this study, as far as thisreview is concerned, is that it covers far more thanBPD, since it includes people with recurrent self-harm and most of these people were notdiagnosed with BPD. In subgroup analysis of thesame trial data the authors report the cost of carefor patients with personality disorders.63 Inpatients with BPD, the average societal costs ofcare were higher for patients in the MACT group(£16,144) than in the TAU group (£14,185).Outcomes were not reported by treatment forBPD.
Studies assessing the cost of therapiesOnly two studies were identified that investigatedthe cost of therapy specifically for BPD patients.The Bateman study70 estimated the annualhealthcare utilisation costs for BPD patientsreceiving either partial hospitalisation or generalpsychiatric care. The setting was the Halliwick DayUnit at St Ann’s Hospital in the UK. Costs were reported in US dollars and have beenconverted back to pounds sterling using theexchange rate quoted in the publication. The year of unit costs is not stated. The mean (SD) cost of care at 18 months was £19,000 (£11,000) in the partial hospitalisation group and £22,000(£18,000) in the general psychiatric care group.
The Hall study71 compared the cost of 1 year ofpsychotherapy in 30 BPD patients with the cost ofthe previous year in which the same patientsreceived no formal psychotherapy. The mean costof care for 1 year of psychotherapy was AUS$ 7309(1998 unit costs). This includes the cost ofinpatient stay, outpatient visits, ambulatory care,diagnostics, medications and psychotherapytreatment.
Further details of these studies can be found inAppendix 7.
Cost-effectiveness and cost–utilityanalysisThe aim of this section is to assess the cost-effectiveness of different psychological therapies inthe treatment of BPD. This assessment of cost-effectiveness is not based on a conventionaldecision-analytic model owing to the complexnature of BPD and the lack of evidence. Theapproach has been to undertake a cost-effectiveness analysis for each of the nine RCTsreviewed in Chapter 3 using a combination of datareported in published papers, trial data sets sentby the investigators and a cost model using datafrom the POPMACT study. Cost-effectiveness hasbeen assessed in terms of cost per parasuicideevent avoided in six trials and cost per QALY hasalso been undertaken in four of the six studies (bymapping BDI results onto the EQ-5D in the caseof three trials; see Appendix 8).
Overall methodsModellingThe NICE guidelines for economic evaluation72
recommend undertaking a formal decision-analytic modelling approach. This was not felt tobe useful in BPD owing to the lack of evidence fora well-defined treatment pathway. The applicationof models to synthesise evidence was also inhibitedby the inability to conduct any meta-analysis of theclinical studies reviewed. As argued in the sectionon clinical effectiveness, the nine RCTs and onenon-RCT were too different in terms of theirpatient populations and interventions to permit aformal meta-analysis. For these reasons it wasdecided to undertake a series of economicevaluations on those trials that contain sufficientdata, namely six of the nine RCTs reviewed inChapter 3.51–53,57–59 The study-specific approachlimits the generalisability of the results and makescomparison between interventions difficult, but itdoes at least provide some evidence on thepotential cost-effectiveness of the interventions(where currently so little exists).
Technique of economic evaluationThe form of economic evaluation was limited bythe data collected in the clinical studies. For all sixRCTs, cost-effectiveness analyses are undertakenusing parasuicidal events as the unit of effect. Asdescribed later in some detail, this is a verylimited measure of outcome since it does notreflect the overall health-related quality of life of
Health Technology Assessment 2006; Vol. 10: No. 35
25
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
the patient. The ideal approach would be toconduct cost–utility analysis, where the outcomesare expressed in the form of QALYs to permitcomparison with other health serviceinterventions. However, only one trial used apreference-based measure of health (POPMACT)and of the remainder, just three used outcomemeasures that could be mapped onto the EQ-5D.A cost-effectiveness analysis using events avoidedhas been undertaken in all six RCTs and acost–utility analysis in four out of the six.
Economic perspectiveThe NICE perspective is limited to the NHS andPersonal Social Service (PSS). However, the highconsumption by BPD patients of othergovernment services, such as the criminal justicesystem, makes a broader government perspectivemore relevant. The base case was therefore takenfrom a government perspective, with NICE andsocietal perspectives being presented in theunivariate sensitivity analyses.
Time horizonThe length of the follow-ups undertaken in thepublished studies limited the time horizon forthese analyses. Most studies were 12 months inlength and this has been applied in all studies.Studies of different durations have been convertedto 12 months by a simple pro-rata approach.
Outcome measureTo compare cost-effectiveness between studies, acommon outcome measure is required. Theoutcome measure most commonly reported acrossthe studies is the number of parasuicide events.The cost-effectiveness analysis by Byford andcolleagues,69 identified from the literature review,was based on the POPMACT study and used thismeasure. However, cost-effectiveness wasexpressed in terms of cost per 1% reduction in theproportion of patients with a self-harm episode.
There are several problems with this outcomemeasure. One is the use of a 1% reduction. This isproblematic because it is a relative measure andmay not mean the same between studies.Therefore, the absolute number of parasuicideevents was used as the main outcome measure,which provides a measure that is more meaningfulto decision-makers making comparisons acrossstudies. A more fundamental problem is the focuson just one outcome rather than a fuller measureof health-related quality of life, particularly sincethe relationship between the patient’s quality oflife and levels of parasuicide is not necessarily
linear. Parasuicide events are nonethelesscomprehensible to clinicians and decision-makersand although there is no willingness-to-pay figurefor this outcome, decision-makers may be able tocome to some judgement about a reasonablefigure.
The other problem with using parasuicide activityor related outcomes such as attempted self-harm isthat they have been defined in slightly differentways and may not be comparable between studies.We have attempted to standardise this wherepossible.
The QALY is used by NICE to undertakecomparisons across programmes and this wouldhave been a better measure to compare the cost-effectiveness of psychological therapies for BPD.In this appraisal, only the Tyrer study58 was foundto have used a preference-based measure of healththat could be used to generate QALYs. Analternative method is to use another self-reportmeasure of health and map that onto apreference-based measure. Three of the trials usedthe BDI, which has been previously mapped ontothe EQ-5D (Appendix B).73 This mappingfunction was applied to BDI data in three trials togenerate QALYs.
The BDI is typically reported at several trial time-points. To summarise these measurements andalso to take into account the between-groupdifferences at baseline, a mean QALY gain (orloss) has been calculated to be the area under thecurve (AUC).
CostsThe costs implications of the interventions aremore than the cost of the psychotherapy providedto patients in the experimental arm. BPD patientsare heavy users of resources across a large range ofservices69 and a successful therapy is likely to haveknock-on implications for the use of services wellbeyond the specific intervention being evaluatedin the trial. All resource consequences have beencosted. Costs were inflated to 2003/04 prices using the Hospital and community health servicespay and prices index (Office for NationalStatistics).74
Costs of psychological therapySessional costsThe costs of the interventions were estimated fromdescriptions provided in the published papers ofthe trials and other available documentation onthe number and type of sessions and the therapistsproviding the therapy.
Cost-effectiveness
26
The main concern with this approach is that theresources devoted to the interventions in the trials,particularly those undertaken in other countries,may not be typical of what would be provided inthe NHS. The dilemma is that in conducting aneconomic evaluation alongside a clinical trial, it isimportant to retain the internal validity from adirect link between the resources costed and thestudy outcomes since a different level of resourceis likely to have achieved a different outcome.However, the trial resource level may beunrealistic. The solution adopted here has been touse published trial data on the numbers ofsessions (individual and group), but to cost thesessions using UK estimates of the types oftherapist likely to be taking the sessions.
Estimates for the types of therapist conductingboth group and individual sessions are based on asurvey of DBT practitioners by the PsychologicalTherapies Research Centre at the University ofLeeds.75 All practitioners registered on the LeedsDBT Practice Research Network (PRN) databasewere sent a postal survey. The aim of the surveywas to address questions of how DBT wasimplemented, which clients and patients receivedDBT, and how services were being monitored andresearched. The percentage breakdown ofprofessionals delivering DBT was clinicalpsychologists (48%), nurses (26%), psychiatrists(8%) and occupational therapists (8%). Theremaining 10% were psychologists,psychotherapists, forensic psychologists or socialworkers (classed as ‘other therapists’ in Table 9).
Unless stated otherwise in the trial publications, itis assumed that two staff members76 deliver thegroup therapy to an average of seven or eightpatients (Rees A, Psychological Therapies ResearchCentre, University of Leeds: personalcommunication) and that sessions last on averagefor 2.25 hours.76 Individual therapy sessions areassumed to last for 1 hour,76 unless indicatedotherwise in the trial publications.
Staff time has been costed using national averagehourly costs for therapists taken from Curtis andNetten (2003),77 which include the fullemployment costs of the therapists, along withrelated costs such as general training (Table 9).The only exception to this is the cost of a clinicalpsychologist, which in Curtis and Netten is £69per hour of clinical contact. This seems to beunrealistically low and appears to be so because itdoes not include the costs of qualifications. Thecost of qualifications is considerable forprofessionals at this level; for example, the cost ofa consultant psychiatrist includes qualificationcosts of £26,000 per year. In the absence of analternative data source the cost of a psychologistwas based on expert opinion (Dent-Brown K,University of Sheffield: personal communication)and was estimated at half the cost of a consultantpsychiatrist.
TrainingThere is an extra cost associated with trainingpeople in the specific therapies evaluated in theRCTs. The cost of DBT training, for example, isbetween £1350 and £1925 per person dependingon the number of staff attending (British IslesDBT training. www.capricorn.uk.net). To includethis in the cost of DBT would require anassumption as to an annuity period and thenumber of patients treated in this period. The resulting annual mean cost per patient oftraining would be small and have little effect onthe overall results, and has therefore not beenincluded.
Telephone consultationAlthough DBT requires therapists to be availablefor telephone consultations, the only study thatreported telephone contact reported nostatistically significant difference between the DBTand TAU groups. Patients receiving TAU in theUK also have telephone contact with theirtherapist. The contrast is one of differentlystructured methods of delivering telephone
Health Technology Assessment 2006; Vol. 10: No. 35
27
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TABLE 9 Unit costs
Personnel and services National average unit cost Service
Consultant psychiatrist £272 Per 1-hour contactCommunity psychiatric nurse £72 Per 1-hour contactClinical psychologist £136 Per 1-hour contactSocial worker £99 Per 1-hour contactOccupational therapist £44 Per 1-hour contactOther therapists £118 Per 1-hour contactGP visit £85 Per 1-hour contactMental health services (inpatient) adult acute care £190 Per bed-dayA&E £83 Per first attendance
contact, rather than one of cost, and so the cost oftelephone contact has not been included in thisanalysis.
Staff supervisionDBT is a team effort, and an integral and essentialpart of the therapy is support for the therapistthrough regular supervision meetings with theother therapists. This constitutes a major costelement in the overall cost of DBT. Of the fourDBT studies, three52,59,78 report the frequency,number of hours and number of staff involved insupervision meetings, and this information hasbeen used to estimate a cost of supervision perpatient.
Staff delivering TAU may also receive supervision;however, no data were found to enable theresource implications to be quantified. To takeTAU staff supervision into account it was assumedthat TAU supervision is less resource-use intensiveby a factor of 0.5 compared with DBT supervision.This supervision cost has been added to all TAUarms. It was also assumed that there is nodifference in the supervision costs between thegroups in the non-DBT studies and the cost ofTAU supervision.
Other resource consequencesThe full range of additional resourceconsequences of BPD patients is shown in Table 10.
For three trials most of the resource-use data wereavailable from the published articles or theindividual-level data provided by theauthors.51,58,68 Resource-use data were convertedinto current UK prices using methods described inthe section below on study-specific methods. Intwo other studies,57,59 resource-use data were onlyavailable for length of inpatient hospital stay, andfor the remaining study52 no resource-use datawere available. It has been shown that inpatientlength of stay is the largest element of resourcecost79 and so the potential for modelling totalresource-use cost from inpatient hospital stay wasinvestigated. The relationship between thenumber of parasuicide events and resource-usecosts was also investigated to estimate a cost forthe study that did not have inpatient data.52
A key component in the costings was a cost modelrelating length of stay and parasuicide events tocosts. This required UK patient-level trial data andthe only trial data available that collected costs inall of the main resource categories (i.e. hospital,community health, social services, voluntary sector,community accommodation, criminal justice andproductivity), which can be combined to representthe perspectives of the government, NICE andsociety, was the Tyrer study.58
In the Tyrer study,58 no significant differences werefound between the MACT and TAU arms and it
Cost-effectiveness
28
TABLE 10 Resources used by BPD patients
Hospital Community Social services Voluntary Community Criminal Productivityservices health services sector accommodation justice
services system
Psychiatric Community Day centre/ Helpline Staffed Police officer Days off workinpatient psychiatry drop-in centre
psychology
Other Community Specialist Advice Unstaffed Prison/policeinpatient counselling education facility cell
A&E Community mental Sheltered Support Psychiatrichealth nurse/team workshop assessment
in custody
Psychiatric/ Community physical Social worker Counsellingpsychology therapyoutpatient
Other Primary care Home help Group therapyoutpatient
Day hospital Other community Other socialhealth services services
Medication
Source: Byford et al.69
was therefore considered appropriate to combineboth arms for the purposes of this investigation.The study had 480 self-harming patientsrandomised and with a reasonable proportionbeing BPD patients (n = 62). The cost modellingwas conducted on the BPD subgroup of patients.Since the Tyrer data also contain the cost oftherapy there may be an element of double-counting with the therapy costs. However, theaverage number of MACT sessions in the trial was so low (4.3) that double-counting of therapycost is expected to have a minimal impact onoverall cost.
A regression analysis was performed with inpatientlength of stay and parasuicide events as theindependent variables and total government cost asthe dependent variable. Both variables were foundto have significant coefficients. The analysis wasrepeated from the NHS and societal perspectives.The government perspective included all resourcesexcept for the voluntary sector services group,unstaffed community accommodation andproductivity cost. For the NICE perspective,hospital services, community health services, socialservices and the cost of staffed accommodation wereincluded. For the societal perspective all items wereincluded. From Table 11, it can be seen thatinpatient stay and parasuicide events account foraround two-thirds of the variation.
As reported above, the Koons study52 did notprovide data on inpatient stay. Therefore aregression model based on parasuicide eventsalone was used to estimate costs (Table 12). Theexplanatory ability of this model is poor(R2 = 0.1); however, in the absence of better datait gives some indication of the expected relativecosts of the treatment and control groups.
AnalysisThe cost-effectiveness results are presented in termsof incremental cost per parasuicide event avoidedand cost per QALY. PSA is used to investigate theimpact of the uncertainty around parameters. Itpermits an analysis of the effect of joint uncertaintyin all of the variables simultaneously. A distributionis attached to the range associated with each of thevariables in the analysis and Monte Carlosimulation simultaneously selects values from thespecified ranges and distributions. The simulationsare run 10,000 times to generate a distribution ofvalues. The results of this have been showngraphically on a cost-effectiveness plane for eachstudy.
The cost and effectiveness distributions for eachintervention are then combined to form a series ofnet benefit distributions, one for each interventionand at each level of willingness-to-pay. Finally, theprobability that the intervention of interest isoptimal is quantified and plotted for every valueof the willingness-to-pay threshold. PSA was usedto examine the probability that the intervention iscost-effective compared with the control arm atdifferent levels of willingness to pay for avoidingparasuicide events and per QALY. This has beenpresented graphically in the form of CEACs andin terms of the specific likelihood of anintervention being cost-effective compared withthe control arm at an arbitrary cut-off of £5000per suicidal event avoided and a cost per QALY of£20,000.
The PSA does not capture all of the uncertainty inthe results. One source of uncertainty is theperspective of the evaluation, which can includeNICE, government or societal. The impact of thisis explored in a univariate analysis. The other
Health Technology Assessment 2006; Vol. 10: No. 35
29
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TABLE 11 Results of the cost regression model based on parasuicide and length of stay (LOS)
R2 Coefficient SE
Government perspective 0.66(Constant) 3,767 818Parasuicide events 55 109Inpatient psychiatric LOS (days) 244 41
NICE perspective 0.76(Constant) 3,136 627Parasuicide events 18 84Inpatient psychiatric LOS (days) 256 32
Societal perspective 0.63(Constant) 13,414 643Parasuicide events 198 86Inpatient psychiatric LOS (days) 162 33
factors explored in the univariate analyses are therelative size of the supervision costs of TAU andthe costing of DBT.
Methods by studyThe methods used to analyse each study varydepending on the availability and source of dataand are therefore presented separately in thissection.
DBT trialsTurner (DBT versus CCT)57
This study attempted to keep the levels of therapyin DBT and CCT equal in terms of therapycontact hours and as a result no significantdifferences were reported between treatment armsin the number of therapy sessions. However, there was a cost difference between DBT and CCT sessions and so the mean number ofindividual therapy sessions for each arm wasneeded, but this was not reported. The numberwas estimated from the reported minimum (49)and maximum (84) number of therapy sessions.Six group therapy sessions were also provided foreach group.
Supervision costs were estimated from the numberof supervisors (two), the number of therapists(four) and the frequency of meetings (weekly intwo separate sessions). The length of the meetingswas not reported and was assumed to be 1 hour.Supervision costs were added to the CCT group bythe method described above.
Resource-use costs were estimated by applying theregression cost model to the number ofparasuicide events and the inpatient psychiatriclength of stay.
The definition of parasuicide is unclear in thisstudy. The authors state that “patients alsomaintained daily logs of suicide urges andattempts”, but the analysis assumed that it iscomparable to the other studies. The BDI wasreported at baseline, 6 months and 12 months,and these data were converted to the EQ-5Dpreference-based index using the mappingfunction and used to calculate QALYs using the AUC.
Linehan (DBT versus TAU)53
Costs were not provided in the published trialpaper for this study; however, costs are availablefrom a cost-effectiveness dissertation by Heard.68
The cost of therapy was reported separately forindividual therapy, group therapy and daytreatment. Resource-use costs were reported forpsychiatric inpatient stay, emergency room visits,physician visits and medical inpatient days. Thecosts reported in the Heard dissertation arereproduced in Table 13.
Heard reported unit costs alongside mean costsand so the mean hours of therapy and resourceuse could be estimated. These resource-use figureshave been recosted using current UK unit costs.Several assumptions are made in this costing: first, that sessional input will be delivered in theUK at the same level as in the Linehan study;secondly, that the efficacy rates will be similargiven the differences in service delivery in the UK compared with the Linehan study; andthirdly, that the amounts of psychiatric inpatientstay and other medical treatment are the same inthe UK as in the USA. Since the Linehan teampioneered DBT, it is reasonable to expect efficacyrates to be higher for them than for DBTdelivered in the UK. The first two assumptions,therefore, would probably result in a larger
Cost-effectiveness
30
TABLE 12 Cost regression model based on parasuicide events alone
R2 Coefficient SE
Government perspective 0.1(Constant) 4,518 1010Parasuicide events 320 124
NICE perspective 0.1(Constant) 3,924 891Parasuicide events 296 110
Societal perspective 0.22(Constant) 13,913 751Parasuicide events 374 92
difference in efficacy between treatments andcontrol than would be achieved in the UK and soexaggerate the difference. For the thirdassumption, inpatient stay is generally consideredto be used less in the USA than in the UK and theanalysis is likely to underestimate the cost savings.The overall impact of these differences isunknown.
Insufficient data were reported on supervisioncosts and therefore the average of the other threeDBT studies was used.
In the Linehan study,53 parasuicides weremeasured using the PHI. Although the BDI wasmeasured in this study, the values at 12 monthswere not reported and so QALYs could not beestimated.
Van den Bosch (DBT versus TAU)59
The annual mean hours of therapy for individualand group sessions were estimated from thenumber of patients in therapy, the proportioncontinuing therapy and the frequency of sessionsreported in the trial publication. In order toinclude the length of time that dropouts receivedtherapy, it is assumed that, on average, theywithdrew halfway through the trial.
Supervision costs were estimated based on thetypes of therapist delivering treatment and thefrequency of supervision meetings reported in thetrial publication. It was assumed that supervisionsessions last for 1 hour. Supervision costs wereadded to the TAU group by the methodsdescribed above.
No other resources were reported in the study, sothese were estimated by applying the regressioncost model to the number of parasuicide eventsand the inpatient psychiatric stay.
Parasuicide was measured with the LPCinstrument. The number of parasuicide events wascalculated from LPC trial data provided by theauthors. The BDI was not used in this study.
Koons (DBT versus TAU)52
The mean hours of individual therapy, grouptherapy and medication management visits werereported in the trial publication. These weremultiplied by the appropriate unit costs describedin the section ‘Costs’ (p. 26). Supervision costswere estimated for the types of therapistdelivering treatment and the frequency and lengthof supervision meetings reported in the trialpublication.
No other resource-use costs were reported, norwas length of stay. Other resource costs weretherefore estimated using the regression costmodel with number of parasuicide events as theonly independent variable. This is a very crudemodel and its estimates should be viewed withextreme caution.
Parasuicide was measured in this trial using thePHI, which is the same instrument used byLinehan’s group.53 The number of parasuicideattempts was reported at 3 and 6 months. For thepurposes of comparison with the other studies,these event rates were multiplied by a factor of 2to represent 12-month event rates. The BDI was
Health Technology Assessment 2006; Vol. 10: No. 35
31
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TABLE 13 Costs (1998 US$) reported by Heard80
DBT TAU
Treatment mode Unit cost Mean SE Mean SE
Outpatient psychotherapyIndividual hours $88 $3,885 $232 $2,915 $927Group hours $35 $1,514 $128 $147 $7Day treatment hours $77 $10 $10 $876 $167
Psychiatric inpatientDays $309 $2,612 $1,086 $12,079 $3,692
Emergency roomVisits $144 $226 $43 $569 $90
Medical treatmentPhysician visits $91 $783 $198 $650 $160Medical inpatient days $360 $360 $212 $1,096 $542
reported at baseline, 3 months and 6 months. Toestimate a QALY over 12 months, the assumptionwas made that the mean BDI scores remainconstant between months 6 and 12.
MBT trials(Bateman MBT versus TAU)51
This study was conducted over an 18-monthperiod. It was converted to a 12-month period prorata to make it comparable to the other studies.The 18-month study results are presented in thesensitivity analysis.
Therapy and other resource-use costs werereported in a cost analysis performed by theauthor. Other resource-use costs reported in thestudy were psychiatric inpatient and outpatientcare, medication and emergency room treatment.Social and other government costs were notincluded. Although twice-weekly supervision oftherapists was provided, the number of therapistsinvolved was not reported. The level ofsupervision appears comparable to DBT and theaverage supervision cost of DBT was thereforeassumed. Supervision costs were added to the TAU group by the method described above.Costs were reported in US dollars and wereconverted back to UK pounds using the exchangerate reported.
Suicide and self-harm acts were measured usingthe Suicide and Self-Harm Inventory. Thisinstrument is available from the trial authors andno further details were reported.
The number of suicide and self-harm events wascalculated from individual-level trial data suppliedby the authors. The BDI was reported at baselineand 3-month intervals thereafter. A mean QALYgain for both groups was estimated for 12 monthspro rata.
Tyrer (MACT versus TAU)58
The intervention therapy, other resource-use costs,parasuicide events and EQ-5D scores wereavailable from trial data supplied by the authors.69
From this data set the reviewers were able toundertake an analysis of the BPD subgroup. Thecost categories used to calculate costs from thegovernment, NICE and societal perspectives arethe same as those described above in theregression cost modelling. Supervision oftherapists was not described in the trialpublication. It was assumed that supervision forboth groups was similar to TAU. The cost of TAUsupervision (estimated by the method describedabove) was therefore added to both groups.
Parasuicide was measured using the PHI, the sameinstrument used by Linehan’s group.53 The EQ-5Dwas directly used in the study and this was used toestimate QALYs.
Cost-effectiveness resultsThe costs of the interventions and their studycontrol arms are shown in Table 14.
The mean therapy costs for DBT across the fourDBT trials were between £10,372 and £16,903 andin all cases exceeded the therapy costs in thecontrol arm, whether TAU or CCT. There was alower length of psychiatric inpatient stay and therewere fewer parasuicide events than in the controlarms. In two of the studies, the decrease ingovernment costs associated with the reduction inuse of services was sufficient to outweigh the costof the additional therapy53,57 and nearly tooutweigh the cost in another study.59 Only in thefourth study was there still a large incremental costfor the DBT arm.
For MBT and MACT, therapy costs were notrecorded separately. The mean total costs of theMBT patients (£18,174) were within the range ofthe DBT intervention groups, while the MACTpatients’ costs were somewhat lower (£9580). Thecosts of the control patients were roughly the samefor the MBT trial, but lower in the POPMACTstudy (£7563). The large cost difference betweenpatients in the POPMACT study and three of theother studies may be partly due to a degree ofdouble counting, since it was not possible toexclude therapy costs from the cost model (sincethe POPMACT data set did not separate theseout). However, overall those categories of costsmost likely to contain therapy, hospital psychiatricoutpatients and community NHS services,together only account for 23% of total cost.Therefore, double-counting does not account forthe order of difference between the three DBTstudies by Turner,57 Linehan53 and van denBosch59 compared with POPMACT. It also cannotaccount for any of the differences from theLinehan and Bateman studies,51,53 where actualother resource-use data collected in the trial wereused. The reason for the lower costs of MACTpatients is not clear.
Tables 15 and 16 present the key constituents ofthe cost-effectiveness analyses: total numbers ofparasuicidal events, total QALYs and total costsassociated with each arm of the trials, along withthe number of events avoided or QALYs gained
Cost-effectiveness
32
and the incremental costs. The incremental cost-effectiveness results are presented in three waysdepending on the results: where the interventionis both cheaper and more effective than thecontrol it is considered to dominate the controlarm; where the intervention is both moreexpensive and less effective than the control it is considered dominated by the control arm;and where the intervention costs more andis more effective the results are presented as anincremental cost per unit of benefit. The detailed results are now described by therapy and study in the same order as the Methodssection.
DBTTurner57
The incremental cost per patient of DBT overCCT was –£5242 and the incremental benefit was9.4 events avoided. At the mean level, DBT wasmore effective and cheaper than CCT and this is reflected in the plot on the cost-effective plane (Figure 2). The CEAC shows that DBT would have a probability of more than 80% ofbeing cheaper and more effective and theprobability that it would be preferable at athreshold of £5000 per event avoided was around85% (Figure 3).
The incremental QALY gain from DBT was 0.12and the domination of DBT is again reflected inthe cost-effectiveness plane (Figure 4). The CEACsuggests that DBT would have a probability of85% of being cheaper and more effective thanCCT and the probability that it would bepreferable at a threshold of £20,000 was around90% (Figure 5).
Linehan53
The incremental cost per patient of DBT overTAU was –£1207, with 26.7 events avoided.Although at the mean level DBT dominated TAU,there is rather more uncertainty surrounding thisresult than for the Turner study. The scatterplot
on the cost-effectiveness plane is clustered aroundall four quadrants (Figure 6). DBT had aprobability of around 53% of being cheaper andmore effective, and the probability that it would bepreferable at a threshold of £5000 was around60% (Figure 7).
van den Bosch59
The incremental cost per patient of DBT overTAU was £724 and the incremental number ofevents avoided 18.1. This results in an incrementalcost per event avoided of £724. The plot on thecost-effectiveness plane suggests a linearrelationship between events avoided and cost(Figure 8), which may reflect the fact that the costmodel used events avoided as one of theindependent variables. The CEAC is very flat,reflecting the uncertainty in this analysis; theprobability that DBT is cost-effective was around65% at all thresholds (Figure 9).
Koons52
The incremental cost per patient of DBT overCCT was £8625, with just 0.2 events avoided. Theincremental cost per event avoided was £43,124(Figure 10). The higher incremental cost may be aconsequence of using the simpler cost modelbased on events avoided, which did not take anyreduction in inpatient cost directly into account.However, the trial report indicates that there waslittle difference between the arms in hospitaladmissions. The probability of being cost-effectiveat £5000 per event avoided was below 40%(Figure 11).
The incremental QALY gain was 0.03 and thisresulted in an incremental cost per QALY of£273,801. This reflected the higher incrementalcost of DBT in this study. The cost-effectivenessplane reflects the high probability that DBT willcost more than TAU (Figure 12). The CEACsuggests that TAU would have a probability ofaround 95% of being the treatment of choice(Figure 13) at a willingness to pay of £20,000.
Health Technology Assessment 2006; Vol. 10: No. 35
33
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TABLE 14 Cost implications of psychological therapy
Intervention arm Control arm
Study Therapy costs Other costs Total Therapy costs Other costs Total
Turner (DBT vs CCT)57 £10,372 £5,371 £15,743 £9,428 £11,557 £20,985Linehan (DBT vs TAU)53 £13,033 £2,658 £15,691 £7,958 £8,941 £16,898van den Bosch (DBT vs TAU)59 £11,996 £5,434 £17,430 £6,060 £10,646 £16,706Koons (DBT vs TAU)52 £16,903 £6,536 £23,439 £8,206 £6,609 £14,815Bateman (MBT vs TAU)51 £18,174 £17,743Tyrer (MACT vs TAU)58 £9,580 £7,563
Cost-effectiveness
34 TA
BLE
15
Cost
per
par
asui
cide
eve
nt a
void
ed
Stud
yPa
rasu
icid
e ev
ents
, mea
n (9
5% C
I)C
ost,
mea
n (9
5% C
I)Ev
ents
In
crem
enta
l C
ost
per
even
t av
oide
dco
stav
oide
d
Inte
rven
tion
arm
Con
trol
arm
Inte
rven
tion
arm
Con
trol
arm
Turn
er (D
BT v
s C
CT
)572.
92 (1
.6 t
o 4.
2)12
.33
(7.8
to
16.8
)£1
5,74
3 (£
14,0
00 t
o 20
,000
)£2
0,98
5 (£
17,0
00 t
o 28
,000
)9.
4–£
5,24
2In
terv
entio
ndo
min
ates
Line
han
(DBT
vs
TAU
)536.
82 (1
.7 t
o 12
)33
.54
(4.3
to
62.8
)£1
5,69
1 (£
13,0
00 t
o 18
,000
)£1
6,89
8 (£
9,00
0 to
25,
000)
26.7
–£1,
207
Inte
rven
tion
dom
inat
es
van
den
Bosc
h (D
BT v
s TA
U)59
16 (0
to
34.6
)34
.1 (9
.4 t
o 58
.8)
£17,
430
(£15
,000
to
23,0
00)
£16,
706
(£11
,000
to
28,0
00)
18.1
£724
£40
Koon
s (D
BT v
s TA
U)52
4 (0
to
8.4)
4.2
(0.7
to
7.7)
£23,
439
(£20
,000
to
27,0
00)
£14,
815
(£11
,000
to
18,0
00)
0.2
£8,6
25£4
3,12
4
Bate
man
(MBT
vs
TAU
)516.
1 (2
.3 t
o 10
)17
.5 (1
0.7
to 2
4.2)
£18,
174
(£16
,000
to
21,0
00)
£17,
743
(£14
,000
to
22,0
00)
11.3
£432
£38
Tyre
r (M
AC
T v
s TA
U)58
4.9
(1.1
to
8.7)
1.7
(0.7
to
2.7)
£9,5
80 (£
6,00
0 to
14,
000)
£7,5
63 (£
4,00
0 to
11,
000)
–3.2
£2,0
17In
terv
entio
n do
min
ated
by
cont
rol
TA
BLE
16
Cost
per
QAL
Y
Stud
yQ
ALY
gai
nC
ost
Incr
emen
tal Q
ALY
In
crem
enta
l cos
tC
ost
per
QA
LY(9
5% C
I)
Inte
rven
tion
arm
Con
trol
arm
Inte
rven
tion
arm
Con
trol
arm
Turn
er (D
BT v
s C
CT
)570.
170.
05£1
5,74
3£2
0,98
50.
12 (–
1.2
to 1
.4)
–£5,
242
Inte
rven
tion
dom
inat
es
Koon
s (D
BT v
s TA
U)52
0.07
0.04
£23,
439
£14,
815
0.03
(–1.
3 to
1.3
)£8
,625
£273
,801
Bate
man
(MBT
vs
TAU
)510.
04–0
.01
£18,
174
£17,
743
0.06
(–1.
2 to
1.3
)£4
32£7
,242
Tyre
r (M
AC
T v
s TA
U)58
0.19
0.14
£9,5
80£7
,563
0.
02 (0
to
0.09
)£2
,107
£84,
032
Health Technology Assessment 2006; Vol. 10: No. 35
35
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
–120,000
–100,000
–80,000
–60,000
–40,000
–20,000
0
20,000
40,000
–5 0 5 10 15 20 25 30 35
Incremental PS events avoided
Incr
emen
tal c
ost (
£)
FIGURE 2 Turner: cost-effectiveness plane of parasuicide (PS) events avoided
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 10,000 20,000 30,000
Cost-effectiveness threshold (£)
Prob
abili
ty c
ost-
effe
ctiv
e
FIGURE 3 Turner: CEAC of PS events avoided
Cost-effectiveness
36
–120,000
–100,000
–80,000
–60,000
–40,000
–20,000
0
20,000
40,000
60,000
–0.05 0.00 0.05 0.10 0.15 0.20 0.25 0.30
Incremental QALYs
Incr
emen
tal c
ost (
£)
FIGURE 4 Turner: cost-effectiveness plane of QALYs
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 20,000 40,000 60,000 80,000 100,000
Cost-effectiveness threshold (£)
Prob
abili
ty c
ost-
effe
ctiv
e
FIGURE 5 Turner: CEAC of QALYs
Health Technology Assessment 2006; Vol. 10: No. 35
37
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
–120,000
–100,000
–80,000
–60,000
–40,000
–20,000
0
20,000
40,000
–200 –100 0 100 200 300 400 500 600 700 800
Incremental PS events avoided
Incr
emen
tal c
ost (
£)
FIGURE 6 Linehan: cost-effectiveness plane of PS events avoided
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 10,000 20,000 30,000
Cost-effectiveness threshold (£)
Prob
abili
ty c
ost-
effe
ctiv
e
FIGURE 7 Linehan: CEAC of PS events avoided
Cost-effectiveness
38
–400,000
–300,000
–200,000
–100,000
0
100,000
200,000
300,000
–250 –200 –150 –100 –50 0 50 100 150 200 250 300
Incremental PS events avoided
Incr
emen
tal c
ost (
£)
FIGURE 8 van den Bosch: cost-effectiveness plane of PS events avoided
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 10,000 20,000 30,000Cost-effectiveness threshold (£)
Prob
abili
ty c
ost-
effe
ctiv
e
FIGURE 9 van den Bosch: CEAC of PS events avoided
MBTBateman51
In the baseline 12-month analysis, the centralestimate of the incremental cost per patient was£432 and number of events avoided was 11.3,resulting in a cost per event avoided of £38. Therewas a strong tendency for the intervention to bemore effective with little cost difference (Figure 14).At a threshold of £5000 there was an 80% chancethat the intervention would be cost-effective(Figure 15).
The cost per QALY was £7242. The cost-effectiveness plane demonstrates the uncertaintyin this analysis (Figure 16). This is also reflected inthe CEAC, which demonstrates that below athreshold of £20,000 there was a 55% chance thatTAU was more cost-effective than partialhospitalisation (Figure 17).
At 18 months 17 parasuicide events were avoided,at an incremental cost of £647, resulting in a costper event avoided of £38, as before, and littlereduction in uncertainty, with an 80% probabilitythat the intervention would be cost-effective at £5000 per event avoided (Figures 18 and 19).The cost per QALY was reduced from £7000 to£3000, but there was considerable uncertaintyaround the result (Figure 20). MBT had a
probability of around 40% to 50% of beingcheaper and more effective than TAU and theprobability that it was cost-effective at a threshold of £20,000 per QALY was around 55%(Figure 21).
MACTTyrer58
The central estimate for the incremental cost ofMACT over TAU was £2017 in BPD patients, with 3.2 more events. MACT was dominated by TAU in BPD. However, the uncertainty is such that the cost-effectiveness could range fromTAU dominating MACT to MACT dominatingTAU (Figure 22). The CEAC demonstrates thatthere was around a 60% chance that TAU would be more cost-effective than MACT (Figure 23).
The incremental QALY gain of MACT was 0.02and this resulted in a cost per QALY of £84,032.The direction of the gain in QALY contradicts theprimary outcome of parasuicide events (whichwere higher in the MACT group). However, therewas considerable uncertainty surrounding thisfigure and this is reflected in the plots on the cost-effectiveness plane (Figure 24). The probability ofbeing cost-effective at £20,000 was around 45%(Figure 25).
Health Technology Assessment 2006; Vol. 10: No. 35
39
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
–100,000
–50,000
0
50,000
100,000
150,000
200,000
–80 –60 –40 –20 0 20 40 60
Incremental PS events avoided
Incr
emen
tal c
ost (
£)
FIGURE 10 Koons: cost-effectiveness plane of PS events avoided
Cost-effectiveness
40
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 10,000 20,000 30,000Cost-effectiveness threshold (£)
Prob
abili
ty c
ost-
effe
ctiv
e
FIGURE 11 Koons: CEAC of PS events avoided
–100,000
–50,000
0
50,000
100,000
150,000
200,000
–0.15 –0.10 –0.05 0.00 0.05 0.10 0.15 0.20 0.25 0.30
Incremental QALYs
Incr
emen
tal c
ost (
£)
FIGURE 12 Koons: cost-effectiveness plane of QALYs
Health Technology Assessment 2006; Vol. 10: No. 35
41
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 20,000 40,000 60,000 80,000 100,000
Cost-effectiveness threshold (£)
Prob
abili
ty c
ost-
effe
ctiv
e
FIGURE 13 Koons: CEAC of QALYs
–60,000
–50,000
–40,000
–30,000
–20,000
–10,000
0
10,000
20,000
30,000
40,000
–150 –100 –50 0 50 100 150 200
Incremental PS events avoided
Incr
emen
tal c
ost (
£)
FIGURE 14 Bateman: cost-effectiveness plane of 12-month PS events avoided
Cost-effectiveness
42
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 10,000 20,000 30,000Cost-effectiveness threshold (£)
Prob
abili
ty c
ost-
effe
ctiv
e
FIGURE 15 Bateman: CEAC of 12-month PS events avoided
–50,000
–40,000
–30,000
–20,000
–10,000
0
10,000
20,000
30,000
40,000
–0.14 –0.12 –0.10 –0.08 –0.06 –0.04 –0.02 0.00 0.02 0.04 0.06 0.08
Incremental QALYs
Incr
emen
tal c
ost (
£)
FIGURE 16 Bateman: cost-effectiveness plane of 12-month QALYs
Health Technology Assessment 2006; Vol. 10: No. 35
43
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 20,000 40,000 60,000 80,000
Cost-effectiveness threshold (£)
Prob
abili
ty c
ost-
effe
ctiv
e
FIGURE 17 Bateman: CEAC of 12-month QALYs
–50,000
–40,000
–30,000
–20,000
–10,000
0
10,000
20,000
30,000
40,000
–150 –100 –50 0 50 100 150 200
Incremental PS events avoided
Incr
emen
tal c
ost (
£)
FIGURE 18 Bateman: cost-effectiveness plane of 18-month PS events avoided
Cost-effectiveness
44
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 10,000 20,000 30,000Cost-effectiveness threshold (£)
Prob
abili
ty c
ost-
effe
ctiv
e
FIGURE 19 Bateman: CEAC of 18-month PS events avoided
–50,000
–40,000
–30,000
–20,000
–10,000
0
10,000
20,000
30,000
40,000
–0.30 –0.20 –0.10 0.00 0.10 0.20 0.30 0.40
Incremental QALYs
Incr
emen
tal c
ost (
£)
FIGURE 20 Bateman: cost-effectiveness plane of 18-month QALYs
Health Technology Assessment 2006; Vol. 10: No. 35
45
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Cost-effectiveness threshold (£)
Prob
abili
ty c
ost-
effe
ctiv
e
0 20,000 40,000 60,000 80,000 100,000
FIGURE 21 Bateman: CEAC of 18-month QALYs
–80,000
–60,000
–40,000
–20,000
0
20,000
40,000
60,000
80,000
–100 –80 –60 –40 –20 0 20 40
Incremental PS events avoided
Incr
emen
tal c
ost (
£)
FIGURE 22 Tyrer: cost-effectiveness plane of PS events avoided
Cost-effectiveness
46
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 10,000 20,000 30,000Cost-effectiveness threshold (£)
Prob
abili
ty c
ost-
effe
ctiv
e
FIGURE 23 Tyrer: CEAC of PS events avoided
–80,000
–60,000
–40,000
–20,000
0
20,000
40,000
60,000
80,000
100,000
0 0.02 0.04 0.06 0.08 0.10 0.12 0.14
Incremental QALYs
Incr
emen
tal c
ost (
£)
FIGURE 24 Tyrer: cost-effectiveness plane of QALYs
Univariate sensitivity analysisNICE economic perspectiveFor those studies where the costing for otherresources was based on the regression model, theNICE perspective did not alter the direction ormagnitude of the cost-effectiveness results. Thismay be due to the fact the regression coefficientsin the government and NICE models had asimilar order of magnitude. The direction andmagnitude of the Tyrer results58 were also similarwhen costs were estimated from the trial data froma NICE perspective.
Societal perspectiveThe societal model has the effect of increasing themagnitude of costs in both arms by a factor ofapproximately 75%. The van den Bosch study59
was the only one in which the societal model
resulted in a change in the direction of the results.Using the societal model the incremental cost inthis study became –£818, and the interventiondominated TAU.
Supervisor costsIn the baseline analysis TAU supervision costswere assumed to be less than DBT supervisioncosts by a factor of 0.5. This factor was increasedto 0.75 (increasing control arm costs) anddecreased to 0.25 (decreasing control arm costs).For the Bateman (MBT) and van den Bosch (DBT)studies,51,59 a factor of 0.75 resulted in theintervention arm becoming cheaper than thecontrol arm and therefore DBT dominated TAU.A factor of 0.25 resulted in small increases in thecost per QALY or events avoided for van denBosch59 and Koons,52 and a three- to four-foldincrease for Bateman.51
Health Technology Assessment 2006; Vol. 10: No. 35
47
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Cost-effectiveness threshold (£)
Prob
abili
ty c
ost-
effe
ctiv
e
0 20,000 40,000 60,000 80,000 100,000
FIGURE 25 Tyrer: CEAC of QALYs
Main results: clinical effectivenessNine RCTs and one non-RCT of moderate to poorquality were identified in the clinical effectivenessreview. Of these ten studies, six show that there issome evidence that psychological therapies forBPD may be effective, whereas the evidence fromfour trials suggests that psychological therapies areno more effective than the alternatives. However,these results should be interpreted with caution asnot all studies were primarily targeted toborderline symptoms and there were considerabledifferences in patient characteristics, comparisongroups and outcomes between the studies.
Main results: cost-effectivenessThe review of published studies identified onecost-effectiveness analysis of data from an RCTcomparing DBT with TAU for the treatment ofBPD. Subjects were women who were clinicallyreferred to a psychotherapy outcome study.Subjects receiving DBT (n = 22) incurredsignificantly higher psychotherapy costs, lowerpsychiatric inpatient costs and lower emergencyroom costs than those receiving TAU (n = 22).The two treatment groups did not differsignificantly with respect to median medical ortotal healthcare costs. The cost-effectivenessmeasures used were cost per week employed andcost per point of global adjustment and nosignificant difference was found in either of thesemeasures for DBT compared with TAU.
The review of published studies also identified aneconomic evaluation of psychological therapies ofpartial relevance to BPD. This was a cost-effectiveness analysis of data from an RCTcomparing MACT with TAU for the treatment ofpeople with recurrent episodes of DSH. There wereno significant differences between the groups inthe total costs across all patients or among thosewith BPD (n = 62). The cost per 1% reduction inthe proportion of patients with a repeat self-harmepisode was £120, with more than 90% chance ofbeing cost-effective, but this analysis was notundertaken for the BPD subgroup. Theincremental mean effect as measured by the EQ-5D instrument was negative for MACT (–0.01118).
The incremental cost per QALY gained from TAUwas therefore £66,000, but the authors argued thatthis was probably a chance finding given that thedifference in EQ-5D was not significant.
A formal decision modelling approach could notbe applied given the complex care pathways forpatients with BPD and the lack of evidence. It wasdecided instead to undertake separate economicevaluations for the six RCTs that had sufficientdata using a combination of data reported inpublished papers, trial data sets sent by theinvestigators, a cost model using data from thePOPMACT study and a utility mapping exercise.Cost-effectiveness was assessed in terms of cost perparasuicide event avoided in all six trials and costper QALY in four of them (this was done bymapping BDI results onto the EQ-5D for threetrials). All results are at 2003/04 prices and for12 months follow-up.
In three of the four DBT trials the interventiondominated the control groups in terms ofparasuicide events (Turner57 and Linehan53) orachieved a cost per event avoided below £50 (vanden Bosch59). However, in a fourth DBT trial theestimated cost per event avoided was £43,124.While the studies by Linehan53 and van denBosch59 seem favourable to DBT in terms of meanincremental cost-effectiveness, the probability ofbeing cost-effective at £5000 per parasuicide eventavoided was around just 60% in each case. Onlytwo DBT trials were subjected to a cost per QALYanalysis, and for one the intervention againdominated (Turner57) and the other had a cost perQALY of £273,801. The PSA showed substantialuncertainty surrounding these results. ForKoons,52 the probability of DBT being cost-effective was very low (at less than 10%). Thismixture of results, high levels of uncertainty andthe limitations described in the next section donot support the cost-effectiveness of DBT,although they suggest that DBT could potentiallybe cost-effective.
The MBT study group achieved a very low costper event avoided, with a probability of being cost-effective at £5000 per parasuicide eventavoided of 80%. While the cost per QALY wasmodest at £7242, there was substantial uncertainty,
Health Technology Assessment 2006; Vol. 10: No. 35
49
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Chapter 5
Discussion
with a probability of being cost-effective at £20,000per QALY of less than 60%. For POPMACT, theBPD subgroup analysis undertaken by theassessment team found that the intervention wasdominated in terms of cost per parasuicide eventavoided. There was an insignificant incrementalQALY gain in BPD, but the associated cost perQALY was £84,032. These assessments of MACTwere both associated with a high degree ofuncertainty, where the probability of being cost-effective was less than 50% in each case.
Assumptions, limitations anduncertaintiesClinical effectivenessThe size of the studies was small. Therefore, a lackof power was a major problem for the majority oftrials. Bias was likely in some trials owing to a lackof blind outcome assessment, unclear allocationconcealment and high dropout rates. Most trialsincluded only women or predominantly women,limiting their generalisability.
A significant number of BPD patients usually donot complete the programme; thus, greatertherapist involvement is necessary.53 The degree oftherapist involvement in most intervention groupswas high. Therapists in experimental groups tendto have a more effective and supportive approach,with flexible relationships with patients, thantherapists in control groups, which may have led tothe lower attrition rates in the intervention arms.
The level of training and supervision of thetherapist as well as therapist beliefs may haveinfluenced the treatment outcomes. The therapistswith special training, previous experience and aparticular interest in working with patients withBPD may have achieved more success in thetreatment than would be achieved in a routinesetting. For example, the trials by Linehan’s groupwere conducted in research clinics, whichspecialised in these patients. Personnel madeconsiderable efforts to keep patients in treatment,which is not usual in normal practice. Althoughthe principle of BPD therapies is based ondeveloping a strong working alliance betweentherapist and client to prevent withdrawal andconsequently to provide a complete treatmentcourse, this requires considerable skills, effort andtime from therapists and may not be a practicalapproach for busy mental health settings.
One trial52 compared two active psychologicaltreatment conditions where the same
pharmacotherapy was administered in bothgroups. Although both groups significantlyimproved at the end of the treatment, there wereno differences between the two treatments and theauthors assumed that the effect might haveoccurred because of the drug. Therefore, it isimportant to consider drug administration instudies examining psychosocial therapies.
The long-term efficacy of psychological therapiesis unclear, since the longest reported follow-up was36 months,51 at which time significant reductionsin clinical symptoms of the intervention arm notonly were maintained but continued to declineand were associated with low hospital admissionrates.
DBT was one of the more successful treatments,with significant positive outcomes in relation toreducing self-harm and improving behaviouralcontrol in women. However, most studies comefrom a single group of investigators whodeveloped the method and have a strongallegiance to it. More recent and moreindependent replication of these trials with alarger sample in Europe found that DBT is notsignificantly more effective than TAU.
There was insufficient evidence to examinewhether any particular type of psychologicaltreatment is more effective than others.
Most of the BPD participants in trials werereferred by tertiary care settings and described as“(severely) parasuicidal” or “substance abusing”,which omitted the rest of the (less severe, non-parasuicidal or non-substance-abusing) BPDpatients. Such patients cannot fully represent theBPD population.
The need for a common scale for major clinicaloutcomes derives from the considerableheterogeneity of measures for assessing similarvariables. Also, there is a need for more generichealth-related quality of life measures.
Little information has been provided in thestudies regarding patient preference. Research isstill needed in these areas.
Finally, the authors are confident that they havenot missed any important RCTs; however, thereview of the non-RCTs might not have beensufficiently thorough. The reason for thisuncertainty is that BPD can be included as asubgroup in general personality disorder,parasuicide/self-harm, suicide attempts, A&E and
Discussion
50
forensic studies. These would require a largenumber of studies to be reviewed and would havebeen beyond the resources available for thisreview.
Cost-effectivenessThe two studies that provided a cost-effectivenessanalysis were of good quality, but the Heardstudy68 had limitations concerning the lack ofimportant cost data and the Byford study69 wasconcerned with DSH and not just BPD. Apublished subgroup analysis did not replicate thefull economic analysis. Indirectly, however, theByford study provided important evidence on thelikely cost-effectiveness of MACT, which was usedin a subgroup analysis undertaken by thisassessment team, and the trial provided data forthe economic analyses undertaken for the otherRCTs presented in this report.
The assessments of cost-effectiveness undertakenfor each of the six trials presented in this reportmust be interpreted with great care. The trialswere conducted in different settings in patientswith varying baseline disease conditions. Inaddition, the methods and types of staff used todeliver the therapy differed, as did the length,frequency and type of sessions delivered. Thesedifferences make comparisons between the trialsproblematic. The comparability of TAU betweenthe trials is also questionable. For two of thestudies, the analysis only uses parasuicide eventsavoided, which limits the generalisability of theresults and makes comparison betweeninterventions difficult. Comparison betweenstudies even using this outcome is limited becausethe studies used different definitions of theoutcome measure.
The methods for assessing outcome used in theeconomic analysis are also subject to limitations.The cost-effectiveness analysis used the number ofparasuicide events avoided and, as reportedearlier, there was some variation between studies in the recording of these events.Furthermore, this is not an outcome that capturesall of the consequences of the interventions for patients and is not a useful outcome fordecision-makers concerned with allocatingresource across programmes. This reviewattempted to estimate QALYs to try to capture a more patient-focused outcome that provides amore generic assessment of benefit. However, only one study directly measured QALYs and for three others a mapping from the BDI to theEQ-5D that was, inevitably, not completelyaccurate.
There are more general concerns with the use ofQALYs in this condition and specifically QALYsthat use the EQ-5D preference-based measure.The EQ-5D may not capture all of theconsequences for health-related quality of life ofpeople with BPD. There have been no validationstudies of EQ-5D in BPD. Suggestions for furtherwork in this area are suggested below.
The methodology to estimate costs differs betweenthe trials. Some trials reported costs in detail,whereas others had to be estimated using simpleregression models. The estimation of resource usefor the Koons study52 is particularly weak, since itwas based on a poor association betweenparasuicide events and total costs. However, thenumber of parasuicide events was similar for botharms in this trial and the use of this modelprobably had little influence on the cost-effectiveness results.
The cost-effectiveness estimates are also dependenton the quality of the trials, which were small andsuffered from high dropout rates. This wouldsuggest that in many cases the results exaggeratethe cost-effectiveness of the interventions. Theremust also be doubts about the generalisability ofthe results from these trials. Most of the DBTstudies, for example, were undertaken in countriesoutside the UK and there must be doubts as towhether the results are transferable to the NHS.These results merely indicate the potential cost-effectiveness of DBT and MBT.
Need for further researchResearch into psychological interventions for BPDhas tended to comprise either uncontrolled studieswhere it is impossible to interpret the findings, orsmall, poor-quality RCTs with high rates ofdropouts that have not been properly followed up.At the same time, little thought has been given tothe needs of contemporary economic analysis,although there are some notable exceptions. Theresults have produced a body of evidence that hasbeen largely inconclusive. BPD is an importantcondition with a number of resource-intensivetherapies available and it should be a priority areafor future research. Here, more detailedsuggestions are made for further research in termsof pragmatic trials and studies to inform economicevaluation.
Pragmatic controlled trialsThe basic evidence of clinical efficacy was poor.Appropriately powered, head-to-head RCTs of
Health Technology Assessment 2006; Vol. 10: No. 35
51
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
psychological therapies are needed. The keyfeatures of these trials include the following:
● Where possible, a trial should have more thanone psychological therapy being compared.
● Studies must be designed with adequate statisticalpower, taking into account expected dropouts.
● Patients from a variety of ethnic and socio-economic backgrounds must be included, withan age and gender mix comparable to thosereceiving treatment in the NHS.
● The level of severity and dysfunction must bewell defined in future surveys and trials.
● The definition of dropout must be standardisedand reduced where possible in the RCTsexamining psychological therapies for BPD.Where patients drop out of therapyconsiderable effort must still be undertaken tocollect data on them.
● The different therapies need to be properlydescribed, including a TAU arm. Medication,for example, must be taken into account.
● Studies comparing active intervention with TAUneed to be designed so that TAU is indeed thatand not minimal intervention to maximise thebenefits associated with interventionpsychological therapies.
● The longest follow-up was for 18 months, and 6months was more common. Given the high costof the interventions, longer term follow-upsshould be undertaken.
● Data should be collected on outcomes,including recognised generic measures ofhealth-related quality of life, as well aspreference-based measures to permitcomparisons across programmes (see below).
● Data should be collected on resource-useservices (see below).
● Research teams should include independentresearchers.
Studies to inform future economicanalysisThe lack of data and complexity of the carepathway of BPD meant that a conventional
economic model that synthesised a range ofevidence could not be constructed for this report.This makes it difficult to conduct an estimate ofthe value of information to provide evidence onthe size of investment needed and the prioritiesfor future research. However, based on theanalyses presented here the authors are able torecommend the following (in addition to theabove on clinical effectiveness).
● Conducting cost-effectiveness analyses was inpart limited by the complexity of the condition,but also the absence of good data on currentpractice. Survey work is needed into currentpractice in terms of pathways of care to begin to be able to model the longer term benefitsusing a more formal decision modellingapproach.
● A survey of current practice and the use of thefull range of services (including number ofsessions attended and type of therapist) bypeople with BPD is needed to inform futureeconomic analyses.
● Full resource-use data must be collected in thecontext of pragmatic clinical trials.
● A psychometric assessment is needed of thevalidity of the EQ-5D and other genericpreference-based measures in BPD.
● If the generic measures are found wanting, thena more condition-specific preference-basedmeasure must be developed that captures theimpact of BPD on people’s lives.
Related to the above researchrecommendationsIt must be recognised that BPD is not ahomogeneous condition and it often occursalongside other psychological co-morbidities.
Research is needed to determine the relationshipof BPD and co-occurring major disorders todevelop an optimal multicomponent programme, which will be targeted not only toBPD-specific symptoms but also to the coexisting problems.
Discussion
52
There is some evidence to support the clinicaleffectiveness of psychological therapies for
BPD.
● There is some evidence that DBT is moreeffective than TAU for the treatment ofchronically parasuicidal and drug-dependentwomen with BPD.
● There is some evidence that DBT-orientatedtherapy is more effective than CCT for thetreatment of BPD.
● There is some evidence that DBT is as effectiveas CVT+12S for the treatment of opioid-dependent women with BDP.
● There is some evidence that MBT is moreeffective than TAU in the treatment of BPD.
● There is good evidence that MACT is no moreeffective than TAU in the treatment of BPD.
● There is some evidence that IGP is no moreeffective than individual MBT for the treatmentof BPD.
● There is some evidence that TFP is moreeffective than TAU.
The overall efficacy of psychological therapies ispromising; however, at this stage the evidence isinconclusive.
In terms of cost-effectiveness, this reviewattempted to examine the cost-effectiveness of theintervention in six RCTs. The mix of resultsbetween the four trials of DBT, along with thehigh levels of uncertainty and the limitations ofthe analyses, do not support the cost-effectivenessof DBT, although they suggest that it has thepotential to be cost-effective. The results for MBTare promising, although again surrounded by ahigh degree of uncertainty, and for MACT theanalysis suggests that the intervention is unlikelyto be cost-effective.
Although the results do not support the cost-effectiveness of any psychological interventionover the rest, or of psychological therapy as awhole, the results do offer the hope that suchinterventions could be cost-effective. What isneeded now are well-designed, pragmatic RCTs ofthe leading therapies in head-to-headcomparisons with appropriate patient-focusedoutcomes (including an appropriate preference-based measure) and resource-use data collection,with formal modelling of the longer termconsequences.
Health Technology Assessment 2006; Vol. 10: No. 35
53
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Chapter 6
Conclusions
The authors wish to thank Dr Ian Kerr(Consultant Psychiatrist, Sheffield), Professor
Kate Davidson (Consultant Clinical Psychologist,Glasgow), Dr Graham Connell-Jones (ConsultantForensic Psychiatrist, Nottingham), Professor PeterTyrer (Professor of Community Psychiatry,London) and Dr Nick Huband (Department ofForensic Mental Health, Nottingham) for acting asclinical advisors to the project. Also many thanksto Sarah Byford, Peter Tyrer and Marco Chiesa forproviding additional useful trial data. The authorsalso wish to thank Gill Rooney and AndreaShippam for their help in preparing andformatting the report.
Contribution of authorsIndra Tumur (Research Fellow) and MichaelFerriter (Research Fellow) compiled the clinicaleffectiveness section. John Brazier (Professor ofHealth Economics) and Michael Holmes (ResearchAnalyst) wrote the economic evaluation section.Glenys Parry (Professor of Applied PsychologicalTherapies) and Kim Dent-Brown (PostdoctoralResearch Fellow) provided the backgroundinformation and Suzy Paisley (Research Scientist)did the literature searching.
Health Technology Assessment 2006; Vol. 10: No. 35
55
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Acknowledgements
1. Tyrer P. Borderline personality disorder: a motleydiagnosis in need of reform. Lancet 1999;354:2095–6.
2. Weissman MM. The epidemiology of personalitydisorders: a 1990 update. J Pers Disord1993;7(suppl 1):44–62.
3. Samuels J, Eaton WW, Bienvenu OJ III, Brown CH,Costa PT Jr, Nestadt G. Prevalence and correlatesof personality disorders in a community sample.Br J Psychiatry 2002;180:536–42.
4. Torgersen S, Kringlen E, Cramer V. Theprevalence of personality disorders in acommunity sample. Arch Gen Psychiatry2001;58:590–6.
5. Coid J. Epidemiology, public health and theproblem of personality disorder. Br J Psychiatry2003;182:s3–10.
6. Swartz HA, Blaze DG, Winfield I. Estimating theprevalence of borderline personality disorder inthe community. J Personal Disord 1990;4:457–572.
7. Widiger TA, Weissman MM. Epidemiology ofborderline personality disorder. Hospital andCommunity Psychiatry 1991;42:1015–21.
8. Ogata SN, Silk NR, Goodrich S, Lohr N.Childhood sexual of physical abuse in adultpatients with borderline personality disorder. Am JPsychiatry 1990;147:1008–13.
9. Shearer SL, Peters CP, Quaytman MS, Ogden RL.Frequency and correlates of childhood sexual andphysical abuse histories and adult femaleborderline patients. Am J Psychiatry1990;147:214–6.
10. Paris J, Zweig-Frank H, Guzder H. Psychologicalrisk factors for borderline personality disorder infemale patients. Compr Psychiatry 1994;35:301–5.
11. Zanarini MC, Williams AA, Lewis RE, Reich RB.Reported pathological childhood experiencesassociated with the development of borderlinepersonality disorder. Am J Psychiatry1997;154:1101–6.
12. Livesley WJ, Jang KL, Vernon PA. Phenotypic andgenetic structure of traits delineating personalitydisorder. Arch Gen Psychiatry 1998;55:941–8.
13. Skodol AE, Siever L, Livesley WJ, Gunderson J,Pfohl B, Widiger TA. The borderline diagnosis II:biology, genetics and clinical course. Biol Psychiatry2002;51:951–63.
14. Carpenter W, Gunderson JP, Strauss J.Considerations of the borderline syndrome: alongitudinal comparative study of borderline andschizophrenic patients. In Borderline personalitydisorders: the concept, the syndrome, the patient. NewYork: International Universities Press; 1997.pp. 231–53.
15. Pope HG Jr, Jonas JM, Hudson JI, Cohen BM,Gunderson JG. The validity of DSM-III borderlinepersonality disorder. A phenomenologic, familyhistory, treatment response, and long-term follow-up study. Arch Gen Psychiatry 1983;40:23–30.
16. Werble B. Second follow-up of borderline patients.Arch Gen Psychiatry 1970;23:3–7.
17. McGlashan TH. The Chestnut Lodge follow-upstudy. I. Follow-up methodology and study sample.Arch Gen Psychiatry 1984;41:573–85.
18. Paris J, Brown R, Nowlis D. Long term follow upof borderline patients in a general hospital. ComprPsychiatry 1987;28:530–5.
19. Stone MH. Long-term outcome in personalitydisorders [review]. Br J Psychiatry1993;162:299–313.
20. Lenzenweger MF, Johnson MD, Willett JB.Individual growth curve analysis illuminatesstability and change in personality disorderfeatures: the longitudinal study of personalitydisorders. Arch Gen Psychiatry 2004;61:1015–24.
21. Zanarini MC, Frankenburg FR, Khera GS,Bleichmar J. Treatment histories of borderlineinpatients. Compr Psychiatry 2001;42:144–50.
22. Huffman JC, Popkin MK, Stern TA. Psychiatricconsiderations in the patient receiving organtransplantation: a clinical case conference. GenHosp Psychiatry 2003;25:484–91.
23. McGlashan TH, Grilo CM, Sanislow CA, Ralevski E,Morey LC, Gunderson JG. Two-year prevalenceand stability of individual DSM-IV criteria forschizotypal, borderline, avoidant andobsessive–compulsive personality disorders: towarda hybrid model of Axis II disorders. Am J Psychiatry2005;162:883–9.
24. Lieb K, Zanarini MC, Schmahl C, Linehan MM,Bohus M. Borderline personality disorder. Lancet2004;364:453–61.
25. Conklin CZ, Westen D. Borderline personalitydisorder in clinical practice. Am J Psychiatry2005;162:867–75.
Health Technology Assessment 2006; Vol. 10: No. 35
57
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
References
26. Clarkin JF, Foelsch PA, Levy KN, Hull JW, DelaneyJC, Kernberg OF. The development of apsychodynamic treatment for patients withborderline personality disorder: a preliminarystudy of behavioral change. J Personal Disord2001;15:487–95.
27. Bateman A, Fonaghy P. Psychotherapy for borderlinepersonality disorder: Mentalization based treatment.Oxford: Oxford University Press; 2004.
28. Linehan MM. Cognitive–behavioral treatment ofborderline personality disorder. New York: GuilfordPress; 1993.
29. Young JE, Klosko JS, Weishaar ME. Schema therapy:a practitioner’s guide. New York: Guilford Press;2003.
30. Evans K, Tyrer J, Catalan U, Schmidt K, Davidson J, Dent P. Manual-assisted cognitive-behaviour therapy (MACT): a randomizedcontrolled trial of a brief intervention withbibliotherapy in the treatment of recurrentdeliberate self-harm. Psychol Med 1999;29:19–25.
31. Davidson K. Cognitive therapy for personalitydisorders: a guide for therapists. Oxford:Butterworth–Heinemann; 2000.
32. Ryle A. Cognitive analytic therapy and borderlinepersonality disorder: the model and the method.Chichester: John Wiley; 1997.
33. Linehan MM, Turner RM. Cognitive-behavioraltreatment of borderline personality disorder[review]. Journal of Family Violence 1996;11:429–32.
34. Boyce P, Carter G, Penrose-Wall J, Wilhelm K,Goldney R. Summary Australian and New Zealandclinical practice guideline for the management ofadult deliberate self-harm (2003). AustralasianPsychiatry 2003;11:150–5.
35. Warren F, McGauley G, Norton K, Dolan B,Preedy-Fayers K, Pickering A. Review of treatmentsfor severe personality disorder. London: Home Office;2003.
36. Lees J, Manning N, Rawlings B. Therapeuticcommunity effectiveness: a systematic internationalreview of therapeutic community treatment for peoplewith personality disorders and mentally disorderedoffenders. CRD Report 17. York: Centre for Reviewsand Dissemination; 1999.
37. Cornah D, Stein K, Stevens A. The therapeuticcommunity method of treatment for borderline personalitydisorder. Southampton: WIHRD; 1997.
38. Binks CA, Fenton M, McCarthy L, Lee T, Adams CE, Duggan C. Psychological therapies forpeople with borderline personality disorder.Cochrane Database Syst Rev 2006;CD005652.
39. Hawton K, Arensman E, Townsend E, Bremner S,Feldman E, Goldney R, et al. Deliberate self harm:
systematic review of efficacy of psychosocial andpharmacological treatments in preventingrepetition. BMJ 1998;317:441–7.
40. van der Sande R, Buskens E, Allart E, van der Graaf Y, van Engeland H. Psychosocialintervention following suicide attempt: asystematic review of treatment interventions. ActaPsychiatr Scand 1997;96:43–50.
41. Perry JC, Banon E, Ianni F. Effectiveness ofpsychotherapy for personality disorders. Am JPsychiatry 1999;156:1312–21.
42. Leichsenring F. The effectiveness ofpsychodynamic therapy. A review using criteria ofevidence-based medicine. Z Psychosom MedPsychother 2002;48:139–62.
43. Sanislow CA, McGlashan TH. Treatment outcomeof personality disorders. Can J Psychiatry1998;43:237–50.
44. Woods P, Richards D. Effectiveness of nursinginterventions in people with personality disorders.J Adv Nurs 2003;44:154–72.
45. Bateman AW, Fonagy P. Effectiveness ofpsychotherapeutic treatment of personalitydisorder [review]. Br J Psychiatry 2000;177:138–43.
46. Clarke M, Oxman AD. The Cochrane reviewershandbook 4.1.4 [updated April 2002]. In TheCochrane Library. Oxford: Update Software; 2002.
47. Oxman AD, Cook DJ, Guyatt GH. Evidence BasedMedicine Working Group. Users’ guides to themedical literature VI. How to use an overview.JAMA 1994;272:1367–71.
48. Guyatt GH, Sackett DL, Cook DJ. Users’ guides tothe medical literature II. How to use an articleabout therapy or prevention. A. Are the results ofthe study valid? JAMA 1993,270:2598–601.
49. Guyatt GH, Sackett DL, Cook DJ. Users’ guides tothe medical literature II. How to use an articleabout therapy or prevention. B. What were theresults and will they help me in caring for mypatients? JAMA 1994;271:59–63.
50. Drummond M, Jefferson TO. Guidelines forauthors and peer reviewers of economicsubmissions to the BMJ. BMJ 1996;313:275–83.
51. Bateman A, Fonagy P. Effectiveness of partialhospitalization in the treatment of borderlinepersonality disorder: a randomized controlledtrial. Am J Psychiatry 1999;156:1563–9.
52. Koons CR, Robins CJ, Tweed JL, Lynch TR,Gonzalez AM, Morse JQ, et al. Efficacy ofdialectical behavior therapy in women veteranswith borderline personality disorder. BehaviorTherapy 2001;32:371–90.
53. Linehan MM, Armstrong HE, Suarez A, Allmon D,Heard HL. Cognitive-behavioral treatment of
References
58
chronically parasuicidal borderline patients. ArchGen Psychiatry 1991;48:1060–4.
54. Linehan MM, Schmidt H III, Dimeff LA, Craft JC,Kanter J, Comtois KA. Dialectical behaviortherapy for patients with borderline personalitydisorder and drug-dependence. Am J Addict1999;8:279–92.
55. Linehan MM, Dimeff LA, Reynolds SK, Comtois KA, Welch SS, Heagerty P, et al.Dialectical behavior therapy versus comprehensivevalidation therapy plus 12-step for the treatmentof opioid dependent women meeting criteria forborderline personality disorder. Drug AlcoholDepend 2002;67:13–26.
56. Munroe-Blum H, Marziali E. A controlled trial ofshort-term group treatment for borderlinepersonality disorder. J Personal Disord1995;9:190–8.
57. Turner RM. Naturalistic evaluation of dialecticalbehavior therapy-oriented treatment forborderline personality disorder. Cognitive andBehavioral Practice 2000;7:413–9.
58. Tyrer P, Thompson S, Schmidt U, Jones V,Knapp M, Davidson K, et al. Randomizedcontrolled trial of brief cognitive behaviourtherapy versus treatment as usual in recurrentdeliberate self-harm: the POPMACT study. PsycholMed 2003;33:969–76.
59. van den Bosch LM, Verheul R, Schippers GM, van den Brink W. Dialectical behavior therapy ofborderline patients with and without substance useproblems. Implementation and long-term effects.Addict Behav 2002;27:911–23.
60. Wilberg T, Friis S, Karterud S, Mehlum L, Urnes O,Vaglum P. Outpatient group psychotherapy: avaluable continuation treatment for patients withborderline personality disorder treated in a dayhospital? A 3-year follow-up study. Nord J Psychiatry1998;52:213–21.
61. Linehan MM, Heard HL, Armstrong HE.Naturalistic follow-up of a behavioral treatment forchronically parasuicidal borderline patients. ArchGen Psychiatry 1993;50:971–4.
62. Bateman A, Fonagy P. Treatment of borderlinepersonality disorder with psychoanalyticallyoriented partial hospitalization: an 18-monthfollow-up. Am J Psychiatry 2001;158:36–42.
63. Tyrer P, Tom B, Byford S, Schmidt U, Jones V,Davidson K, et al., POPMACT Group. Differentialeffects of manual assisted cognitive behaviortherapy in the treatment of recurrent deliberateself-harm and personality disturbance: thePOPMACT study. J Personal Disord 2004;18:102–16.
64. Verheul R, van den Bosch LM, Koeter MW,De Ridder MA, Stijnen T, van den Brink W.Dialectical behaviour therapy for women with
borderline personality disorder: 12-month,randomised clinical trial in The Netherlands. Br JPsychiatry 2003;182:135–40.
65. Lackner JM, Mesmer C, Morley S. Psychologicaltreatments for irritable bowel syndrome: asystematic review and meta analysis. J Consult ClinPsychol 2004;72:1100–13.
66. Hollis S, Campbell F. What is meant by intention-to-treat analysis? Survey published randomisedcontrolled trials. BMJ 1999;319:670–4.
67. Linehan MM, Kanter JW, Comtois KA. Dialecticalbehavior therapy for borderline personalitydisorder. In Janowsky D (editor), Psychotherapyindications and outcomes. Washington, DC: AmericanPsychiatric; 1993. pp. 93–118.
68. Heard HL. Cost-effectiveness of dialecticalbehavior therapy for borderline personalitydisorder. Dissertation Abstracts International SectionB: Sciences and Engineering 2000;61(6B):3278.
69. Byford S, Knapp M, Greenshields J, Ukoumunne OC, Jones V, Thompson S, et al. Cost-effectiveness of brief cognitive behaviour therapyversus treatment as usual in recurrent deliberateself-harm: a decision-making approach. PsycholMed 2003;33:977–86.
70. Bateman A, Fonagy P. Health service utilizationcosts for borderline personality disorder patientstreated with psychoanalytically oriented partialhospitalization versus general psychiatric care. Am J Psychiatry 2003;160:169–71.
71. Hall J, Caleo S, Stevenson J, Meares R. Aneconomic analysis of psychotherapy for borderlinepersonality disorder patients. J Ment Health PolicyEcon 2001;4:3–8.
72. National Institute for Clinical Excellence. NICEGuide to the Methods of Technology Appraisal.London: NICE; 2004.
73. Kaltenthaler E, Brazier J, De Nigris E, Tumur I,Ferriter M, Beverley C, et al. Computerisedcognitive behaviour therapy for depression andanxiety update: a systematic review and economic evaluation. Health Technol Assess2006;10(33).
74. Office for National Statistical. Hospital andcommunity health services pay and prices index.London: ONS; 2005.
75. Rees A, Barkham M. DBT practice research network:DBT treatment implementation survey: feedback topractitioners. PTRC Memo No. 495. Leeds:Psychological Therapies Research Centre,University of Leeds; 2001.
76. Palmer R. Dialectical behaviour therapy forborderline personality disorder. Advances inPsychiatric Treatment 2002;8:16.
Health Technology Assessment 2006; Vol. 10: No. 35
59
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
77. Curtis L, Netten A, Unit costs of health and socialcare. Canterbury: PRSSU, University of Kent;2003.
78. Turner RM. Understanding dialectical behaviortherapy. Clinical Psychology: Science and Practice2000;7:95–8.
79. Gabbard GO, Lazar SG, Hornberger J, Spiegel D.The economic impact of psychotherapy: a review.Am J Psychiatry 1997;154:147–55.
80. Heard HL. Cost-effectiveness of dialecticalbehavior therapy in the treatment of borderlinepersonality disorder. Dissertation AbstractsInternational: Section B: The Sciences and Engineering2000;61:3278–358.
81. Allard R, Marshall M, Plante MC. Intensive follow-up does not decrease the risk of repeat suicideattempts. Suicide Life Threat Behav 1992;22:303–14.
82. Chiesa M, Fonagy P, Holmes J, Drahorad C.Residential versus community treatment ofpersonality disorders: a comparative study of threetreatment programs. Am J Psychiatry2004;161:1463–70.
83. Clarkin JF, Levy KN, Lenzenweger MF, Kernberg OF. The Personality DisordersInstitute/Borderline Personality Disorder ResearchFoundation randomized control trial forborderline personality disorder: rationale,methods, and patient characteristics. J PersonalDisord 2004;18:52–72.
84. Manning SY. The effects of a cognitive-behavioraltreatment on females with borderline personalitydisorder. Dissertation Abstracts International Section A:Humanities and Social Sciences 1997;57:2880.
85. Piper WE, Rosie JS, Azim HF, Joyce AS. Arandomized trial of psychiatric day treatment forpatients with affective and personality disorders.Hospital and Community Psychiatry 1993;44:757–63.
86. Simpson EB, Yen S, Costello E, Rosen K, Begin A,Pistorello J, et al. Combined dialectical behaviortherapy and fluoxetine in the treatment ofborderline personality disorder. J Clin Psychiatry2004;65:379–85.
87. Sloane RB, Staples FR, Cristol AH, Yorkston NJ.Short-term analytically oriented psychotherapyversus behavior therapy. Am J Psychiatry1975;132:373–7.
88. Springer T, Lohr NE, Buchtel HA, Silk KR. A preliminary report of short-term cognitive-behavioral group therapy for inpatients withpersonality disorders. Journal of PsychotherapyPractice and Research 1996;5:57–71.
89. Stiwne D. Group psychotherapy with borderlinepatients: contrasting remainers and dropouts.Group 1994;18:37–45.
90. Winston A, Laikin M, Pollack J, Samstag LW,McCullough L, Muran JC. Short-termpsychotherapy of personality disorders. Am JPsychiatry 1994;151:190–4.
91. Bohus M, Haaf B, Stiglmayr C, Pohl U, Bohme R,Linehan M. Evaluation of inpatient dialectical-behavioral therapy for borderline personalitydisorder – a prospective study. Behav Res Ther2000;38:875–87.
92. Bohus M, Haaf B, Simms T, Limberger MF,Schmahl C, Unckel C, et al. Effectiveness ofinpatient dialectical behavioral therapy forborderline personality disorder: a controlled trial.Behav Res Ther 2004;42:487–99.
93. Brobyn S, Goren S, Lego S. The borderlinepatient: systemic versus psychoanalytic approach.Arch Psychiatr Nurs 1987;1:172–82.
94. Battegay R, Klaui C. Analytically oriented grouppsychotherapy with borderline patients as long-term crisis management. Crisis: Journal of CrisisIntervention and Suicide 1986;7:94–110.
95. Buzov I, Persic-Brida M. Rascjepljivanje iprojektivna identifikacija u kratkoj dinamskojpsihoterapiji granicnog poremecaja licnosti[Splitting and projective identification in shortdynamic psychotherapy of patients with aborderline personality disorder]. SocijalnaPsihijatrija 1985;13:21–30.
96. Chiesa M, Drahorad C, Longo S. Earlytermination of treatment in personality disordertreated in a psychotherapy hospital: quantitativeand qualitative study. Br J Psychiatry2000;177:107–11.
97. Clarkin JF, Marziali E, Munroe-Blum H. Groupand family treatments for borderline personalitydisorder. Hospital and Community Psychiatry1991;42:1038–43.
98. Clarkin JF, Hull J, Yeomans F, Kakuma T, Cantor J.Antisocial traits as modifiers of treatment responsein borderline inpatients. Journal of PsychotherapyPractice and Research 1994;3:307–12.
99. Damman G, Kachele H. Outcomes ofpsychodynamic treatment of borderline disorders.Nervenheilkunde 2001;20:31–7.
100. Dolan B, Warren F, Norton K. Change inborderline symptoms one year after therapeuticcommunity treatment for severe personalitydisorder. Br J Psychiatry 1997;171:274–9.
101. Hirvas J. [Practical experience with psychoanalyticmanagement of borderline personality disorder][in Finnish]. Duodecim 1987;103:1387–90.
102. Karterud S, Kvarstein E, Pedersen G. Severelydisturbed borderline patients need more thanshort-term day hospital treatment. TherapeuticCommunities 2004;25:120–30.
References
60
103. Kent ML, Hartstone MD. Trial of a structuredcognitive behaviour group therapy program forpatients with borderline personality disorder. AustN Z J Psychiatry 2000;34:A36.
104. Kern RS, Kuehnel TG, Teuber J, Hayden JL.Multimodal cognitive-behavior therapy forborderline personality disorder with self-injuriousbehavior. Psychiatr Serv 1997;48:1131–3.
105. Koenigsberg HW. A comparison of hospitalizedand nonhospitalized borderline patients. Am JPsychiatry 1982;139:1292–7.
106. Lopez D, Cuevas P, Gomez A, Mendoza J.Transference-focused psychotherapy for borderlinepersonality disorder. A study with 44 femalepatients. Salud Mental 2004;27:44–54.
107. Meares R, Stevenson J, Comerford A.Psychotherapy with borderline patients: I. Acomparison between treated and untreatedcohorts. Aust N Z J Psychiatry 1999;33:467–72.
108. Pfitzer F, Rosen E, Esch E, Held T. [Inpatientpsychiatric treatment of borderline patients] [inGerman]. Nervenarzt 1990;61:294–300.
109. Rathus JH, Miller AL. Dialectical behavior therapyadapted for suicidal adolescents. Suicide Life ThreatBehav 2002;32:146–57.
110. Ryle A, Golynkina K. Effectiveness of time-limitedcognitive analytic therapy of borderline
personality disorder: factors associated withoutcome. British Journal of Medical Psychology2000;73:197–210.
111. Schane M, Kovel V. Family therapy in severeborderline personality disorders. InternationalJournal of Family Psychiatry 1988;9:241–58.
112. Ushijima S. [Treatment of patients with borderlinepersonality disorder with dynamic psychotherapy][in Japanese]. Seishin Shinkeigaku Zasshi1994;96:676–83.
113. Wildgoose A, Clarke S, Waller G. Treatingpersonality fragmentation and dissociation inborderline personality disorder: a pilot study ofthe impact of cognitive analytic therapy. BritishJournal of Medical Psychology 2001;74:47–55.
114. Carkhuff RR, Berenson BG. Teaching as treatment.Amherst, MA: Human Resources DevelopmentPress; 1976.
115. Richards A, Barkham B, Cahill J, Richards D,Williams C, Heywood P. PHASE: a randomised,controlled trial of supervised self-help cognitive-behavoural therapy in primary care. Br J Gen Pract2003;53:764–70.
116. Brazier JE, Kolotkin RL, Crosby RD, Williams GR.Estimating a preference-based single index for theImpact of Weight on Quality of Life-LiteInstrument (IWQOL-Lite) from the SF-6D. ValueHealth 2004;7:490–8.
Health Technology Assessment 2006; Vol. 10: No. 35
61
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
This appendix contains information on thesources and keyword strategies used in the
identification of studies.
Electronic databases searchedAddiction Abstracts
ASSIA (Applied Social Sciences Index andAbstracts)
CareData
CDSR (Cochrane Database of Systematic Reviews)
CENTRAL (Cochrane Central Register ofControlled Trials)
Diss Abs (Dissertation Abstracts)
EconLIT
EMBASE
IBSS (International Bibliography of the SocialSciences)
Index to Theses
ISIP (Institute for Science InformationProceedings)
MEDLINE
NHS DARE (NHS Database of Abstract of Reviewsof Effectiveness)
NHS EED (NHS Economic Evaluation Database)
HTA (Health Technology Assessment Database)
NCJRSA (National Criminal Justice ReferenceService Abstracts)
OHE HEED (Office of Health Economics HealthEconomic Evaluations Database)
PsycINFO
PUBMED
Soc Abs (Sociological Abstracts)
SSA (Social Services Abstracts)
UKOP (United Kingdom Official Publications)
WOS (Web of Science)
Sources consulted via the WorldWide WebAHRQ (Agency for Healthcare Research andQuality)
AIHW (Australian Institute of Health and Welfare)
AHFMR (Alberta Heritage Foundation for MedicalResearch)
APA (American Psychiatric Association)
APA (American Psychological Association)
Bandolier
BPD Research Foundation
BIGSPD (British and Irish Group for the Study ofPersonality Disorders)
Campbell Collaboration
CCOHTA (Canadian Co-ordinating Office forHealth Technology Assessment)
CCT (Controlled Clinical Trials)
CEMH (Centre for the Economics of MentalHealth)
CenterWatch
CHE (Centre for Health Economics)
Chestnut Lodge Hospital
CRD (Centre for Reviews and Dissemination)
DoH PRP (Department of Health Policy ResearchProgramme)
DACEHTA (Danish Centre for Evaluation andHealth Technology Assessment)
DPHE (Department of Public Health andEpidemiology, University of Birmingham)
DTB (Drug and Therapeutics Bulletin)
Harvard CEA Registry (Harvard Cost EffectivenessAnalysis Registry)
HCNA (Health Care Needs Assessmentepidemiological reviews)
HEBE (Health Boards Executive)
HERC (Health Economics Research Centre)
Health Technology Assessment 2006; Vol. 10: No. 35
63
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Appendix 1
Identification of studies
HERG (Health Economics Research Group)
HERU (Health Economics Research Unit)
Home Office
HSPSCB (High Security Psychiatric ServicesCommissioning Board)
HSRU (Health Services Research Unit)
ICSI (Institute for Clinical Services Improvement)
INAHTA Clearing House (International Networkof Associations for Health Technology Assessment)
Institute of Psychiatry
ISSPD (International Society for the Study ofPersonality Disorders)
MIHSR (Monash Institute for Health ServicesResearch)
MIND (National Association for Mental Health)
mRCT (Meta Registers of RCTs)
MSAC (Medical Services Advisory Committee)
NGC (National Guideline Clearinghouse)
NPC (National Prescribing Centre)
NCCHTA (National Co-ordinating Centre forHealth Technology Assessment)
NHS QIS (NHS Quality Improvement, Scotland)
NHS R&D Programmes (including forensic mentalhealth)
NHSC (National Horizon Scanning Centre)
NIH (National Institutes of Health)
NIH Clinical Trials Database
NIMHE (National Institute for Mental Health inEngland)
North of England Guidelines
NSF Mental Health (National Service Frameworkfor Mental Health)
NZHTA (New Zealand Health TechnologyAssessment)
PDI (Personality Disorders Institute)
PSSRU, Kent (Personal and Social ServicesResearch Unit)
RAND Corporation
RCP (Royal College of Physicians)
RCPsych (Royal College of Psychiatrists)
SBU (Swedish Health Technology Assessment)
SIGN (Scottish Intercollegiate Guidelines)
SPR (Society for Psychotherapy Research)
Thames Valley Initiative
Therapeutics Initiative (Vancouver)
WPA (World Psychiatric Association)
Appendix 1
64
All searches were undertaken in March 2005.
Addiction Abstracts1996 onwardsMetaPress version
Personality disorder or personality disorders orborderline
ASSIA1987 onwardsVia Cambridge Scientific Abstracts (CSA)
Clinical effectiveness searchLast Search Query: ((((DE="Borderline personalitydisorder") or (DE="Personality disorders") or(otherwise specified) or (axis ii) or(behavi*ral dyscontrol) or bpc or (cluster b) or(borderline and(personality or disorder*)) or (severe personalitydysfunction) or(unstable personality) or ((dissocial or dramatic oremotional* orerratic or flamboyant or impulsivity or instability)within 3(personality or disorder*))) and((DE=("Psychotherapy" or "Analyticalpsychotherapy" or "Child analyticalpsychotherapy" or "Art therapy" or"Behaviour therapy" or "Aversion therapy" or"Cognitive behaviourtherapy" or "Covert sensitization" or"Selfreevaluation therapy" or"Stress inoculation training" or "Verbal satiation" or"Contingencycontracts" or "Habit reversal" or "Implosivetherapy" or "Interruptionprompting" or "Stimulus control" or "Subconsciousretraining" or"Behavioural psychotherapy" or "Cognitivebehavioural psychotherapy" or"Bibliotherapy" or "Brief therapy" or "Solutionsbased brief therapy" or"Child psychotherapy" or "Posttraumatic childtherapy" or
"Psychoanalytic child psychotherapy" or "Cognitivepsychotherapy" or"Countertransference"or "Couple therapy" or "Systemic couple therapy"or "Dialogicalpsychotherapy" or "Drama therapy" or "Duotherapy" or "Existentialpsychotherapy" or "Experiential psychotherapy" or"Experimentalpsychotherapy" or "Family therapy" or "Behaviourfamily therapy" or"Brief family therapy" or "Cognitive behaviourfamily therapy" or"Contextual therapy" or "Developmental familytherapy" or "Family playtherapy" or "Medical family therapy" or "Multiplefamily therapy groups"or "Structural family therapy" or "Systemic familytherapy" or "Feministtherapy" or "Forensic psychotherapy" or "Gestalttherapy" or "Grouppsychotherapy" or "Analytical grouppsychotherapy" or "Sociotherapygroups" or "Forensic group psychotherapy" or"Psychodynamic grouppsychotherapy" or "Individual psychotherapy" or"Interpersonalpsychotherapy" or "Milieu therapy" or "Mother-Infant psychotherapy" or"Multimodal therapy" or "Music therapy" or"Primal therapy" or"Psychoanalytic supportive psychotherapy" or"Psychodrama" or"Psychodynamic therapy" or "Brief psychodynamictherapy" or"Psychosynthesis" or "Psychotherapeutictechniques" or "Mirroring" or"Rational-Emotive therapy" or "Reality therapy" or"Social economytherapy" or "Supportive psychotherapy" or"Therapeutic communities" or"Transactional analysis" or "Transference" or "Self-Object transference"or "Validation therapy")) or(DE=("Psychotherapeutic techniques" or"Mirroring")) or (DE=("Psychotherapists" or"Analyticalpsychotherapists"or "Child psychotherapists" or "Grouppsychotherapists" or "Analytical
Health Technology Assessment 2006; Vol. 10: No. 35
65
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Appendix 2
Database keyword strategies
group psychotherapists")) or (DE="Psychologicalservices") or(DE="Psychological intervention") or(DE=("Psychoanalysis" or"Castrationanxiety" or "Psychological splitting" or "Seductiontheory" or"Selfobjects" or "Selfpsychology")) or(DE=("Psychoanalysts" or "Socialwork psychoanalysts")) or (DE=("Group therapy"or "Brief group therapy"or "Cognitive group therapy" or "Feminist grouptherapy" or "Group focalconflict theory" or "Psychoanalytic group therapy"or "Psychoeducationalgroup therapy" or "Sensitivity training" or"Crosscultural sensitivitytraining")) or psychotherap* or boscot or cat orcbt or dbt or(democratic within 2 communit*) or (therapeuticcommunit*) or (hendersonhospital*) or psychoanaly* or psycho-analy* orpsycho-therap* or ipt ormact or popmact or linehan or stepps or (crisisintervention) or((therap* or treatment* or strateg* or approach*or system* orintervention* or program* or oriented or focus*or framework) within 2(analytic or autogenic or behavi*r* or bio-cognitive or biocognitive orbrief or dynamic* or cognitive or client cent*redor outpatient orindividual or validation or day patient ordialectic* or eclectic orexpressive or family or inpatient or insight orintensive orinterpersonal or interpretive or long term orlongterm or intermittentormanuali?ed or mentali?ation or partialhospitali?ation or psychodynamic*or psycho-dynamic* or supportive or talk* or timelimited or short termor transference or framework orpsychoeducational or psychological orpsychosocial)))))
Cost-effectiveness searchLast Search Query: ((DE="Borderline personalitydisorder") or(DE="Personality disorders") or (otherwisespecified) or (axis ii) or(behavi*ral dyscontrol) or bpc or (cluster b) or(borderline and(personality or disorder*)) or (severe personalitydysfunction) or
(unstable personality) or ((dissocial or dramatic oremotional* orerratic or flamboyant or impulsivity or instability)within 3(personality or disorder*))) and ((cost* oreconomic* or qaly*) or(quality adjusted))
CareData1993 onwardsElectronic Library for Social Care
Borderline in tiBorderline in abPersonality disorder not borderline in tiPersonality disorder not borderline in ab
Cochrane Library (CDSR andCENTRAL)Issue 1, 2005Wiley version
Borderline in All Fields and personality in All Fields
Dissertation Abstracts1861 onwardsProQuest
Borderline personality and (therapy or treatmentor psychotherapy)Borderline personality and (cost* or economic* orqaly* or quality adjusted)
EconLIT1969 onwardsSilverPlatter WebSPIRS Version 4.3
#1 borderline and (personality or disorder*)
EMBASE1980 onwardsSilverPlatter WebSPIRS Version 4.3
Clinical effectiveness search#44 #18 and #43#43 #19 or #20 or #21 or #22 or #23 or #24 or#25 or #26 or #27 or #28 or #29 or #30 or #31or #32 or #33 or #34 or #35 or #36 or #37 or#38 or #39 or #42#42 #40 near2 #41
Appendix 2
66
#41 analytic or autogenic or behavio?r* or bio-cognitive or biocognitive or brief or dynamic* orcognitive or client cent?red or outpatient orindividual or validation or day patient or dialectic*or eclectic or expressive or family or inpatient orinsight or intensive or interpersonal or interpretiveor long term or longterm or intermittent ormanuali?ed or mentali?ation or partialhospitali?ation or psychodynamic* or psycho-dynamic* or supportive or talk* or time limited orshort term or transference or framework orpsychoeducational or psychological or psychosocial#40 therap* or treatment* or strateg* orapproach* or system* or intervention* orprogram* or oriented or focus* or framework#39 crisis intervention*#38 counsel*#37 stepps#36 popmact#35 linehan#34 mact#33 ipt#32 psychoanaly* or psycho-analy*#31 henderson hospital*#30 therapeutic communit*#29 democratic near2 communit*#28 dbt#27 cbt#26 cat#25 boscot#24 psychotherap* or psycho-therap*#23 'counseling-' / all subheadings inDEM,DER,DRM,DRR#22 explode 'psychodynamics-' / all subheadingsin DEM,DER,DRM,DRR#21 'psychological-and-psychiatric-procedures-techniques-and-concepts' / all subheadings inDEM,DER,DRM,DRR#20 'psychoanalysis-' / all subheadings inDEM,DER,DRM,DRR#19 explode 'psychotherapy-' / all subheadings inDEM,DER,DRM,DRR#18 #1 or #2 or #3 or #4 or #5 or #6 or #7 or#8 or #9 or #10 or #11 or #12 or #13 or #14or #15 or #16 or #17#17 instability near3 (personality or disorder*)#16 impulsivity near3 (personality or disorder*)#15 flamboyant near3 (personality or disorder*)#14 erratic near3 (personality or disorder*)#13 emotional* near3 (personality or disorder*)#12 dramatic near3 (personality or disorder*)#11 dissocial near3 (personality or disorder*)#10 unstable personality#9 severe personality dysfunction#8 borderline and (personality or disorder*)#7 cluster b#6 bpc
#5 behavio?ral dyscontrol#4 axis ii#3 otherwise specified #2 'personality-disorder' / all subheadings inDEM,DER,DRM,DRR#1 'borderline-state' / all subheadings inDEM,DER,DRM,DRR
Cost-effectiveness search#20 #18 and #19#19 explode 'economic-aspect' / all subheadings inDEM,DER,DRM,DRR#18 #1 or #2 or #3 or #4 or #5 or #6 or #7 or#8 or #9 or #10 or #11 or #12 or #13 or #14or #15 or #16 or #17#17 instability near3 (personality or disorder*)#16 impulsivity near3 (personality or disorder*)#15 flamboyant near3 (personality or disorder*)#14 erratic near3 (personality or disorder*)#13 emotional* near3 (personality or disorder*)#12 dramatic near3 (personality or disorder*)#11 dissocial near3 (personality or disorder*)#10 unstable personality#9 severe personality dysfunction#8 borderline and (personality or disorder*)#7 cluster b#6 bpc#5 behavio?ral dyscontrol#4 axis ii#3 otherwise specified #2 'personality-disorder' / all subheadings inDEM,DER,DRM,DRR#1 'borderline-state' / all subheadings in DEM,D
IBSS1951 onwardsVia BIDS (Bath Information and Data Services)
Clinical effectiveness search((Borderline and (personality or disorder*)) orpersonality disorder*) and (treatment* or therap*or psychotherapy*)
Cost-effectiveness search((Borderline and (personality or disorder*)) orpersonality disorder*) and (cost* or economic* orqaly* or quality adjusted)
Index to Theses1716 onwardsExpert Information
(borderline and (personality or disorder*)) or(personality disorder*)
Health Technology Assessment 2006; Vol. 10: No. 35
67
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
ISIP1990 onwardsWeb of Knowledge
Clinical effectiveness search#1 TS=borderline and TS=(personality or
disorder*)#2 TS=(otherwise specified or axis ii or
behavio*ral dyscontrol or bpc or borderlinestate or cluster b or severe personalitydysfunction or unstable personality orpersonality disorder*)
#3 TS=(dissocial* or dramatic or emotional* orerratic or flamboyant or impulsivity orinstability) same TS=(personality or disorder*)
#4 #1 or #2 or #3#5 TS=(psychotherap* or psycho-therap* or
psychoanaly* or psycho-analy* or counsel*ingor boscot or cat or cbt or democraticcommunit* or therapeutic communit* orhenderson hospital* or ipt or mact or linehanor popmact or stepps or crisis intervention)
#6 TS=(therap* or treatment*) sameTS=(analytic or autogenic or behavio*r* orbio-cognitive or biocognitive or brief ordynamic* or cognitive or client cent*red orindividual or validation or dialectic* or eclecticor expressive or family or group or insight orintensive or interpersonal or interpretive orlong term or longterm or intermittent ormanuali?ed or mentali?ation or partialhospitali?ation or psychodynamic* or psycho-dynamic* or supportive or talk* or timelimited or short term or transference orframework or psychoeducational orpsychological or psychosocial)
#7 #5 or #6#8 #4 and #7
Cost-effectiveness search#1 TS=borderline and TS=(personality or
disorder*)#2 TS=(otherwise specified or axis ii or
behavio*ral dyscontrol or bpc or borderlinestate or cluster b or severe personalitydysfunction or unstable personality orpersonality disorder*)
#3 TS=(dissocial* or dramatic or emotional* orerratic or flamboyant or impulsivity orinstability) same TS=(personality or disorder*)
#4 #1 or #2 or #3#5 TS=(cost* or economic* or qaly* or quality
adjusted)#6 #4 and #5
MEDLINE1966 onwardsOvid Online version 9.3
Clinical effectiveness search1 Borderline Personality Disorder/2 Personality Disorders/3 otherwise specified.tw.4 axis ii.tw.5 behavio?ral dyscontrol.tw.6 BPC.tw.7 cluster b.tw.8 (borderline and (personality or disorder$)).tw.9 severe personality dysfunction.tw.
10 unstable personality.tw.11 (dissocial adj3 (personality or disorder$)).tw.12 (dramatic adj3 (personality or disorder$)).tw.13 (emotional$ adj3 personality disorder$).tw.14 (erratic adj3 (personality or disorder$)).tw.15 (flamboyant adj3 (personality or disorder$)).tw.16 (impulsivity adj3 (personality or disorder$)).tw.17 (instability adj3 (personality or disorder$)).tw.18 or/1-1719 exp Psychotherapy/20 Psychoanalysis/21 exp Psychological Techniques/22 Counseling/23 (psychotherap$ or psycho-therap$).tw.24 boscot.tw.25 cat.tw.26 cbt.tw.27 dbt.tw.28 (democratic adj2 communit$).tw.29 therapeutic communit$.tw.30 henderson hospital$.tw.31 (psychoanaly$ or psycho-analy$).tw.32 ipt.tw.33 mact.tw.34 linehan.tw.35 popmact.tw.36 stepps.tw.37 counsel$.tw.38 crisis intervention.tw.39 (therap$ or treatment$ or strateg$ or
approach$ or system$ or intervention$ orprogram$ or oriented or focus$ orframework).tw.
40 (analytic or autogenic or behavio?r$ or bio-cognitive or biocognitive or brief or dynamic$or cognitive or client cent?red or outpatient orindividual or validation or day patient ordialectic$ or eclectic or expressive or family orinpatient or insight or intensive orinterpersonal or interpretive or long term orlongterm or intermittent or manuali?ed ormentali?ation or partial hospitali?ation or
Appendix 2
68
psychodynamic$ or psycho-dynamic$ orsupportive or talk$ or time limited or shortterm or transference or framework orpsychoeducational or psychological orpsychosocial).tw.
41 ((therap$ or treatment$ or strateg$ orapproach$ or system$ or intervention$ orprogram$ or oriented or focus$ or framework)adj2 (analytic or autogenic or behavio?r$ or bio-cognitive or biocognitive or brief or dynamic$ orcognitive or client cent?red or outpatient orindividual or validation or day patient ordialectic$ or eclectic or expressive or family orinpatient or insight or intensive or interpersonalor interpretive or long term or longterm orintermittent or manuali?ed or mentali?ation orpartial hospitali?ation or psychodynamic$ orpsycho-dynamic$ or supportive or talk$ or timelimited or short term or transference orframework or psychoeducational orpsychological or psychosocial)).tw.
42 or/19-38,4143 18 and 42
Cost-effectiveness search1 Borderline Personality Disorder/2 Personality Disorders/3 otherwise specified.tw.4 axis ii.tw.5 behavio?ral dyscontrol.tw.6 BPC.tw.7 cluster b.tw.8 (borderline and (personality or disorder$)).tw.9 severe personality dysfunction.tw.
10 unstable personality.tw.11 (dissocial adj3 (personality or disorder$)).tw.12 (dramatic adj3 (personality or disorder$)).tw.13 (emotional$ adj3 personality disorder$).tw.14 (erratic adj3 (personality or disorder$)).tw.15 (flamboyant adj3 (personality or disorder$)).tw.16 (impulsivity adj3 (personality or disorder$)).tw.17 (instability adj3 (personality or disorder$)).tw.18 or/1-1719 Economics/20 exp "Costs and cost analysis"/21 Economic value of life/22 exp Economics, hospital/23 exp Economics, medical/24 Economics, nursing/25 exp models, economic/26 Economics, pharmaceutical/27 exp "Fees and charges"/28 exp Budgets/29 ec.fs.30 (cost or costs or costed or costly or costing$).tw.31 (economic$ or pharmacoeconomic$ or price$
or pricing).tw.
32 Quality-adjusted life years/33 (qaly or qalys).af.34 (quality adjusted life year or quality adjusted
life years).af.35 or/19-3436 18 and 35
NCJRSA1975 onwardsVia CSA
Clinical effectiveness search(((personality disorder*) or (borderline and(personality ordisorder*))) and (treatment* or therap* orpsychotherap*))
Cost-effectiveness search((personality disorder*) or (borderline and(personality ordisorder*))) and ((cost* or economic* or qaly*) or(quality adjusted))
NHS DARE, NHS EED, HTADate coverage not known (approx. 1994–2005)CRD website version
Borderline and personality
OHE HEEDDate coverage not knownCD-ROM version
Borderline or personality
PsycINFO1887 onwardsSilverPlatter WebSPIRS Version 4.3
Clinical effectiveness search#44 #19 and #43#43 #20 or #21 or #22 or #23 or #24 or #25 or#26 or #27 or #28 or #29 or #30 or #31 or #32or #33 or #34 or #35 or #36 or #37 or #38 or#39 or #42#42 #40 near2 #41#41 analytic or autogenic or behavio?r* or bio-cognitive or biocognitive or brief or dynamic* orcognitive or client cent?red or outpatient or
Health Technology Assessment 2006; Vol. 10: No. 35
69
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
individual or validation or day patient ordialectic* or eclectic or expressive or family orinpatient or insight or intensive or interpersonalor interpretive or long term or longterm orintermittent or manuali?ed or mentali?ation orpartial hospitali?ation or psychodynamic* orpsycho-dynamic* or supportive or talk* or timelimited or short term or transference orframework or psychoeducational or psychologicalor psychosocial#40 therap* or treatment* or strateg* orapproach* or system* or intervention* orprogram* or oriented or focus* or framework#39 crisis intervention*#38 counsel*#37 stepps#36 popmact#35 linehan#34 mact#33 ipt#32 psychoanaly* or psycho-analy*#31 henderson hospital*#30 therapeutic communit*#29 democratic near2 communit*#28 dbt#27 cbt#26 cat#25 boscot#24 psychotherap* or psycho-therap*#23 explode 'Psychotherapeutic-Processes' inMJ,MN#22 explode 'Psychotherapeutic-Techniques' inMJ,MN#21 'Cognitive-Therapy' in MJ,MN#20 explode 'Psychotherapy-' in MJ,MN#19 #1 or #2 or #3 or #4 or #5 or #6 or #7 or#8 or #9 or #10 or #11 or #12 or #13 or #14or #15 or #16 or #17 or #18#18 instability near3 (personality or disorder*)#17 impulsivity near3 (personality or disorder*)#16 flamboyant near3 (personality or disorder*)#15 erratic near3 (personality or disorder*)#14 emotional* near3 (personality or disorder*)#13 dramatic near3 (personality or disorder*)#12 dissocial near3 (personality or disorder*)#11 unstable personality#10 severe personality dysfunction#9 borderline and (personality or disorder*)#8 cluster b#7 bpc#6 behavio?ral dyscontrol#5 axis ii#4 otherwise specified#3 'Personality-Disorders' in MJ,MN#2 'Borderline-States' in MJ,MN#1 'Borderline-Personality' in MJ,MN
Cost-effectiveness search#19 and #20#20 cost* or economic* or qaly* or qualityadjusted#19 #1 or #2 or #3 or #4 or #5 or #6 or #7 or#8 or #9 or #10 or #11 or #12 or #13 or #14or #15 or #16 or #17 or #18#18 instability near3 (personality or disorder*)#17 impulsivity near3 (personality or disorder*)#16 flamboyant near3 (personality or disorder*)#15 erratic near3 (personality or disorder*)#14 emotional* near3 (personality or disorder*)#13 dramatic near3 (personality or disorder*)#12 dissocial near3 (personality or disorder*)#11 unstable personality#10 severe personality dysfunction#9 borderline and (personality or disorder*)#8 cluster b#7 bpc#6 behavio?ral dyscontrol#5 axis ii#4 otherwise specified#3 'Personality-Disorders' in MJ,MN#2 'Borderline-States' in MJ,MN#1 'Borderline-Personality' in MJ,MN
PUBMEDSeptember 2004 onwardsVersion not known
Clinical effectiveness search#5 Search #3 and #4 Limits: 180 Days#4 Search treatment* or therap* or
psychotherap* or psycho-therap*#3 Search #1 or #2#2 Search personality disorder*#1 Search borderline and (personality or
disorder*)
Cost-effectiveness search#5 Search #3 and #5 Limits: 180 Days#4 Search cost* or economic* or qaly* or quality
adjusted#3 Search #1 or #2#2 Search personality disorder*#1 Search borderline and (personality or
disorder*)
SSA1980 onwardsVia CSA
Appendix 2
70
Clinical effectiveness search(((personality disorder*) or (borderline and(personality ordisorder*))) and (treatment* or therap* orpsychotherap*))
Cost-effectiveness search((personality disorder*) or (borderline and(personality ordisorder*))) and ((cost* or economic* or qaly*) or(quality adjusted))
Soc Abs1963 onwardsVia CSA
Clinical effectiveness search(((personality disorder*) or (borderline and(personality ordisorder*))) and (treatment* or therap* orpsychotherap*))
Cost-effectiveness search((personality disorder*) or (borderline and(personality ordisorder*))) and ((cost* or economic* or qaly*) or(quality adjusted))
UKOP1980 onwardsThe Stationery Office
Borderline or personality
WOS1981 onwardsWOK
Clinical effectiveness search
#1 TS=borderline and TS=(personality ordisorder*)
#2 TS=(otherwise specified or axis ii orbehavio*ral dyscontrol or bpc or borderlinestate or cluster b or severe personalitydysfunction or unstable personality orpersonality disorder*)
#3 TS=(dissocial* or dramatic or emotional* orerratic or flamboyant or impulsivity orinstability) same TS=(personality or disorder*)
#4 #1 or #2 or #3#5 TS=(psychotherap* or psycho-therap* or
psychoanaly* or psycho-analy* or counsel*ingor boscot or cat or cbt or democraticcommunit* or therapeutic communit* orhenderson hospital* or ipt or mact or linehanor popmact or stepps or crisis intervention)
#6 TS=(therap* or treatment*) sameTS=(analytic or autogenic or behavio*r* orbio-cognitive or biocognitive or brief ordynamic* or cognitive or client cent*red orindividual or validation or dialectic* or eclecticor expressive or family or group or insight orintensive or interpersonal or interpretive orlong term or longterm or intermittent ormanuali?ed or mentali?ation or partialhospitali?ation or psychodynamic* or psycho-dynamic* or supportive or talk* or timelimited or short term or transference orframework or psychoeducational orpsychological or psychosocial)
#7 #5 or #6#8 #4 and #7
Cost-effectiveness search#1 TS=borderline and TS=(personality or
disorder*)#2 TS=(otherwise specified or axis ii or
behavio*ral dyscontrol or bpc or borderlinestate or cluster b or severe personalitydysfunction or unstable personality orpersonality disorder*)
#3 TS=(dissocial* or dramatic or emotional* orerratic or flamboyant or impulsivity orinstability) same TS=(personality or disorder*)
#4 #1 or #2 or #3#5 TS=(cost* or economic* or qaly* or quality
adjusted)#6 #4 and #5
Health Technology Assessment 2006; Vol. 10: No. 35
71
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Health Technology Assessment 2006; Vol. 10: No. 35
73
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Appendix 3
Evidence tables for BPD studies
This appendix contains the evidence tables with data extracted from the 10 studies included in thisreview.
TABLE 17 Studies included in the review
Study Funding, location Intervention Study type Patient population
Bateman and NR, UK Partial hospitalisation RCT Patients with severe Fonagy, 199951 (MBT) parasuicidal BPD
Koons et al., VA Research Advisory Group grant, DBT RCT Women Veterans with 200152 USA BPD
Linehan et al., Grant MH34486, National Institute of DBT RCT Chronically parasuicidal 199153 Mental Health, Bethesda, USA women with BPD
Linehan et al., Grant DA08674, National Institute of DBT RCT Substance-abusing 199954 Drug Abuse, Bethesda, USA women with BPD
Linehan et al., Grant DA 08674, National Institute of DBT RCT Heroin-dependent 200255 Drug Abuse, National Institute of women with BPD
Health, USA
Munroe-Blum Ontario Mental Health Foundation; Time-limited IGP RCT Patients with BPDand Marziali, Grant 88-87-89, grants 6606-4232-MH 199556 and 6606-4232-64, National Health
Research and Development Program, Canada
Turner, 200057 NR, USA DBT-orientated RCT Patients with BPDtherapy
Tyrer et al., Medical Research Council, UK MACT RCT Patients with recurrent 200358 DSH (including BPD)
van den Bosch Province of Noord-Holland and ZAO DBT RCT Female BPI with or et al., 200259 Health Insurance Company in without comorbid (Verheul et al., Amsterdam, The Netherlands substance abuse200364)
Wilberg et al., Norwegian Research Council, S and JP Group psychotherapy Non-random Patients with BPD199860 Sommer’s Foundation and controlled
Maja-Jonn-Nilsen’s Foundations study (naturalistic follow-up study)
Appendix 3
74 TA
BLE
18
Stud
y ch
arac
teris
tics:
RCT
s
Stud
yD
escr
ipti
on o
f psy
chot
hera
pySt
udy
qual
ity
Co-
ther
apy
or
Com
para
tor
Sam
ple
size
med
icat
ion
Bate
man
and
Fon
agy,
1999
51M
BT c
onsis
ted
of: (
1) o
nce-
wee
kly
indi
vidu
alps
ycho
anal
ytic
al p
sych
othe
rapy
; (2)
thr
ice-
wee
kly
grou
p an
alyt
ical
psy
chot
hera
py;
(3) o
nce-
a-w
eek
expr
essiv
e th
erap
y or
ient
ated
tow
ards
psy
chod
ram
a te
chni
ques
; (4)
a w
eekl
yco
mm
unity
mee
ting;
(5) o
nce-
a-m
onth
mee
ting
with
cas
e ad
min
istra
tor;
(6) m
edic
atio
n re
view
by t
he r
esid
ent
psyc
hiat
rist
Rand
omisa
tion
met
hod:
NR
Blin
ded
asse
ssm
ent:
NR
Pow
er c
alcu
latio
n: N
RLo
ss t
o fo
llow
-up:
yes
Lack
ner Q
ualit
y Ra
ting
Scal
e ite
ms
Patie
nt s
elec
tion:
12/
14In
terv
entio
n: 4
/8O
utco
me
mea
sure
men
t: 7/
12D
ata
pres
enta
tion
and
stat
istic
alan
alys
is: 6
/12
Inve
stig
ator
: 0/4
Misc
ella
neou
s: 5
/8To
tal:
34/5
8
Ant
idep
ress
ants
and
antip
sych
otic
dru
gspr
escr
ibed
as
appr
opria
te;
poly
phar
mac
y w
asdi
scou
rage
d. U
ncle
arw
heth
er t
his
rela
tes
tobo
th g
roup
s
TAU
: (1)
reg
ular
psyc
hiat
ric r
evie
ww
ith a
sen
ior
psyc
hiat
rist
whe
nne
cess
ary;
(2
) inp
atie
ntad
miss
ion
asap
prop
riate
; (3
) out
patie
nt a
ndco
mm
unity
follo
w-u
p(e
very
2 w
eeks
by
aco
mm
unity
psyc
hiat
ric n
urse
)
Rand
omise
d: n
= 4
4 (M
BT n
= 2
2,
TAU
n =
22)
Ana
lyse
d: n
= 3
8(M
BT n
= 1
9,
TAU
n =
19)
Koon
s et
al.,
200
152D
BT b
ased
on
bala
nce
and
synt
hesis
of
acce
ptan
ce o
f the
pat
ient
as
he/s
he is
cur
rent
ly,us
ing
valid
atio
n st
rate
gies
, with
the
att
empt
to
get
the
patie
nt t
o ch
ange
, usin
g be
havi
our
ther
apy
stra
tegy
(for
det
ails
see
Line
han
et a
l.,19
9153
)
Rand
omisa
tion
met
hod:
com
pute
rised
ran
dom
num
ber
gene
ratio
nBl
inde
d as
sess
men
t: N
RPo
wer
cal
cula
tion:
yes
Loss
to
follo
w-u
p: y
es
Lack
ner Q
ualit
y Ra
ting
Scal
e ite
ms
Patie
nt s
elec
tion:
12/
14In
terv
entio
n: 4
/8O
utco
me
mea
sure
men
t: 8/
12D
ata
pres
enta
tion
and
stat
istic
alan
alys
is: 8
/12
Inve
stig
ator
: 2/4
Misc
ella
neou
s: 3
/8To
tal:
37/5
8
Phar
mac
othe
rapy
: SSR
Is,
moo
d st
abili
ser
and/
orlo
w-d
ose
neur
olep
tic
TAU
: 60
min
utes
of
wee
kly
indi
vidu
alth
erap
y w
ith a
clin
icia
n. P
atie
nts
wer
e al
so o
ffere
don
e or
mor
e of
seve
ral s
uppo
rtiv
ean
dps
ycho
educ
atio
nal
grou
ps t
hat
they
coul
d at
tend
Rand
omise
d n
= 2
8(D
BT n
= 1
3,
TAU
n =
15)
Ana
lyse
d: n
= 2
0(D
BT n
= 1
0,TA
U n
= 1
0) cont
inue
d
Health Technology Assessment 2006; Vol. 10: No. 35
75
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TA
BLE
18
Stud
y ch
arac
teris
tics:
RCT
s (c
ont’d
)
Stud
yD
escr
ipti
on o
f psy
chot
hera
pySt
udy
qual
ity
Co-
ther
apy
or
Com
para
tor
Sam
ple
size
med
icat
ion
Line
han
et a
l., 1
99153
DBT
is a
man
ualis
ed 1
2-m
onth
tre
atm
ent
that
com
bine
s fo
ur m
odul
es: (
1) w
eekl
y in
divi
dual
cogn
itive
beh
avio
ural
psy
chot
hera
py s
essio
nsw
ith t
he p
rimar
y th
erap
ist; (
2) w
eekl
y sk
ills
trai
ning
gro
ups
last
ing
for
2–2.
5h
per
sess
ion;
(3) w
eekl
y su
perv
ision
and
con
sulta
tion
mee
tings
for
the
ther
apist
s; a
nd (4
) tel
epho
neco
nsul
tatio
n, w
here
pat
ient
s ar
e en
cour
aged
to
obta
in c
oach
ing
in t
he a
pplia
nce
of n
ew e
ffect
ive
skill
s by
tel
epho
ning
the
ir pr
imar
y th
erap
ists
eith
er d
urin
g or
out
side
offic
e ho
urs.
Indi
vidu
alth
erap
y fo
cuse
s pr
imar
ily o
n m
otiv
atio
nal i
ssue
s,in
clud
ing
the
mot
ivat
ion
to s
tay
aliv
e an
d to
sta
yin
tre
atm
ent.
Gro
up t
hera
py t
each
es s
elf-
regu
latio
n an
d ch
ange
ski
lls, a
nd s
elf a
nd o
ther
acce
ptan
ce s
kills
. Am
ong
the
cent
ral p
rinci
ples
isD
BT’s
sim
ulta
neou
s fo
cus
on a
pply
ing
both
acce
ptan
ce a
nd v
alid
atio
n st
rate
gies
and
cha
nge
(beh
avio
ural
) str
ateg
ies
to a
chie
ve a
syn
thet
ic(d
iale
ctic
al) b
alan
ce in
pat
ient
func
tioni
ng
Rand
omisa
tion
met
hod:
part
icip
ants
wer
e m
atch
ed o
nth
e nu
mbe
r of
life
time
para
suic
ides
and
psy
chia
tric
hosp
italis
atio
n, a
ge a
nd g
ood
vspo
or c
linic
al p
rogn
osis
Blin
ded
asse
ssm
ent:
blin
d in
depe
nden
t as
sess
orPo
wer
cal
cula
tion:
NR
Loss
to
follo
w-u
p: y
es
Lack
ner Q
ualit
y Ra
ting
Scal
e ite
ms
Patie
nt s
elec
tion:
9/1
4In
terv
entio
n: 4
/8O
utco
me
mea
sure
men
t: 9/
12D
ata
pres
enta
tion
and
stat
istic
alan
alys
is: 6
/12
Inve
stig
ator
: 2/4
Misc
ella
neou
s: 4
/8To
tal:
34/5
8
Part
icip
ants
had
to
term
inat
e ot
her
indi
vidu
alps
ycho
ther
apie
s an
dta
per
off p
sych
otro
pic
med
icat
ion
(ant
idep
ress
ants
,ne
urol
eptic
s, a
nxio
lytic
s,an
ticon
vulsa
nts,
seda
tives
, lith
ium
)
TAU
: con
trol
part
icip
ants
wer
egi
ven
alte
rnat
ive
ther
apy
refe
rral
s,us
ually
by
the
orig
inal
refe
rral
sou
rce,
from
whi
ch t
hey
coul
dch
oose
Rand
omise
d: n
=63
(DBT
n=
32,
TAU
n=
31)
Ana
lyse
d: n
=44
(DBT
n=
22,
TAU
n=
22)
Follo
wed
up:
n=
39(D
BT n
=19
, TA
U n
=20
) cont
inue
d
Appendix 3
76 TA
BLE
18
Stud
y ch
arac
teris
tics:
RCT
s (c
ont’d
)
Stud
yD
escr
ipti
on o
f psy
chot
hera
pySt
udy
qual
ity
Co-
ther
apy
or
Com
para
tor
Sam
ple
size
med
icat
ion
Line
han
et a
l., 1
99954
DBT
with
rep
lace
men
t m
edic
atio
n. “
Seve
ral
mod
ifica
tions
and
add
ition
s w
ere
adde
d to
stan
dard
DBT
for
use
with
sub
stan
ce a
busin
gpo
pula
tion.
A n
ew s
et o
f “at
tach
men
t st
rate
gies
”w
ere
adde
d to
DBT
. The
se s
trat
egie
s co
nsist
edof
a s
et o
f org
anise
d in
terv
entio
ns d
esig
ned
toin
crea
se t
he p
ositi
ve b
alan
ce o
f the
rapy
and
the
ther
apist
, as
wel
l as
to r
each
out
to
and
bric
kba
ck ‘l
ost’
patie
nts.
” A
‘tra
nsiti
onal
mai
nten
ance
’re
plac
emen
t m
edic
atio
n ph
arm
acot
hera
pypr
otoc
ol w
as a
dded
for
indi
vidu
als
with
stim
ulan
t or
opi
ate
depe
nden
ce
Rand
omisa
tion
met
hod:
part
icip
ants
wer
e m
atch
ed o
nag
e, s
ever
ity o
f dru
gde
pend
ence
, rea
dine
ss t
och
ange
and
glo
bal a
djus
tmen
t Bl
inde
d as
sess
men
t: bl
ind
inde
pend
ent
asse
ssor
Pow
er c
alcu
latio
n: N
RLo
ss t
o fo
llow
-up:
Yes
Lack
ner Q
ualit
y Ra
ting
Scal
e ite
ms
Patie
nt s
elec
tion:
12/
14In
terv
entio
n: 4
/8O
utco
me
mea
sure
men
t: 8/
12D
ata
pres
enta
tion
and
stat
istic
alan
alys
is: 8
/12
Inve
stig
ator
: 2/4
Misc
ella
neou
s: 6
/8To
tal:
40/5
8
Seve
ral m
odifi
catio
ns a
ndad
ditio
ns w
ere
adde
d to
stan
dard
DBT
for
use
with
sub
stan
ce-a
busin
gpo
pula
tion
Four
mon
ths
of d
rug
mai
nten
ance
(to
prov
ide
time
for
skill
sac
quisi
tion)
, 4 m
onth
s of
drug
tap
erin
g (fo
r sk
ills
stre
ngth
enin
g) a
nd 4
mon
ths
of n
o dr
ugre
plac
emen
t (fo
r sk
ills
gene
ralis
atio
n). I
llici
tst
imul
ants
wer
e re
plac
edw
ith m
etha
done
. The
max
imum
dos
e of
met
hylp
heni
date
giv
enw
as 2
0 m
g da
ily, a
nd t
hem
axim
um d
ose
ofm
etha
done
giv
en w
as
70 m
g da
ily
TAU
par
ticip
ants
wer
e ei
ther
ref
erre
dto
alte
rnat
ive
subs
tanc
e ab
use
and/
or m
enta
l hea
lthco
unse
llors
and
prog
ram
mes
in t
heco
mm
unity
, or
allo
wed
to
cont
inue
with
the
ir in
divi
dual
psyc
hoth
erap
ists
ifth
ey w
ere
rece
ivin
gse
rvic
e at
the
tim
e of
pret
reat
men
tas
sess
men
t
n=
28(D
BT n
=12
, TA
U n
=16
) cont
inue
d
Health Technology Assessment 2006; Vol. 10: No. 35
77
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TA
BLE
18
Stud
y ch
arac
teris
tics:
RCT
s (c
ont’d
)
Stud
yD
escr
ipti
on o
f psy
chot
hera
pySt
udy
qual
ity
Co-
ther
apy
or
Com
para
tor
Sam
ple
size
med
icat
ion
Line
han
et a
l., 2
00255
DBT
for
subs
tanc
e ab
user
s. T
reat
men
t re
quire
sa
synt
hesis
of v
alid
atio
n to
str
engt
hen
self-
trus
tva
lidat
ion,
red
uce
fear
of s
elf-
gene
rate
d(in
trin
sical
ly m
otiv
ated
) res
pons
e pa
tter
ns, a
ndm
aint
ain
wor
king
alli
ance
, beh
avio
ur t
hera
py t
ote
ach
emot
iona
l reg
ulat
ion,
sel
f-va
lidat
ion
and
skilf
ul r
espo
nses
to
prob
lem
s in
livi
ng, a
nd t
oex
tingu
ish o
r pu
nish
BPD
beh
avio
urs
(incl
udin
gill
icit
drug
use
), an
d di
alec
tics
to c
ount
erac
t rig
idan
d ex
trem
e re
spon
se p
atte
rns
Rand
omisa
tion
met
hod:
part
icip
ants
wer
e m
atch
ed o
nse
verit
y, d
rug
depe
nden
ce,
pres
ence
/abs
ence
of P
D a
ndgl
obal
ass
essm
ent
of fu
nctio
ning
Blin
ded
asse
ssm
ent:
blin
din
depe
nden
t as
sess
orPo
wer
cal
cula
tion:
NR
Loss
to
follo
w-u
p: y
es
Lack
ner Q
ualit
y Ra
ting
Scal
e ite
ms
Patie
nt s
elec
tion:
11/
14In
terv
entio
n: 4
/8O
utco
me
mea
sure
men
t: 8/
12D
ata
pres
enta
tion
and
stat
istic
alan
alys
is: 8
/12
Inve
stig
ator
: 4/4
Misc
ella
neou
s: 6
/8To
tal:
41/5
8
Opi
ate-
repl
acem
ent
med
icat
ion:
LA
AM
(met
hado
ne a
ltern
ativ
e);
dosin
g: 9
0/90
/130
mg;
the
max
imum
sch
edul
ew
as 1
10/1
10/1
80m
g
CVT
+12
S re
quire
s a
synt
hesis
of v
alid
atio
nto
str
engt
hen
self-
trus
t, re
duce
fear
of
self-
gene
rate
dre
spon
se p
atte
rns,
decr
ease
aro
usal
,in
crea
se t
heex
perie
nce
ofco
ntro
l, an
d m
aint
ain
wor
king
alli
ance
and
fello
wsh
ip o
f sim
ilar
com
mun
ity s
uch
as12
-Ste
p to
val
idat
ebo
th s
ense
of s
elf a
ndre
cove
ry
n=
23(D
BT n
=11
,C
VT+
12S
n=
12)
Mun
roe-
Blum
and
Mar
zial
i, 19
9556
IGP
base
d on
man
ualis
ed t
rain
ing
proc
edur
e; it
sst
rate
gies
ref
lect
an
inte
rper
sona
l gro
uptr
eatm
ent
appr
oach
gea
red
to a
ddre
ssin
g a
cent
ral f
eatu
re o
f the
BPD
Rand
omisa
tion
met
hod:
rand
omise
d ov
er fi
ve w
aves
(unc
lear
)Bl
inde
d as
sess
men
t: in
depe
nden
tas
sess
orPo
wer
cal
cula
tion:
yes
Loss
to
follo
w-u
p: y
es
Lack
ner Q
ualit
y Ra
ting
Scal
e ite
ms
Patie
nt s
elec
tion:
11/
14In
terv
entio
n: 4
/8O
utco
me
mea
sure
men
t: 7/
12D
ata
pres
enta
tion
and
stat
istic
alan
alys
is: 6
/12
Inve
stig
ator
: 2/4
Misc
ella
neou
s: 4
/8To
tal:
34/5
8
NR
Indi
vidu
alps
ycho
dyna
mic
psyc
hoth
erap
yco
nsist
ing
of o
pen-
ende
d, in
divi
dual
,dy
nam
icps
ycho
ther
apy;
the
typi
cal m
odel
of
trea
tmen
t of
fere
d to
patie
nts
with
BPD
Rand
omise
d n
=79
(IGP
n=
38,
indi
vidu
al n
=41
)
Ana
lyse
d: n
=48
(IGP
n=
22;
indi
vidu
al n
=26
)
cont
inue
d
Appendix 3
78 TA
BLE
18
Stud
y ch
arac
teris
tics:
RCT
s (c
ont’d
)
Stud
yD
escr
ipti
on o
f psy
chot
hera
pySt
udy
qual
ity
Co-
ther
apy
or
Com
para
tor
Sam
ple
size
med
icat
ion
Turn
er, 2
00057
DBT
-orie
ntat
ed t
hera
py w
hich
had
tw
om
odifi
catio
ns m
ade
to L
ineh
an’s
DBT
app
roac
h:(1
) psy
chod
ynam
ic t
echn
ique
s w
ere
inco
rpor
ated
to
conc
eptu
alise
the
pat
ient
’sbe
havi
oura
l, em
otio
nal a
nd c
ogni
tive
rela
tions
hip
sche
ma;
(2) i
n or
der
to k
eep
trea
tmen
tco
nditi
ons
equa
l with
reg
ard
to c
linic
al c
onta
ctho
urs,
the
aut
hors
did
not
run
a s
epar
ate
DBT
skill
s tr
aini
ng g
roup
, but
pro
vide
d sk
ills
durin
gth
e co
urse
of i
ndiv
idua
l the
rapy
Rand
omisa
tion
met
hod:
NR
Blin
ded
asse
ssm
ent:
blin
din
depe
nden
t as
sess
orPo
wer
cal
cula
tion:
NR
Loss
to
follo
w-u
p: n
one
Lack
ner Q
ualit
y Ra
ting
Scal
e ite
ms
Patie
nt s
elec
tion:
12/
14In
terv
entio
n: 4
/8O
utco
me
mea
sure
men
t: 8/
12D
ata
pres
enta
tion
and
stat
istic
alan
alys
is: 8
/12
Inve
stig
ator
: 0/4
Misc
ella
neou
s: 5
/8To
tal:
37/5
8
19 p
atie
nts
wer
e ta
king
pres
crib
ed p
sych
otro
pic
med
icat
ions
at
the
begi
nnin
g of
the
stu
dy.
The
re w
as n
o co
nsist
ent
patt
ern
of m
edic
atio
nty
pes
CC
T b
ased
on
Car
nuff
’s m
odel
s of
CC
T, e
mph
asisi
ngem
phat
icun
ders
tand
ing
of t
hepa
tient
’s s
ense
of
alon
enes
s an
dpr
ovid
ing
asu
ppor
tive
atm
osph
ere
for
indi
vidu
atio
n. C
CT
did
not
use
ast
ruct
ured
age
nda.
Inst
ead,
the
rapi
sts
inst
ruct
ed p
atie
nts
toex
pres
s w
hat
was
on
thei
r m
inds
at
each
sess
ion
n=
24(D
BT n
=12
, TA
U n
=12
)
Tyre
r et
al.,
200
358M
AC
T is
70-
page
boo
klet
, offe
ring
up t
o se
ven
trea
tmen
t se
ssio
ns w
ith a
the
rapi
st w
ho h
asbe
en t
rain
ed in
adv
ance
in t
he M
AC
T m
etho
ds
Rand
omisa
tion
met
hod:
str
atifi
edby
cen
tre
and
base
line
para
suic
ide
risk
scor
e an
dal
loca
ted
rand
omly
by
tele
phon
e/fa
xBl
inde
d as
sess
men
t: N
RPo
wer
cal
cula
tion:
yes
Loss
to
follo
w-u
p: y
es
Lack
ner Q
ualit
y Ra
ting
Scal
e ite
ms
Patie
nt s
elec
tion:
12/
14In
terv
entio
n: 4
/8O
utco
me
mea
sure
men
t: 8/
12D
ata
pres
enta
tion
and
stat
istic
alan
alys
is: 1
2/12
Inve
stig
ator
: 2/4
Misc
ella
neou
s: 4
/8To
tal:
42/5
8
NR
TAU
pat
ient
sno
rmal
ly r
ecei
ve a
nin
itial
psy
chia
tric
asse
ssm
ent
follo
wed
by p
sych
iatr
icou
tpat
ient
car
e,oc
casio
nal d
ay-
patie
nt c
are
orre
ferr
al b
ack
to t
heG
P, de
pend
ing
on t
hear
rang
emen
ts o
f the
hosp
ital.
If pa
tient
sw
ere
alre
ady
unde
rps
ychi
atric
car
e an
yfu
rthe
r tr
eatm
ent
was
per
mitt
ed, a
part
from
MA
CT
n=
480
(MA
CT
n=
239,
TAU
n=
241)
Of t
hese
, PD
n=
391,
BPD
n=
67 cont
inue
d
Health Technology Assessment 2006; Vol. 10: No. 35
79
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TA
BLE
18
Stud
y ch
arac
teris
tics:
RCT
s (c
ont’d
)
Stud
yD
escr
ipti
on o
f psy
chot
hera
pySt
udy
qual
ity
Co-
ther
apy
or
Com
para
tor
Sam
ple
size
med
icat
ion
van
den
Bosc
h et
al.,
2002
59D
BT (f
or d
etai
ls se
e Li
neha
n et
al.,
199
153)
Rand
omisa
tion
met
hod:
min
imisa
tion
met
hod
mat
ched
by a
ge, a
lcoh
ol, d
rug
and
soci
alpr
oble
ms
Blin
ded
asse
ssm
ent:
inde
pend
ent
asse
ssor
sa
Pow
er c
alcu
latio
n: N
RLo
ss t
o fo
llow
-up:
yes
Lack
ner Q
ualit
y Ra
ting
Scal
e ite
ms
Patie
nt s
elec
tion:
12/
14In
terv
entio
n: 4
/8O
utco
me
mea
sure
men
t: 8/
12D
ata
pres
enta
tion
and
stat
istic
alan
alys
is: 9
/12
Inve
stig
ator
: 2/4
Misc
ella
neou
s: 5
/8To
tal:
40/5
8
Thr
ee-q
uart
ers
of t
hepa
tient
s re
port
ed u
se o
fm
edic
atio
n fr
om o
ne o
rm
ore
of t
he fo
llow
ing
cate
gorie
s:be
nzod
iaze
pine
s, S
SRIs
(DBT
52%
, TA
U 6
1%),
tric
yclic
ant
idep
ress
ants
,m
ood
stab
ilise
rs a
ndne
urol
eptic
s
TAU
con
siste
d of
clin
ical
man
agem
ent
from
the
orig
inal
refe
rral
sou
rce
(add
ictio
n tr
eatm
ent
cent
res
n=
11,
psyc
hiat
ric s
ervi
ces
n=
20).
Patie
nts
atte
nded
no
mor
eth
an t
wo
sess
ions
per
mon
th w
ith a
psyc
holo
gist
, aps
ychi
atris
t or
aso
cial
wor
ker
Rand
omise
d: n
=58
(DBT
n=
27,
TAU
n=
31)
Ana
lyse
d: n
=47
(DBT
n=
23,
TAU
n=
24)
a In
depe
nden
t as
sess
ors
wer
e no
t in
form
ed a
bout
the
tre
atm
ent
cond
ition
of t
he in
terv
iew
ees;
how
ever
, pat
ient
s m
ight
hav
e gi
ven
the
info
rmat
ion
abou
t tr
eatm
ent.
LAA
M, l
evo-
alph
a ac
etyl
met
hado
l; SS
RI, s
elec
tive
sero
toni
n re
upta
ke in
hibi
tors
.
Appendix 3
80 TA
BLE
19
Stud
y ch
arac
teris
tics:
non
-RCT
Stud
yD
escr
ipti
on o
f psy
chot
hera
pySt
udy
qual
ity
Co-
ther
apy
or
Com
para
tor
Sam
ple
size
med
icat
ion
Wilb
erg
et a
l.,19
9860
Day
hos
pita
l tre
atm
ent
and,
pos
tdisc
harg
e,gr
oup
anal
ytic
al t
hera
py. T
he d
ay t
reat
men
tpr
ogra
mm
e co
nsist
s of
a c
ombi
natio
n of
indi
vidu
al a
nd g
roup
psy
chot
hera
pies
and
phar
mac
othe
rapy
, con
duct
ed in
acc
orda
nce
with
the
rape
utic
com
mun
ity p
rinci
ples
. The
subs
eque
nt o
utpa
tient
gro
up t
hera
py w
asco
nduc
ted
in a
ccor
danc
e w
ith g
roup
anal
ytic
al p
rinci
ples
Com
para
tor
and
drop
outs
desc
ribed
. Bas
elin
e co
mpa
rabi
lity
not
just
ified
. Ana
lyse
s do
nere
tros
pect
ivel
y
Lack
ner Q
ualit
y Ra
ting
Scal
e ite
ms
Patie
nt s
elec
tion:
9/1
4In
terv
entio
n: 4
/8O
utco
me
mea
sure
men
t: 8/
12D
ata
pres
enta
tion
and
stat
istic
alan
alys
is: 3
/12
Inve
stig
ator
: 2/4
Misc
ella
neou
s: 4
/8To
tal:
30/5
8
NR
Day
hos
pita
ltr
eatm
ent
and,
post
disc
harg
e, T
AU
“ran
ging
from
no
trea
tmen
t to
psyc
hoth
erap
y tw
ice
wee
kly”
n =
43
(tre
atm
ent
grou
p n
= 1
2,co
ntro
l gro
up
n =
31)
Health Technology Assessment 2006; Vol. 10: No. 35
81
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TA
BLE
20
Ther
apy
deta
ils: R
CTs
Stud
yR
ecru
itm
ent
Num
ber
of s
essi
ons
Leng
th o
f ses
sion
sT
hera
pist
con
tact
P
rofe
ssio
nal
betw
een
sess
ions
back
grou
nd o
f the
rapi
st
Bate
man
and
Fona
gy, 1
99951
Patie
nts
from
gen
eral
psyc
hiat
ric s
ervi
ces
refe
rred
durin
g 19
93 a
nd 1
994
(1) O
nce-
wee
kly
indi
vidu
alps
ycho
anal
ytic
al p
sych
othe
rapy
; (2
) thr
ice-
wee
kly
grou
p an
alyt
ical
psyc
hoth
erap
y; (3
) onc
e-a-
wee
kex
pres
sive
ther
apy
orie
ntat
edto
war
ds p
sych
odra
ma
tech
niqu
es;
(4) a
wee
kly
com
mun
ity m
eetin
g;
(5) o
nce-
a-m
onth
mee
ting
with
cas
ead
min
istra
tor;
(6) m
edic
atio
n re
view
by t
he r
esid
ent
psyc
hiat
rist
(1) 1
hou
r; (2
) 1 h
our
each
; (3
) 1ho
ur; (
4) 1
hour
;(5
)1ho
ur. T
he a
vera
ge le
ngth
of
stay
was
1.4
5 ye
ars
TAU
: (1)
reg
ular
psy
chia
tric
,re
view
by
seni
or p
sych
iatr
istw
hen
nece
ssar
y (t
wic
e pe
rw
eek)
; (2)
inpa
tient
adm
issio
n as
appr
opria
te (a
dmiss
ion
rate
90%
, ave
rage
sta
y 11
.6 d
ays)
;(3
) out
patie
nt a
nd c
omm
unity
follo
w-u
p (v
isit
ever
y 2
wee
ksby
a c
omm
unity
psy
chia
trist
)
Part
ial h
ospi
talis
atio
n:th
erap
ies
and
patie
nt–s
taff
cont
act
wer
e or
gani
sed
inac
cord
ance
with
the
psyc
hoan
alyt
ical
mod
el
All
ther
apy
was
giv
en b
yps
ychi
atric
nur
ses
who
wer
em
embe
rs o
f the
par
tial
hosp
italis
atio
n pr
ogra
mm
e’s
team
, but
who
had
no
form
alps
ycho
ther
apy
qual
ifica
tion
Koon
s et
al.,
2001
52D
urha
m V
A M
edic
al C
entr
ean
d Ve
tera
ns R
eadj
ustm
ent
and
Cou
nsel
ling
Cen
tre
inon
e ca
se
DBT
: wee
kly
for
6 m
onth
s;
TAU
: wee
kly
indi
vidu
al t
hera
py fo
r6
mon
ths
DBT
: 90
min
utes
/wee
k;
TAU
: 60
min
utes
per
wee
kTe
leph
one
calls
with
the
prim
ary
ther
apist
(whe
n ne
eded
)
DBT
: psy
chia
trist
, tw
ops
ycho
logi
sts,
a c
linic
al s
ocia
lw
orke
r an
d a
clin
ical
nur
sesp
ecia
list
in p
sych
iatr
y.C
linic
ians
had
a m
ean
of 8
.2ye
ars
of c
linic
al e
xper
ienc
e.A
ll ex
cept
one
att
ende
din
tens
ive
trai
ning
in D
BT
TAU
: thr
ee p
sych
olog
ists,
two
resid
ent
psyc
hiat
rists
,tw
o cl
inic
al s
ocia
l wor
kers
and
a cl
inic
al n
urse
spe
cial
istin
psy
chia
try.
The
y ha
d a
mea
n of
10.
6 ye
ars
of c
linic
alex
perie
nce.
Fou
r de
scrib
edth
emse
lves
as
cogn
itive
beha
viou
ral,
two
asps
ycho
dyna
mic
and
tw
o as
ecle
ctic
in t
heir
prim
ary
orie
ntat
ion
cont
inue
d
Appendix 3
82 TA
BLE
20
Ther
apy
deta
ils: R
CTs
(con
t’d)
Stud
yR
ecru
itm
ent
Num
ber
of s
essi
ons
Leng
th o
f ses
sion
sT
hera
pist
con
tact
P
rofe
ssio
nal
betw
een
sess
ions
back
grou
nd o
f the
rapi
st
Line
han
et a
l.,19
9153
Part
icip
ants
wer
e cl
inic
ally
refe
rred
and
vol
unta
rily
enro
lled
in t
he s
tudy
DBT
: ind
ivid
ual t
hera
py w
assc
hedu
led
for
wee
kly
1-ho
urse
ssio
n. G
roup
the
rapy
was
sche
dule
d on
ce e
ach
wee
k fo
r2.
5ho
urs
(for
1 ye
ar);
cont
rol
grou
p: N
R
Indi
vidu
al s
essio
n 1-
hour
; gro
upse
ssio
n 2.
5 ho
urs
Tele
phon
e co
ntac
tw
ith t
he in
divi
dual
ther
apist
bet
wee
nse
ssio
ns w
as p
art
ofD
BT. T
he g
roup
ther
apist
did
not
acce
pt t
elep
hone
calls
from
pat
ient
s,an
d pa
tient
cris
esw
ere
refe
rred
to
the
indi
vidu
al t
hera
pist
Expe
rienc
ed g
radu
ate
psyc
holo
gy s
tude
nts,
ther
apist
s w
ith m
aste
r’s
leve
ltr
aini
ng a
nd c
linic
alps
ycho
logi
sts
Line
han
et a
l.,19
9954
Part
icip
ants
wer
e re
ferr
ed t
oth
e pr
ogra
mm
e by
are
acl
inic
ians
Wee
kly
indi
vidu
al p
sych
othe
rapy
(1ho
ur),
grou
p sk
ills
trai
ning
ses
sion
(2ho
urs
plus
a 1
5-m
inut
e w
ind-
dow
n) (f
or 1
yea
r)
Indi
vidu
al s
essio
n 1-
hour
; gro
upse
ssio
n 2
hour
s pl
us a
15-
min
ute
win
d-do
wn
Skill
s co
achi
ngte
leph
one
calls
with
the
prim
ary
ther
apist
(whe
n ne
eded
)
DBT
: tw
o ps
ycho
logi
sts,
one
psyc
hiat
rist
and
two
mas
ter’
sle
vel c
linic
ians
TAU
: of 1
5 th
erap
ists
inte
rvie
wed
, fiv
e w
ere
psyc
hoth
erap
ists,
eig
htm
aste
r’s
leve
l the
rapi
sts
and
two
had
no o
r un
know
nhe
alth
deg
ree
(tw
obe
havi
oura
l, th
ree
cogn
itive
,fiv
e C
CT
or
supp
ortiv
e, o
neps
ycho
dyna
mic
, one
ecl
ectic
)
cont
inue
d
Health Technology Assessment 2006; Vol. 10: No. 35
83
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TA
BLE
20
Ther
apy
deta
ils: R
CTs
(con
t’d)
Stud
yR
ecru
itm
ent
Num
ber
of s
essi
ons
Leng
th o
f ses
sion
sT
hera
pist
con
tact
P
rofe
ssio
nal
betw
een
sess
ions
back
grou
nd o
f the
rapi
st
Line
han
et a
l.,20
0255
Part
icip
ants
wer
e re
crui
ted
from
men
tal h
ealth
clin
ics,
need
le e
xcha
nge
prog
ram
mes
, sub
stan
ceab
use
clin
ics,
met
hado
nem
aint
enan
ce c
linic
s an
d no
n-pr
ofit
HIV
/AID
S pr
even
tion
orga
nisa
tions
tre
atin
gun
ders
erve
d m
inor
itypo
pula
tions
DBT
or
CVT
+12
S pl
us a
n op
iate
med
icat
ion
for
appr
oxim
atel
y 1
year
(48–
56 w
eeks
)
DBT
: ind
ivid
ual D
BT40
–90
min
utes
per
wee
k; g
roup
skill
s tr
aini
ng 1
50m
inut
es p
erw
eek;
indi
vidu
al s
kills
coa
chin
g30
min
utes
per
wee
k(r
ecom
men
ded)
; 12S
and
oth
ersu
ppor
tive
grou
p m
eetin
gs(r
ecom
men
ded)
;D
BT c
ase
man
agem
ent
thro
ugho
ut 1
yea
r(4
8–56
wee
ks)
CVT
+12
S: in
divi
dual
CVT
+12
S40
–90
min
utes
per
wee
k; ‘1
2-an
d-12
’ Nar
cotic
s A
nony
mou
sgr
oup
120
min
ute
per
wee
k;12
S sp
onso
r m
eetin
g; 1
2Sm
eetin
g (r
ecom
men
ded)
;C
VT+
12S
case
man
agem
ent
(as
need
ed) t
hrou
ghou
t 1
year
(48–
56w
eeks
)
Tele
phon
eco
nsul
tatio
n an
dcr
isis
inte
rven
tions
Five
the
rapi
sts
(one
mal
e,fo
ur fe
mal
e; t
hree
DBT
and
two
CVT
+12
S): t
wo
doct
oral
leve
l and
one
mas
ter’
s le
vel b
ehav
iour
ther
apist
del
iver
ed D
BT, a
ndtw
o m
aste
r’s
leve
l the
rapi
sts
with
che
mic
al d
epen
denc
yce
rtifi
catio
n an
d 12
Sex
perie
nce
deliv
ered
CVT
+12
S. E
ach
ther
apist
had
a m
inim
um o
f 8 m
onth
trai
ning
, and
had
sup
ervi
sed
trai
ning
clie
nts
in t
heir
resp
ectiv
e m
odal
ities
bef
ore
seei
ng t
heir
rese
arch
clie
nts.
The
rapi
sts
in e
ach
cond
ition
met
wee
kly
with
sup
ervi
sors
to d
iscus
s ca
se m
ater
ials
and
revi
ew s
essio
n vi
deot
apes
Mun
roe-
Blum
and
Mar
zial
i, 19
9556
Part
icip
ants
wer
e re
crui
ted
from
the
inpa
tient
and
outp
atie
nt p
sych
iatr
y un
its o
fth
e te
achi
ng h
ospi
tals
of a
larg
e C
anad
ian
urba
nun
iver
sity
IGP:
30
sess
ions
of t
reat
men
t(2
5w
eekl
y se
ssio
ns fo
llow
ed b
y fiv
etw
ice
wee
kly
sess
ions
lead
ing
into
term
inat
ion)
Each
ses
sion
sche
dule
d fo
r1.
5ho
urs
NR
All
ther
apist
s w
ere
trai
ned
tous
e IG
P
cont
inue
d
Appendix 3
84 TA
BLE
20
Ther
apy
deta
ils: R
CTs
(con
t’d)
Stud
yR
ecru
itm
ent
Num
ber
of s
essi
ons
Leng
th o
f ses
sion
sT
hera
pist
con
tact
P
rofe
ssio
nal
betw
een
sess
ions
back
grou
nd o
f the
rapi
st
Turn
er, 2
00057
Patie
nts
from
loca
l hos
pita
lem
erge
ncy
serv
ices
wer
ere
ferr
ed t
o th
e co
mm
unity
men
tal h
ealth
out
patie
ntcl
inic
Min
imum
49
and
max
imum
84
sess
ions
Six
grou
p se
ssio
ns w
ere
prov
ided
to
patie
nts
in b
oth
trea
tmen
tco
nditi
ons
CC
T: t
reat
men
t w
as s
ched
uled
for
12 m
onth
s. S
essio
ns w
ere
sche
dule
dtw
ice
per
wee
k w
hen
poss
ible
NR
In C
CT
gro
up d
urin
gth
e cr
isis
man
agem
ent
phas
e,th
e th
erap
ist m
etpa
tient
s as
ofte
n as
thre
e tim
es a
wee
k
The
rapi
sts
cond
uctin
g bo
thtr
eatm
ents
had
an
aver
age
22ye
ars
of e
xper
ienc
e, w
ithth
eore
tical
bac
kgro
unds
infa
mily
sys
tem
, clie
nt-c
entr
edan
d ps
ycho
dyna
mic
trea
tmen
ts. D
BT t
rain
ing
last
ed fo
r 3
mon
ths
(12
sess
ions
, eac
h se
ssio
nla
sted
for
90 m
inut
es)
Tyre
r et
al.,
200
358Pa
rtic
ipan
ts w
ere
recr
uite
din
nin
e A
&E
depa
rtm
ents
infiv
e U
K s
tudy
cen
tres
:G
lasg
ow, E
dinb
urgh
,N
ottin
gham
, Wes
t Lo
ndon
and
Sout
h Lo
ndon
Up
to fi
ve s
essio
ns w
ithin
3m
onth
saf
ter
a se
lf-ha
rm e
piso
de, w
ith t
heop
tion
of t
wo
addi
tiona
l boo
ster
sess
ions
with
in 6
mon
ths
NR
Betw
een
sess
ions
,th
e m
anua
l act
s as
an
aide
-mém
oire
and
can
be u
sed
for
hom
ewor
k ta
sks
byth
e pa
tient
NR
van
den
Bosc
het
al.,
2002
59Re
ferr
als
orig
inat
ed fr
omad
dict
ion
trea
tmen
t se
rvic
es,
psyc
hiat
ric h
ospi
tals,
cen
tres
for
men
tal h
ealth
care
,in
depe
nden
tly w
orki
ngps
ycho
logi
sts
and
psyc
hiat
rists
, GPs
and
sel
f-re
ferr
al
Wee
kly
sess
ion
durin
g 12
mon
ths
DBT
: 2–2
.5ho
urs;
TA
U: n
om
ore
than
tw
o se
ssio
ns p
erm
onth
(len
gth
of s
essio
ns n
otre
port
ed)
Tele
phon
e ca
lls w
ithth
e pr
imar
y th
erap
ist(w
hen
need
ed)
Four
psy
chia
trist
s an
d 12
clin
ical
psy
chol
ogist
s (t
wo
with
mas
ter’
s de
gree
s an
don
e a
PhD
). G
roup
tra
inin
gw
as c
ondu
cted
in t
hree
sepa
rate
gro
ups
led
join
tly b
yso
cial
wor
kers
and
clin
ical
psyc
holo
gist
s. A
cor
e gr
oup
of t
hree
the
rapi
sts
was
sen
tto
Sea
ttle
to
be t
rain
ed in
DBT
Health Technology Assessment 2006; Vol. 10: No. 35
85
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TA
BLE
21
Ther
apy
deta
ils: n
on-R
CT
Stud
yR
ecru
itm
ent
Num
ber
of s
essi
ons
Leng
th o
f ses
sion
sT
hera
pist
con
tact
P
rofe
ssio
nal
betw
een
sess
ions
back
grou
nd o
f the
rapi
st
Wilb
erg
et a
l.,19
9860
Patie
nts
bein
g tr
eate
d at
the
Day
Uni
t, U
llevå
l Uni
vers
ityH
ospi
tal
Onc
e a
wee
k fo
r a
mea
n pe
riod
of12
mon
ths
(ran
ge 1
–33
mon
ths)
1.5
hour
sN
R“S
ix o
f the
eig
ht t
hera
pist
sw
ere
in g
roup
ana
lytic
altr
aini
ng a
t th
e tim
e”
Appendix 3
86 TA
BLE
22
Stud
y sit
e, fo
llow
-up
and
incl
usio
n/ex
clus
ion
crite
ria: R
CTs
Stud
ySt
udy
site
Leng
th o
f N
umbe
rs lo
st t
o R
easo
ns fo
r lo
ss t
o In
clus
ion
crit
eria
Excl
usio
n cr
iter
iafo
llow
-up
follo
w-u
pfo
llow
-up/
drop
out
cont
inue
d
Bate
man
and
Fona
gy, 1
99951
Hal
liwic
kPs
ycho
ther
apy
Uni
t, w
hich
is p
art
of t
he g
ener
alps
ychi
atric
serv
ices
, UK
(PH
)
18 m
onth
sO
f 44
elig
ible
pat
ient
s,6
drop
ped
out
durin
gth
e st
udy.
No
loss
to
follo
w-u
p in
PH
gro
upan
d th
ree
in T
AU
With
in t
he fi
rst
mon
th t
hree
pat
ient
sfr
om t
he c
ontr
ol g
roup
cro
ssed
ove
rin
to t
he P
H g
roup
afte
r se
rious
sui
cide
atte
mpt
. With
in 6
mon
ths
thre
e pa
tient
sdr
oppe
d ou
t fr
om t
he P
H g
roup
(rea
son
NR)
BPD
pat
ient
s w
ho s
core
d 7
orm
ore
on (1
) SC
ID fo
r D
SM-II
I-R a
nd (2
) the
Dia
gnos
tic In
terv
iew
for
BPD
Schi
zoph
reni
a, b
ipol
ardi
sord
er, s
ubst
ance
misu
se,
men
tal i
mpa
irmen
t or
orga
nic
brai
n di
sord
er
Koon
s et
al.,
2001
52M
edic
al C
entr
e,U
SA (o
utpa
tient
sett
ing)
No
follo
w-u
pTw
o di
d no
t at
tend
the
first
app
oint
men
t, on
edr
oppe
d ou
t af
ter
first
appo
intm
ent,
two
inTA
U a
nd t
hree
in D
BTdr
oppe
d ou
t of
trea
tmen
t af
ter
atte
ndin
g m
ore
than
one
sess
ion
Loss
of t
rans
port
atio
n an
d di
stan
cefr
om t
he m
edic
al c
entr
e. O
ne d
ropp
edou
t w
hen
she
real
ised
she
wou
ld o
nly
be p
aid
for
asse
ssm
ents
, not
for
atte
ndin
g tr
eatm
ent
Wom
en v
eter
ans
who
met
DSM
-III-R
crit
eria
for
BPD
Schi
zoph
reni
a, b
ipol
ardi
sord
er, s
ubst
ance
depe
nden
ce o
r an
tisoc
ial
pers
onal
ity d
isord
er
Line
han
et a
l.,19
9153
Rese
arch
clin
ic,
USA
(out
patie
ntse
ttin
g)
1 ye
ar19
(dro
pout
s du
ring
the
trea
tmen
t pe
riod)
Five
(dur
ing
1 ye
ar o
ffo
llow
-up)
Ten
(DBT
n=
5, T
AU
n=
5) d
ropp
edou
t du
ring
pret
reat
men
t as
sess
men
t; Se
ven
(DBT
n=
3, T
AU
n=
4)dr
oppe
d ou
t ow
ing
to r
efus
al o
r in
abili
tyto
mee
t st
udy
cond
ition
s; t
wo
(DBT
)qu
it af
ter
four
or
few
er s
essio
ns
(1) S
core
d at
leas
t 7,
out
of a
max
imum
of 1
0, o
n th
e D
IB;
(2) M
et D
MS-
III c
riter
ia fo
rBP
D; (
3) h
ad a
t le
ast
two
inci
denc
es o
f par
asui
cide
in t
hepa
st 5
yea
rs, w
ith o
ne d
urin
gth
e pa
st 8
wee
ks; (
4) w
omen
aged
18–
45ye
ars
Men
; par
ents
who
met
DSM
-III c
riter
ia fo
rsc
hizo
phre
nia,
bip
olar
diso
rder
, sub
stan
cede
pend
ence
or
men
tal
reta
rdat
ion
Line
han
et a
l.,19
9954
Rese
arch
clin
ic,
USA
(out
patie
ntse
ttin
g)
16 m
onth
s (4
mon
ths
post
-tr
eatm
ent)
12Si
x (D
BT n
=1,
TA
U n
=5
)pa
rtic
ipan
ts d
ropp
ed b
efor
e or
imm
edia
tely
afte
r pr
etre
atm
ent
asse
ssm
ent;
two
(DBT
) par
ticip
ants
drop
ped
out
by t
he s
ixth
ses
sion;
tw
o(D
BT) p
rovi
ded
no d
ata
afte
rpr
etre
atm
ent;
one
drop
ped
out
oftr
eatm
ent
afte
r th
e six
th s
essio
n; o
nedi
ed o
f acc
iden
tal d
rug
over
dose
dur
ing
the
4-m
onth
ass
essm
ent
Part
icip
ants
who
met
crit
eria
for
BPD
on
both
PD
E an
d th
eSC
ID-II
and
met
crit
eria
for
subs
tanc
e us
e di
sord
er fo
rop
iate
s, c
ocai
ne,
amph
etam
ines
, sed
ativ
es,
hypn
otic
s, a
nxio
lytic
s or
poly
subs
tanc
e us
e di
sord
er o
nth
e SC
ID
Part
icip
ants
who
met
crite
ria fo
r sc
hizo
phre
nia
and
othe
r ps
ycho
ticdi
sord
er, o
r bi
pola
r m
ood
diso
rder
on
the
SCID
, or
men
tal r
etar
datio
n on
the
Peab
ody
Pict
ure
Voca
bula
ryTe
st–R
evise
d
Health Technology Assessment 2006; Vol. 10: No. 35
87
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TA
BLE
22
Stud
y sit
e, fo
llow
-up
and
incl
usio
n/ex
clus
ion
crite
ria: R
CTs
(con
t’d)
Stud
ySt
udy
site
Leng
th o
f N
umbe
rs lo
st t
o R
easo
ns fo
r lo
ss t
o In
clus
ion
crit
eria
Excl
usio
n cr
iter
iafo
llow
-up
follo
w-u
pfo
llow
-up/
drop
out
cont
inue
d
Line
han
et a
l.,20
0255
Rese
arch
clin
ic,
USA
(out
patie
ntse
ttin
g)
16 m
onth
s(4
mon
ths
post
-tr
eatm
ent)
DBT
: 36%
loss
to
follo
wup C
VD+
12S:
no
loss
to
follo
w-u
p
NR
(1) W
omen
age
d 18
–45
year
s;(2
) dia
gnos
is of
BPD
acc
ordi
ngto
tw
o st
ruct
ured
inte
rvie
ws:
PDE
and
SCID
-II; (
3) d
iagn
osis
of c
urre
nt o
piat
e de
pend
ence
acco
rdin
g to
SC
ID-I
Bipo
lar
diso
rder
, psy
chos
is,se
izur
e di
sord
er, o
r m
enta
lre
tard
atio
n; p
regn
ancy
or
any
othe
r m
edic
al c
ondi
tion
in w
hich
the
use
of o
piat
e-re
plac
emen
t m
edic
atio
n w
asco
ntra
indi
cate
d; in
dica
tions
of c
oerc
ion
(e.g
. cou
rtor
dere
d/ag
ency
ord
ered
to
reta
in h
ousin
g)
Mun
roe-
Blum
and
Mar
zial
i,19
9556
Out
patie
ntho
spita
ls, C
anad
a(o
utpa
tient
sett
ing)
24 m
onth
s31
with
drew
at
the
poin
t of
ran
dom
isatio
nN
R(1
) Men
and
wom
en a
ged
18–6
5 ye
ars;
(2) h
ad a
t le
ast
one
prio
r ps
ychi
atric
con
tact
and
met
BPD
crit
eria
on
the
DIB
, cut
-off
scor
e ≥
7
Lang
uage
diff
icul
ty;
neur
olog
ical
impa
irmen
t or
men
tal r
etar
datio
n; a
prim
ary
diag
nosis
of a
lcoh
olor
dru
g ad
dict
ion;
phy
sical
diso
rder
s w
ith k
now
nps
ychi
atric
con
sequ
ence
s
Turn
er, 2
00057
Out
patie
nt c
linic
,U
SA (o
utpa
tient
sett
ing)
12 m
onth
sD
BT: n
= 4
, C
CT:
n=
6
NR
(one
pat
ient
in D
BT g
roup
ret
urne
daf
ter
5-w
eeks
bre
ak)
(1) M
et d
iagn
ostic
crit
eria
for
BPD
; (2)
gav
e in
form
edco
nsen
t; (3
) acc
epte
d ra
ndom
assig
nmen
t
Schi
zoph
reni
a,sc
hizo
affe
ctiv
e di
sord
er,
bipo
lar
diso
rder
, org
anic
men
tal d
isord
er o
r m
enta
lre
tard
atio
n
Tyre
r et
al.,
2003
58Pa
tient
s fr
om t
heho
spita
ls in
Gla
sgow
,Ed
inbu
rgh,
Not
tingh
am, W
est
Lond
on a
nd S
outh
Lond
on b
etw
een
May
198
8 an
dA
pril
2000
, UK
(out
patie
ntse
ttin
g)
12 m
onth
sM
AC
T n
= 4
0,
TAU
n =
38
MA
CT
TA
Un
= 3
n =
5 d
ied
n =
14
n =
13
not
trac
edn
= 9
n =
10
refu
sed
asse
ssm
ent
n =
4n
= 0
with
drew
n =
3n
= 6
did
not
att
end
n =
7n
= 4
oth
er r
easo
n
(1) P
atie
nts
who
had
had
apr
evio
us e
piso
de o
f DSH
; (2
) gav
e in
form
ed w
ritte
nco
nsen
t; (3
) did
not
req
uire
inpa
tient
psy
chia
tric
tre
atm
ent;
(4) a
ged
betw
een
16 a
nd65
year
s, (5
) pre
sent
ed t
o A
&E
afte
r an
epi
sode
of D
SH
Psyc
hotic
or
bipo
lar
diso
rder
; prim
ary
diag
nosis
of s
ubst
ance
dep
ende
nce;
insu
ffici
ent
know
ledg
e of
Engl
ish; t
empo
rary
resid
ence
Appendix 3
88 TA
BLE
22
Stud
y sit
e, fo
llow
-up
and
incl
usio
n/ex
clus
ion
crite
ria: R
CTs
(con
t’d)
Stud
ySt
udy
site
Leng
th o
f N
umbe
rs lo
st t
o R
easo
ns fo
r lo
ss t
o In
clus
ion
crit
eria
Excl
usio
n cr
iter
iafo
llow
-up
follo
w-u
pfo
llow
-up/
drop
out
van
den
Bosc
het
al.,
200
259Je
lline
k A
ddic
tion
Trea
tmen
t C
entr
ein
Am
ster
dam
,T
he N
ethe
rland
s(o
utpa
tient
sett
ing)
18 m
onth
sO
f 58
elig
ible
pat
ient
s,six
drop
ped
out
durin
gth
e st
udy
DBT
: n=
4 re
fuse
d to
sta
rt t
reat
men
tTA
U: n
=2
did
not
acce
ptTA
U c
ondi
tion
(1) M
et D
SM-IV
crit
eria
for
BPD
; (2)
cur
rent
ly in
outp
atie
nt p
sych
iatr
ic o
rsu
bsta
nce
abus
e tr
eatm
ent;
(3) w
omen
age
d 18
–70
year
s;(4
) res
iden
ce w
ithin
40-
kmci
rcle
aro
und
Am
ster
dam
DSM
-IV c
riter
ia fo
r bi
pola
rdi
sord
er o
r (c
hron
ic)
psyc
hotic
diso
rder
;in
suffi
cien
t co
mm
and
ofD
utch
lang
uage
; sev
ere
cogn
itive
impa
irmen
ts
TA
BLE
23
Stud
y sit
e, fo
llow
-up
and
incl
usio
n/ex
clus
ion
crite
ria: n
on-R
CT
Stud
ySt
udy
site
Leng
th o
f N
umbe
rs lo
st t
o R
easo
ns fo
r lo
ss t
o In
clus
ion
crit
eria
Excl
usio
n cr
iter
iafo
llow
-up
follo
w-u
pfo
llow
-up/
drop
out
Wilb
erg
et a
l.,19
9860
Day
Psy
chia
tric
Uni
t in
Uni
vers
ityH
ospi
tal,
Oslo
,N
orw
ay
An
aver
age
of 3
4 m
onth
spo
stdi
scha
rge
from
day
hosp
ital
Non
e. R
etro
spec
tive
stud
yN
RM
et D
SM-II
I/DSM
-IIIR
crit
eria
for
BPD
Patie
nts
with
sch
izot
ypal
pers
onal
ity d
isord
er;
patie
nts
who
had
sta
yed
for
less
tha
n 3
wee
ks a
t th
e da
yho
spita
l
Health Technology Assessment 2006; Vol. 10: No. 35
89
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TA
BLE
24
Patie
nt c
hara
cter
istic
s: R
CTs
Stud
yM
etho
ds fo
r di
agno
sis
Age
Gen
der
Ethn
icit
yEd
ucat
ion/
soci
o-ec
onom
ic
Pati
ent
hist
ory
Bas
elin
e co
mpa
rabi
lity
of d
isor
der
(yea
rs)
back
grou
nd
cont
inue
d
Bate
man
and
Fona
gy, 1
99951
(1) S
CID
for
DSM
-III;
(2) D
iagn
ostic
Inte
rvie
wfo
r BP
D
16–6
5F
n=
22M
n=
16N
RC
olle
ge: P
H n
= 7
, TA
U n
= 3
Une
mpl
oyed
: PH
n=
19,
TA
U n
= 1
9Si
ngle
: PH
n=
17,
TA
U n
= 1
6Sh
elte
red
acco
mm
odat
ion:
PH
n=
3, T
AU
n=
2W
ith fa
mily
of o
rigin
: PH
n=
6,
TAU
n=
3A
lone
: PH
n=
10,
TA
U n
= 1
4
Early
mat
erna
l los
s:
PH n
= 1
0, T
AU
n=
14
Repo
rted
sex
ual a
buse
: PH
n=
7, T
AU
n=
8Re
port
ed r
ape:
PH
n=
5,
TAU
n=
2Re
port
ed p
hysic
al a
buse
: PH
n=
9, T
AU
n=
8
The
re w
ere
no s
igni
fican
tdi
ffere
nces
on
any
of t
heba
selin
e m
easu
res
betw
een
the
two
grou
ps
Koon
s et
al.,
2001
52BP
D a
nd a
ntiso
cial
pers
onal
ity d
isord
erse
ctio
ns o
f the
SC
ID-II
for
DSM
-III-R
for
Axi
s II,
and
the
subs
tanc
e ab
use,
bipo
lar
diso
rder
, and
schi
zoph
reni
a se
ctio
ns o
fth
e SC
ID fo
r D
SM-II
I-Rfo
r A
xis
II
Mea
n 35
(21–
46)
FC
auca
sian
75%
;A
fric
an–A
mer
ican
25%
Col
lege
80%
; bac
helo
r’s
degr
eeor
equ
ival
ent
20%
; inc
ome
>$2
0,00
0 pa
70%
; liv
es w
ithpa
rtne
r 55
%
75%
had
a li
fetim
e hi
stor
yof
par
asui
cide
, def
ined
as
any
inte
ntio
nal s
elf-
inju
ry,
incl
udin
g su
icid
e at
tem
pts;
40%
rep
orte
d pa
rasu
icid
albe
havi
our
in t
he 6
mon
ths
befo
re t
he s
tudy
. 55%
had
at le
ast
one
lifet
ime
psyc
hiat
ric a
dmiss
ion;
25%
had
an in
patie
nt p
sych
iatr
icad
miss
ion
in t
he p
ast
6m
onth
s. A
ll ha
d at
leas
ton
e ps
ychi
atric
out
patie
ntvi
sit in
the
pre
viou
s 6
mon
ths.
25%
met
crit
eria
for
subs
tanc
e ab
use,
but
not
depe
nden
ce; 6
0% r
epor
ted
sexu
al a
buse
bef
ore
the
age
of 1
3, 6
5% r
epor
ted
bein
gba
tter
ed b
y a
part
ner
and
85%
rep
orte
d be
ing
rape
das
an
adul
t, 46
% w
hile
on
activ
e m
ilita
ry d
uty
Non
e of
the
var
iabl
esdi
ffere
d sig
nific
antly
betw
een
grou
ps
Appendix 3
90 TA
BLE
24
Patie
nt c
hara
cter
istic
s: R
CTs
(con
t’d)
Stud
yM
etho
ds fo
r di
agno
sis
Age
Gen
der
Ethn
icit
yEd
ucat
ion/
soci
o-ec
onom
ic
Pati
ent
hist
ory
Bas
elin
e co
mpa
rabi
lity
of d
isor
der
(yea
rs)
back
grou
nd
cont
inue
d
Line
han
et a
l.,19
9153
(1) S
core
d at
leas
t 7
out
of a
max
imum
of 1
0 on
the
DIB
; (2)
DM
S-III
crite
ria fo
r BP
D; (
3) h
adat
leas
t tw
o in
cide
nces
of
para
suic
ide
in t
he p
ast
5ye
ars,
with
one
dur
ing
the
past
8 w
eeks
18–4
5F
NR
NR
Patie
nts
with
a h
istor
y of
para
suic
ide
and
psyc
hiat
richo
spita
lisat
ion
on t
hegr
ound
of B
PD
Part
icip
ants
wer
em
atch
ed o
n th
e nu
mbe
rof
life
time
para
suic
ides
and
psyc
hiat
richo
spita
lisat
ion,
age
, and
good
vs.
poo
r cl
inic
alpr
ogno
sis
Line
han
et a
l.,19
9954
Part
icip
ants
wer
e gi
ven
asc
reen
ing
inte
rvie
w t
hat
incl
uded
SC
ID fo
r D
SM-II
I and
the
PD
E
Mea
n30
.4±
6.6
(18–
45)
FEu
rope
an 7
8%;
Afr
ican
–Am
eric
an7%
; Lat
ina
4%;
othe
r 11
%
Hig
h sc
hool
gra
duat
e 22
%,
colle
ge g
radu
ate
63%
; inc
ome:
<$5
000
54%
, $50
00–1
9.99
935
%, $
20,0
00 <
12%
; sin
gle
63%
74%
met
SC
ID c
riter
ia fo
rsu
bsta
nce
depe
nden
ce fo
rm
ore
than
one
dru
g, 5
8%fo
r cu
rren
t co
cain
e ab
use
orde
pend
ence
, and
52%
for
alco
hol d
epen
denc
e. E
ight
part
icip
ants
prim
arily
abu
sed
coca
ine,
six
opi
ates
, fou
rm
ariju
ana,
one
met
ham
phet
amin
e, o
neha
luci
noge
ns, a
nd o
ne b
oth
coca
ine
and
met
ham
phet
amin
e.Pa
rtic
ipan
ts c
omm
only
suffe
red
Axi
s I n
on-
subs
tanc
e us
e di
sord
er(7
9% li
fetim
e; 5
0% c
urre
nt)
and
post
-tra
umat
ic s
tres
sdi
sord
er (3
8% c
urre
nt a
ndlif
etim
e). 1
2% w
ere
diag
nose
d w
ith a
ntiso
cial
pers
onal
ity d
isord
er
Part
icip
ants
wer
em
atch
ed o
n ag
e, s
ever
ityon
dru
g de
pend
ence
(bas
ed o
n SC
ID r
atin
gs),
read
ines
s to
cha
nge
and
glob
al a
djus
tmen
t (A
xis
V,D
SM-IV
) usin
g a
min
imisa
tion
rand
omas
signm
ent
proc
edur
e
Health Technology Assessment 2006; Vol. 10: No. 35
91
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TA
BLE
24
Patie
nt c
hara
cter
istic
s: R
CTs
(con
t’d)
Stud
yM
etho
ds fo
r di
agno
sis
Age
Gen
der
Ethn
icit
yEd
ucat
ion/
soci
o-ec
onom
ic
Pati
ent
hist
ory
Bas
elin
e co
mpa
rabi
lity
of d
isor
der
(yea
rs)
back
grou
nd
cont
inue
d
Line
han
et a
l.,20
0255
(1) P
DE;
(2) S
CID
-II fo
rD
SM-IV
; (3)
opi
ate
depe
nden
ce a
ccor
ding
to
SCID
-I
Mea
n36
.1±
7.3
(18–
45)
FC
auca
sian
66%
;A
fric
an–A
mer
ican
26%
; Mix
ed 4
%
Div
orce
d 52
%, m
arrie
d 4%
;ne
ver
been
mar
ried
44%
; hig
hsc
hool
96%
; bus
ines
s/te
chni
cal
scho
ol 4
8%; c
olle
ge 4
%;
grad
uate
or
prof
essio
nal s
choo
lw
ithou
t gr
adua
ting
18%
;em
ploy
ed 5
2%
Dep
ende
nce
on: c
ocai
ne52
%, s
edat
ives
13%
,ca
nnab
is 8.
7%, a
lcoh
ol 2
6%
Co-
mor
bidi
ty: m
ajor
depr
essiv
e di
sord
er o
rdy
sthy
mia
39%
, anx
iety
diso
rder
52%
, eat
ing
diso
rder
18%
65%
of t
he s
ampl
e re
port
eda
hist
ory
of a
t le
ast
one
suic
ide
atte
mpt
or
inte
ntio
nal s
elf-
inju
ry
No
signi
fican
t be
twee
n-gr
oup
diffe
renc
es w
ere
dete
cted
for
diag
nose
s,le
vel o
f gen
eral
func
tioni
ng o
r pa
rasu
icid
eac
ts p
rior
to t
reat
men
t
Mun
roe-
Blum
and
Mar
zial
i,19
9556
DIB
18–5
2F
n =
89;
M
n=
21
NR
Mos
t ha
d co
mpl
eted
hig
h sc
hool
.M
ost
wer
e cu
rren
tly u
nmar
ried,
but
had
had
a pr
ior
mar
riage
or
live-
in r
elat
ions
hip.
Ove
r ha
lf ha
dch
ildre
n, w
ho o
ften
wer
e no
tre
sidin
g w
ith t
hem
. Mos
t ha
d a
hist
ory
of s
ome
form
of
empl
oym
ent,
albe
it in
term
itten
t
One
-thi
rd o
f the
sam
ple
expe
rienc
ed s
igni
fican
tbe
havi
oura
l and
soc
ial
dysf
unct
ion
in t
he 6
mon
ths
befo
re t
reat
men
t; ab
out
half
had
repo
rted
pro
blem
s w
ithth
e la
w a
nd s
ubst
ance
abus
e. 8
5% r
epor
ted
prob
lem
s w
ith im
pulse
cont
rol.
All
part
icip
ants
use
dm
enta
l hea
lth/s
ocia
l ser
vice
s
The
re w
ere
nost
atist
ical
ly s
igni
fican
tdi
ffere
nces
on
clin
ical
and
soci
o-de
mog
raph
ic fa
ctor
sbe
twee
n tw
o gr
oups
or
betw
een
patie
nts
who
rem
aine
d in
the
stu
dy a
ndth
ose
who
with
drew
Turn
er, 2
00057
(1) 9
0-m
inut
e sc
reen
ing
inte
rvie
w b
ased
on
the
DIB
and
SC
ID fo
r D
SM-
III; (
2) P
DE
to c
ross
-va
lidat
e th
e BP
Ddi
agno
sis a
nd d
eter
min
eth
e pr
esen
ce o
fad
ditio
nal A
xis
IIdi
sord
ers
22 (18–
27)
F n
= 1
9;
M n
= 5
Cau
casia
nn
= 1
9;A
fric
an–A
mer
ican
n=
4;
Asia
n–A
mer
ican
n
= 1
Ave
rage
leve
l of e
duca
tion
13.3
year
s (r
ange
12–
16)
23 p
atie
nts
met
crit
eria
for
aco
-mor
bid
Axi
s I d
isord
er;
eigh
t pa
tient
s ha
d a
hist
ory
of b
rief p
sych
otic
or
para
noid
epi
sode
s
Base
line
diffe
renc
e w
asno
t st
atist
ical
ly s
igni
fican
t,�
2 (1) =
2.2
7, p
= 0
.132
Appendix 3
92 TA
BLE
24
Patie
nt c
hara
cter
istic
s: R
CTs
(con
t’d)
Stud
yM
etho
ds fo
r di
agno
sis
Age
Gen
der
Ethn
icit
yEd
ucat
ion/
soci
o-ec
onom
ic
Pati
ent
hist
ory
Bas
elin
e co
mpa
rabi
lity
of d
isor
der
(yea
rs)
back
grou
nd
Tyre
r et
al.,
2003
58IC
D-1
0 di
agno
stic
crite
ria o
f the
per
sona
lity
sche
dule
Mea
n 32
M 3
2%W
hite
90%
(Eng
lish,
Sco
ttish
,W
elsh
)
Mar
ital s
tatu
s: s
ingl
e 55
%,
mar
ried
24%
, oth
er (d
ivor
ced,
sepa
rate
d, w
idow
ed) 2
1%; l
ivin
gal
one
34%
Sim
ple
diso
rder
26%
, diff
use
diso
rder
16%
Year
s sin
ce fi
rst
ever
para
suic
ide
8.8
(SD
8.7
)T
here
wer
e no
impo
rtan
tdi
ffere
nces
in b
asel
ine
char
acte
ristic
s be
twee
ntw
o gr
oups
and
tho
sew
ith a
nd w
ithou
t 12
-mon
th a
sses
smen
t
van
den
Bosc
het
al.,
200
259D
SM-IV
dia
gnos
is fo
rBP
D S
CID
-II, s
cree
ning
devi
ce (P
DQ
-4+
).Su
bsta
nce
abus
epr
oble
ms
wer
e as
sess
edw
ith E
urop
ASI
Mea
n34
.9(1
8–70
)
FD
utch
nat
iona
lity
97%
Educ
atio
n: D
BT 1
2.6
year
s,
TAU
13.
6ye
ars
Une
mpl
oyed
: DBT
26%
, TA
U 1
6%N
ever
mar
ried:
DBT
56%
, TA
U 6
8%Li
ving
alo
ne: D
BT 3
3%,
TAU
39%
Disa
bilit
y pe
nsio
n: D
BT 5
6%,
TAU
61%
Num
ber
of B
PD c
riter
ia(m
ean)
: DBT
n=
7.3
, TA
U n
= 7
.3H
istor
y of
sui
cide
att
empt
s:D
BT n
= 7
0%,
TAU
n=
71%
Hist
ory
of s
elf-
mut
ilatio
n:D
BT n
= 9
3%,
TAU
n=
94%
Life
time
self-
mut
ilatio
n ac
ts(m
edia
n): D
BT n
= 1
3.1,
TAU
n=
14.
4A
ddic
tive
prob
lem
s:
DBT
n=
59%
, TA
U n
= 5
2%
Ave
rage
num
ber
of d
ays
inre
siden
tial t
reat
men
ts in
the
past
4 y
ears
= 7
4 da
ys p
erye
ar; a
vera
ge n
umbe
r of
adm
issio
ns in
the
pas
t4
year
s ra
nged
from
four
to
58
The
re w
as n
o sig
nific
ant
diffe
renc
e be
twee
ntr
eatm
ent
cond
ition
s on
soci
o-de
mog
raph
icva
riabl
es, n
umbe
r of
DSM
-IV c
riter
ia fo
r BP
D,
hist
ory
of s
uici
deat
tem
pts,
num
ber
of s
elf-
mut
ilatin
g ac
ts, o
rpr
eval
ence
of c
linic
ally
signi
fican
t al
coho
l and
/or
drug
use
pro
blem
s
F, fe
mal
e; m
, Mal
e.
Health Technology Assessment 2006; Vol. 10: No. 35
93
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TA
BLE
25
Patie
nt c
hara
cter
istic
s: n
on-R
CT
Stud
yM
etho
ds fo
r di
agno
sis
Age
Gen
der
Ethn
icit
yEd
ucat
ion/
soci
o-ec
onom
ic
Pati
ent
hist
ory
Bas
elin
e co
mpa
rabi
lity
of d
isor
der
(yea
rs)
back
grou
nd
Wilb
erg
et a
l.,19
9860
DSM
-III (
DSM
-III-R
),H
SRS,
SC
ID-I
and
SCID
-II
Trea
tmen
tgr
oup:
mea
n 27
±5
Con
trol
grou
p:m
ean
32±
9
Trea
tmen
tgr
oup:
unc
lear
,th
e te
xt s
tate
sth
at t
here
was
only
one
mal
eou
t of
12
in t
hegr
oup,
but
the
sam
ple
char
acte
ristic
sta
ble
show
sth
at 9
2% o
fth
e gr
oup
wer
em
ale
Con
trol
gro
up:
71%
M,
(22/
31)
NR
Mea
n nu
mbe
r of
mon
ths
in w
ork
in t
he la
st y
ear
befo
re a
dmiss
ion:
trea
tmen
t gr
oup
8±3
and
for
the
cont
rol g
roup
7±
5
Thr
ee o
f the
pat
ient
s in
the
trea
tmen
t gr
oup
(25%
) and
20
in t
he c
ontr
ol g
roup
(65%
) had
ever
bee
n m
arrie
d
Trea
tmen
t gr
oup:
nin
epa
tient
s ha
d pr
evio
usho
spita
lisat
ions
; nin
e ha
d a
hist
ory
of s
uici
de a
ttem
pts;
thre
e ha
d an
anx
iety
diso
rder
, eig
ht a
moo
ddi
sord
er a
nd 8
a s
ubst
ance
use
diso
rder
Con
trol
gro
up: 1
7 pa
tient
sha
d pr
evio
usho
spita
lisat
ions
; 13
had
ahi
stor
y of
sui
cide
att
empt
s;13
had
an
anxi
ety
diso
rder
,12
a m
ood
diso
rder
and
20
a su
bsta
nce
use
diso
rder
The
con
trol
gro
up w
assig
nific
antly
you
nger
tha
nth
e tr
eatm
ent
grou
p(p
<0.
001)
and
had
less
ofte
n be
en m
arrie
d(p
<0.
05).
The
tre
atm
ent
grou
p ha
d a
signi
fican
tlylo
nger
sta
y in
the
day
hosp
ital (
11 v
s 6
mon
ths,
p<
0.05
)
Appendix 3
94 TA
BLE
26
Out
com
es a
nd a
naly
sis in
form
atio
n: R
CTs
Stud
yO
utco
mes
In
stru
men
ts
Mea
sure
men
t pe
riod
sIT
T a
naly
sis
Bate
man
and
Fon
agy,
199
951(1
) Eng
agin
g th
e pa
tient
s in
tre
atm
ents
; (2)
red
uctio
nin
gen
eral
psy
chia
tric
sym
ptom
s; (3
) dec
reas
e in
the
num
ber
of s
elf-
dest
ruct
ive
acts
and
sui
cide
att
empt
s;(4
) im
prov
emen
t in
soc
ial a
nd in
terp
erso
nal f
unct
ions
;(5
) pre
vent
ion
of r
elia
nce
on p
rolo
nged
hos
pita
l sta
ys
(1) S
uici
de a
nd S
elf-
Har
m In
vent
ory;
(2) S
CL-
90-R
; (3)
BD
I; (4
) Spi
elbe
rgSt
ate-
Trai
t A
nxie
ty In
vent
ory;
(5
) mod
ified
SA
S; (6
) Inv
ento
ry o
fIn
terp
erso
nal P
robl
ems
Base
line,
3, 6
, 9, 1
2, 1
5an
d 18
mon
ths
No;
cro
ss-o
ver
and
drop
out
patie
nts
wer
e no
t in
clud
edin
the
ana
lyse
s
Koon
s et
al.,
200
152(1
) Whe
ther
res
earc
h th
erap
ist o
f cur
rent
stu
dy c
ould
cond
uct
DBT
with
ade
quat
e ad
here
nce;
(2) o
utco
mes
of t
he D
BT a
re s
uper
ior
to t
hose
of u
sual
car
e in
the
sam
e se
ttin
g an
d sy
stem
: par
asui
cide
, sui
cida
l ide
atio
nan
d ho
pele
ssne
ss, m
ood
and
emot
ion,
diss
ocia
tion,
psyc
hiat
ric in
patie
nt a
dmiss
ion,
BPD
crit
eria
(1) P
HI;
(2) B
eck
Scal
e fo
r Su
icid
eId
eatio
n; (3
) BH
S; (4
) BD
I; (5
) HA
M-
D; (
6) H
ARS
; (7)
Spi
elbe
rg A
nger
Expr
essio
n Sc
ale;
(8) D
ES
Pret
reat
men
t, 3
and
6m
onth
sN
o
Line
han
et a
l., 1
99153
(1) P
aras
uici
de; (
2) m
aint
enan
ce in
the
the
rapy
; (3
) psy
chia
tric
inpa
tient
tre
atm
ent
(1) P
HI;
(2) s
elf-
repo
rt fo
rm o
f the
Scal
e fo
r Su
icid
e Id
eato
rs; (
3) B
DI;
(4) B
HS;
(5) R
easo
ns fo
r Li
ving
Inve
ntor
y, S
urvi
val a
nd C
opin
g Sc
ale
Pret
reat
men
t, 4,
8 a
nd
12 m
onth
sN
o
Line
han
et a
l., 1
99954
(1) A
dapt
atio
n of
the
orig
inal
DBT
man
ual f
or a
popu
latio
n of
sub
stan
ce-a
busin
g w
omen
with
BPD
;(2
) com
paris
on o
f its
effi
cacy
to
a co
ntro
l (TA
U)
cond
ition
: dru
g ab
use;
tre
atm
ent
initi
atio
n, e
xpos
ure
and
rete
ntio
n; p
sych
opat
holo
gy
(1) S
truc
ture
d cl
inic
al in
terv
iew
s;
(2) u
rine
anal
yses
; (3)
tre
atm
ent
hist
ory
inte
rvie
w; (
4) P
HI;
(5) S
HI;
(6) S
AS;
(7) L
ongi
tudi
nal I
nter
view
Follo
w-U
p Ev
alua
tion
base
sch
edul
e;(8
) GSA
; (9)
GA
S; (1
0) S
piel
berg
Stat
e–Tr
ait
Ang
er E
xpre
ssio
nIn
vent
ory
Pret
reat
men
t, 4,
8, 1
2 an
d 16
mon
ths
Yes
Line
han
et a
l., 2
00255
(1) D
ecre
ase
in o
piat
e de
pend
ence
; (2)
dec
reas
e in
BPD
sym
ptom
s: t
reat
men
t in
itiat
ion,
exp
osur
e an
dre
tent
ion;
dru
g us
e ou
tcom
es; p
sych
opat
holo
gy
(1) U
rine
anal
yses
Inte
rvie
w a
nd s
elf-
repo
rt m
easu
res:
(2) T
LFB;
(3) P
HI;
(4) S
HI;
(5) S
AS;
(6) L
ongi
tudi
nal I
nter
view
Fol
low
-up
Eval
uatio
n; (7
) GA
S; (8
) GSA
; (9
) GA
F; (1
0) B
SI
Pret
reat
men
t, 4,
8, 1
2an
d 16
mon
ths
Yes
cont
inue
d
Health Technology Assessment 2006; Vol. 10: No. 35
95
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TA
BLE
26
Out
com
es a
nd a
naly
sis in
form
atio
n: R
CTs
(con
t’d)
Stud
yO
utco
mes
In
stru
men
ts
Mea
sure
men
t pe
riod
sIT
T a
naly
sis
Mun
roe-
Blum
and
Mar
zial
i,19
9556
(1) B
ehav
iour
s re
late
d to
soc
ial d
ysfu
nctio
n;
(2) b
ehav
iour
s re
late
d to
soc
ial p
erfo
rman
ce;
(3) s
ympt
om s
tatu
s
(1) O
BI; (
2) S
AS;
(3) B
DI;
(4) H
SC-9
0Ba
selin
e, 6
, 12,
18
and
24m
onth
sN
o
TA
BLE
27
Out
com
es a
nd a
naly
sis in
form
atio
n: n
on-R
CTs
Stud
yO
utco
mes
In
stru
men
ts
Mea
sure
men
t pe
riod
sIT
T a
naly
sis
Wilb
erg
et a
l., 1
99860
GSI
, Glo
bal S
ympt
om In
dex.
Redu
ctio
n of
sub
stan
ce m
isuse
, sui
cide
att
empt
s,m
enta
l sta
teH
SRS,
GSI
At
adm
issio
n, d
ischa
rge
and
follo
w-u
pN
o
Turn
er, 2
00057
(1) S
uici
de/s
elf-
harm
ing
beha
viou
r; (2
) em
otio
nal
dysr
egul
atio
n; (3
) im
pact
of t
reat
men
t in
dica
tors
; (4
) hos
pita
lisat
ion
(1) H
AM
-D; (
2) B
PRS;
(3) T
arge
tBe
havi
our
Ratin
gs; (
4) B
DI;
(5) B
AI;
(6) B
eck
Scal
e fo
r Su
icid
e Id
eatio
n
Pret
reat
men
t, 6
and
12m
onth
sYe
s
Tyre
r et
al.,
200
358(1
) Par
asui
cide
eve
nts
(incl
udin
g su
icid
e) in
the
follo
win
g ye
ar; (
2) u
ptak
e of
MA
CT
ses
sions
; (3
) sym
ptom
impr
ovem
ents
(DSH
, anx
iety
,de
pres
sion
and
soci
al fu
nctio
ning
) at
1 ye
ar; (
4) O
ther
outc
omes
(1) H
AD
S; (2
) GA
F; (3
) SFQ
; (4)
EQ
-5D
; (5)
CSR
I; (6
) PH
IBa
selin
e, 6
and
12m
onth
sYe
s
van
den
Bosc
h et
al.,
200
259(1
) Tre
atm
ent
rete
ntio
n; (2
) hig
h-ris
k be
havi
our,
incl
udin
g su
icid
al, s
elf-
mut
ilatin
g an
d se
lf-da
mag
ing
impu
lsive
beh
avio
urs;
(3) w
heth
er t
he e
ffica
cy o
f DBT
is m
odifi
ed b
y ba
selin
e se
verit
y of
par
asui
cide
(1) B
PDSI
(sem
i-str
uctu
red
inte
rvie
w).
BPD
SI c
onsis
ts o
f nin
ese
ctio
ns, o
ne fo
r ea
ch o
f the
DSM
-IVcr
iteria
for
BPD
. The
par
asui
cide
sect
ion
incl
udes
thr
ee it
ems;
the
impu
lsive
sec
tion
incl
udes
11
item
s;(2
) LPC
Base
line,
11,
12,
33,
44
and
52 w
eeks
; 18
mon
ths
follo
w-u
p
Yes
Appendix 3
96 TA
BLE
28
Resu
lts o
f rep
orte
d ou
tcom
es (
psyc
holo
gica
l sym
ptom
s an
d in
terp
erso
nal a
nd s
ocia
l fun
ctio
ning
: RCT
s
Stud
yR
esul
tsO
ther
out
com
e in
form
atio
n
cont
inue
d
Bate
man
and
Fon
agy,
1999
51(1
) Spi
elbe
rg In
vent
ory
scor
e
Base
line
End-
poin
t (1
8 m
onth
s)Fo
llow
-up
(36
mon
ths)
Mea
nSD
Adj
uste
d m
ean
Mea
nSD
Adj
uste
d m
ean
Mea
nSD
Adj
uste
d m
ean
Stat
ePH
68.4
7.0
–52
.5
11.5
51.3
32.6
5.9
32.3
TAU
63.2
6.8
–65
.59.
366
.652
.410
.352
.9
Trai
tPH
66.5
6.1
–56
.89.
155
.234
.46.
134
.1TA
U62
.09.
9–
61.0
7.6
36.0
42.7
10.1
43.4
(2) B
eck
Dep
ress
ion
Scal
e Base
line
End-
poin
t (1
8 m
onth
s)Fo
llow
-up
(36
mon
ths)
Mea
nSD
Adj
uste
d m
ean
Mea
nSD
Adj
uste
d m
ean
Mea
nSD
Adj
uste
d m
ean
PH36
.07.
6–
20.6
7.0
20.3
11.9
3.3
11.9
TAU
34.9
7.4
–35
.27.
435
.720
.410
.520
.6
(3) G
loba
l Sev
erity
Inde
x sc
ore
Base
line
End-
poin
t (1
8mon
ths)
Follo
w-u
p (3
6 m
onth
s)
Mea
nSD
Adj
uste
d m
ean
Mea
nSD
Adj
uste
d m
ean
Mea
nSD
Adj
uste
d m
ean
PH2.
500.
58–
2.10
0.82
2.1
0.8
0.6
0.8
TAU
2.30
0.71
–2.
400.
702.
42.
00.
52.
0
(4) P
ositi
ve S
ympt
om t
otal
sco
re
Base
line
End-
poin
t (1
8 m
onth
s)Fo
llow
-up
(36
mon
ths)
Mea
nSD
Adj
uste
d m
ean
Mea
nSD
Adj
uste
d m
ean
Mea
nSD
Adj
uste
d m
ean
PH74
.114
.5–
70.7
17.3
72.0
40.6
19.6
40.2
TAU
72.3
15.2
–73
.115
.073
.574
.59.
675
.3
(5) I
n th
e PH
gro
up, t
here
was
a cl
ear
redu
ctio
n of
sui
cide
atte
mpt
s fr
om 9
4.7%
on
adm
issio
n to
5.3
%
(mea
n 0.
16) a
t 18
mon
ths.
Thi
s tr
end
was
hig
hly
signi
fican
t (K
enda
ll’s
W=
0.59
,�
2=
33.5
, df=
3, p
<0.
001)
.N
o sig
nific
ant
tren
d fo
r th
eco
ntro
l gro
up (K
enda
ll’s
W=
0.04
, �2
=2.
4, d
f=3,
ns).
Gro
up d
iffer
ence
s em
erge
d by
6 m
onth
s. T
he n
umbe
r of
indi
vidu
als
who
wer
e no
long
erpa
rasu
icid
al w
as s
igni
fican
tlygr
eate
r in
the
PH
gro
up t
han
inth
e TA
U g
roup
by
12 m
onth
s(�
2=
4.3,
df =
1, p
<0.
05)
Health Technology Assessment 2006; Vol. 10: No. 35
97
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TA
BLE
28
Resu
lts o
f rep
orte
d ou
tcom
es (
psyc
holo
gica
l sym
ptom
s an
d in
terp
erso
nal a
nd s
ocia
l fun
ctio
ning
: RCT
s
Stud
yR
esul
tsO
ther
out
com
e in
form
atio
n
cont
inue
d
Koon
s et
al.,
200
152Pr
eM
idPo
stPr
e–m
idPr
e–po
stPr
e vs
Pos
t G
roup
×F
Fef
fect
siz
eaT
ime
F 2,3
6
Mea
nSD
Mea
nSD
Mea
nSD
Para
suic
ides
pas
t 3
mon
ths
DBT
5.1
13.2
1.6
3.7
0.4
1.3
2.5
4.75
†0.
352.
44†
TAU
0.7
1.3
1.1
2.3
12.
20.
030.
010.
28Su
icid
al id
eatio
nD
BT36
.213
.534
.913
.526
.28
0.89
9.64
*0.
983.
71*
TAU
44.6
11.4
41.9
13.3
41.5
14.3
1.61
2.89
0.54
Hop
eles
snes
sD
BT11
.96.
70.
47.
55.
15.
31.
6317
.08*
*1.
318.
03**
TAU
13.6
6.8
127.
814
.27.
31.
250.
3–0
.18
Ham
ilton
Dep
ress
ion
DBT
29.7
13.7
24.7
10.1
17.1
5.7
4.67
†12
.40*
*1.
120.
71TA
U32
.69.
731
.111
.324
.37.
80.
889.
09*
0.95
Beck
Dep
ress
ion
DBT
22.8
11.1
21.3
13.4
13.4
7.5
0.25
9.35
*0.
963.
70*
TAU
34.7
14.6
2714
.629
.317
.725
.40*
*5.
93*
0.77
Ham
ilton
Anx
iety
DBT
18.4
7.3
18.1
8.4
19.1
7.5
0.02
0.13
–0.3
11.
32TA
U27
.79.
325
.810
.732
.212
.41.
071.
79–0
.42
Ang
er in
DBT
22.9
5.7
19.3
5.4
17.3
46.
19*
11**
1.04
1.71
TAU
20.5
4.7
18.2
5.4
19.2
6.2
6.99
**0.
310.
17A
nger
out
DBT
18.2
5.7
17.3
4.8
14.5
3.9
0.67
13.3
8**
1.16
5.89
**TA
U17
.25.
814
.63.
117
.96.
13.
160.
16–0
.12
Diss
ocia
tion
DBT
22.3
15.2
2016
.213
.212
0.79
13**
1.13
1.21
TAU
4122
.429
.522
.530
.623
.33.
052.
40.
48BP
D c
riter
iaD
BT6.
81.
13.
61.
679
.45*
*2.
830.
79b
TAU
6.7
0.8
4.2
2.3
12.6
4**
1.13
† p <
0.1
; *p
< 0
.05;
**p
< 0
.01.
aEf
fect
siz
e co
mpu
ted
as (m
ean
pre-
mea
n po
st) (
SD p
re-p
ost)
.b
BPD
crit
eria
wer
e as
sess
ed o
nly
at p
retr
eatm
ent
and
post
trea
tmen
t. D
egre
es o
f fre
edom
for
Fte
st a
re (1
,18)
.
The
pro
port
ion
of p
atie
nts
with
any
adm
issio
ns d
urin
g th
epr
ior
3 m
onth
s w
as r
elat
ivel
ylo
w a
t pr
etre
atm
ent,
and
neith
er g
roup
sho
wed
signi
fican
t ch
ange
in t
his
prop
ortio
n by
the
end
of
trea
tmen
t
Appendix 3
98 TA
BLE
28
Resu
lts o
f rep
orte
d ou
tcom
es (
psyc
holo
gica
l sym
ptom
s an
d in
terp
erso
nal a
nd s
ocia
l fun
ctio
ning
: RCT
s (c
ont’d
)
Stud
yR
esul
tsO
ther
out
com
e in
form
atio
n
cont
inue
d
Line
han
et a
l., 1
99153
(1) P
aras
uici
deLi
kelih
ood
of a
ny p
aras
uici
de:
DBT
63.
6%, T
AU
95.
5%, z
2.26
, p <
0.0
05
Med
ical
risk
sco
res:
DBT
mea
n 9.
21, S
D 8
.22,
n =
14
TAU
mea
n 17
.86,
SD
20.
94, n
= 2
1(t
= 1
.70,
df =
28.
01, p
< 0
.05)
Part
icip
ants
who
rec
eive
d D
BT h
ad a
med
ian
of 1
.5 p
aras
uici
de a
cts
per
year
com
pare
d w
ith n
ine
acts
per
yea
r fo
rco
ntro
l par
ticip
ants
(2) M
aint
enan
ce in
the
rapy
DBT
pat
ient
s w
ere
signi
fican
tly m
ore
likel
y to
sta
rt in
divi
dual
the
rapy
tha
n w
ere
cont
rol p
artic
ipan
ts, a
ll of
who
m w
ere
refe
rred
for
trea
tmen
t (1
00%
and
73%
for
part
icip
ants
ass
igne
d to
DBT
and
con
trol
par
ticip
ants
, res
pect
ivel
y, z
= 2
.75,
p <
0.0
03)
(3) P
sych
iatr
ic in
patie
nt t
reat
men
tN
umbe
r of
hos
pita
lised
par
ticip
ants
DBT
(n)
Con
trol
(n)
zp
0–4
mon
ths
69
2.54
<0.
005
4–8
mon
ths
57
8–12
mon
ths
37
1.49
<0.
10Ye
ar8
121.
70<
0.05
(4) T
hrou
ghou
t th
e fo
llow
-up
year
, bot
h th
e pa
rasu
icid
ere
peat
rat
e (D
BT 2
6%,
TAU
60%
, z =
2.1
2, p
< 0
.01)
and
the
likel
ihoo
d of
any
psyc
hiat
ric h
ospi
talis
atio
n(D
BT 1
1%, T
AU
40%
, z
= 1
.43,
p <
0.0
7) w
ere
low
er fo
r pa
rtic
ipan
tsco
mpl
etin
g D
BT t
han
for
part
icip
ants
rem
aini
ng in
TA
U
(5) P
sych
iatr
ic in
patie
nt d
ays
durin
g th
e 18
–24-
mon
thpe
riod
wer
e lo
wer
for
part
icip
ants
com
plet
ing
trea
tmen
t
(6) A
t th
e en
d of
the
24-
mon
thtim
e-po
int,
part
icip
ants
com
plet
ing
DBT
wer
e ra
ted
signi
fican
tly h
ighe
r on
ove
rall
soci
al a
djus
tmen
t by
the
inte
rvie
wer
Health Technology Assessment 2006; Vol. 10: No. 35
99
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TA
BLE
28
Resu
lts o
f rep
orte
d ou
tcom
es (
psyc
holo
gica
l sym
ptom
s an
d in
terp
erso
nal a
nd s
ocia
l fun
ctio
ning
: RCT
s (c
ont’d
)
Stud
yR
esul
tsO
ther
out
com
e in
form
atio
n
cont
inue
d
Line
han
et a
l., 1
99954
(1) D
rug
abus
e (IT
T a
naly
ses,
n =
28)
DBT
TAU
Mea
nSD
Mea
nSD
F-va
lue
Effe
ct s
ize
(p-v
alue
)
Pret
reat
men
t as
sess
men
t0.
360.
260.
220.
28
Pret
reat
men
t to
4 m
onth
s0.
630.
340.
320.
373.
160.
80(p
< 0
.05)
4–8
mon
ths
0.62
0.35
0.38
0.34
1.50
0.65
8–12
mon
ths
0.67
0.38
1.39
0.44
1.67
0.64
(n=
8)
Year
tot
al0.
630.
330.
350.
342.
830.
93(n
= 1
0)(p
< 0
.05)
12–1
6 m
onth
s0.
940.
170.
580.
364.
040.
59(p
< 0
.05)
(2) T
reat
men
t in
itiat
ion,
exp
osur
e an
d re
tent
ion:
DBT
mea
n 43
.14±
10.6
7; T
AU
mea
n 31
.6±
27.8
8, F
1,15
= 1
.07,
ns
Whe
n ca
se m
anag
emen
t ho
urs
are
excl
uded
from
the
se a
naly
ses,
DBT
par
ticip
ants
rec
eive
sig
nific
antly
mor
eps
ycho
ther
apy
than
do
TAU
par
ticip
ants
D
BT: m
ean
43.1
4±10
.67;
TA
U: m
ean
21.8
8±32
.32,
F1,
15=
2.0
7, p
< 0
.05.
TA
U p
artic
ipan
ts r
arel
y pa
rtic
ipat
ed in
gro
upps
ycho
ther
apy
(3) P
artic
ipan
ts a
s a
grou
p sh
owed
sig
nific
ant
redu
ctio
ns o
ver
time
on fr
eque
ncy
of p
aras
uici
de e
piso
des
and
stat
e an
dtr
ait
ange
r
(4) P
sych
opat
holo
gy. T
here
wer
e no
bet
wee
n-gr
oup
diffe
renc
es o
n ot
her
outc
ome
mea
sure
s (e
.g. p
aras
uici
deep
isode
s, G
SA, G
AS
or a
nger
)du
ring
trea
tmen
t or
at
the
12-m
onth
pos
t-tr
eatm
ent
follo
w-u
p
At
the
16-m
onth
follo
w-u
pas
sess
men
t D
BT p
artic
ipan
tssh
owed
bet
ter
soci
al a
ndgl
obal
adj
ustm
ent,
with
signi
fican
tly lo
wer
(bet
ter)
scor
es o
n th
e G
SA:
DBT
: mea
n 2.
25±
0.75
; TA
U:
mea
n 2.
92±
0.71
, F1,
12=
3.9
8,p
<0.
05 fo
r be
st w
eek
scor
es
DBT
: mea
n 3.
04±
0.89
; TA
U: m
ean
3.74
±0.
67,
F 1,1
2=
2.94
, p=
0.0
56 fo
r la
stm
onth
sco
res
and
high
ersc
ores
on
the
GA
S
DBT
: mea
n 69
±12
; TA
U: m
ean
49±
10,
F 1,1
2=
22.2
4, p
<0.
001
for
best
wee
k sc
ores
DBT
: mea
n 62
±10
; TA
U: m
ean
44±
10,
F 1,1
2=
22.
19, p
< 0
.001
for
last
wee
k sc
ores
Appendix 3
100 TA
BLE
28
Resu
lts o
f rep
orte
d ou
tcom
es (
psyc
holo
gica
l sym
ptom
s an
d in
terp
erso
nal a
nd s
ocia
l fun
ctio
ning
: RCT
s (c
ont’d
)
Stud
yR
esul
tsO
ther
out
com
e in
form
atio
n
cont
inue
d
Line
han
et a
l., 2
00255
(1) T
reat
men
t in
itiat
ion,
exp
osur
e an
d re
tent
ion.
DBT
CVT
+12
S
Year
tot
alM
ean
SDM
ean
SDt 2
1p
Mea
n nu
mbe
r of
indi
vidu
al s
essio
ns r
ecei
ved
33.2
20.4
33.0
9.6
2.62
<0.
05N
umbe
r of
ski
lls g
roup
ses
sions
26.6
15.9
10.8
12.8
(2) U
rinan
alys
isBy
wee
k 52
(end
of 1
2 m
onth
s); D
BT p
artic
ipan
ts h
ad a
sig
nific
antly
low
er p
erce
ntag
e of
opi
ate-
posit
ive
urin
anal
yses
tha
nC
VT+
12S
part
icip
ants
(t =
2.3
2, p
< 0
.02)
At
16 m
onth
s pe
rcen
tage
of p
ositi
ve u
rinan
alys
es in
bot
h co
nditi
ons:
DBT
27%
, CVT
+12
S 33
%
(3) S
elf-
repo
rt
Year
tot
alD
BTC
VT+
12S
Mea
n (%
)SD
Mea
n (%
)SD
Self-
repo
rtH
eroi
n66
.26
67.0
191
.14
91.8
6C
ocai
ne83
.12
80.9
378
.90
72.0
8A
mph
etam
ines
99.9
799
.92
100.
0010
0.00
Barb
itura
tes
99.9
799
.92
99.9
299
.84
Seda
tives
99.8
299
.39
99.8
399
.48
Uri
nana
lyse
sH
eroi
n46
.43
68.3
453
.30
83.4
0C
ocai
ne70
.86
80.6
672
.27
78.5
8A
mph
etam
ines
88.0
789
.87
91.7
288
.68
Barb
itura
tes
88.4
690
.11
92.4
588
.14
Seda
tives
86.3
689
.24
90.9
288
.22
(4) P
sych
opat
holo
gyin
bot
h gr
oups
:
Pret
reat
men
t12
mon
ths
Mea
nSD
Mea
nSD
Zp
BSI
1.78
711.
170.
603.
17<
0.00
2BS
I (16
-mon
th fo
llow
-up)
GA
S37
.65.
647
.410
.73.
59<
0.00
1
BSI (
16-m
onth
follo
w-u
p): z
= 1
.76,
p <
0.0
8, m
ean
0.98
±0.
74).
(5) G
loba
l adj
ustm
ent
impr
ovem
ents
wer
em
aint
aine
d, b
ut d
id n
otim
prov
e fu
rthe
r. T
he in
cide
nts
of p
aras
uici
dal b
ehav
iour
and
psyc
hiat
ric o
r dr
ug-r
elat
edvi
sits
to e
mer
genc
y ro
oms
and
inpa
tient
uni
ts w
ere
low
, but
ther
e w
ere
no s
igni
fican
tbe
twee
n-co
nditi
on d
iffer
ence
s
Health Technology Assessment 2006; Vol. 10: No. 35
101
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TA
BLE
28
Resu
lts o
f rep
orte
d ou
tcom
es (
psyc
holo
gica
l sym
ptom
s an
d in
terp
erso
nal a
nd s
ocia
l fun
ctio
ning
: RCT
s (c
ont’d
)
Stud
yR
esul
tsO
ther
out
com
e in
form
atio
n
cont
inue
d
Mun
roe-
Blum
and
Mar
zial
i, 19
9556
The
re w
ere
no s
igni
fican
t fin
ding
s w
ith r
espe
ct t
o th
e m
ajor
out
com
e va
riabl
es u
nder
stu
dy; h
owev
er, b
oth
trea
tmen
tsdi
d ex
perie
nce
signi
fican
t im
prov
emen
ts o
ver
time
refle
cted
on
beha
viou
ral i
ndic
ator
s, s
ocia
l adj
ustm
ent,
glob
alsy
mpt
oms
and
depr
essio
n
Coh
ort
anal
ysis
of v
aria
nce
Scal
e m
eans
(SD
) at
thre
e po
ints
in t
ime
OBI
SAS
HSC
-90
BDI
Pret
reat
men
t32
.01
(10.
88)
2.13
(0.4
2)1.
76 (0
.68)
25.9
(9.8
9)12
mon
ths
30.9
9 (1
2.67
)1.
91 (0
.50)
1.26
(0.6
9)18
.4 (1
2.46
)24
mon
ths
23.6
1 (1
0.58
)1.
89 (0
.59)
1.03
(0.7
8)14
.6 (1
2.29
)na
4843
4546
F 2.9
4=
10F 2
.84
= 7
F 2.8
8=
16
F 2.9
0=
17
0.76
b0.
040.
420.
93
a N
umbe
rs v
ary
beca
use
of m
issin
g da
ta.
b p
= 0
.000
1.
The
re w
ere
no s
igni
fican
tdi
ffere
nces
in o
utco
me
betw
een
the
two
trea
tmen
t-tim
e ex
posu
re g
roup
s; t
he lo
wex
posu
re g
roup
mad
e ga
ins
com
para
ble
to t
hose
of t
hehi
gh e
xpos
ure
grou
p on
all
ofth
e ou
tcom
e m
easu
res
Turn
er, 2
00057
(1) M
easu
res
CC
T (n
= 1
2)D
BT (n
= 1
2)
Pret
reat
men
t6
mon
ths
12 m
onth
sPr
etre
atm
ent
6 m
onth
s12
mon
ths
Ratin
g of
par
asui
cide
7.25
(0.7
5)4.
33 (1
.92)
4.25
(2.1
8)7.
17 (0
.83)
2.08
(2.0
2)1.
50 (1
.98)
Beck
Sui
cide
Idea
tion
Scal
e23
.53
(3.3
4)13
.33
(9.7
9)11
.58
(9.2
1)24
.08
(3.7
3)2.
83 (3
.49)
3.83
(8.0
3)N
umbe
r of
sui
cide
/13
.58
(3.3
4)6.
75 (5
.97)
5.58
(5.2
8)14
.08
(3.7
3)2.
17 (1
.95)
0.75
(1.2
3)se
lf-ha
rm a
ttem
pts
Ratin
g of
impu
lsive
ness
7.58
(0.5
1)6.
67 (0
.78)
6.08
(1.0
8)7.
42 (0
.51)
5.83
(0.8
3)4.
58 (1
.62)
Ratin
g of
ang
er7.
08 (0
.90)
5.92
(0.7
9)5.
67 (1
.15)
7.33
(0.6
5)5.
00 (1
.21)
4.67
(1.3
0)BD
I27
.75
(6.1
1)24
.75
(4.9
4)24
.08
(5.5
5)27
.58
(5.3
0)18
.08
(7.9
1)14
.92
(8.2
6)H
AM
-D17
.42
(4.4
6)13
.67
(2.9
3)12
.58
(3.9
0)20
.75
(4.3
3)8.
58 (6
.58)
7.50
(5.9
6)BA
I20
.42
(3.4
5)17
.08
(5.8
2)14
.83
(6.3
4)19
.25
(3.5
5)12
.58
(4.8
9)10
.17
(6.5
3)BP
RS30
.83
(6.0
0)25
.83
(8.4
0)25
.33
(3.9
4)30
.33
(6.5
6)18
.42
(7.3
3)18
.17
(7.9
0)H
ospi
talis
atio
n da
ys10
.00
(8.1
1)10
.75
(16.
27)
13.0
0 (1
5.34
)10
.20
(3.3
7)2.
67 (6
.58)
0.75
(1.9
6)
(2) A
sses
smen
t of
tre
atm
ent
cred
ibili
ty
The
ana
lysis
foun
d no
signi
fican
t di
ffere
nce
betw
een
trea
tmen
ts r
egar
ding
cred
ibili
ty, (
t 22
= 1
.63,
p
= 0
.116
)
(3) A
sses
smen
t of
hel
ping
allia
nce
diffe
renc
es
Hel
ping
Rel
atio
nshi
pQ
uest
ionn
aire
(HRQ
) sho
wed
no s
igni
fican
t di
ffere
nce
betw
een
trea
tmen
ts
(F1,
16=
1.1
0, p
= 0
.31)
Appendix 3
102 TA
BLE
28
Resu
lts o
f rep
orte
d ou
tcom
es (
psyc
holo
gica
l sym
ptom
s an
d in
terp
erso
nal a
nd s
ocia
l fun
ctio
ning
: RCT
s (c
ont’d
)
Stud
yR
esul
tsO
ther
out
com
e in
form
atio
n
cont
inue
d
Tyre
r et
al.,
200
358Re
sults
rel
ated
to
BPD
are
onl
y re
port
ed w
here
it is
pos
sible
to
diffe
rent
iate
(1) P
ropo
rtio
ns o
f tot
al 4
30 p
atie
nts
expe
rienc
ing
a re
peat
sel
f-ha
rm (p
aras
uici
de) e
piso
de o
ver
1ye
ar, i
ncid
ence
rat
e an
dfr
eque
ncy
of s
elf-
harm
epi
sode
s in
tho
se s
epar
ated
by
cate
gory
of P
D (I
CD
-10
cate
gorie
s)
Para
suic
ide
even
ts d
urin
g fo
llow
-up
BPD
PD
Diff
use
No
PD(n
= 6
7)(n
= 1
17)
(n=
69)
(n=
39)
Non
e30
(44.
8%)
57 (4
8.7%
)31
(44.
9%)
31 (7
9.5%
)A
t le
ast
one
37 (5
5.2%
)60
(51.
3%)
38 (5
5.1%
)8
(20.
5%)
Sam
ple
sizea
(n=
60)
Inci
denc
e ra
tea
Firs
t ep
isode
per
per
son-
year
follo
w-u
p0.
762
0.63
40.
659
0.22
925
th p
erce
ntile
tim
e to
par
asui
cide
eve
nt (d
ays)
a89
132
162
Freq
uenc
y of
sel
f-ha
rm (p
er y
ear)
3.15
aT
hese
figu
res
refe
r to
the
orig
inal
400
pat
ient
s fo
r w
hom
det
aile
d in
form
atio
n on
firs
t se
lf-ha
rm w
as a
vaila
ble.
(2) P
ropo
rtio
n of
430
pat
ient
s ex
perie
ncin
g a
repe
at s
elf-
harm
(par
asui
cide
) epi
sode
ove
r 1
year
MA
CT
(PD
)TA
U (P
D)
Para
suic
ide
even
ts d
urin
g fo
llow
-up
0–6
mon
ths
6–12
mon
ths
0–6
mon
ths
6–12
mon
ths
(n=
91)
(n=
88)
(n=
90)
(n=
83)
Inci
denc
e ra
tea
Non
e57
(62.
2%)
55 (6
2.5%
)49
(54.
4%)
61(7
3.5%
)A
t le
ast
one
34 (3
7.4%
)33
(37.
5%)
41 (4
5.6%
)22
(26.
5%)
Firs
t ep
isode
per
per
son-
year
follo
w-u
p0.
584
0.71
25th
per
cent
ile t
ime
to p
aras
uici
de e
vent
(day
s)a
212
89Fr
eque
ncy
of s
elf-
harm
(per
6-m
onth
per
iod)
1.65
1.18
3.74
(1.6
3b )3.
61 (0
.88b )
Freq
uenc
y of
sel
f-ha
rm (p
er y
ear)
2.84
7.36
(2.5
4b )
a T
hese
figu
res
refe
r to
the
orig
inal
400
pat
ient
s fo
r w
hom
det
aile
d in
form
atio
n on
firs
t se
lf-ha
rm w
as a
vaila
ble.
b Ex
clud
ing
patie
nts
with
420
epi
sode
s.
(3) T
he p
sych
omet
ricas
sess
men
t ou
tcom
es s
how
edno
diff
eren
ce b
etw
een
MA
CT
and
TAU
Health Technology Assessment 2006; Vol. 10: No. 35
103
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TA
BLE
28
Resu
lts o
f rep
orte
d ou
tcom
es (
psyc
holo
gica
l sym
ptom
s an
d in
terp
erso
nal a
nd s
ocia
l fun
ctio
ning
: RCT
s (c
ont’d
)
Stud
yR
esul
tsO
ther
out
com
e in
form
atio
n
van
den
Bosc
h et
al.,
2002
59
(Ver
haul
et
al.,
2003
64)
(1) T
reat
men
t re
tent
ion
Sign
ifica
ntly
mor
e pa
tient
s w
ho w
ere
rece
ivin
g D
BT (n
= 1
7; 6
3%) t
han
patie
nts
in t
he c
ontr
ol g
roup
(n =
7; 2
3%)
cont
inue
d in
the
rapy
with
the
sam
e th
erap
ist fo
r th
e en
tire
year
(�2 1
= 9
.70,
p =
0.0
02)
(2) H
igh-
risk
beha
viou
rsT
he fr
eque
ncy
and
cour
se o
f sui
cida
l beh
avio
urs
wer
e no
t sig
nific
antly
diff
eren
t ac
ross
tre
atm
ent
grou
ps (t
1,13
7=
0.0
3,p
=0.
866)
and
the
inte
ract
ion
betw
een
time
and
trea
tmen
t co
nditi
on (t
1,16
6=
0.2
2, p
= 0
.639
) did
not
rea
ch s
tatis
tical
signi
fican
ceSe
lf-m
utila
ting
beha
viou
rs in
DBT
pat
ient
s gr
adua
lly d
imin
ished
ove
r th
e tr
eatm
ent
year
, whe
reas
pat
ient
s in
TA
U g
roup
grad
ually
det
erio
rate
d: a
sig
nific
ant
effe
ct w
as T
ime
x Tr
eatm
ent
cond
ition
(t1,
44.4
= 1
0.24
, p =
0.0
03),
but
not
for
trea
tmen
t co
nditi
on a
lone
(t1,
69.1
= 3
.80;
p=
0.0
55).
The
mos
t co
mm
on s
elf-
mut
ilatin
g ac
ts w
ere
cutt
ing,
bur
ning
,pr
icki
ng a
nd h
ead
bang
ing
(3) I
mpa
ct o
f DBT
on
seve
rity
of s
ubst
ance
use
pro
blem
s at
18-
mon
th fo
llow
-up
Euro
pASI
item
(a)
DBT
, mea
n±SD
TAU
, mea
n±SD
Com
paris
on a
t 18
-mon
th fo
llow
-up
corr
ecte
d fo
r ba
selin
eb
Base
line
Follo
w-u
paBa
selin
eFo
llow
-upa
Fp
(n=
27)
(n =
20)
(n =
31)
(n =
24)
Day
s ≥
5 dr
inks
pas
t 7.
1±10
.36.
1±9.
86.
2±9.
23.
8±7.
80.
90.
34m
onth
s 0–
30
Day
s m
edic
atio
n us
e pa
st
14.2
±14
.07.
9±12
.213
.5±
14.5
11.5
±13
.90.
40.
54m
onth
s 0–
30
Day
s ca
nnab
is us
e pa
st
6.5±
11.2
9.2±
13.3
2.3±
5.8
5.9±
11.5
0.1
0.73
mon
ths
0–30
Day
s al
coho
l pro
blem
s pa
st
8.7±
12.3
7.0±
11.3
9.0±
12.9
6.7±
11.3
00.
89m
onth
s 0–
30
Day
s dr
ug p
robl
ems
past
8.
1±11
.49.
5±13
.29.
0±12
.64.
5±10
.02
0.17
mon
ths
0–30
Seve
rity
of a
lcoh
ol
2.7±
2.3
2.8±
2.6
3.0±
2.5
2.4±
2.1
1.1
0.31
prob
lem
s 0–
9
Seve
rity
of d
rug
prob
lem
s 0–
93.
3±2.
02.
8±2.
23.
6±2.
32.
3±1.
80.
50.
47a
Follo
w-u
p sc
ores
at
18 m
onth
s sin
ce s
tart
of t
reat
men
t.b
Usin
g th
e G
ener
al L
inea
r M
odel
Mod
ule
of S
PSS
8.0,
with
Eur
opA
SI s
core
s at
18-
mon
th fo
llow
-up
as d
epen
dent
varia
bles
, tre
atm
ent
cond
ition
as
fixed
fact
or a
nd b
asel
ine
scor
es o
n Eu
ropA
SI a
s co
varia
tes.
(4) I
mpa
ct o
f bas
elin
e se
verit
yon
effe
ctiv
enes
sFo
r su
icid
al b
ehav
iour
an
alm
ost
signi
fican
t ef
fect
was
evid
ent
for
the
thre
e-w
ayin
tera
ctio
n te
rm T
ime
xTr
eatm
ent
x C
ondi
tion
(t1,
170
= 4
.81,
p=
0.0
29),
indi
catin
g a
tren
d to
war
dsgr
eate
r ef
fect
iven
ess
of D
BT in
seve
rely
affe
cted
indi
vidu
als.
For
self-
mut
ilatin
g be
havi
ours
asig
nific
ant
effe
ct w
as e
vide
ntfo
r th
e 3-
way
inte
ract
ion
term
Tim
e ×
Seve
rity
×Tr
eatm
ent
cond
ition
(t1,
404
= 1
6.82
,p
=0.
000)
and
the
inte
ract
ion
term
Sev
erity
×Tr
eatm
ent
cond
ition
(t1,
67.6
= 9
.63,
p=
0.00
3), i
ndic
atin
g th
atD
BT w
as s
uper
ior
to T
AU
inth
e hi
gh-s
ever
ity g
roup
, but
not
for
thei
r lo
wer
sev
erity
coun
terp
arts
. No
diffe
rent
ial
effe
ctiv
enes
s w
as fo
und
for
self-
dam
agin
g im
pulsi
vity
Appendix 3
104 TA
BLE
29
Resu
lts o
f rep
orte
d ou
tcom
es (
psyc
holo
gica
l sym
ptom
s an
d in
terp
erso
nal a
nd s
ocia
l fun
ctio
ning
): no
n-RC
Ts
Stud
yR
esul
tsO
ther
out
com
e in
form
atio
n
Wilb
erg
et a
l., 1
99860
(1) S
tatu
s di
men
sions
at
adm
issio
n, d
ischa
rge
and
follo
w-u
p
Tota
l sam
ple
(n=
43)
Trea
tmen
t gr
oup
(n=
12)
Con
trol
gro
up (n
= 3
1)
HSR
S ad
miss
ion
38.6
±5.
136
.9±
5.1
39.2
±5.
1H
SRS
disc
harg
e43
.5±
7.8
47.3
±5.
442
.0±
8.1*
HSR
S fo
llow
-up
51.2
±11
.957
.4±
9.0
48.8
±12
.1*
GSI
adm
issio
n1.
84±
0.55
(n =
42)
1.67
±0.
481.
92±
0.56
(n =
30)
GSI
disc
harg
e1.
28±
0.65
(n =
35)
1.11
±0.
491.
37±
0.72
(n =
23)
GSI
follo
w-u
p1.
37±
0.67
(n =
36)
1.02
±0.
501.
55±
0.67
(n =
24)
*Re
hosp
italis
atio
n13
(33%
) (n
= 4
0)1
(8%
)12
(43%
) (n
= 2
8)Su
icid
e at
tem
pts
6 (1
5%) (
n =
44)
1 (8
%)
5 (1
8%) (
n =
28)
Rem
issio
n fr
om s
ubst
ance
use
diso
rder
13/2
5 (5
2%)
6/8
(75%
)7/
17 (4
1%)
*p <
0.0
5 (t
-tes
t, tw
o-ta
iled)
(2) M
ultip
le r
egre
ssio
n m
odel
s fo
r H
SRS
at fo
llow
-up
(95%
CI)
Inde
pend
ent
varia
ble
BSE
B�
Min
.M
ax.
Sign
ifica
nce
HSR
S at
adm
issio
n0.
638
0.32
60.
274
–0.0
231.
298
0.05
8C
ontin
uous
tre
atm
ent
0.19
60.
010
0.26
6–0
.006
0.39
80.
057
Wor
k be
fore
adm
issio
n0.
899
0.32
70.
365
0.23
81.
561
0.00
9O
utpa
tient
gro
up t
hera
py7.
537
3.46
40.
288
0.52
414
.549
0.03
6
R2=
0.3
9, F
= 6
.15,
sig
nific
ance
F =
0.0
006
(3) M
ultip
le r
egre
ssio
n m
odel
s fo
r G
SI a
t fo
llow
-up
(95%
CI)
Inde
pend
ent
varia
ble
BSE
B�
Min
.M
ax.
Sign
ifica
nce
GSI
at
adm
issio
n0.
162
0.17
80.
140
–0.2
010.
525
0.37
0C
ontin
uous
tre
atm
ent
0.00
60.
007
0.14
2–0
.008
0.02
10.
362
% o
f fol
low
-up
perio
d on
med
icat
ion
0.00
50.
003
0.32
00.
0002
0.01
00.
041
Out
patie
nt g
roup
the
rapy
–0.5
520.
214
–0.4
10–0
.989
–0.1
150.
015
R2=
0.3
5, F
= 4
.08,
sig
nific
ance
F=
0.0
09
Health Technology Assessment 2006; Vol. 10: No. 35
105
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TA
BLE
30
Patie
nt p
refe
renc
es a
nd c
oncl
usio
ns: R
CTs
Stud
yPa
tien
t pr
efer
ence
, sat
isfa
ctio
n an
d C
oncl
usio
nsac
cept
abili
ty o
f tre
atm
ent
Bate
man
and
Fon
agy,
199
951N
RPa
tient
s tr
eate
d w
ith P
H fo
r 18
mon
ths
show
ed s
igni
fican
t im
prov
emen
t on
bot
h sy
mpt
omat
ic a
ndcl
inic
al m
easu
res.
Tre
atm
ent
was
effe
ctiv
e fo
r bo
th m
en a
nd w
omen
Koon
s et
al.,
200
152N
RT
his
stud
y de
mon
stra
tes
that
DBT
can
be
cond
ucte
d w
ith r
easo
nabl
y go
od a
dher
ence
by
a gr
oup
of t
hera
pist
s at
a s
ite in
depe
nden
t of
the
tre
atm
ent’s
dev
elop
er. T
he t
reat
men
t w
as a
ssoc
iate
dw
ith c
linic
ally
sig
nific
ant
chan
ges
in t
he s
ympt
oms
and
func
tions
of b
orde
rline
pat
ient
s, c
hang
esth
at w
ere
signi
fican
tly g
reat
er t
han
thos
e as
soci
ated
with
TA
U o
n a
num
ber
of m
easu
res.
Effi
cacy
of D
BT is
not
lim
ited
only
to
patie
nts
with
rec
urre
nt s
uici
dal a
nd s
elf-
inju
rious
beh
avio
ur
Line
han
et a
l., 1
99153
No
trea
tmen
t sp
ecifi
c ga
ins
in g
ener
alsa
tisfa
ctio
n de
spite
sig
nific
ant
impr
ovem
ents
inan
ger
redu
ctio
n an
d so
cial
adj
ustm
ent
(1) T
he a
utho
rs fo
und
a sig
nific
ant
redu
ctio
n in
the
freq
uenc
y an
d m
edic
al r
isk o
f par
asui
cida
lbe
havi
our
amon
g pa
tient
s w
ho r
ecei
ved
DBT
com
pare
d w
ith t
hat
for
cont
rol p
artic
ipan
ts
(2) D
BT e
ffect
ivel
y re
tain
ed p
atie
nts
in t
hera
py
(3) D
ays
of p
atie
nts’
psy
chia
tric
hos
pita
lisat
ion
wer
e fe
wer
for
part
icip
ants
who
rec
eive
d D
BT t
han
for
cont
rol p
artic
ipan
ts
(4) D
BT w
as n
ot d
iffer
entia
lly e
ffect
ive
in im
prov
ing
patie
nts’
dep
ress
ion,
hop
eles
snes
s, s
uici
deid
eatio
n or
rea
sons
for
livin
g
Line
han
et a
l., 1
99954
NR
(1) T
he a
utho
rs fo
und
a sig
nific
ant
redu
ctio
n in
sub
stan
ce a
buse
am
ong
part
icip
ants
ass
igne
d to
DBT
com
pare
d w
ith t
hose
ass
igne
d to
TA
U
(2) D
BT m
ore
effe
ctiv
ely
reta
ined
par
ticip
ants
in t
reat
men
t
(3) I
mpr
ovem
ents
in s
ocia
l and
glo
bal a
djus
tmen
t in
the
DBT
con
ditio
n w
ere
obse
rved
and
reac
hed
signi
fican
ce c
ompa
red
with
TA
U a
t fo
llow
-up
Line
han
et a
l., 2
00255
cont
inue
d
NR
(1) B
oth
trea
tmen
ts w
hen
com
bine
d w
ith L
AA
M w
ere
effe
ctiv
e in
red
ucin
g op
iate
use
and
mai
ntai
ning
the
red
uctio
n du
ring
the
4-m
onth
follo
w-u
p pe
riod
(2) C
VT+
12S
was
rem
arka
bly
effe
ctiv
e in
mai
ntai
ning
par
ticip
ants
in t
reat
men
t; 10
0% s
taye
d fo
rth
e en
tire
year
(3) P
artic
ipan
ts a
ssig
ned
to D
BT w
ere
signi
fican
tly m
ore
accu
rate
in s
elf-
repo
rtin
g op
iate
use
tha
nw
ere
thos
e as
signe
d to
CVT
+12
S
Appendix 3
106 TA
BLE
30
Patie
nt p
refe
renc
es a
nd c
oncl
usio
ns: R
CTs
(con
t’d)
Stud
yPa
tien
t pr
efer
ence
, sat
isfa
ctio
n an
d C
oncl
usio
nsac
cept
abili
ty o
f tre
atm
ent
Mun
roe-
Blum
and
Mar
zial
i,19
9556
NR
IGP
show
ed s
igni
fican
t im
prov
emen
ts o
n al
l maj
or o
utco
mes
(as
wel
l as
TAU
) at
follo
w-u
p. T
hegr
oup
ther
apy
appe
ars
mor
e co
st-e
ffect
ive
than
indi
vidu
al t
hera
py
Turn
er, 2
00057
NR
DBT
-orie
ntat
ed t
hera
py c
an b
e m
ore
effe
ctiv
e th
an p
rovi
ding
onl
y th
e su
ppor
tive
com
pone
nts
ofps
ycho
ther
apy.
DBT
-orie
ntat
ed p
atie
nts
show
ed g
reat
er im
prov
emen
ts t
han
patie
nts
rece
ivin
gC
CT
on
mea
sure
s of
sui
cide
and
sel
f-ha
rm b
ehav
iour
, sui
cide
idea
tion,
dep
ress
ion,
impu
lsive
ness
,an
ger
and
glob
al p
sych
olog
ical
func
tioni
ng, a
nd a
red
uctio
n in
day
s sp
ent
in p
sych
iatr
ic h
ospi
tals
Tyre
r et
al.,
200
358N
RT
he r
esul
ts s
how
ed n
o di
ffere
nce
with
reg
ard
to t
he o
utco
mes
bet
wee
n th
e tw
o gr
oups
.H
owev
er, t
he M
AC
T in
terv
entio
n w
as c
heap
er, l
ed t
o so
mew
hat
few
er e
piso
des
of s
elf-
harm
and
was
ass
ocia
ted
with
few
er s
uici
des
than
in t
he T
AU
gro
up
van
den
Bosc
h et
al.,
200
259Bo
th s
ubgr
oups
(BPD
and
BPD
with
sub
stan
ceab
use)
of p
atie
nts
appe
ared
to
get
alon
g ea
sily
by t
he s
econ
d w
eek.
Thr
ough
the
disc
ussio
n of
hom
ewor
k, t
he s
ubst
ance
-abu
sing
and
non-
subs
tanc
e-ab
usin
g pa
rtic
ipan
ts r
ealis
ed t
hat
they
shar
ed m
ost
of t
he e
ssen
tial b
orde
rline
prob
lem
s. A
ll pa
rtic
ipan
ts ju
dged
the
prog
ram
me
as v
alid
atin
g an
d he
lpfu
l. T
hey
felt
ackn
owle
dged
as
bord
erlin
e pa
tient
s an
d ju
dged
the
trea
tmen
t as
ver
y im
port
ant.
Sess
ion
atte
ndan
ce fo
r th
e to
tal g
roup
was
81%
(1) D
BT h
ad a
sub
stan
tially
low
er 1
2-m
onth
att
ritio
n ra
te t
han
TAU
(2) D
BT r
esul
ted
in g
reat
er r
educ
tions
in s
elf-
mut
ilatin
g be
havi
ours
and
sel
f-da
mag
ing
impu
lsive
acts
tha
n TA
U
(3) T
he b
enef
icia
l im
pact
on
the
freq
uenc
y of
sel
f-m
utila
ting
beha
viou
rs w
as fa
r m
ore
pron
ounc
edin
par
ticip
ants
who
rep
orte
d hi
gher
bas
elin
e fr
eque
ncie
s th
an in
tho
se r
epor
ting
low
er b
asel
ine
freq
uenc
ies
TA
BLE
31
Patie
nt p
refe
renc
es a
nd c
oncl
usio
ns: n
on-R
CT
Stud
yPa
tien
t pr
efer
ence
, sat
isfa
ctio
n an
d C
oncl
usio
nsac
cept
abili
ty o
f tre
atm
ent
Wilb
erg
et a
l., 1
99860
NR
“The
res
ults
sup
port
the
clin
ical
exp
erie
nce
that
a t
reat
men
t m
odel
com
bini
ng d
ay t
reat
men
t an
dou
tpat
ient
gro
up p
sych
othe
rapy
may
be
favo
urab
le fo
r se
lect
ed p
atie
nts
with
BPD
”
Health Technology Assessment 2006; Vol. 10: No. 35
107
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Appendix 4
Excluded studies
TABLE 32 Excluded studies
Study Population Reason for exclusion
RCTsAllard et al., 199281 Suicide attempters No BPD subgroup analysisChiesa et al., 200482 PD No BPD subgroup analysisClarkin et al., 200483 BPD Ongoing trial, no evaluable dataEvans et al., 199930 DSH No BPD subgroup analysis Manning, 199784 BPD Abstract, no evaluable dataPiper et al., 199385 PD No BPD subgroup analysisSimpson et al., 200486 BPD Combination of drug with DBTSloane et al., 197587 PD No BPD subgroup analysisSpringer et al., 199688 PD No BPD subgroup analysisStiwne et al., 199489 BPD Therapist assessmentWinston et al.,199490 PD No BPD subgroup analysis
Non-RCTsBohus et al., 200091 BPD DBTBohus et al., 200492 BPD DBTBrobin et al., 198793 BPD Case studyBattegay and Klaui, 198694 BPD Qualitative studyBuzov et al., 198595 BPD Qualitative studyChiesa et al., 200096 PD Qualitative studyClarkin et al., 199197 BPD ReviewClarkin et al., 199498 BPD No comparison Clarkin et al., 200126 BPD No comparisonDamman et al., 200199 BPD ReviewDolan et al., 1997100 PD Not eligible comparison (admitted vs not admitted patients)Hirvas, 1987101 BPD No comparisonKarterud et al., 2004102 BPD Not eligible comparison (day hospital vs Bateman’s MBT)Kent and Hartstone, 2000103 BPD Abstract, no dataKern et al., 1997104 BPD Case seriesKoenigsberg, 1982105 BPD Diagnostic paperLopez et al., 2004106 BPD No comparisonMeares et al., 1999107 BPD Not eligible comparison (treatment vs waiting list)Pfitzer et al., 1990108 BPD No comparisonRathus and Miller, 2002109 Suicide attempters AdolescentsRyle and Golynkina, 2000110 BPD No comparisonSchane and Kovel, 1988111 BPD Case seriesUshijima, 1994112 BPD Case seriesWildgoose et al., 2001113 BPD No comparison
Health Technology Assessment 2006; Vol. 10: No. 35
109
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Appendix 5
Consensus trial quality ratings according to Lackner’s quality checklist
TABLE 33 Quality ratings
Bateman Koons Linehan Linehan Linehan Munroe-Blum Turner Tyrer van den Wilberg and et al., et al., et al., et al., and Marziali, 200057 et al., Bosch et al.,
Fonagy, 200152 199153 199954 200255 199556 200358 et al., 199860
199951 200259
Q1 2 2 2 2 2 2 2 2 2 2Q2 2 2 2 2 2 2 2 2 2 2Q3 2 2 2 2 2 2 2 2 2 0Q4 0 0 0 0 0 0 0 0 0 0Q5 2 2 2 2 1 2 2 2 2 2Q6 2 2 0 2 2 1 2 2 2 1Q7 2 2 1 2 2 2 2 2 2 2Q8 2 2 2 2 2 2 2 2 2 2Q9 2 2 2 2 2 2 2 2 2 2Q10 0 0 0 0 0 0 0 0 0 0Q11 0 0 0 0 0 0 0 0 0 0Q12 2 2 2 2 2 2 2 2 2 2Q13 2 2 2 2 2 2 2 2 2 2Q14 0 1 1 1 1 0 1 0 0 0Q15 1 2 2 1 1 1 1 2 2 2Q16 0 0 0 0 0 0 0 0 0 0Q17 2 1 2 2 2 2 2 2 2 2Q18 0 2 0 0 0 2 0 2 0 0Q19 0 0 0 0 0 2 0 2 1 0Q20 2 2 2 2 2 1 2 1 2 1Q21 2 2 2 2 2 0 2 2 2 1Q22 2 2 2 2 2 1 2 2 2 1Q23 0 0 0 2 2 0 2 2 2 0Q24 0 2 2 2 2 2 0 2 2 2Q25 0 0 0 0 2 0 0 0 0 0Q26 2 0 0 2 2 2 2 2 2 1Q27 2 2 2 2 2 2 2 2 2 1Q28 0 0 0 0 0 0 0 0 0 2Q29 1 1 2 2 2 0 1 0 1 0Total 34 37 34 40 41 34 37 41 40 30
Q1. Were exclusion criteria specified and numberof exclusion/refusals reported?
Q2. Were formal diagnostic criteria used toconfirm [BPD] and/or inclusion criteriaspecified?
Q3. What was the method (e.g. randomisation)and adequacy of the method by whichpatients were allocated to treatment arms?
Q4. Was allocation concealed from those involvedin patient recruitment?
Q5. Were patients comparable on prognosticvariables, and were statistical proceduresused to adjust for differences in analyses?
Q6. How well were sample demographics andclinical characteristics described?
Q7. What was the source and representativenessof participants?
Q8. Was compliance with experimentalprocedure (e.g. attendance and checks foradherence with behavioral assignments)conducted?
Q9. How clearly was the content of therapeuticand control conditions (e.g. manualisedtreatment procedures) operationalised?
Q10. Were participants blind to treatmentallocation, and, if so, was integrity test
Appendix 5
110
conducted? (replaced with credibilitycriterion)
Q11. Were therapy credibility and expectancy forimprovement assessed? (Credibility criterion)
Q12 .Were the objectives and main outcomesspecified a priori?
Q13. Were outcome measures clearly describedand/or psychometrically sound outcomemeasure used?
Q14. Was a blind assessor used, and, if so, wasintegrity of blinding tested?
Q15. Were the number and reasons for withdrawalby group recorded?
Q16. Were details on side effects recorded bygroup?
Q17. What was the planned duration of trialincluding follow-up?
Q18. Were power calculations stated a priori?Q19. Was sample size (number per group)
adequate?
Q20. Were appropriate statistical analysesconducted (including correction for multipletests where applicable)?
Q21. Did presented results include data forreanalysis of main outcomes (e.g. pointestimates and measures of variability foreach primary outcome such as standarddeviations, 95% confidence interval)?
Q22. Were conclusions justified?Q23. Were withdrawals included in analyses?Q24. Was declaration of interests (e.g. source of
funding) stated?Q25. Was declaration of allegiance to therapy
stated?Q26. Was a post-treatment follow-up conducted
for all groups?Q27. Were cointerventions avoided or equal across
conditions?Q28. Were consecutive participants recruited?Q29. Was concurrent drug use recorded?
Health Technology Assessment 2006; Vol. 10: No. 35
111
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Appendix 6
British Medical Journal checklist for economic evaluations50
TABLE 34 Heard study68
Item Assessment
The study design1. The research question is stated The economic hypothesis was that DBT is more cost-effective
than TAU in the treatment of BPD
2. The economic importance of the research The disease is costly in terms of resources used. An intervention question is stated that reduces outcomes may impact on these costs
3. The viewpoint(s) of the analysis are clearly A societal perspective was taken, including hospital inpatient and stated and justified outpatient visits and physician visits. Costs incurred in the
community or the criminal justice system were not included
4. The rationale for choosing the alternative TAU is a relevant measure for evaluating the opportunity cost of programmes or interventions compared is stated the new treatment
5. The alternatives being compared are clearly DBT and TAU therapy is clearly describeddescribed
6. The form of economic evaluation used is stated Cost-effectiveness
7. The choice of form of economic evaluation is Yesjustified in relation to the questions addressed
Data collection8. The source(s) of effectiveness estimates are used Yes
as stated
9. Details of the design and results of effectiveness Brief description of design, no resultsstudy are given (if based on a single study)
10. Details of the method of synthesis or meta-analysis Based on a single trialof estimates are given (if based on an overview of a number of effectiveness studies)
11. The primary outcome measure(s) for the economic Yes (employment or global functioning cost-effectiveness ratios)evaluation are clearly stated
12. Methods to value health states and other Health states not valuedbenefits are stated
13. Details of the subjects from whom valuations NAwere obtained are given
14. Productivity changes (if included) are reported Yesseparately
15. The relevance of productivity changes to the Yesstudy question is discussed
16. Quantities of resources are reported separately Yesfrom their unit costs
17. Methods for the estimation of quantities and unit Yescosts are described
continued
Appendix 6
112
TABLE 34 Heard study68 (cont’d)
Item Assessment
18. Currency and price data are recorded Yes
19. Details of currency of price adjustments for Yes (Employment Cost Index used to inflate prices to 1999 inflation or currency conversion are given levels)
20. Details of any model used are given No model used
21. The choice of model used and the key parameters NAon which it is based are justified
Analysis and interpretation of results22. Time horizon of costs and benefits is stated Trial period (12 months)
23. The discount rate(s) is stated NA
24. The choice of rate(s) is justified NA
25. An explanation is given if costs or benefits are NAnot discounted
26. Details of statistical tests and confidence intervals t-Test, ordinary least squares regression, bootstrapping. CIs for are given for stochastic data overall results not given, only SDs for individual resources
reported
27. The approach to sensitivity analysis is given One-way and PSA
28. The choice of variables for sensitivity analysis Yesis justified
29. The ranges over which the variables are varied Local costs converted to national costs. No other sensitivity are stated analysis on costs reported
30. Relevant alternatives are compared Yea
31. Incremental analysis is reported Yes
32. Major outcomes are presented in a disaggregated Yesas well as aggregated form
33. The answer to the study question is given Yes
34. Conclusions follow from the data reported Yes
35. Conclusions are accompanied by the appropriate Yescaveats
Health Technology Assessment 2006; Vol. 10: No. 35
113
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
TABLE 35 Byford study69
Item Assessment
Study design1. The research question is stated The economic hypothesis was that MACT is more cost-effective
than TAU in the treatment of DSH
2. The economic importance of the research question The disease is costly in terms of resources used. An intervention is stated that reduces outcomes may impact on these costs
3. The viewpoint(s) of the analysis are clearly stated A broad societal perspective was taken, including hospital, and justified community and social services and the criminal justice system
4. The rationale for choosing the alternative TAU is a relevant measure for evaluating the opportunity cost of programmes or interventions compared is stated the new treatment
5. The alternatives being compared are clearly MACT is not well described here. However, it is described in described the trial publication58
6. The form of economic evaluation used is stated Cost-effectiveness
7. The choice of form of economic evaluation is Yesjustified in relation to the questions addressed
Data collection8. The source(s) of effectiveness estimates are used Yes
as stated
9. Details of the design and results of effectiveness Brief description of design, no resultsstudy are given (if based on a single study)
10. Details of the method of synthesis or meta-analysis Based on a single trialof estimates are given (if based on an overview of a number of effectiveness studies)
11. The primary outcome measure(s) for the Yes (cost per QALY)economic evaluation are clearly stated
12. Methods to value health states and other benefits Yes (EQ-5D)are stated
13. Details of the subjects from whom valuations were Yesobtained are given
14. Productivity changes (if included) are reported Yesseparately
15. The relevance of productivity changes to the study Yesquestion is discussed
16. Quantities of resources are reported separately Yesfrom their unit costs
17. Methods for the estimation of quantities and unit Yescosts are described
18. Currency and price data are recorded Yes
19. Details of currency of price adjustments for inflation Yes (Hospital and Community Health Services Pay and Prices or currency conversion are given Index)
20. Details of any model used are given No model used
21. The choice of model used and the key parameters NAon which it is based are justified
continued
Appendix 6
114
TABLE 35 Byford study69 (cont’d)
Item Assessment
Analysis and interpretation of results22. Time horizon of costs and benefits is stated Trial period (12 months)
23. The discount rate(s) is stated NA
24. The choice of rate(s) is justified NA
25. An explanation is given if costs or benefits are not NAdiscounted
26. Details of statistical tests and confidence intervals t-Test, ordinary least squares regression, bootstrapping. CIs for are given for stochastic data overall results not given, only SDs for individual resources
reported
27. The approach to sensitivity analysis is given One-way and PS
28. The choice of variables for sensitivity analysis Yesis justified
29. The ranges over which the variables are varied Local costs converted to national costs. No other sensitivity are stated analysis on costs reported
30. Relevant alternatives are compared Yes
31. Incremental analysis is reported Yes
32. Major outcomes are presented in a disaggregated as Yeswell as aggregated form
33. The answer to the study question is given Yes
34. Conclusions follow from the data reported Yes
35. Conclusions are accompanied by the appropriate Yescaveats
Health Technology Assessment 2006; Vol. 10: No. 35
115
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Appendix 7
Case studies
Appendix 7
116 TA
BLE
36
Tabl
e of
cos
t st
udie
s
Stud
ySu
bjec
tsSe
ttin
gSt
udy
leng
thTr
eatm
ents
Res
ourc
es c
oste
dC
osts
, mea
n (S
D)
Com
men
t
Bate
man
, 200
370BP
DH
alliw
ick
Uni
t, 18
mon
ths
(a) P
artia
l hos
pita
lisat
ion;
Inpa
tient
/out
patie
nt
(a) £
19,3
64 (£
10,9
66)
Hea
lthca
re u
tilisa
tion
of a
ll BP
D p
atie
nts
UK
(b) g
ener
al p
sych
iatr
ic
psyc
hiat
ry, p
artia
l (b
) £21
,969
(£17
,985
)w
ho p
artic
ipat
ed in
a p
revi
ous
tria
l of
serv
ices
(TA
U)
hosp
italis
atio
n,
part
ial h
ospi
tal t
reat
men
t co
mpa
red
with
m
edic
atio
nTA
U w
as a
sses
sed
usin
g in
form
atio
nfr
om c
ase
note
s an
d se
rvic
e pr
ovid
ers.
Cos
ts w
ere
com
pare
d fo
r th
e 6
mon
ths
befo
re t
reat
men
t, 18
mon
ths
oftr
eatm
ent
and
an 1
8-m
onth
follo
w-u
ppe
riod
Hal
l, 20
0171
BPD
Wes
tmea
d 12
mon
ths
No
form
al
Any
hos
pita
l A
vera
ge c
ost
per
The
sta
ted
aim
of t
his
stud
y w
as t
o H
ospi
tal,
psyc
hoth
erap
y ad
miss
ion
patie
nt A
US$
7309
cond
uct
a pr
elim
inar
y co
st–b
enef
it st
udy
Aus
tral
ia12
mon
ths
befo
re
of t
he e
ffect
of o
utpa
tient
psy
chot
hera
py,
with
12
mon
ths
of
twic
e a
wee
k fo
r 1
year
, in
30 b
orde
rline
ps
ycho
ther
apy
patie
nts
. How
ever
, thi
s is
not
a tr
ueco
st–b
enef
it st
udy
This mapping uses data from a recentlypublished RCT of supervised self-help CBT in
primary care for patients with depression115 thatincorporated EQ-5D and CORE (ClinicalOutcomes in Routine Practice). These patientswere recruited from 17 primary healthcare teamsin the NHS. It used CORE rather than the BDI,but CORE is a depression-specific questionnairethat is similar in many ways to the BDI and it hasbeen mapped onto the BDI by the developer ofthe CORE (Barkham: personal communication).The mapping function was fitted to these data toprovide BDI data on each case.
This provided 62 patients with predicted BDIscores and EQ-5D data. The BDI score has been
fitted to EQ-5D data using a simple linear modelby ordinary least squares. The model produced inSPSS is summarised in Tables 37–39.
The interpretation of the constant at over 1 onlypresents problems for patients with BDI scoresbelow 5.0, which does not arise in the mean BDIscores to which it is applied in the studies reportedhere. More complex models such as curvilinearones did not greatly improve the fit of the model.
A better model would have fitted the original itemresponses (e.g. Brazier116), but because these BDIscores were derived from the CORE, item-leveldata were not available. The adjusted R-squaredsuggests at best a moderate fit and, moreimportantly, its dependence on the BDI meansthat many of the quality of life consequences forpeople suffering from BPD are not considered.The quality of life score probably under-representsthe impact of BPD Nonetheless, the BDI doesfocus on affect, so at least reflects the way in whichit impacts on the feelings and happiness of peoplewith BPD.
Health Technology Assessment 2006; Vol. 10: No. 35
117
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Appendix 8
Mapping BDI to EQ-5D
TABLE 37 Model summarya
Model R R2 Adjusted SE of the R2 estimate
1 0.534b 0.285 0.273 0.262183
a Dependent variable: EQ-5D overall utility (tariff).b Predictors: (constant), BDIPRED.
TABLE 38 ANOVAa
Model Sum of squares df Mean square F Significance
1 Regression 1.645 1 1.645 23.937 0.000b
Residual 4.124 60 0.069Total 5.770 61
a Dependent variable: EQ-5D overall utility (tariff).b Predictors: (constant), BDIPRED.
TABLE 39 Coefficientsa
Model Unstandardised coefficients Standardised coefficients t Significance
B SE �
1 (Constant) 1.110 0.111 9.954 0.000BDI score –0.021 0.004 –0.534 –4.893 0.000
a Dependent variable: EQ-5D overall utility (tariff).
Health Technology Assessment 2006; Vol. 10: No. 35
133
Health Technology AssessmentProgramme
Prioritisation Strategy GroupMembers
Chair,Professor Tom Walley, Director, NHS HTA Programme,Department of Pharmacology &Therapeutics,University of Liverpool
Professor Bruce Campbell,Consultant Vascular & GeneralSurgeon, Royal Devon & ExeterHospital
Dr Edmund Jessop, MedicalAdvisor, National Specialist,Commissioning Advisory Group(NSCAG), Department ofHealth, London
Professor Jon Nicholl, Director,Medical Care Research Unit,University of Sheffield, Schoolof Health and Related Research
Dr John Reynolds, ClinicalDirector, Acute GeneralMedicine SDU, RadcliffeHospital, Oxford
Dr Ron Zimmern, Director,Public Health Genetics Unit,Strangeways ResearchLaboratories, Cambridge
Director, Professor Tom Walley, Director, NHS HTA Programme,Department of Pharmacology &Therapeutics,University of Liverpool
Deputy Director, Professor Jon Nicholl,Director, Medical Care ResearchUnit, University of Sheffield,School of Health and RelatedResearch
HTA Commissioning BoardMembers
Programme Director, Professor Tom Walley, Director, NHS HTA Programme,Department of Pharmacology &Therapeutics,University of Liverpool
Chair,Professor Jon Nicholl,Director, Medical Care ResearchUnit, University of Sheffield,School of Health and RelatedResearch
Deputy Chair, Professor Jenny Hewison,Professor of Health CarePsychology, Academic Unit ofPsychiatry and BehaviouralSciences, University of LeedsSchool of Medicine
Dr Jeffrey AronsonReader in ClinicalPharmacology, Department ofClinical Pharmacology,Radcliffe Infirmary, Oxford
Professor Deborah Ashby,Professor of Medical Statistics,Department of Environmentaland Preventative Medicine,Queen Mary University ofLondon
Professor Ann Bowling,Professor of Health ServicesResearch, Primary Care andPopulation Studies,University College London
Dr Andrew Briggs, PublicHealth Career Scientist, HealthEconomics Research Centre,University of Oxford
Professor John Cairns, Professorof Health Economics, PublicHealth Policy, London School ofHygiene and Tropical Medicine,London
Professor Nicky Cullum,Director of Centre for EvidenceBased Nursing, Department ofHealth Sciences, University ofYork
Mr Jonathan Deeks, Senior Medical Statistician,Centre for Statistics inMedicine, University of Oxford
Dr Andrew Farmer, SeniorLecturer in General Practice,Department of Primary Health Care, University of Oxford
Professor Fiona J Gilbert,Professor of Radiology,Department of Radiology,University of Aberdeen
Professor Adrian Grant,Director, Health ServicesResearch Unit, University ofAberdeen
Professor F D Richard Hobbs,Professor of Primary Care &General Practice, Department ofPrimary Care & GeneralPractice, University ofBirmingham
Professor Peter Jones, Head ofDepartment, UniversityDepartment of Psychiatry,University of Cambridge
Professor Sallie Lamb, Professor of Rehabilitation,Centre for Primary Health Care, University of Warwick
Professor Stuart Logan,Director of Health & SocialCare Research, The Peninsula Medical School, Universities of Exeter &Plymouth
Dr Linda Patterson, Consultant Physician,Department of Medicine,Burnley General Hospital
Professor Ian Roberts, Professorof Epidemiology & PublicHealth, Intervention ResearchUnit, London School ofHygiene and Tropical Medicine
Professor Mark Sculpher,Professor of Health Economics,Centre for Health Economics,Institute for Research in theSocial Services, University of York
Dr Jonathan Shapiro, SeniorFellow, Health ServicesManagement Centre,Birmingham
Ms Kate Thomas,Deputy Director,Medical Care Research Unit,University of Sheffield
Ms Sue Ziebland,Research Director, DIPEx,Department of Primary HealthCare, University of Oxford,Institute of Health Sciences
Current and past membership details of all HTA ‘committees’ are available from the HTA website (www.hta.ac.uk)
© Queen’s Printer and Controller of HMSO 2006. All rights reserved.
Health Technology Assessment Programme
134
Diagnostic Technologies & Screening PanelMembers
Chair,Dr Ron Zimmern, Director ofthe Public Health Genetics Unit,Strangeways ResearchLaboratories, Cambridge
Ms Norma Armston,Lay Member, Bolton
Professor Max BachmannProfessor of Health Care Interfaces, Department of Health Policy and Practice,University of East Anglia
Professor Rudy BilousProfessor of Clinical Medicine &Consultant Physician,The Academic Centre,South Tees Hospitals NHS Trust
Dr Paul Cockcroft, Consultant MedicalMicrobiologist and ClinicalDirector of Pathology,Department of ClinicalMicrobiology, St Mary'sHospital, Portsmouth
Professor Adrian K Dixon,Professor of Radiology,University Department ofRadiology, University ofCambridge Clinical School
Dr David Elliman, Consultant Paediatrician/Hon. Senior Lecturer,Population Health Unit, Great Ormond St. Hospital,London
Professor Glyn Elwyn,Primary Medical Care Research Group,Swansea Clinical School,University of Wales Swansea
Mr Tam Fry, HonoraryChairman, Child GrowthFoundation, London
Dr Jennifer J Kurinczuk,Consultant ClinicalEpidemiologist,National PerinatalEpidemiology Unit, Oxford
Dr Susanne M Ludgate, MedicalDirector, Medicines &Healthcare Products RegulatoryAgency, London
Professor William Rosenberg,Professor of Hepatology, LiverResearch Group, University ofSouthampton
Dr Susan Schonfield, Consultantin Public Health, SpecialisedServices Commissioning NorthWest London, HillingdonPrimary Care Trust
Dr Phil Shackley, SeniorLecturer in Health Economics,School of Population andHealth Sciences, University ofNewcastle upon Tyne
Dr Margaret Somerville, PMSPublic Health Lead, PeninsulaMedical School, University ofPlymouth
Dr Graham Taylor, ScientificDirector & Senior Lecturer,Regional DNA Laboratory, TheLeeds Teaching Hospitals
Professor Lindsay WilsonTurnbull, Scientific Director,Centre for MR Investigations &YCR Professor of Radiology,University of Hull
Professor Martin J Whittle,Associate Dean for Education,Head of Department ofObstetrics and Gynaecology,University of Birmingham
Dr Dennis Wright, Consultant Biochemist &Clinical Director, Pathology & The KennedyGalton Centre, Northwick Park & St Mark’sHospitals, Harrow
Pharmaceuticals PanelMembers
Chair,Dr John Reynolds, ChairDivision A, The John RadcliffeHospital, Oxford RadcliffeHospitals NHS Trust
Professor Tony Avery, Head of Division of PrimaryCare, School of CommunityHealth Services, Division ofGeneral Practice, University ofNottingham
Ms Anne Baileff, ConsultantNurse in First Contact Care,Southampton City Primary CareTrust, University ofSouthampton
Professor Stirling Bryan,Professor of Health Economics,Health Services Management Centre,University of Birmingham
Mr Peter Cardy, ChiefExecutive, Macmillan CancerRelief, London
Professor Imti Choonara,Professor in Child Health,Academic Division of ChildHealth, University ofNottingham
Dr Robin Ferner, ConsultantPhysician and Director, WestMidlands Centre for AdverseDrug Reactions, City HospitalNHS Trust, Birmingham
Dr Karen A Fitzgerald,Consultant in PharmaceuticalPublic Health, National PublicHealth Service for Wales,Cardiff
Mrs Sharon Hart, Head of DTB Publications, Drug &Therapeutics Bulletin, London
Dr Christine Hine, Consultant inPublic Health Medicine, SouthGloucestershire Primary CareTrust
Professor Stan Kaye,Cancer Research UK Professor of Medical Oncology,Section of Medicine, The Royal Marsden Hospital,Sutton
Ms Barbara Meredith,Lay Member, Epsom
Dr Andrew Prentice, SeniorLecturer and ConsultantObstetrician & Gynaecologist,Department of Obstetrics &Gynaecology, University ofCambridge
Dr Frances Rotblat, CPMPDelegate, Medicines &Healthcare Products RegulatoryAgency, London
Professor Jan Scott, Professor of Psychological Treatments,Institute of Psychiatry,University of London
Mrs Katrina Simister, AssistantDirector New Medicines,National Prescribing Centre,Liverpool
Dr Richard Tiner, MedicalDirector, Medical Department,Association of the BritishPharmaceutical Industry,London
Dr Helen Williams,Consultant Microbiologist,Norfolk & Norwich UniversityHospital NHS Trust
Current and past membership details of all HTA ‘committees’ are available from the HTA website (www.hta.ac.uk)
Therapeutic Procedures PanelMembers
Chair, Professor Bruce Campbell,Consultant Vascular andGeneral Surgeon, Departmentof Surgery, Royal Devon &Exeter Hospital
Dr Aileen Clarke,Reader in Health ServicesResearch, Public Health &Policy Research Unit, Barts &the London School of Medicine& Dentistry, London
Dr Matthew Cooke, Reader inA&E/Department of HealthAdvisor in A&E, WarwickEmergency Care andRehabilitation, University ofWarwick
Dr Carl E Counsell, ClinicalSenior Lecturer in Neurology,Department of Medicine andTherapeutics, University ofAberdeen
Ms Amelia Curwen, ExecutiveDirector of Policy, Services andResearch, Asthma UK, London
Professor Gene Feder, Professorof Primary Care R&D,Department of General Practiceand Primary Care, Barts & theLondon, Queen Mary’s Schoolof Medicine and Dentistry,London
Professor Paul Gregg,Professor of OrthopaedicSurgical Science, Department ofGeneral Practice and PrimaryCare, South Tees Hospital NHSTrust, Middlesbrough
Ms Bec Hanley, Co-Director,TwoCan Associates,Hurstpierpoint
Ms Maryann L Hardy, Lecturer, Division ofRadiography, University ofBradford
Professor Alan Horwich,Director of Clinical R&D,Academic Department ofRadiology, The Institute ofCancer Research, London
Dr Simon de Lusignan,Senior Lecturer, Primary Care Informatics,Department of CommunityHealth Sciences,St George’s Hospital MedicalSchool, London
Professor Neil McIntosh,Edward Clark Professor of Child Life & Health,Department of Child Life &Health, University of Edinburgh
Professor James Neilson,Professor of Obstetrics andGynaecology, Department ofObstetrics and Gynaecology,University of Liverpool
Dr John C Pounsford,Consultant Physician,Directorate of Medical Services,North Bristol NHS Trust
Karen Roberts, NurseConsultant, Queen ElizabethHospital, Gateshead
Dr Vimal Sharma, ConsultantPsychiatrist/Hon. Senior Lecturer,Mental Health Resource Centre,Cheshire and Wirral PartnershipNHS Trust, Wallasey
Dr L David Smith, ConsultantCardiologist, Royal Devon &Exeter Hospital
Professor Norman Waugh,Professor of Public Health,Department of Public Health,University of Aberdeen
Health Technology Assessment 2006; Vol. 10: No. 35
135Current and past membership details of all HTA ‘committees’ are available from the HTA website (www.hta.ac.uk)
Health Technology Assessment Programme
136Current and past membership details of all HTA ‘committees’ are available from the HTA website (www.hta.ac.uk)
Expert Advisory NetworkMembers
Professor Douglas Altman,Director of CSM & CancerResearch UK Med Stat Gp,Centre for Statistics inMedicine, University of Oxford,Institute of Health Sciences,Headington, Oxford
Professor John Bond,Director, Centre for HealthServices Research, University ofNewcastle upon Tyne, School ofPopulation & Health Sciences,Newcastle upon Tyne
Mr Shaun Brogan, Chief Executive, RidgewayPrimary Care Group, Aylesbury
Mrs Stella Burnside OBE,Chief Executive, Office of theChief Executive. TrustHeadquarters, AltnagelvinHospitals Health & SocialServices Trust, Altnagelvin AreaHospital, Londonderry
Ms Tracy Bury, Project Manager, WorldConfederation for PhysicalTherapy, London
Professor Iain T Cameron,Professor of Obstetrics andGynaecology and Head of theSchool of Medicine,University of Southampton
Dr Christine Clark,Medical Writer & ConsultantPharmacist, Rossendale
Professor Collette Clifford,Professor of Nursing & Head ofResearch, School of HealthSciences, University ofBirmingham, Edgbaston,Birmingham
Professor Barry Cookson,Director, Laboratory ofHealthcare Associated Infection,Health Protection Agency,London
Professor Howard Cuckle,Professor of ReproductiveEpidemiology, Department ofPaediatrics, Obstetrics &Gynaecology, University ofLeeds
Dr Katherine Darton, Information Unit, MIND – The Mental Health Charity,London
Professor Carol Dezateux, Professor of PaediatricEpidemiology, London
Mr John Dunning,Consultant CardiothoracicSurgeon, CardiothoracicSurgical Unit, PapworthHospital NHS Trust, Cambridge
Mr Jonothan Earnshaw,Consultant Vascular Surgeon,Gloucestershire Royal Hospital,Gloucester
Professor Martin Eccles, Professor of ClinicalEffectiveness, Centre for HealthServices Research, University ofNewcastle upon Tyne
Professor Pam Enderby,Professor of CommunityRehabilitation, Institute ofGeneral Practice and PrimaryCare, University of Sheffield
Mr Leonard R Fenwick, Chief Executive, Newcastleupon Tyne Hospitals NHS Trust
Professor David Field, Professor of Neonatal Medicine,Child Health, The LeicesterRoyal Infirmary NHS Trust
Mrs Gillian Fletcher, Antenatal Teacher & Tutor andPresident, National ChildbirthTrust, Henfield
Professor Jayne Franklyn,Professor of Medicine,Department of Medicine,University of Birmingham,Queen Elizabeth Hospital,Edgbaston, Birmingham
Ms Grace Gibbs, Deputy Chief Executive,Director for Nursing, Midwifery& Clinical Support Services, West Middlesex UniversityHospital, Isleworth
Dr Neville Goodman, Consultant Anaesthetist,Southmead Hospital, Bristol
Professor Alastair Gray,Professor of Health Economics,Department of Public Health,University of Oxford
Professor Robert E Hawkins, CRC Professor and Director ofMedical Oncology, Christie CRCResearch Centre, ChristieHospital NHS Trust, Manchester
Professor Allen Hutchinson, Director of Public Health &Deputy Dean of ScHARR,Department of Public Health,University of Sheffield
Dr Duncan Keeley,General Practitioner (Dr Burch& Ptnrs), The Health Centre,Thame
Dr Donna Lamping,Research Degrees ProgrammeDirector & Reader in Psychology,Health Services Research Unit,London School of Hygiene andTropical Medicine, London
Mr George Levvy,Chief Executive, MotorNeurone Disease Association,Northampton
Professor James Lindesay,Professor of Psychiatry for theElderly, University of Leicester,Leicester General Hospital
Professor Julian Little,Professor of Human GenomeEpidemiology, Department ofEpidemiology & CommunityMedicine, University of Ottawa
Professor Rajan Madhok, Medical Director & Director ofPublic Health, Directorate ofClinical Strategy & PublicHealth, North & East Yorkshire& Northern Lincolnshire HealthAuthority, York
Professor David Mant, Professor of General Practice,Department of Primary Care,University of Oxford
Professor Alexander Markham, Director, Molecular MedicineUnit, St James’s UniversityHospital, Leeds
Dr Chris McCall, General Practitioner, TheHadleigh Practice, Castle Mullen
Professor Alistair McGuire,Professor of Health Economics,London School of Economics
Dr Peter Moore, Freelance Science Writer, Ashtead
Dr Sue Moss, Associate Director,Cancer Screening EvaluationUnit, Institute of CancerResearch, Sutton
Mrs Julietta Patnick, Director, NHS Cancer ScreeningProgrammes, Sheffield
Professor Tim Peters,Professor of Primary CareHealth Services Research,Academic Unit of PrimaryHealth Care, University ofBristol
Professor Chris Price, Visiting Chair – Oxford, ClinicalResearch, Bayer DiagnosticsEurope, Cirencester
Professor Peter Sandercock,Professor of Medical Neurology,Department of ClinicalNeurosciences, University ofEdinburgh
Dr Eamonn Sheridan,Consultant in Clinical Genetics,Genetics Department,St James’s University Hospital,Leeds
Dr Ken Stein,Senior Clinical Lecturer inPublic Health, Director,Peninsula TechnologyAssessment Group, University of Exeter
Professor Sarah Stewart-Brown, Professor of Public Health,University of Warwick, Division of Health in theCommunity Warwick MedicalSchool, LWMS, Coventry
Professor Ala Szczepura, Professor of Health ServiceResearch, Centre for HealthServices Studies, University ofWarwick
Dr Ross Taylor, Senior Lecturer, Department ofGeneral Practice and PrimaryCare, University of Aberdeen
Mrs Joan Webster, Consumer member, HTA –Expert Advisory Network
How to obtain copies of this and other HTA Programme reports.An electronic version of this publication, in Adobe Acrobat format, is available for downloading free ofcharge for personal use from the HTA website (http://www.hta.ac.uk). A fully searchable CD-ROM is alsoavailable (see below).
Printed copies of HTA monographs cost £20 each (post and packing free in the UK) to both public andprivate sector purchasers from our Despatch Agents, York Publishing Services.
Non-UK purchasers will have to pay a small fee for post and packing. For European countries the cost is£2 per monograph and for the rest of the world £3 per monograph.
You can order HTA monographs from our Despatch Agents, York Publishing Services by:
– fax (with credit card or official purchase order) – post (with credit card or official purchase order or cheque)– phone during office hours (credit card only).
Additionally the HTA website allows you either to pay securely by credit card or to print out yourorder and then post or fax it.
Contact details are as follows:York Publishing Services Email: [email protected] Box 642 Tel: 0870 1616662YORK YO31 7WX Fax: 0870 1616663UK Fax from outside the UK: +44 1904 430868
NHS libraries can subscribe free of charge. Public libraries can subscribe at a very reduced cost of £100 for each volume (normally comprising 30–40 titles). The commercial subscription rate is £300 per volume. Please contact York Publishing Services at the address above. Subscriptions can only bepurchased for the current or forthcoming volume.
Payment methods
Paying by chequeIf you pay by cheque, the cheque must be in pounds sterling, made payable to York PublishingDistribution and drawn on a bank with a UK address.
Paying by credit cardThe following cards are accepted by phone, fax, post or via the website ordering pages: Delta, Eurocard,Mastercard, Solo, Switch and Visa. We advise against sending credit card details in a plain email.
Paying by official purchase orderYou can post or fax these, but they must be from public bodies (i.e. NHS or universities) within the UK.We cannot at present accept purchase orders from commercial companies or from outside the UK.
How do I get a copy of HTA on CD?
Please use the form on the HTA website (www.hta.ac.uk/htacd.htm). Or contact York PublishingServices (see contact details above) by email, post, fax or phone. HTA on CD is currently free of chargeworldwide.
The website also provides information about the HTA Programme and lists the membership of the variouscommittees.
HTA
Health Technology Assessm
ent 2006;Vol. 10: No. 33
Com
puterised cognitive behaviour therapy for depression and anxiety update
Computerised cognitive behaviour therapy for depression and anxietyupdate: a systematic review and economic evaluation
E Kaltenthaler, J Brazier, E De Nigris, I Tumur, M Ferriter, C Beverley, G Parry, G Rooney and P Sutcliffe
Health Technology Assessment 2006; Vol. 10: No. 33
HTAHealth Technology AssessmentNHS R&D HTA Programme
The National Coordinating Centre for Health Technology Assessment,Mailpoint 728, Boldrewood,University of Southampton,Southampton, SO16 7PX, UK.Fax: +44 (0) 23 8059 5639 Email: [email protected]://www.hta.ac.uk ISSN 1366-5278
FeedbackThe HTA Programme and the authors would like to know
your views about this report.
The Correspondence Page on the HTA website(http://www.hta.ac.uk) is a convenient way to publish
your comments. If you prefer, you can send your comments to the address below, telling us whether you would like
us to transfer them to the website.
We look forward to hearing from you.
September 2006