ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2017

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1360

Long-term perspectives onmusculoskeletal pain

Health care utilization and integration of behavioralmedicine treatment into physical therapy

CHRISTINA EMILSON

ISSN 1651-6206ISBN 978-91-513-0044-3urn:nbn:se:uu:diva-328363

Dissertation presented at Uppsala University to be publicly examined in A7:107a,Biomedicinsk centrum, Husargatan 3, Uppsala, Friday, 13 October 2017 at 09:00 for thedegree of Doctor of Philosophy. The examination will be conducted in Swedish. Facultyexaminer: Docent Ida Flink (Institutionen för juridik, psykologi och socialt arbete, Örebrouniversitet).

AbstractEmilson, C. 2017. Long-term perspectives on musculoskeletal pain. Health careutilization and integration of behavioral medicine treatment into physical therapy. DigitalComprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 1360.83 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-513-0044-3.

There are currently no effective methods for treating and preventing chronic pain. The aim ofthis thesis was to study prognostic factors for health care utilization, and the long-term outcomesof tailored behavioral medicine treatment for patients with musculoskeletal pain. Another aimwas to increase knowledge about physical therapists’ assessment and analysis of patients’ painconditions and to investigate the potential of subgrouping patients based on prognostic factors.

Methods: In Study I, a prospective population-based cohort was followed over 21 years. Datafrom three measure points were analyzed: 1995 (n=2425), 2007 (n=1582) and 2016 (n=1184).Study II was a 10-year follow-up of randomized controlled trial (n=97), comparing tailoredbehavioral medicine treatment and exercise-based physical therapy. In Study III, a descriptiveand explorative design was applied, using data from video-recordings of 12 physical therapists.In study IV, assignment to three subgroups based on the Örebro Musculoskeletal Pain ScreeningQuestionnaire was validated against reference instruments, and the stability between two pointsof measurement was investigated in patients (n=40) who were seeking primary health care dueto musculoskeletal pain.

Results: Chronic pain, female gender and high age predict high health care utilization over21 years, and a trajectory of stable high health care utilization over the entire period. Thedifferences between groups in favor for tailored behavioral medicine treatment reported at post-treatment and after two years, were not maintained at the 10-year follow-up. A majority ofthe physical therapists assessed factors for poor prognosis. The analyses were mainly basedon biomedical assessments and none of the physical therapists included behavioral factors.Subgroup assignment according to the Örebro Musculoskeletal Pain Screening Questionnaireappears to be valid and stable over time.

Conclusion: Prognostic factors such as chronic pain and female gender need to be consideredwhen allocating health care resources and planning treatment to improve long-term outcomes.The treatment should also be tailored based on individual functional behavioral analyses ofkey behaviors and on patient´s biomedical and psychosocial condition, including strategiesfor maintenance of behavioral changes. Evidence-based methods for integrating behavioralmedicine treatment into physical therapy need to be further evaluated and improved.

Keywords: Chronic pain, health care utilization, behavioral medicine, stratified treatment,prognostic factors, physical therapy, primary care

Christina Emilson, Department of Neuroscience, Box 593, Uppsala University, SE-75124Uppsala, Sweden.

© Christina Emilson 2017

ISSN 1651-6206ISBN 978-91-513-0044-3urn:nbn:se:uu:diva-328363 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-328363)

To Axel, Louise and Sofie

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Emilson C, Åsenlöf P, Demmelmaier I, Bergman S. (2017) As-

sociation between health care utilization and musculoskeletal pain. (2017). A 21-year follow-up of a population cohort. Sub-mitted

II Emilson C, Demmelmaier I, Bergman S, Lindberg P, Denison E, Åsenlöf P. A 10-year follow-up of tailored behavioural treatment and exercise-based physiotherapy for persistent musculoskeletal pain. Clinical Rehabilitation. 2016; 31:186-196.

III Emilson C, Åsenlöf P, Pettersson S, Bergman S, Sandborgh M,

Martin C, Demmelmaier I. Physical therapists´ assessments, analyses and use of behavior change techniques in initial consul-tations on musculoskeletal pain: direct observations in primary health care. BMC musculoskeletal disorders. 2016; 17:316.

IV Emilson C, Demmelmaier I, Pettersson S, Bergman S, Åsenlöf P. (2017). Concurrent validity and stability of subgroup assign-ment to three levels of pain condition severity in patients with musculoskeletal pain. Submitted

Reprints were made with permission from the respective publishers.

Contents

Introduction ................................................................................................... 13 Definitions of pain .................................................................................... 13 Epidemiology of pain ............................................................................... 14 Pain physiology ........................................................................................ 14 Psychological factors in pain .................................................................... 15 Factors in poor prognosis ......................................................................... 16 Consequences of pain ............................................................................... 17

Disability ............................................................................................. 17 Health care utilization and sickness-related absence ........................... 17 Pain and suffering ................................................................................ 18

Theoretical perspectives ........................................................................... 18 Respondent and operant learning theory ............................................. 18 Social cognitive theory ........................................................................ 19

Pain management ..................................................................................... 20 Clinical guidelines and practice ........................................................... 20 Models for clinical analysis of pain conditions ................................... 21 Exercise-based physical therapy .......................................................... 23 Psychological treatment ....................................................................... 23 Integration of behavioral medicine into physical therapy ................... 24 Stratified treatment .............................................................................. 26

Rationale for this thesis ............................................................................ 27

Aims .............................................................................................................. 29 Specific aims ............................................................................................ 29

Methods ........................................................................................................ 30 Design ...................................................................................................... 30 Ethical considerations .............................................................................. 31 Participants and procedures ...................................................................... 31

Study I.................................................................................................. 31 Study Ⅱ ................................................................................................ 31 Study Ⅲ and Ⅳ ................................................................................... 33

Data collection .......................................................................................... 33 Pain-related Disability (studies Ⅱ and Ⅳ) ........................................... 34 Pain prevalence and pain location (study I) ......................................... 34 Pain intensity and pain control (studies Ⅱ and Ⅳ) .............................. 34

Health care utilization (study I) ........................................................... 35 Health-related quality of life (study I) ................................................. 35 Sickness-related absence (study Ⅱ)...................................................... 35 Self-reported benefit (study Ⅱ) ............................................................ 35 Fear and avoidance (studies Ⅱ and Ⅳ) ................................................ 36 Pain catastrophizing (study Ⅳ) ........................................................... 36 Prognostic factors for poor outcome (study Ⅳ) .................................. 36 Observable clinical behavior (study Ⅲ) .............................................. 37

Data management and analysis ................................................................ 38 Study I.................................................................................................. 38 Study Ⅱ ................................................................................................ 39 Study Ⅲ ............................................................................................... 39 Study Ⅳ ............................................................................................... 40

Results ........................................................................................................... 41 Follow-up of a population-based cohort .................................................. 41

Prediction of high health care utilization in 2016 ................................ 41 Trajectories of health care utilization .................................................. 42

Follow-up of tailored behavioral medicine treatment .............................. 45 Pain-related disability .......................................................................... 45 Secondary outcomes ............................................................................ 45

Observation of physical therapists in primary health care ....................... 47 Description of clinical behavior........................................................... 47 Variation in clinical behavior .............................................................. 48

Subgroup assignment based on the ÖMPSQ ............................................ 49 Concurrent validity of subgroup assignment ....................................... 50 Stability of subgroup assignment ......................................................... 51

Discussion ..................................................................................................... 52 General discussion .................................................................................... 52 Prognostic factors for health care utilization ............................................ 53 Behavioral medicine and physical therapy ............................................... 54

Long-term outcome ............................................................................. 55 Assessment of prognostic factors ........................................................ 55 Strategies for integration...................................................................... 57

Methodological considerations................................................................. 58 Longitudinal studies ............................................................................ 58 Video analysis ...................................................................................... 59 Validity of subgroup assignment ......................................................... 60

Conclusions ................................................................................................... 61 Clinical implications ................................................................................ 61 Future research ......................................................................................... 62

Svensk sammanfattning ................................................................................ 63

Acknowledgements ....................................................................................... 65

References ..................................................................................................... 67

Abbreviations

IASP International Association for the Study of Pain ICD International Classification of Diseases WHO Worlds Health Organization CWP Chronic widespread pain CRP Chronic regional pain NCP Non chronic pain CNS Central nervous system SCT Social cognitive theory CBT Cognitive behavioral therapy ÖMPSQ Örebro Musculoskeletal Pain Screening Questionnaire SBT STarT Back Tool BCT Behavior change technique RCT Randomized controlled trial TBT Tailored behavioral medicine treatment EBT Exercise-based physical therapy treatment PDI Pain Disability Index NRS Numerical Rating Scale TSK Tampa Scale for Kinesiophobia PCS Pain Catastrophizing Scale SF36 Short Form-36 Health Survey

11

About this thesis

A majority of all individuals in the general population will be affected by mus-culoskeletal pain at some point during their lifetime. For some of them, the pain becomes chronic, affecting both physical and psychosocial aspects of life and leading to great suffering for the individual. Despite advances in pain re-search, within both the biomedical and psychological areas, there are currently no effective methods for treating and preventing chronic pain, which is a chal-lenge for both health care and society. Chronic pain is complex, involving biological, psychosocial and behavioral factors. In recent decades, psychoso-cial aspects and active strategies for health behavioral change have increased in pain research and treatment, and are the main focus of this thesis.

In the four included studies, an integrated biopsychosocial perspective on pain was applied. The long-term perspective has been studied not only regard-ing the course of chronic pain associated with prognostic factors and health care utilization but also in relation to the long-term outcome of a behavioral medicine interventions. Another focus is the potential for the integration of behavioral medicine into physical therapy practice regarding the assessment, analysis and treatment of patients with musculoskeletal pain conditions.

13

Introduction

Definitions of pain The international Association for the Study of Pain (IASP) has defined pain as “An unpleasant sensory and emotional experience associated with actual or potential tissue damage, or described in terms of such damage”1. This defini-tion is multidimensional and includes physical, psychological and social as-pects related to pain. The definition of chronic pain according to the IASP is that pain must “persist past the normal time of healing”1.

The definition of chronic pain is often described based on a duration of at least three months of persistent or recurrent pain2. The new version of the In-ternational Classification of Diseases (ICD) of the Worlds Health Organiza-tion (WHO)3 includes diagnostic codes for chronic pain conditions and has been developed to suite different types of chronic pain.

The definition of acute and sub-acute pain is a duration of pain up to three months, but the differentiation of the duration between acute and subacute pain is inconsistently described in the literature4,5. For the definition of recur-rent pain, a criterion of minimum pain duration of 24 hours has been suggested to signify a new pain episode6. Recovery has been defined as a pain-free pe-riod of at least one month6.

Chronic pain can also be differentiated into chronic regional pain (CRP) or chronic widespread pain (CWP) according to the American College of Rheu-matology criteria for the Classification of Fibromyalgia regarding the distri-bution of pain7.

A prognostic approach to define chronic pain has also been described8,9, focusing mainly on the psychological risk factors for future disabling pain.

In study I the definitions CRP and CWP were used to categorize the pres-ence of chronic pain in the general population. If the criteria for CWP were not met the pain was defined as CRP. Study Ⅱ, Ⅲ and Ⅳ included patients with nonmalignant musculoskeletal pain, defined as sub-acute, recurrent or persistent. In study Ⅳ prognostic criteria for the classification of pain condi-tion severity were also applied.

14

Epidemiology of pain The prevalence of musculoskeletal pain is between 10% and 48% of the gen-eral population in Western countries, depending of the type of prevalence that has been investigated and the type of pain that is reported. The 12-month prevalence of chronic pain for at least 3 months has reported to be between 35% and 48%10,11, of CWP to be between 11% and 20%10,11 , and of chronic pain for at least six months to be 19%12. The point prevalence of chronic mus-culoskeletal pain has been estimated to between 23% and 28%13,14. A system-atic review of the global prevalence of low back pain revealed that the point prevalence was estimated to be 12%, the 1-month prevalence was estimated to be 23%, and the 12-month prevalence was estimated to be 38%, with varied definitions of chronic pain15. The overall point prevalence of pain has been reported to be approximately 50%16, that of chronic widespread pain to be 5%17. A population study based on a U.S national health interview survey re-ported a prevalence of chronic pain of 19%18.

The prevalence of chronic pain is higher among women11,13,15-19, increase with age13,15-17,19, and has been associated with socioeconomic factors, i.e., low levels of education and income, sickness-related absence and immigrant sta-tus12,16,17,19,20 10. Psychological factors such as depression, anxiety, catastro-phizing beliefs about pain and fear of movement have also been associated with the prevalence of chronic pain18-21.

Although the prognosis for musculoskeletal pain is uncertain, clinical guidelines indicate that most musculoskeletal pain conditions will recover within a few weeks22,23, but approximately 10% of those who experience a new episode of pain do not recover and develop into a recurrent or chronic pain condition. A process of recurring episodes of pain has been reported as one of the main characteristics of low back pain24,25. A cohort-study in Aus-tralia in patients in whom pain persisted for ≥3 months, approximately 40% were reported to be recovered within 12 months after onset26.

Pain physiology In the transition from acute to chronic pain, a number of physiological changes occur in the central nervous system (CNS). In acute nociceptive pain, free nerve endings will respond to mechanical, chemical or thermal nociceptive stimulation in the periphery and conduct the nerve signal in the primary so-matosensory neuron27. Two types of nerve fibers are involved in the transmis-sion of the pain signal from the peripheral to the dorsal horn of the spinal cord, the rapid myelinated Aẟ-fibers and the thin slower unmyelinated C-fibers. Aẟ-fibers provides a sharp localization of pain, while C-fibers give a diffuse and dull pain sensation. In the dorsal horn, synaptic transmission from the primary neuron to the secondary neuron will occur, and an afferent projection to higher

15

centers in the CNS occurs; the somatosensory cortex involves the sensory ex-periences of pain, such as localization, duration and intensity, and the limbic structures are involved in the affective or emotional components of pain. The neurophysiological mechanisms of pain also involve endogenous pain modu-lation deriving from the brainstem and can be activated by higher central struc-tures27,28.

Central sensitization is a complex phenomenon that is associated with the development and maintenance of chronic pain. The IASP has defined the term central sensitization as “Increased responsiveness of nociceptive neurons in the central nervous system to their normal or subthreshold afferent input”1. Changes in the CNS also include long term potentiation, dysfunction in the descending disinhibition system, increased temporal summation, and facilita-tion of cognitive-affective mechanisms, which has been associated with CWP, fatigue and sleep disturbance29,30.

Psychological factors in pain Psychological factors have been associated with pain experiences and the de-velopment of chronic pain, disability and poor outcome. The most important factors that describe and explain pain perceptions and the association with treatment outcome are targeted in this thesis. Emotional factors including depression, anxiety and fear, have been associ-ated with long-term disability and pain, sickness absence and poor treatment outcome for pain31-33. Cognitive factors include beliefs and expectations related to pain, negative thoughts, and catastrophizing. Negative thoughts and expectation regarding the pain condition and recovery have been described as unhelpful coping strat-egies32. Pain catastrophizing is defined as a negative mental set during an ac-tual or anticipated experience34, including irrational expectations about future events, which is reported to have a significant impact on pain experiences, depression, disability32,35,36, and delayed recovery from chronic pain37. A high level of pain catastrophizing has also been associated with anxiety, fear and avoidance38. Pain behavior is behavior associated with pain symptoms, such as avoidance of activity, and is influenced by emotions and cognitions and can be reinforced by contextual factors and consequences32,39. The fear-avoidance model of pain explains how disability, depression and dis-use develop as a consequence of prolonged avoidance with fear of pain as a regulation factor40-42. Catastrophizing is one of the key components in the fear-avoidance model. The model also suggests that if the fear-avoidance beliefs related to pain decrease and the individual uses more active strategies and confronts the pain, this will lead to recovery. The fear-avoidance model has been extensively used in pain research41,42. The role of pain-related fear as a

16

consequence rather than an antecedent of the development of pain severity has been suggested43. Instead, pain intensity would be a predictor of pain-related fear and disability.

Factors in poor prognosis Biomedical, psychological and socioeconomic factors have all been associ-ated with chronic disabling pain, but there is no consensus about which one of these factors has the greatest impact on the development of chronic pain and treatment outcomes. A combination of these factors is suggested to explain the course of chronic pain and its consequences, and prognostic factors may also differ between pain conditions44. Early identification of such factors is important to provide adequate assessment and treatment.

Biomedical factors such as high pain intensity, multiple pain sites, long pain duration and baseline disability, have been reported to be prognostic fac-tors for chronic disabling pain conditions45,46. Chronic regional pain has been found as a prognostic factor for the development of chronic widespread pain10,47. Serious medical conditions and symptoms, i.e. malignancy, fractures and cauda equine syndrome, defined as “red flags”, are factors that need to be managed promptly based on their etiology48.

Psychological factors, such as depression, anxiety, fear-avoidance beliefs and catastrophizing, low expectations for recovery are factors associated with chronic disabling pain and delayed recovery, also defined as “yellow flags”49. Yellow flags are important factors for whether treatment will succeed and are recommended to be identified and targeted49.

In addition to red and yellow flags, definitions of other prognostic factors associated with musculoskeletal pain have been described49. Orange flags are psychiatric symptoms, e.g., clinical depression and personality disorder, blue flags are defined as perceptions about the relationship between work and health and black flags are related to systemic or contextual factors, e.g., insur-ance and health care systems. These additional flags are not targeted in this thesis.

Socioeconomic factors, i.e., female gender, high age, low education and employment status, and geographic and cultural aspects have also been re-ported as prognostic factors associated with poor treatment outcome, sickness-related absence and chronic pain19. Being an immigrant has been reported to be a predictor of CWP and fibromyalgia50.

The potential predictive factors included in study I were chronic pain and socioeconomic factors. The intervention that was followed up in study Ⅱ, was targeting yellow flags, such as fear of movement and self-efficacy. Red and yellow flags were targeted in studies Ⅲ and Ⅳ.

17

Consequences of pain Disability Chronic pain is associated with several negative consequences for the individ-ual. Disability is defined as an umbrella term including impairments, activity limitations, and participation restrictions, according to the World Health Or-ganization’s International Classification of Functioning, Disability and Health (ICF)51. The term involves biomedical, psychological and environmental fac-tors that affect the health condition in the individual. Pain-related disability is one of the main consequences of musculoskeletal pain both in the short and long perspectives, and it is important in the transition from acute to chronic pain4,45.

Health care utilization and sickness-related absence Musculoskeletal pain is a common reason why patients seek health care in the general population, but many different factors explain the large variation related to health care seeking. Major determinants of health care utilization are female gender, higher age13,52,53, degree of disability53, and fear-avoid-ance beliefs related to pain54. Low education level is also associated with higher health care utilization, except for consultations to physiotherapists for which no such association has been found52. Regarding chronic pain, the fre-quency and intensity of pain are associated with health care seeking16. Indi-viduals with CWP, chronic low back pain55, and fibromyalgia56,57 are re-ported to have a high health care utilization compared to individuals with other musculoskeletal pain conditions, who use no or little health care52,53,57.

It is also interesting to study patterns of health care utilization during the course of chronic pain over time. To describe changing patterns over time, a person-centred model for identifying different trajectories has been used58. Trajectories for health care utilization due to pain may vary in the population over time, which is sparsely reported. Regarding chronic low back pain, dif-ferent trajectories of persistent or fluctuating pain with different intensities and frequencies have been identified and described59, which can have impli-cations for health care utilization over time. A few longitudinal population studies have reported that both chronic13 and acute musculoskeletal pain (<2 weeks duration)60 could have a considerable impact on health care utilization in the longer term. Knowledge of the factors that predict specific trajectories for health care utilization associated with the course of pain can contribute to better understanding of individual conditions for managing pain. Such knowledge is also valuable when allocating health care resources and differentiating treatment methods.

Sickness-related absence is also a major consequence of musculoskeletal pain61. In Sweden, musculoskeletal disorders were the second most common

18

cause, accounting for about 22% of all sickness-related absence (>14 days) in 201662.

Pain and suffering Chronic pain is not only a condition explained by physiological and psycho-logical mechanisms but also causes great suffering for the individual affecting many aspect of life63. Suffering is a reaction to the consequences of pain and is associated with the person’s perception of the meaning and the impact of pain on his or her life, e.g., threats to one’s self-concept, self-esteem and in-tegrity, helplessness, and psychological and economic distress. It is important to understand how these factors affect the individual and how they may influ-ence the treatment outcome. Poor health-related quality of life has also been reported to be associated with the development and experiences of chronic pain50,64,65.

Theoretical perspectives In this thesis, a biopsychosocial approach is applied to investigate and explain the complexity of musculoskeletal pain and its impact on the individual in the longer perspective. The biopsychosocial model of health is multidimensional including an integration of biological, psychological and social factors66,67. The model is widely used, become fundamental in pain research and treat-ment68, and aims to explain how these factors influence the experience of pain for the individual. The biopsychosocial model has been criticized for being too general and not describing the relationship between the different factors in the model68,69. To integrate the behavioral components related to pain re-sponses, a bio-behavioral view has been suggested63.

There are many behavioral change theories described in the literature aim-ing to explain behavioral learning and behavioral change70. Some of the most important and frequently used theories in pain research are respondent and operant learning theories, social cognitive theory (SCT), and self-efficacy the-ory, which have informed the study designs, interventions and assessments in this thesis.

Respondent and operant learning theory Respondent learning is described as learning by association63. An uncondi-tioned stimulus (i.e., pain) elicit a biological reflex (i.e., fear), which is called an unconditioned response to indicate that no learning is involved in the pro-cess. If preceded by a neutral stimulus (e.g. heavy lifting), a learning process starts where the neutral stimulus becomes conditioned and associated with a

19

conditioned response (i.e. fear). Hence, the heavy lifting then acts as a condi-tioned stimulus that is associated with the conditioned response in terms of fear. In patients with musculoskeletal pain, the learning mechanism of fear and avoidance behavior can be explained by respondent conditioning but can also be driven and maintained by operant learning mechanisms. An example of pain treatment based on respondent conditioning is graded exposure, con-sisting of gradual exposure to fearful situations identified by the patient71.

Operant learning theory72 describe the mechanisms of how behaviors can be learned and changed by their consequences. The frequency of the behavior can be modified both by positive and negative reinforcement or punishment. Reinforcing consequences increase the likelihood that a behavior will occur, while punishing consequences may cause the behavior to become less frequent in the future. During the last decades, the operant learning theory has more frequently been used to explain pain behaviors39. Operant discrimination re-fers to a specific situation in which the reinforcement of a behavior is contin-gent on the presence of a certain stimulus and can be defined as an individual´s ability to differentiate between a learned voluntary response and an irrelevant non-learned response.

Maladaptive pain behaviors can be changed by reinforcement, to increase the frequency of a desired target behavior, or by punishment, to decrease the likelihood of an unwanted behavior to occur by using reducers. Attention to a specific activity goal instead of pain and using positive reinforcements when the desired behavior occurs can increase the probability of a behavioral change. Respondent and operant learning mechanisms often interact simulta-neously in everyday activities63,73, such as avoidance of a fearful situation. One example of how operant learning can be used in the management of mus-culoskeletal pain is graded activity, where operant conditioning is used to re-inforce healthy behaviors, such as improving physical activity tolerance.

Social cognitive theory SCT74, which aims to provide a framework for the understanding of human thoughts and behavior, is derived from social cognitive learning75. The central core in this theory is the causal model of triadic reciprocal determinism, de-scribing how personal, behavioral and environmental factors interact as deter-minants of each other in the process of behavioral change. SCT describes functions of human behavior in a number of key concepts: 1) determinants of a behavior, such as outcome expectations; 2) observational learning by mod-eling others’ behavior; and 3) self-regulation, referring to the individuals’ abil-ity to motivate or regulate a behavior on the basis on personal standards and to evaluate of their own behavior through self-monitoring, goal setting, feed-back or self-reward. Perceived self-efficacy is also described as an important concept in the SCT that influence the behavior and behavioral change.

20

Several constructs in SCT correspond to behavioral change interventions and to several specific behavioral change techniques that are used in pain in-terventions, i.e., self-monitoring, modelling or demonstrating a behavior, planning social support, goal setting and learning to use prompts and cues to action76.

Self-efficacy theory Self-efficacy theory77 describes central psychological mechanisms in behavior change, and the individual`s beliefs about their ability to perform a particular activity or behavior in a specific situation. Self-efficacy is based on four sources of beliefs: personal experiences of success, vicarious experiences, ver-bal persuasion about capability, and emotional individual states. The percep-tion of self-efficacy influence individuals’ intention to action, the effort they are willing to do, and the confidence with which they approach action. The self-efficacy concept is central in several theories of health behavior and be-havioral change, and is described as a key factor for success in behavioral change interventions. In musculoskeletal pain self-efficacy is an important predictor of several behaviors and has been identified as a key determinant of physical activity78. Self-efficacy has also been found to be a mediator between pain and pain-related disability79,80, and it is an important factor to consider in pain management.

Pain management Clinical guidelines and practice The current management of musculoskeletal pain includes different treat-ments, such as medication, exercise and psychological therapies. In interna-tional guidelines for the management of musculoskeletal pain, a biopsychoso-cial model is recommended, and the overall goals are to reduce pain, disability and psychosocial suffering, and to increase the patient’s participation, work-ability and health-related quality of life23,81. In chronic pain, and in CWP in particular, a more comprehensive pain management intervention is needed that focuses on the integration of the biomedical and psychosocial aspects of pain82.

Treatment recommended for acute pain are as follows: 1) assessment of relevant red and yellow flags in initial consultations; 2) provision of adequate information about the pain condition and reassurance that rapid recovery is expected in most patients; 3) advice on being physically active (after having excluded red flags), returning to normal activities, and avoiding passive treat-ment strategies such as rest in bed; and 4) if necessary, prescription of medi-cation for pain relief22,23. Stay-active advice has been reported to be equally effective as more extensive physical therapy, including manual therapy and

21

exercise-based treatment83-85. In patients seeking care for musculoskeletal pain, red flags are identified in less than 1%, and for a majority of patients, physical activity can be recommended86.

Treatments recommended for chronic pain conditions are as follows: 1) su-pervised exercised-based therapy; 2) short-term use of medication or manual therapies; 3) cognitive behavioral therapy; and 4) multimodal rehabilita-tion22,23.

Multimodal rehabilitation programs based on a biopsychosocial model and involving several health care professions are recommended to patients with complex chronic pain conditions, including more extensive integrated pain management with psychosocial and biomedical interventions87,88. Multimodal rehabilitation has reported to improve the functional status, work ability, qual-ity of life89-91, and in psychological variables, i.e., distressed mood, and pain-related fear92. The evidence is still limited regarding the effects of specific treatment components or what characterizes patients who benefit from such treatments87.

Models for clinical analysis of pain conditions In clinical practice, an analysis of a pain condition is the basis for all treatment and can be performed according to different models to understand and explain the probable relationship between different aspects of pain.

Historically, a biomedical model93, based on biological factors, has domi-nated clinical practice. This model is derived from a reductionistic and dual-istic principle to explain a health condition or phenomena, e.g., separation be-tween biological and psychosocial factors. The definition of biomedical anal-ysis used in this thesis is based mainly on physical assessment and physical impairments or limitations. To understand and explain the determinants of pain conditions, the biomedical model has proved to be insufficient, and a more broad model that includes psychosocial aspects is requested66.

A biopsychosocial model of health67,94, is based on an integration of bio-logical, psychological and social factors and has been used to understand and explain chronic pain (figure 1). In the biopsychosocial model, the functional relationship between the three components, and a specific behavior is not iden-tified and described.

A functional behavioral analysis is based on the biopsychosocial model and has been defined as “identification of important, controllable, and causal functional relationship applicable to specific target behaviors for an individ-ual95. The A-B-C format describes a hypothetical functional relationship be-tween the main components of operant learning principles72, including ante-cedent cues (A) that precede a behavior (B) and consequences of the behavior (C) (figure 2). The antecedents do not cause a behavior but set the context in which the behavior occurs. Reinforcing or punishing consequences increase or decrease the probability of the behavior being more or less frequent in the

22

Figure 1. Illustration of a biopsychosocial model of health, including biological, psychological and social aspects.

Figure 2. Illustration of a functional behavioral analysis in the A-B-C format, in-cluding the functional relationship between antecedents, behavior and conse-quences.

Health

23

future, and they also influence the way the antecedents set the context for whether a behavior will occur. In this thesis, clinical analyses of pain condi-tions have been categorized based on these three models.

Exercise-based physical therapy Physical therapists are often the first primary health care providers for patients seeking care for musculoskeletal pain and have an important role in the as-sessment, analysis, and treatment. According to clinical guidelines, a biopsy-chosocial model including psychosocial factors is recommended to be inte-grated into physical therapy clinical practice for patients with musculoskeletal pain23,96,97.

Physical activity or exercise-based therapy is one of the most frequently used physical therapy treatments for chronic pain81. Physical activity has been defined as bodily movement produced by skeletal muscles that results in en-ergy expenditure98. The definition of physical exercise is a planned, structured physical activity with the purpose of improving or maintaining physical fit-ness98. Exercise-based therapy and physical activity, i.e., performing activities of daily living, are reported to be effective in chronic pain conditions regard-ing reduced pain severity and improved physical function compared to no in-tervention, minimal care or other manual therapies 81,97,99. Aerobic and re-sistance exercises are recommended, but evidence is lacking regarding the most effective type of exercise and dosage for chronic pain conditions100. Mo-tor control exercise for non-specific low back pain, commonly used in clinical practice, has been reported to be effective but not superior compared with other exercise-based treatments101.

There is consensus for individualized, supervised exercise, based on pa-tient’s pain condition severity, goals and preferences97.

Psychological treatment All patients with musculoskeletal pain are affected by psychological factors to some extent. Most health care providers can learn to use psychologically informed patient-centered communication, e.g., active listening, assessment of psychosocially factors, providing advice and assurance in pain manage-ment. Other patients need advanced psychological treatment that require more highly educated and trained health care providers, such as clinical psycholo-gists67.

In pain management, cognitive-behavioral therapy (CBT) and operant treatments, are the most common psychological therapies102. The CBT aims to identify harmful cognitions regarding pain and disability and involving a systematic application of psychological principles to support patients to change cognitions, negative emotions and behaviors63. The operant treatment

24

involves the removal of positive reinforcement of pain behaviors and the pro-motion of healthy behaviors (e.g. physical activity)39. When comparing oper-ant and cognitive therapies for low back pain, little or no differences was re-ported regarding pain and depression, which had equally effects as exercise-based treatment102.

Graded activity treatment, based on operant mechanisms, involves a grad-ual increase in exercise and activity tolerance based on a quota system103. Graded activity has been successfully used in patients with pain-related disa-bility to improve the function in daily activities104,105, while other studies have reported equal improvements when comparing graded activity and exercise-based therapies106,107.

Exposure treatment in vivo or graded exposure, based on the principle of respondent conditioning, has been developed for the treatment of pain-related fear71,108. In graded exposure treatment, the patient is exposed to a specific situation identified as fearful and as a barrier for the rehabilitation and pro-ceeds hierarchically as the patient’s fear in the specific situation decreases. Studies comparing graded activity and graded exposure treatment have re-ported equally effects regarding decrease in pain intensity and disability, and increase in activity level109-112 and were slightly more effective than a minimal intervention (e.g., stay-active advice and information)110. Another review and meta-analysis113 reported insufficient evidence for the impact of graded activ-ity on disability but the methods in many of the included studies were poorly described, which reduced the possibility of drawing valid conclusions.

Integration of behavioral medicine into physical therapy Behavioral medicine is defined as an interdisciplinary field integrating behav-ioral, psychosocial, and biomedical science, and knowledge and techniques relevant to the understanding of health and illness, and is applied in health promotion, prevention, diagnosis, treatment, and rehabilitation114. Integration of behavioral medicine, also defined as psychologically informed physical therapy96,115,116, require education and skills-training for physical therapists to change their clinical behavior. Traditionally, physical therapists are educated and trained in accordance with a biomedical approach, based on physical as-sessments and treatments to reduce pain symptoms, such as manual and exer-cise-based therapies and have limited education in psychological treatment116. However, in recent years the knowledge and understanding about the im-portance of psychosocial factors in pain management has increased among physical therapists. Studies have concluded that physical therapists can be suc-cessfully educated and trained to deliver behavioral medicine treatment and psychologically informed interventions, resulting in improved understanding, interpretation and skills, but comprehensive training and supervision may be crucial for success104,117-119.

25

The evidence of positive treatment outcomes in patients or whom behav-ioral medicine treatment was implemented is still limited120,121. A structured model for tailored behavioral medicine treatment was developed in a random-ized controlled trial in patients with musculoskeletal pain provided by physi-cal therapists in primary health care122. A functional behavioral analysis was performed based on individual activity goals identified by the patient, fol-lowed by an individually tailored behavioral treatment protocol, and compared with exercise-based physical therapy in a previous RCT104. Individually be-havioral treatment was reported to be more effective than exercise treatment regarding pain-related disability. The results were maintained two years after the intervention123. Supervised exercised-based physical therapy treatment in-cluding a behavioral approach has been reported to decrease pain-related dis-ability and pain catastrophizing in patients with chronic whiplash-associated disorders (WAD)124, and improve physical function in patients with knee os-teoarthritis125.

The content of behavioral change interventions has traditionally been poorly defined and described in previous studies126,127. To address this prob-lem, a hierarchically organized taxonomy of behavior change techniques (BCTs) has been developed128. The taxonomy includes 93 well defined BCTs grouped into 16 clusters129 and has been used in behavioral change research both in single and more complex interventions76. A BCT is an active compo-nent of an intervention that is used to support a specific health behavior change, e.g., increase physical activity, healthy eating or decrease sedentary behavior. Systematic reviews indicate that BCTs targeting self-regulation, e.g., goal-setting and self-monitoring of behavior, are the most salient for in-fluencing and maintaining health behaviors130-133. The BCTs that were most frequently identified in chronic pain interventions are “Instruction on how to perform the behavior” and “demonstration of the behavior, behavior prac-tice”, while BCTs associated with self-regulation were not as frequently used76. Limited treatment effects of specific BCTs on target behaviors have been reported133-135. In other studies, positive effects were found regarding BCTs promoting physical activity and healthy eating133,136,137. Combinations of BCTs have also been reported to be effective, e.g., “Provide information about behavior” and “Prompt intention formation”138.

There is still a challenge to integrate behavioral medicine treatment into clinical practice139-141. Physical therapists have been reported to be unprepared or uncomfortable in using psychological strategies in their clinical prac-tice139,142,143. Another challenge may be the patient’s beliefs and expectations, particularly of physical examination, adequate biomedical diagnosis and man-ual treatment of the pain symptoms, which can have an impact on the clinical approach and selection of methods. A few studies on physical therapists’ ex-periences of using behavioral medicine strategies have reported enhanced knowledge and changes in attitudes and beliefs142,144. An important aspect is that interviews and self-report questionnaires have been used in most of these

26

studies120, but observed clinical behaviors regarding what the physical thera-pists actually do have rarely been studied. In other studies the content of the assessment and intervention are not specifically described121,141, which make the clinical behavior difficult to evaluate.

Screening for prognostic factors is another important assessment for phys-ical therapists at the initial consultation. A previous study revealed that assess-ment of red flags was integrated into physical therapy145, but the questioning for identification was suggested to be more specific, e.g., regarding previous malignancy and other serious conditions. Audio recording of brief telephone consultations in primary health care indicated that physical therapists can im-prove their assessment of yellow flags146, but it is not fully integrated in a systematic way into clinical practice147,148.

The purpose of an integrated physical therapy treatment in musculoskeletal pain, would be to provide individually tailored treatment based on biomedical, psychosocial and behavioral factors related to the pain condition and by tar-geting the prognostic factors for poor treatment outcome.

Stratified treatment Different methods for differentiation of treatment according to the patients’ needs have been described in the literature. To provide adequate treatment and for optimizing the treatment effects in patients with musculoskeletal pain, stratified care has been suggested149,150. Stratified care involves target-ing treatments based on specific subgroups of patients. The subgroups can be based on 1) prognostic factors (e.g., psychosocial factors) for poor outcome, 2) diagnosis based on underlying pain mechanisms or pathoanatomical causes, and 3) the patients’ expected treatment responsiveness (e.g., exercise or manual therapy treatments)150.

In pain conditions, the treatment is suggested to target not only prognostic factors for poor outcome but also other factors that have a negative impact on treatment outcome, such as pain condition severity. Target treatment by matching patients to an intervention based on their psychological risk profile was performed by Bergbom et.al,151. The definition of matching was described as targeted treatment based on psychological risk factors and individual deter-minants of outcome152.

One screening instrument is the Örebro Musculoskeletal Pain Screening Questionnaire (ÖMPSQ), a multidimensional instrument assessing pain vari-ables, disability, fear-avoidance beliefs, catastrophizing, and depressive symptoms153. Risk profiles based on the ÖMPSQ have previously been vali-dated using standard reference measures assessing fear-avoidance beliefs and depressive symptoms in patients with musculoskeletal pain151 and measures assessing disability, catastrophizing and fear-avoidance in patients with low back pain154. However, the definitions of the risk profiles differed between

27

these studies, and the assignment to subgroups with three levels of pain con-dition severity needs to be further validated against reference instruments. To identify prognostic risk factors, a screening tool is useful in clinical practice.

Tailored behavioral treatment is based on individual functional behavioral analyses of key behaviors and health outcomes of interest32,35.

Another method for differentiating treatment based on the severity of the pain condition is the stepped-care design155,156. This method provides treat-ment of differing intensity or extent, meaning that more intensive treatments are reserved only for those patients who do not benefit from the simpler first-step treatment, e.g., stay-active advice and adequate information about the pain condition. The clinical and economic advantages have the potential for deriving the greatest benefits from available health resources. The Stepped care design has rarely been used in pain research155, but it was planned for future interventions based on studies Ⅲ and Ⅳ in this thesis.

Rationale for this thesis The prevalence of musculoskeletal pain is high in the population and is a com-mon reason for poor health and extensive health care utilization. Biomedical, psychosocial and behavioral factors are important in pain management, but knowledge about methods for assessment and treatment and how to integrate a behavioral medicine treatment model into clinical practice is still unclear.

Several cross-sectional and short-term follow-up studies of prognostic fac-tors for health care utilization have previously been reported, but the long-term perspective and trajectories for health care utilization among individuals with musculoskeletal pain in the general populations have not been investi-gated previously.

Increasing evidence is reported about the importance of integrating psycho-logical factors into pain management to improve the treatment outcome. Pos-itive effects of tailored behavioral medicine treatment, provided by physical therapists, have been reported regarding pain-related disability and pain inten-sity, but the long-term effects have not been investigated. Knowledge about the outcomes and consequences in the long run is needed to improve future pain interventions.

Physical therapists are one of the first health care providers who meet pa-tients seeking primary health care for musculoskeletal pain and therefore have a significant role regarding the assessment, analysis and treatment of these patients. In recent years, awareness and knowledge about the importance of behavioral medicine into pain treatment has increased but has not fully been integrated in physical therapy and needs to be further investigated.

Assessment of prognostic factors for poor prognosis and recovery is rec-ommended in clinical guidelines. To identify patients at risk not only for poor outcome but also for stratification of treatment based on prognostic factors,

28

screening instruments for assessing these characteristics are important. The method of subgroup assignment with different levels of pain condition based on the ÖMPSQ is promising for this purpose but needs to be further investi-gated.

29

Aims

The aim of this thesis was to study prognostic factors for health care utiliza-tion, and long-term outcomes of individual behavioral medicine treatment for patients with musculoskeletal pain. Another aim was to increase knowledge about physical therapists’ assessment and analysis of patients’ pain conditions and to investigate the potential of subgrouping patients based on prognostic factors for future outcomes.

Specific aims I To describe and predict health care utilization associated with muscu-

loskeletal pain in a 21-year follow-up of a population cohort. The sec-ondary aim was to describe trajectories of health care utilization over the entire period.

II To investigate whether the previously reported beneficial effects of

tailored behavioral medicine treatment in a physical therapy context were sustained 10 years after completion of treatment. The primary outcome was pain-related disability.

III To explore variation and describe physical therapists´ observed behav-

iors regarding the assessment of red and yellow flags, analysis of a pain condition and use of BCTs when performing initial consultations in patients with musculoskeletal pain in primary health care.

IV To investigate the concurrent validity of subgroup assignment based

on the ÖMPSQ and reference measures of pain-related disability, fear of movement and/ or (re)-injury, and pain-catastrophizing. The sec-ondary aim was to investigate the stability of the subgroups over time in patients seeking primary health care due to musculoskeletal pain.

30

Methods

Design This thesis includes four studies (table 1). Study I was a population-based co-hort (n= 2425 at the baseline) over 21 years including five measure points. Study Ⅱ was a 10-year follow-up of a randomized controlled trial (RCT) that initially included 97 patients with musculoskeletal pain in primary health care. In study Ⅲ, video recordings of initial consultations with 12 primary health care physical therapists were analyzed and described. In study Ⅳ, a cross-sec-tional and prospective design was used based on a sample of patients with musculoskeletal pain (n=40) in primary health care.

Table 1. Overview of design, participants, variables, data collection and analysis in the four studies.

Study 1 Study Ⅱ Study Ⅲ Study Ⅳ

Design Population cohort study over 21 years

10-year follow-up of an RCT

Descriptive and explorative

Cross-sectional and prospective

Participants Individuals, aged 20-74, from a gen-eral population (n=1184 in 2016)

Patients, aged 18-65, with MSP in primary care (n=43)

PTs in primary health care (n=12)

Patients aged 18-65, with MSP in primary care (n=40)

Variables HCU, pain preva-lence, quality of life

Pain-related disa-bility, pain inten-sity, pain control, fear of movement, sickness-absence

Assessment of red and yellow flags, analysis of a pain condition, and use of BCTs.

Subgroup assess-ment based on the ÖMPSQ, pain-re-lated disability, fear of move-ment, pain catastrophizing

Data collec-tion

Self-reported questionnaire

Self-reported questionnaire, reg-ister data on sick-ness absence

Video-recordings Self-reported questionnaire

Analysis Prediction of cate-gorical outcome variables

Between-group comparisons

Deductive and in-ductive analyses

Analysis of con-current validity and intra-rater agreement

MSP= Musculoskeletal pain, PT= Physical therapist, HCU= Health care utilization, BCT= Behavior change technique taxonomy, RCT= Randomized controlled trial, ÖMPSQ= Örebro musculoskeletal pain screening questionnaire

31

Ethical considerations Ethical approval was obtained for all studies included in the thesis. Study I was approved by the Ethics Research Committee, Faculty of Medicine, Uni-versity of Lund Sweden (Dnr:1995 (LU389-94). The computerized registra-tion was approved by the Swedish Data Inspection Board. Study Ⅱ was ap-proved by the regional ethics review board in Uppsala, Sweden (Dnr: 01-34899 for the original study, Dnr: 2012/512 for the follow-up). Studies Ⅲ and Ⅳ were approved by the regional ethic review board in Uppsala, Swe-den (Dnr: 2012/448). Participants were given oral and written information about the respective study, and informed consent was obtained from each participant. In study Ⅲ, personal data were coded so that participants´ names and workplaces could not be associated with a specific case.

Participants and procedures Study I A representative sample of 3928 subjects, every 18th man and woman age 20-74 years, from two municipalities in the south of Sweden was selected from the computerized Swedish national population register. Data were collected at five measure points over a 21-year period (figure 3) by self-reported question-naires consisting of the prevalence of chronic pain during the last 12 months, health care utilization, demographic variables and general health. Participants who responded to the baseline measure in 1995 (n=2425)10 received follow-up surveys in 1998, 2003, 2007 and 2016. Respondents to the follow-up in 2016 was 1184. The questionnaire was sent out by regular mail, and at the last measure point in 2016, participants also had the option to answer by a web-based survey. Two reminders were sent out on each occasion.

Study Ⅱ The participants in the original RCT104 (n=122) were recruited from three physical therapy primary care centers from 2003 to 2004. The eligible pa-tients, aged 18-65 years, were seeking care for sub-acute, recurrent, or persis-tent musculoskeletal pain. The RCT included two active treatments provided by physical therapists: 1) a tailored behavioral medicine treatment protocol (TBT) and 2) an exercise-based physical therapy protocol (EBT). The physical therapists were trained to deliver treatments according to the protocols before the study start. The TBT was tailored according to individual behavioral goals and functional behavioral analysis95. The EBT was based on the best available evidence of physical exercise in chronic musculoskeletal pain.

32

Of the 97 participants who completed the RCT, 92 were eligible for the 10-year follow-up-study. The participants were traced through contact infor-mation collected during the original study, which was updated against the Swedish population register. Two had died, and another three could not be found in the register. For the 10-year follow-up, presented in this thesis, data were collected either by a postal survey or a web-based survey, depending on the participants’ preferences. Two reminders were sent out after two and six weeks.

Figure 3. Flow diagram of the surveys in the study cohort (study I).

33

Study Ⅲ and Ⅳ Study Ⅲ and Ⅳ are based on initial data from a planned RCT with a stepped care design, including behavioral medicine interventions based on both prog-nostic and individual factors in patients with musculoskeletal pain provided by physical therapists in primary health care157. The RCT aimed to study meth-ods for the identification of prognostic factors and subgroups with different pain condition severity, treatment effects on patients, and physical therapists´ clinical behaviors. In this thesis, investigations of physical therapists’ clinical behavior (study Ⅲ), and methods for assessments of subgroup assignment (study Ⅳ) were studied.

In study Ⅲ, 12 experienced physical therapists from six primary care cen-ters in three different counties in Sweden participated. The participants were females, aged 39-57 years (mean:50), recruited from five primary health care centers in different parts of Sweden. They had between 10 and 35 years in the profession (mean:19), and half of the participants were further educated in behavioral medicine or cognitive behavioral therapy. Background data and study-specific questions about attitudes towards a behavioral medicine model for clinical reasoning were collected by self-reported questionnaires. Data on their clinical behavior during initial consultations with patients seeking care for musculoskeletal pain were collected by video recording.

In study Ⅳ, the participants were 40 patients, aged 19-64 years (mean:42) seeking physical therapy for low back, neck or widespread pain in primary health care who were recruited from five primary care centers in three differ-ent counties in Sweden. The exclusion criteria were insufficient skills in Swe-dish language and red flags, i.e., identified serious spinal, malignant or sys-temic disorders or symptoms of severe depression. Data on background vari-ables, pain intensity and duration, pain control, general health, pain-related disability, anxiety and depression, fear avoidance beliefs, and pain catastro-phizing, were collected by self-reported questionnaires at two time points: 1) before the first consultation with the physiotherapists at the primary care cen-ter and 2) two to three weeks after the first consultation. The questionnaires were administered by a researcher and were answered by the participants by a web-based survey or a postal survey.

Data collection An overview of the variables and questionnaires used in studies I-Ⅱ and Ⅳ is presented in table 2. Data collection in study Ⅲ was performed by video-re-cordings.

34

Table 2. Variables and instruments used in studies I-Ⅱ and Ⅳ in this thesis. Variables Study 1 Study Ⅱ Study Ⅳ

Pain prevalence x

Health care utilization x

Pain intensity and pain control (NRS) x x

Sickness-related absence x

Self-reported benefit x

Instruments

Pain Disability Index (PDI) x x

Short Form-36 Health Survey (SF 36) x

Tampa Scale for Kinesiophobia (TSK) x x

Pain Catastrophizing Scale (PCS) x

Örebro Musculoskeletal Pain Screening Questionnaire (ÖMPSQ)

x

NRS= Numerical rating scale

Pain-related Disability (studies Ⅱ and Ⅳ) The Pain Disability Index (PDI)158 is a questionnaire assessing pain-related disability and social role functioning regarding seven areas: family and home responsibilities, recreation, social role activities, occupation, sexual behavior, self-care and life support activity. Each item ranges from 0-10 (total score 0-70. The PDI is considered has acceptable validity and test-retest reliabil-ity159,160. The Swedish version of the PDI79 was used in studies Ⅱ and Ⅳ.

Pain prevalence and pain location (study I) The prevalence of chronic pain (>3 months) during the last 12 months was as-sessed by specific items included in a self-reported questionnaire. Based on a pain drawing of the body with 18 predefined regions, pain was categorized into three categories: non-chronic pain (NCP), chronic widespread pain (CWP) according to the American College of Rheumatology 1090 criteria for fibromyalgia7, or chronic regional pain (CRP) if the criteria for CWP were not met.

Pain intensity and pain control (studies Ⅱ and Ⅳ) The maximum pain intensity (study Ⅱ) and the average pain intensity (study Ⅳ) experienced over the past two weeks were scored on a numerical rating scale (NRS)161 ranging from 0= no pain to 10= worst pain imaginable. The perceived pain control considering the participants´ strategies to manage pain during an average day was scored on the NRS ranging from 0= no control to 10= total control.

35

Health care utilization (study I) Questions about health care utilization during the last 12 months (study I) were responded to in a multiple-choice format with four categories for each health care provider: 0 consultations, 1 consultation, 2-5 consultations and >5 con-sultations. The ordinal data on health care utilization included consultations with physiotherapists, general practitioners, rheumatologists, orthopaedic sur-geons/general surgeons, pain physicians, emergency room physicians, occu-pational physicians and other physicians.

Health-related quality of life (study I) The standard version of the Short Form-36 Health Survey (SF 36)162 (study I) measures health-related quality of life across eight generic health concepts, covering physical, psychosocial and health variables. It is considered generic and represents basic human functions and well-being. The questionnaire con-sists of 35 items grouped into the following eight categories: physical func-tioning, role function – physical aspects, bodily pain, general health percep-tion, vitality, social functioning, role function – emotional aspects, and mental health. For each item, raw scores were coded, summed and transformed into a subscale according to a scoring algorithm162 ranging from 0-100, where a higher score indicates better health. The Swedish version of the SF-36 that was used in the current study has been reported to be reliable and valid in general adult and elderly populations163-165.

Sickness-related absence (study Ⅱ) Data on sickness-related absence periods were gathered from the Swedish So-cial Insurance Agency Register during the entire period from three months prior to baseline to the 10-year follow-up. The register includes data on sick-ness-related compensation periods (i.e., days of sick leave and disability pen-sion) exceeding 14 days, and different degrees of compensation from 25% to 100% of the full-time compensation are also available.

Self-reported benefit (study Ⅱ) Self-reported benefit from the intervention and confidence in solving future problems were measured by two study-specific ordered-category items. The first question was “To what degree have you had benefits from what you learned from treatment?” A four-point item ranging from 1 = “no benefit” to 4 = “very large benefit” was used. The second question was “How confident are you that you can cope with future problem situations?” A five-point item ranging from 1 = “very confident” to 5 = “very unconfident” was used.

36

Fear and avoidance (studies Ⅱ and Ⅳ) The Tampa Scale for Kinesiophobia (TSK)166 was used to assess fear of move-ment and/or re-injury. The scale consists of 17 items (total score 17-68), and a cut-off score of ≥37 indicates high levels of fear-avoidance beliefs. Re-sponses are made on a 4-point Likert scale ranging from 1=strongly disagree to 4=strongly agree. Four of the items are reverse-scored. The Swedish version of the questionnaire (TSK-SV) was used in this study and has been reported to be valid and reliable167.

Pain catastrophizing (study Ⅳ) The Pain Catastrophizing Scale (PCS)168 was used to assess pain catastrophiz-ing. The scale consists of 13 items (total score 0-52) describing thoughts and emotions related to pain, with a cut-off score of ≥30 for clinically relevant catastrophizing 169. The PCS comprises three components of catastrophizing: 1) rumination, reflecting ruminative thoughts and worry; 2) magnification, re-flecting magnification of negative experiences of pain; and 3) helplessness, reflecting inability to cope with painful situations. Responders are instructed to rate their thoughts and feelings related to pain using a 5-point Likert scale ranging from 0=not at all to 4=all the time, where a high score indicates a high degree of catastrophizing. The PCS has been used in physical therapy settings and has good psychometric properties, including internal reliability and con-struct validity170.

Prognostic factors for poor outcome (study Ⅳ) The Örebro Musculoskeletal Pain Screening Questionnaire (ÖMPSQ)171 was developed and used to screen for psychological risk factors for long-term pain-related disability153. The ÖMPSQ consists of 25 items, of which 21 are scored (total score 3-210). The questionnaire has good reliability and predictive va-lidity for musculoskeletal pain153,172,173. Different high-risk cut-off scores have been reported in different pain conditions and settings149,174,175. A higher cut-off score, i.e., ≥112, is considered to have high specificity to identify patients at high risk for poor outcome, but with lower sensitivity175. A cut-off of ≥90 has been reported to have lower specificity but higher sensitivity, which in-creases the ability to identify a majority of patients at high risk of developing long-term disability153. Another purpose of the ÖMPSQ may be to identify potentially modifiable prognostic factors in patients with musculoskeletal pain as a basis for stratified treatment. A few studies have reported psychometric properties of subgroup assignment based on the ÖMPSQ151,154,176 but used dif-ferent methods for categorizing the subgroups.

37

Observable clinical behavior (study Ⅲ) Data on the physical therapists´ clinical behaviors were collected by video re-cordings during initial consultations with patients seeking care for musculoskel-etal pain; the patients were scheduled according to the usual routines. The video recordings were performed at the primary care center where the participants were working. The video camera was placed in a fixed position using a single viewpoint with focus on the physical therapist during the consultation, and the researcher was an observer. Before the video recordings began, the researcher ensured that patients were comfortable with the camera’s presence, and oral in-formed consent was obtained from every patient. The video recordings lasted 19-69 minutes (mean 39 minutes) and occurred between June 2013 and January 2014.

Observation protocols Four domains were targeted for the observation. An observation protocol was developed to identify the participants’ specific behaviors regarding the assess-ment of red and yellow flags and analyzing clinical problems. Red flags: To be defined as having assessed red flags, the physical therapists had to ask at least one question specifically about severe symptoms and other severe diagnoses, traumas or spinal pathology in combination with a question about general health. Yellow flags: The yellow flags were categorized into three domains: 1) beliefs, appraisals, and judgments; 2) emotional responses; and 3) pain behavior, in-cluding pain coping strategies49. To be defined as having assessed yellow flags, physical therapists had to ask at least one question about specific factors according to the protocol to make it possible to identify or exclude the pres-ence of yellow flags. Analysis of the clinical problem was categorized as one of three models: a) func-tional behavioral, b) bio-psychosocial, or c) biomedical. For an analysis to be categorized as functional and behavioral, the pain problem had to be defined in behavioral terms, and functional relationships between individual and contex-tual factors had to be identified and discussed with the patient, for instance, how a patient’s fear of increased pain resulted in avoidance behaviors that were neg-atively reinforced by subsequent reductions in fear, which in turn impeded par-ticipation in important activities77,95. For an analysis to be categorized as bio-psychosocial, both biomedical and psychosocial factors had to be integrated. A biomedical analysis had to be based on solely physical assessments, physical impairments or limitations and the physical environment. The assessment of BCTs was measured using the BCT taxonomy129. The BCTs that were observed in the video observations were identified and labeled ac-cording to the definitions in the taxonomy. The overall inter-coder reliability

38

and test-retest reliability has been reported to be good, and the taxonomy pro-vides a reliable method for coding behavior change techniques in intervention studies177,178.

Data management and analysis Statistical analyses were performed using the IBM, Statistical Package for the Social Science (SPSS) software version 24.0 (SPSS Inc., Chicago, IL, USA). The statistical methods used in studies 1-Ⅳ are presented in table 3.

Table 3. Statistical methods used in studies I-Ⅳ. Methods Study 1 Study Ⅱ Study Ⅲ Study Ⅳ

Mean and standard deviation (SD) x x x

Mean and range x x

Frequencies and percentage x x x x

Median and interquartile range (IQR) x x

Pearson´s chi-squared test x

Fisher´s exact test x x x

McNemar´s test x

Independent-sample t-test x

Mann-Whitney U-test x

Binary logistic regression analysis x

Weighted kappa analysis x

Deductive and inductive video analysis x

Study I Data on health care utilization were dichotomized at each time point to either high or low health care utilization. High utilization was based on >5 consulta-tions with at least one health care provider, or on ≥1 consultation with at least 3 different health care providers. All other combinations were considered to be low health care utilization. Data on health care utilization for physical ther-apists and for general practitioners were selected for separate analyses. A bi-nary logistic regression analysis was used to calculate odds ratios and 95% confidence intervals for the association between potential predictive baseline variables and health care utilization in 2016.

Trajectories for health care utilization were identified based on low or high total health care utilization at each point of measurement. Three points of measurements were included in the analyses: 1995, 2007 and 2016. In case of missing data in 2007, data from 2003 were imputed to represent 2007 data in

39

the analyses. Comparisons between different trajectories were performed us-ing a chi-square test or Fisher’s exact test for categorical data and with inde-pendent-sample t-tests for numerical data. Comparisons of pain prevalence, i.e., NCP, CRP, and CWP, and health care utilizations in 1995, 2007 and 2016 were performed with McNemar´s test.

Study Ⅱ Between-group analyses of the change scores for the differences between the scores from each variable at post-treatment and at the10-year follow-up data, were analyzed by Mann-Whitney U-tests.

Regarding the total days of sickness-related absence from the register data, three periods were analyzed: from three months prior to baseline until base-line, from baseline until post-treatment, and post-treatment until the 10-year follow-up. The data on sickness absence were dichotomized into two catego-ries (i.e., absence/ no absence) and were analyzed with Fisher´s exact test. The follow-up questions about self-reported benefit were analyzed with Fisher´s exact test.

Study Ⅲ The video recordings were analyzed using a combined deductive and induc-tive approach (figure 4). An initial deductive analysis was performed based on a protocol with predefined criteria for physical therapists’ assessment of red and yellow flags, definitions of three different models of analysis and a taxonomy of BCTs129. Sequences demonstrating assessment, analyses and use of BCTs in all 12 video recordings were identified and transcribed. The con-tent of the transcriptions was analyzed and organized into subgroups based on the predefined criteria and coding of specific BCTs.

To describe the variation in the physical therapists’ clinical practice, all the selected video sequences were repeatedly reviewed to a) identify the physical therapists’ different ways of assessing, analyzing and using BCTs and b) iden-tify physical therapists’ different ways of interact with their patients during the consultations. Four of the 12 cases were deemed to illustrate the most clin-ically relevant variations according to the clinical behavior that were targeted in the video-sequences and are presented with brief descriptions of each one’s background, narratives and quotations.

40

Figure 4. Overview of the steps of the video-based analysis in study Ⅲ. BCT= Behav-ior change techniques.

Study Ⅳ Data on the total scores of the measurements included in the concurrent valid-ity analyses against the reference instruments were dichotomized as high or low based on previously established cut-off scores for each instrument: for the ÖMPSQ, ≥90; for the TSK, ≥37; and for the PSC, ≥30. No standardized cut-off score for the PDI was found in previous research. A cut-off score of ≥25 for high pain-related disability was therefore used based on the pooled mean scores on the PDI reported in previous studies on three primary health care samples of patients with musculoskeletal pain79,104.

The concurrent validity of the subgroups based on the ÖMPSQ in compar-ison to the reference measurements (PDI, TSK and PCS), including data from the first measurement point (baseline), were analyzed with Fisher´s exact test.

The test-retest reliability of the subgroup assignment was performed with quadratic weighted kappa analyses based on the assumption that the three pre-defined groups differed in pain severity according to their graded level (from low to high)179. The weighted kappa analysis was calculated using Vassar Stats

180. The values of the weighted K-coefficient (Kw) were interpreted as follows: below 0.20 as poor agreement, 0.21-0.40 as fair, 0.41-0.60 as moderate, 0.61-0.80 as good, and >0.80 as very good179.

STEP 1:A first review of all video-recordings based on the protocol with pre-defined criteria and the taxonomy of BCTs.

STEP 2:Relevant video-sequences for the three domains in all 12 video-recordings were identified and transcribed.

STEP 3: Transcriptions were analyzed and organized, including a refined coding of specific BCTs, resulting in a summarizing description of all 12 case.

STEP 4: The video-sequences were reviewed several times for the in-ductive analysis and illustrative cases were identified to describe the variation.

41

Results

Follow-up of a population-based cohort Of the 1882 eligible participants, the total number of respondents to the fol-low-up questionnaire in 2016 was 1184 (63%). Attrition was due to actively declining participation (13%), not responding to the questionnaire (85%), and death of the participant (2%). The study sample consisted of 675 (57%) women and 509 (43%) men, and these rates did not differ significantly com-pared to the baseline data from 1995, which included 1285 (53%) women and 1140 (47%) men.

Significant differences were found between those who responded and those who did not respond to the follow-up in 2016. Among those who responded, there were more females in the 1st and 2nd age quartiles and more individuals who reported either NCP or CWP at baseline. Immigrants were more likely to be non-responding. Regarding the trajectory analysis, 44% (n=1064) of all participants in 1995 (n=2425) responded to all of the three measures (1995, 2007 and 2016).

The prevalence of chronic pain (CRP and CWP) increased from 1995 to 2007 and from 1995 to 2016, and the NCP group decreased. No differences were found regarding data on total health care utilization between the points of measurements. Regarding consultations to physiotherapists, differences were found between 1995 and 2007, during which high health care utilization decreased, whereas low utilization increased. Increased frequencies of consul-tations with general practitioners were found from 1995 to 2016 and from 2007 to 2016.

Prediction of high health care utilization in 2016 Significant associations were found between baseline variables and high or low health care utilization in 2016 regarding gender, age, pain prevalence, health-related quality of life, and levels of health care utilization. Further bi-nary logistic regression analyses of the predictors of high health care utiliza-tion in 2016 were performed (table 4). Female gender was a predictor of high total health care utilization based on physiotherapist consultations but not on general practitioner consultations. High age was a predictor of high health care utilization based on general practitioner consultations. Both CRP and CWP predicted high health care utilization 21 years after baseline according to total

42

health care utilization and consultations with physiotherapists and general practitioners, respectively. Being an immigrant did not predict high health care utilization in 2016. Low scores on health-related quality of life (SF36) at base-line regarding all domains were associated with high total health care utiliza-tion and with consultations with general practitioners. Low scores in the four domains of bodily pain, general health, vitality and mental health were asso-ciated with high health care utilization based on physiotherapist consultations. The results of the logistic regression analysis are presented with ORs and 95% confidence intervals (table 4).

Table 4. Predictive variables at baseline 1995 for high total health care utilization, for physical therapists, and for general practitioners at the 21-year follow-up.

Variables 1995

Total health care utilization n=1130 OR (95% CI) p

Physical therapist n=1070 OR (95% CI) p

General practitioner n=1107 OR (95% CI) p

Gender Male 1 1 Female 2.0 (1.4-3.0) <0.01 1.9 (1.2-3.1) 0.01 1.3 (0.7-2.3) 0.36 Immigrant No 1 1 1 Yes 1.6 (1.0-2.7) 0.07 0.9 (0.5-2.0) 0.87 1.6 (0.8-3.4) 0.21 Age 20-30 1 1 1 32-47 1.0 (0.6-1.7) 0.87 1.2 (0.7-2.1) 0.54 0.5 (0.2-1.2) 0.12 48-59 1.2 (0.7-1.8) 0.69 0.8 (0.4-1.5) 0.50 0.7 (0.4-1.6) 0.43 60-75 1.6 (0.9-2.9) 0.12 0.4 (0.2-1.2) 0.12 2.2 (1.0-4.8) 0.04 Pain-group NCP 1 1 1 CRP 1.8 (1.2-2.6) <0.01 1.9 (1.1-3.1) 0.02 2.0 (1.1-3.8) 0.03 CWP 3.2 (1.9-5.1) <0.01 3.9 (2.1-7.1) <0.01 4.8 (2.4-9.5) <0.01

NCP= Non chronic pain, CRP= Chronic regional pain, CWP= Chronic widespread pain

Trajectories of health care utilization The trajectories illustrate different patterns of total health care utilization from baseline to the 21-year follow-up. Four distinct trajectories were identified based on visual analyses of the descriptive data: stable high (n=23), stable low (n=744), increasing (n=108), and decreasing (n=107) trajectories. One group fluctuated between high and low levels of health care utilization without a specific direction of trajectory (n=88) and was not included in the analyses.

The stable low group was characterized by low health care utilization at all three points of measurement. A majority of the subjects in the low group re-ported NCP at baseline (80%), and there was an equal distribution between genders.

43

The stable high group was characterized by high levels of health care utili-zation at all measurement time points. This group consisted of significantly more women (74%) compared to the stable low and the increasing groups. A majority of participants in the stable high group reported CWP at baseline (61%).

The increasing group was characterized by low levels of health care utili-zation at baseline followed by an increasing trend. A majority of the subjects were women in the increasing group (72%). Approximately half of the partic-ipants in the increasing group reported CRP or CWP at baseline.

The decreasing group was characterized by high levels at baseline followed by a decreasing trend. Nearly 80% of all the participants reported CRP or CWP at baseline in the decreasing group. Associations between baseline characteristics and trajectories of health care utilization are presented in table 5.

44

45

Follow-up of tailored behavioral medicine treatment Data from a total of 43 (44%) participants in the original study group (n=97), 20 in the TBT group and 23 in the EBT group were collected at the 10-year follow-up. Thirty-three (77%) of the participants were women, and the mean age was 42.8 years. There were no evident differences in pain-related disabil-ity, pain intensity, pain control, fear of movement/(re)injury or baseline vari-ables immediately post-treatment between those who completed the 10-year follow-up and those who did not.

Pain-related disability No significant differences between the groups were found regarding the change in pain-related disability scores (PDI) from immediately post-treat-ment to 10 years after treatment (table 6). The differences reported at the two-year follow-up123 were levelled out at the 10-year follow-up (figure 5).

Secondary outcomes Pain intensity and pain control: A significant difference in the change-scores from immediately post-treatment to 10 years after treatment was found for the maximum pain intensity (table 6). The TBT group reported increased levels of pain, whereas the EBT group reported decreased levels (figure 6). In total, 24 (56%) of the participants in the 10-year follow-up reported a score of five or higher for the pain intensity on the NRS. No differences were found be-tween the two groups regarding pain control or fear (table 6). Fear and avoidance: No differences were found between the two groups re-garding fear of movement and/or re-injury (table 6). Sickness-related absence: The presence of sickness-related absence at three months prior to baseline, at baseline, post-treatment, and at the 10-year follow-up did not differ between the groups, except for three months prior to the in-tervention (p=0.05). In terms of the total days of sickness-related absence (>14 days), there were significant more days of absence in the exercise-based phys-ical therapy group both three months prior to the intervention (p=0.02) and at the 10-year follow-up (p=0.03). Self-reported benefit: Regarding the two questions about “perceived benefit of the former treatment”, and “confidence in coping with future problem situ-ations”, no significant differences were found between the groups (p= 0.63).

46

Figure 5. Pain-related disability at baseline, post treatment, at the 2-year follow-up, and at the 10-year follow-up, measured with the Pain Disability Index (PDI).

Figure 6. The maximum pain intensity at baseline, post treatment, at the 2-year fol-low-up, and at the 10-year follow-up, measured with the Numerical Rating Scale (NRS).

47

Table 6. Differences in change scores from between group-analyses of pain-related disability, maximum pain intensity, pain control, and fear of movement/(re)injury immediately at post-treatment and at the 10-year follow-up.

Outcome Measurement n= 43

Post- treatment Mdn (IQR)

10-year follow-up Mdn (IQR)

Change score Mdn (IQR)

Between-group p

Disability (PDI) TBT EBT

4.5 (11.7) 15.0 (25.0)

10.0 (16.5) 15.0 (22.0)

2.5 (16.8) 0.0 (11.0)

0.17

Max pain intensity (NRS) TBT EBT

3.0 (4.0) 8.0 (5.0)

6.0 (5.7) 6.0 (7.0)

1.0 (5.0) 0.0 (5.3)

0.03

Pain control (NRS) TBT EBT

8.0 (4.5) 7.0 (3.0)

7.5 (3.8) 7.0 (4.0)

0.0 (6.0) 0.0 (4.3)

0.49

Fear of movement (TSK) TBT EBT

25.0 (9.8) 28.0 (12.0)

28.5 (10.0) 29.0 (12.0)

2.3 (10.6) 2.0 (11.0)

0.25

TBT= Tailored behavioral treatment. EBT= Exercise-based physical therapy. PDI= Pain Disability Index (0-70), low scores indicate low disability. NRS= Numerical Rating Scale (0-10); low scores indicate low pain intensity/ pain control. TSK= Tampa Scale for Kinesiophobia (17-68); low scores indicate low fear.

Observation of physical therapists in primary health care Description of clinical behavior The results of the deductive analysis of the observed clinical behavior during the initial consultations with the physical therapists are described with fre-quencies and content according to the observation protocols (table 7). Red flags were assessed in nine out of 12 observations but were not identified in any of them. Yellow flags were assessed in eight out of 12 observations. In four of the con-sultations yellow flags regarding pain behaviors, coping strategies and emo-tional responses were identified. Model for analysis: Biomedical analyses were performed in ten out of 12 ob-servations. A bio-psychosocial analysis was performed by one of the physical therapists. None of the analyses met the criteria for a functional behavioral analysis. In one case, no analysis was performed at all. Use of behavior change techniques: BCTs were used by all of the physical therapists, mainly to facilitate physical activity. The most frequently used BCTs concerned information about health consequences, instruction on how

48

to perform the behavior, i.e., physical exercise, and feedback on behavior. BCTs concerning goals and planning were observed in two of the 12 consul-tations.

Table 7. Description of the 12 physical therapists’ observed clinical behaviors. Clinical behaviors Frequency n

Total n=12

Assessment of red flags 9

Assessment of yellow flags 8

Functional behavioral analysis 0

Biopsychosocial analysis 1

Biomedical analysis 10

No analysis performed 1

Use of BCTs 12

BCT= Behavior change techniques

Variation in clinical behavior The result of the inductive analysis of the video-observations revealed great variation in the physical therapists’ clinical practice regarding their assess-ments, model used for analysis, use of BCTs, and how they interacted with their patients. Four of the cases were selected to illustrate these variations, and there were various degrees of patient involvement regarding their clinical problem and treatment discussion. In case 1, the patient was involved in com-munication to a great extent in the combination of biopsychosocial analysis. In case 2, the patient was involved in the assessments, but the psychosocial factors were not integrated into the analysis or treatment. In cases 3 and 4, the physical therapists used closed questions and were more focused on the bio-medical symptoms to various degree. Quotes from the selected cases and a detailed description of all 12 cases are presented in study Ⅲ. Case 1: High degree of patient involvement and a biopsychosocial analysis This case was selected to demonstrate a physical therapist who performed a bio-psychosocial analysis and guided her patient, who was seeking care for headache, in a behavioral goal setting. The physical therapist was experienced in her profession and educated in both motivational interviews and manual treatments. She assessed red and yellow flags and identified passive coping strategies and worries related to a stressful family situation. A bio-psychoso-cial analysis was performed based on the physical examination and the iden-tified yellow flags. The patient was involved in the communication regarding sharing information, making decisions, and responsibility for assessment and

49

treatment. This was the only case that included a BCT regarding behavioral goal setting. Case 2: Some patient involvement and a biomedical analysis This case was selected to demonstrate the assessment and identification of yellow flags that were not included in the analysis of the patients´ clinical problem or related to the use of BCTs. The physical therapist had many years of experience in the profession, was further educated in physical activity and in motivations interviews on a basic level. The patient, seeking care for shoul-der pain, was assessed for red and yellow flags, and negative thoughts about recovery and treatment outcome, avoiding behavior related to physical activ-ity was identified by the physical therapist. She performed a biomedical anal-ysis based only on the physical examination, despite the identified yellow flags. Case 3: Minor patient involvement and a biomedical analysis This case was selected to demonstrate how a few BCTs can be used in a ped-agogical and structural way to promote home-based exercise. The physical therapist had many years in the profession and was educated in the manual treatment of pain. The analysis of the patient, seeking care for rehabilitation after a hip replacement surgery, was based solely on findings from the physi-cal examination. The communication was directed by the physical therapist and focused on biomedical factors. The patient was involved in the discussion about the home-exercises. Case 4: Very low degree of patient involvement and no analysis This case was selected to demonstrate a traditional biomedical model of clin-ical reasoning in which the physical therapist was active and the patient, seek-ing care for pain of the elbow and shoulder, played a passive role. The com-munication was strictly directed by the physical therapist and consisted mainly of closed questions. The physical therapist had many years in the profession and was well educated in manual treatment and cognitive behavioral therapy for physical therapists. Red flags, but not yellow flags, were assessed. A gen-eral biomedical hypothesis regarding the relationship between pain and work was performed, but it was not specifically based on the patient’s clinical prob-lem or assessments. A few BCTs were used, mainly to promote physical ex-ercise.

Subgroup assignment based on the ÖMPSQ A total of 40 participants from five primary health care centers completed the first questionnaire at baseline (table 8). The second questionnaire after two to three weeks was completed by 32 participants. Reasons for dropping out were red flags (n=2), not interested in participating in the study, or refraining from consultation due to reduced pain or other reasons (n=6).

50

Concurrent validity of subgroup assignment The subgroup assignment based on the ÖMPSQ was associated with scores on the PDI and the TSK (table 9). Being in the low-severity subgroup was associated with low scores on the PDI and the TSK, 2) being in the moderate-severity subgroup was associated with high scores on the PDI and low scores on the TSK, and 3) being in the high-severity subgroup was associated with high scores on both the PDI and the TSK. No association was found between the subgroup assignment and scores on the PCS (table 9).

Table 8. Baseline characteristics of the participants (n=40) Variable

Age (year), mean (SD) 42 (10) Female n (%) 25 (63) Nationality (born in Sweden), n (%)

34 (85)

Duration of pain, n (%) <3 months 12 (30) ≥3 months

28 (70)

Presence of pain, n (%) Persistent 26 (65) Intermittent

14 (35)

Location of pain, n (%) Neck 3 (7.5) Low back 18 (45) Widespread pain 18 (45) Shoulder 1 (2.5)

Pain intensity (average) NRS 0-10, mean (SD) 7 (2) Occupation, n (%) Working

26 (65)

Sick leave 6 (15) Unemployed 4 (10) Maternity leave 3 (7.5) Student 1 (2.5) ÖMPSQ scale 3-210, mean (SD) 114 (31) PDI scale 0-70, mean (SD) 35 (16) TSK scale 17-68, mean (SD) 40 (8) PCS scale 0-52, mean (SD) 30 (12)

NRS= Numerical Rating Scale, ÖMPSQ= Örebro Musculoskeletal Pain Screening Questionnaire, PDI= Pain Disability Index, TSK= Tampa Scale for Kinesiophobia, PCS= Pain Catastrophizing Scale

51

Table 9. Concurrent validity of subgroup assignment based on the Örebro Musculo-skeletal Screening Questionnaire (ÖMPSQ), and measures of the Pain Disability In-dex (PDI), the Tampa Scale for Kinesiophobia (TSK) and the Pain Catastrophizing Scale (PCS) (n=40). Measure Low

severity subgroup n (%)

Moderate severity subgroup n (%)

High severity subgroup n (%)

df

Fisher´s exact test

p-value

PDI (0-70) 2 13.3 0.001 Low <25 7 (78) 2 (20) 2 (9.5) High ≥25 2 (22) 8 (80) 19 (90.5) TSK (17-68)

2

12.1

0.002

Low <37 5 (56) 7 (70) 3 (14) High ≥37 4 (44) 3 (30) 18 (86) PCS (0-52)

2

3.6

0.20

Low <30 6 (67) 7 (70) 8 (38) High ≥30 3 (33) 3 (30) 13 (62)

Stability of subgroup assignment The stability of the subgroup assignment between the two points of measure-ment for the subgroup assignment, as calculated with the weighted kappa-co-efficient, was 0.51 (95% CI: 0.22-0.81), which was considered moderate (Kw0.41-0.60).

52

Discussion

General discussion This thesis includes various aspects of the management of musculoskeletal pain conditions, which are important for further development of the field. In studies I and Ⅱ the long term perspective was considered: a 21-years follow-up of a cohort study regarding chronic pain and health care utilization which adds new information about the long-term impact of chronic pain on health care utilization, and a 10-years follow-up of an RCT including tailored behav-ioral treatment and exercise-based treatment in primary health care, contrib-uting valuable knowledge about the development of pain and disability in a primary health care sample. In study Ⅲ, the potential and challenges for the integration of behavioral medicine treatment into physical therapy were ex-plored and described. In study Ⅳ, an instrument for subgrouping patients based on psychological prognostic factors was evaluated for use in stratified treatment.

Chronic pain is a major problem in the general population, but for some subgroups, it is a problem that extends over a 21-year period and is associated with high healthcare utilization. Even individuals with musculoskeletal pain who participated in an RCT comparing TBT and EBT still reported moderate pain intensity 10 years after the intervention. Although our findings did not correspond with high levels of disability, chronic pain has been reported to be associated with great physical and psychosocial suffering for the individual63. The results from the long-term follow-ups indicate that the treatment of mus-culoskeletal pain for specific subgroups has not been sufficiently effective in the long run, and further efforts are needed to improve the long-term out-comes.

Most patients with musculoskeletal pain seek primary care initially, where physical therapists are often the first health care providers. In study Ⅲ, a ma-jority of the physical therapists assessed red and yellow flags, in line with clinical guidelines, but they did not integrate the psychosocial findings into the analysis and treatment, indicating a need for further efforts to integrate behavioral medicine into physical therapy.

Psychological prognostic factors for a poor prognosis have been reported as significant for targeting treatment and can be identified with screening tools developed for this purpose. Our results indicate that the ÖMPSQ can be used to target treatment to different subgroups based on psychological factors, but

53

it needs to be complemented with an assessment of the pain condition, includ-ing pain catastrophizing and individual characteristics.

Prognostic factors for health care utilization Health care utilization has been associated with several pain conditions, and with chronic pain in particular. Our findings that chronic pain predicted high health care utilization 21 years later, indicated that a specific vulnerable group, characterized by CWP, female gender and high age, is more likely to utilize health care resources in the long run. This result may also indicate that treat-ment has not been successful for individuals who still experience chronic pain for such a long period of time. The results from previous research have been inconsistent regarding predictors of high health care utilization. Similar to our findings, several studies have reported that chronic pain13,14,16,55,181,182 and fe-male gender13,52,53,181 predicted high health care utilization. In contrast, no dif-ferences regarding gender were found in a Swedish population-based study by Andersson et.al.182. Old age has also been found to predict high health care utilization16,52,181, which is in line with our findings, while no association be-tween age and health care utilization was found in a study by Ferreira et.al.53. Regional musculoskeletal pain in the acute stage (≤4 weeks duration) has also been associated as a predictor of health care utilization after 18 months183, and after 20-years60. Health care utilization may vary between different pain pop-ulations and the prevalence has been investigated and analyzed using different methods, which can explain the different results. However, the long follow-up period and the identification of trajectories of health care utilization over the entire period is unique to our study.

Individuals with chronic pain are expected to have an increased health care utilization compared to those without pain, but our findings emphasize the importance of identification of predictors for health care utilization early in the course of pain to target the treatment against these variables. Psychological predictors for high health care utilization, i.e. fear and avoidance, anxiety, and negative beliefs regarding consequences and outcome have been reported54,183, as well as level of disability53. These factors are also important variables to consider in the management of musculoskeletal pain, but were not targeted in study I.

The prevalence of chronic pain increased over a 21-year follow-up period in the general population, which was also expected, considering the aging study population. Interestingly, despite the increased prevalence of pain, no corresponding increase was seen in health care utilization during the same pe-riod. Among individuals with low back pain, less than 60% were seeking care for their pain condition53, and in an elderly population (≥62 years) there were less than 50%184, similar to our results. Individuals with chronic pain may have adapted to their pain condition over time, resulting in a decreased impact on

54

function in daily activities and on health care utilization. In a study on patients with fibromyalgia, increased health care utilization was reported the first six months after the diagnosis, followed by stabilization over a three-year pe-riod56, which supports the explanation of our findings.

Behavioral medicine and physical therapy In recent years a paradigm shift in physical therapy has been discussed regard-ing the management of chronic pain toward an integrated model including bi-omedical, psychosocial and behavioral factors116. One example of an inte-grated intervention is the tailored behavioral medicine treatment intervention evaluated in study Ⅱ, which was based on a treatment protocol including graded activity. Significant improvements were found in the TBT group re-garding pain-related disability, pain intensity and pain control104. Positive treatment effects of graded activity regarding disability and pain variables in musculoskeletal pain have also been reported in other studies106,185, and im-proved adherence to exercise and a higher level of physical activity have been reported in patients with osteoarthritis105. Other studies have reported equally effects of graded activity treatment compared with exercise-based treat-ment106,107,186, in contrast to the findings at post-treatment and at the two-year follow-up in our sample. Treatment effects of behavioral medicine interven-tions in patients with musculoskeletal pain are still sparsely reported, and there is no consensus about the beneficial effects of behavioral medicine treatments versus more traditional treatments, i.e., exercise-based therapy and manual therapy treatments. There is also limited evidence regarding which patients would benefit from which type of treatment. When applied in a physical ther-apy context, it is important that physical therapists have the knowledge and skills regarding assessments of physical and psychosocial factors, identifica-tion of patients at risk for poor prognosis, and what impact those factors have on behaviors that are important to the patient. This information is crucial for the analysis of the patient's clinical problem and as a basis for the choice of treatment. Identification of psychological variables such as a high level of fear and avoidance beliefs and catastrophizing behavior has reported to be im-portant when allocating a specific treatment, e.g., graded exposure; otherwise, the effect may fail, or at worst, lead to a deterioration of the treatment out-come187. To succeed, the treatment has to be tailored based on individual func-tional behavioral analyses of key behaviors and to the patient´s biomedical and psychosocial condition, but it must also be targeted based on prognostic factors.

55

Long-term outcome Few studies on behavioral medicine interventions applied in physical therapy have reported treatment effects longer than 12 months after the intervention. The purpose of behavioral change interventions in chronic pain is to improve the physical and psychosocial function of the individual through learning new strategies, i.e., coping with barriers to recovering. Another goal is to acquire strategies for maintaining of behavioral change, and preventing a relapse to maladaptive behaviors with pain-related disability as a consequence. That is, an important reason why the long-term effect of such interventions has to be investigated to provide valuable information when designing future research on pain interventions.

Previous studies have been reported that positive effects of behavioral med-icine treatment regarding disability, pain intensity and pain control were main-tained two years123 and five years188 after the intervention. In study Ⅱ, no be-tween-group differences remained at the 10-year follow-up. The total number of sickness-related days of absence, included as an outcome variable in the 10-year follow-up, was found to be low in both groups. A limitation regarding the register data on sickness-related absence was that we did not have infor-mation about short-term absence (<14 days). There may be several explana-tions for our findings. We had no data on co-morbidities, important life events or received pain treatments that may have affected participants' experience of their pain condition after such a long time.

The main outcome in study Ⅱ was pain-related disability, which is one of the most important consequences related to chronic pain. Although no differ-ences between the groups could be detected in the follow-up of the previous RCT, the participants in both groups still reported a moderate level of pain-related disability 10 years after the intervention (figure 5). The median scores on the PDI were 10.0 for the TBT group and 15.0 for the EBT group. The scores were below the cut-off score of ≥25 for high disability used in this the-sis but above the scores for recovery of ≤4189. The moderate levels of pain-related disability at the 10-year follow-up did not correspond to high levels of pain intensity, which were reported to be moderate to high in both groups. The results indicate that the participants were still affected by their pain but did not experience such a great impact on the function in daily activities. As previ-ously discussed, adaption to the pain condition may have occurred over time, resulting in a decreased level of self-reported disability.

Assessment of prognostic factors To improve the assessments of psychological prognostic factors in a system-atic way, the use of screening tool in physical therapy is suggested. Assess-ment of prognostic factors for long-term disability and poor outcome is rec-ommended in clinical guidelines22,23. In study Ⅳ, assignment to subgroups

56

with low, moderate and high pain condition severity based on the ÖMPSQ was found to have acceptable concurrent validity and stability, considering the small study sample. In line with our findings, assignment to prognostic sub-groups based on the ÖMPSQ has been used in previous studies149,151,154 and has been found to be a valid and reliable screening tool for this purpose. How-ever, the definitions of the subgroups in these studies differed somewhat from our classification. The assignment to the three subgroups used a cut-off score of ≥ 90 for the high subgroup, which has not been previously validated against the PDI, TSK and PCS. However, if the ÖMPSQ is to be used to stratify treat-ment based on the different prognostic subgroups, further studies are required to investigate the effect of treatment targeted against these subgroups in pa-tients with musculoskeletal pain.

A similar screening tool is the STarT Back Tool (SBT), which was devel-oped to identify prognostic subgroups of patients with back pain in primary health care190. The purpose of the SBT is to categorize patients with modifia-ble prognostic factors into low-, medium- and high-risk subgroups for use as basis for treatment. Comparisons between the ÖMPSQ and the SBT have re-ported equally good concurrent validity154,176 and similar ability to discrimi-nate patients according to standard reference instruments154,191. The correla-tion between the ÖMPSQ and the SBT regarding assignment to low-, moder-ate- and high-risk subgroups has been reported to have moderate agree-ment154,176. Although both instruments are suggested to be used for assignment to different subgroups in clinical practice, there are differences to consider. The SBT is short and easy to use, comprising only modifiable prognostic fac-tors and was developed for patients with back pain, while the ÖMPSQ pro-vides more information about pain characteristics and background variables, and has also been used for other musculoskeletal pain conditions and was therefore considered to be best suited for our purpose. Valid and reliable screening tools are important to identify which patients will benefit from a specific type of treatment, i.e., minimal interventions, exercise-based therapy, graded activity or graded exposure, according to prognostic factors.

Stratified treatment by target prognostic factors based on the subgroup level has been reported150, and a few studies have reported the effects of stratified treatment for back pain provided in primary health care. Promising results of stratified treatment based on prognostic subgroups, compared with non-strat-ified treatment, e.g., exercise-based physical therapy or manual treatment, have been reported in patients with low back pain regarding disability, psy-chological and pain variables192-194. In contrast, another study on patients with back and/ or neck pain did not find that the received stratified treatment cor-responded to the different risk subgroups, and a majority of the patients re-ceived “moderate-risk” treatment regardless of risk profile195. Matched treat-ment based on psychological risk profiles did not have a better outcome than unmatched treatment196. Based on the findings of study Ⅳ, the ÖMPSQ can

57

be used for stratified treatment but has to be validated in patients with different levels of pain condition severity.

A mean score of 35 on the PDI was reported in study Ⅳ, indicating a high level of pain-related disability. The scores was higher than the baseline score for the participants in study Ⅱ despite the similar clinical setting and the inclu-sion criteria104. Similar with our results, a study consisting of patients in mul-tidisciplinary pain rehabilitation programs with a variety of musculoskeletal pain conditions, a mean PDI score of 37.7 has been reported 197, indicating that our sample was not representative of patients seeking primary health for pain conditions.

Strategies for integration To integrate behavioral medicine treatment into physical therapy, it is urgent to investigate the current opportunities and challenges in clinical practice, which have been reported in previous studies. In study Ⅲ, new knowledge about experiences with physical therapists’ observed behavior was added, confirming previous studies reports of increasing awareness about the im-portance of integrating psychosocial factors into clinical prac-tice139,140,143,147,148. A finding observed in our study was that the majority of the participants assessed red and yellow flags, although they did not integrate their findings into the analysis and treatment. In contrast, an observational study of physical therapists’ use of cognitive behavioral principles in pain treatment142 reported that most of the physical therapists did not assess psychosocial fac-tors. Another finding, similar to our results, was that the use of goal-setting was under-utilized by a majority of the physical therapists.

Another finding of our study was the frequent use of BCTs by all partici-pants. Not surprisingly, BCTs targeting activity behaviors, such as “instruc-tions on how to perform a behavior”, “demonstration of the behavior”, and “informing about health consequences” of a specific behavior, were most commonly used, which was similar to previous results76. Techniques associ-ated with self-regulation, e.g., goal setting and self-monitoring of a target be-havior, were not used by the majority of the participants, in line with previous results76,142, while other studies have reported that physical therapists used goal- setting in their clinical practice104,147. The taxonomy of behavior change techniques129 was used only to identify, organize and describe the observed behaviors in the video-observation in study Ⅲ, regardless of whether they were based on biomedical or psychosocial findings. The intention of the ob-served BCTs and their relation to the findings from the assessments were not investigated. When integrating behavioral medicine treatment into physical therapy, the BCTs could also be used in the design and planning of behavioral change interventions by the selection of BCTs that have been reported to have the best effects on specific behaviors133,136-138 and applied within a theoretical

58

framework, e.g., SCT or the Transtheoretical Model of Behavior Change (TTM)198, to improve the treatment outcomes76.

Several aspects need to be considered when behavioral medicine is to be integrated into physical therapy: understanding of the impact of psychological factors on the consequences of pain and recovery, adequate skills training (i.e., assessments, analysis, patient-centered communication and treatments), and supervision and support during the process116-118. Increased understanding and knowledge about psychosocial factors as well as improved competence and confidence regarding the assessment and delivery of treatment have been re-ported after education and training in cognitive behavioral therapy and behav-ioral medicine of physical therapists117-119,144,199. Important factors to consider in those studies in which the integration of such treatment has been success-fully reported are that substantial time has been allocated for education, con-tinuous supervision and follow-up throughout the study. These strategies are considered costly and time consuming and could be difficult to implement in clinical practice. Although a paradigm shift regarding pain management among physiotherapists has been highlighted116, the more traditional biomed-ical approach still dominates, and changes in attitudes, beliefs and expecta-tions of the physical therapists’ role as a professional are needed in health care. If behavioral medicine treatment is to be integrated into physical therapy, it is important to implement change strategies on different levels; the individual level, e.g., clinical behavior; the environmental level, e.g., attitudes on the workplace and patients’ expectations; and on policy level, e.g., health eco-nomic costs and priorities.

To successfully change clinical behavior, an integration of behavioral med-icine into academic physical therapy education has been requested116. At the undergraduate level, such integration has begun but has to be further devel-oped, even in continuing education in physical therapy.

Methodological considerations Longitudinal studies Long term follow-ups have both strengths and weaknesses to be considered200. The most important strengths in the population cohort, study I, is the time line between the predictive baseline variables and the health care utilization at re-peated measures over a 21-year follow-up period, which provided useful in-formation about health care utilization in relation to the prevalence of chronic pain. Long-term follow-ups of rehabilitation interventions are often requested, but are difficult to perform and sparsely reported. In study Ⅱ, the effects of treatment aimed at changing target behaviors in chronic pain conditions were studied, and it was of great interest to investigate the adoption and mainte-

59

nance of these changes in the longer perspective and to describe the pain con-dition severity in the remaining participants after such a long time. Pain-re-lated disability was the main outcome, and a few secondary variables for which a significant difference was reported at post-treatment and at the 2-year follow-up were selected to be included in the 10-year follow-up. Sickness-related absence was included as a new variable to contribute additional infor-mation about the long-term outcome. A limitation was that the data on sick-ness-related absence were based only on register data, and no information about short periods of absence was available.

A weakness of both follow-ups was the rate of attrition. In study I, it was a plausible threat to the external validity, with an increased risk of selection bias in the remaining sample with a higher response rate for females, individuals with NCP or CWP and high age, which limits the possibility of generalizing the results to the general population. In study Ⅱ, the attrition rate raised im-portant issues regarding the validity of the data at the 10-year follow-up. The study was not sufficiently powered to detect meaningful differences after such a long time. There was also a plausible threat to the external validity consid-ering the attrition. However, no differences were reported regarding the drop-outs, baseline characteristics or outcome measures between the two treatment groups, and no evident differences in the baseline characteristics were identi-fied between those who responded to the follow-up and those who did not. Considering the long time span, the response rate of 63% in study 1 was good, and the response rate of nearly 50% in study Ⅱ was acceptable. The criteria for acceptable loss to follow-up in RCTs and cohort studies vary between 20% and 50% in the literature201,202.

The data on health care utilization in study I were self-reported from the last 12 months and were dichotomized into either low or high, based on the number of consultation to physicians and physical therapists. Higher rates of consultations regarding self-reported data compared to register data has been found52, which may has influenced the results. A combination of self-reported data and register data may provide the most valid data on health care utiliza-tion.

Video analysis In study Ⅲ, a combination of deductive and inductive analyses was used. The strengths were that the content in the consultation could be investigated with both quantitative and qualitative methods, which provided important infor-mation about the clinical practice during an initial consultation. Video obser-vation provides both audio and visual information about a specific situation, which is a strength, but it does not allow a further extended investigation of thoughts and intentions. The variation in the physical therapists’ clinical prac-tice regarding assessment, analysis, use of BCTs and the communication with their patients was illustrated in four selected cases. A weakness of this method

60

was that the data collected from 12 video observations were not completely and deeply analyzed. Thematic analysis has been used in previous interview studies of physical therapists clinical practice143,148,199,203 and would have been an alternative qualitative method to identify and describe the variation in the physical therapists’ clinical practice. However, in video-based research on health care providers quantitative coding and frequencies of coding behaviors are included in the analysis method204. A combination of video-observation and interview may provide valuable information about both overt and covert behaviors in a specific situation. We complemented our video-data with brief interviews with the physical therapists, which validated the findings from the video-recording.

Validity of subgroup assignment The associations between the assignment to subgroup with low, moderate and high pain condition severity and the dichotomized scores on three reference instruments were analyzed with Fischer’s exact test. Based on the results of these analyses, two instruments (PDI and TSK) were demonstrated to be as-sociated with assignment of the subgroups, but not the PCS. A limitation of the study was the small study sample, which did not make it feasible to per-form univariate comparisons between the subgroups or to use cluster analysis to identify different subgroups. Considering that the current method of sub-group assignment based on the ÖMPSQ has not been reported previously, the results can be used as a first indication that the psychometric properties are acceptable. Further investigation of the predictive and discriminative validity and the stability of the subgroup assignment is needed in different pain popu-lations and with a larger sample size, with the purpose of being used for strat-ified treatment.

61

Conclusions

Chronic pain, particularly CWP, has implications for health care utiliza-tion over 21-years. A trajectory of high health care utilization was identi-fied in a small group in the general population, characterized by CWP and female gender, which indicates that the treatment of chronic pain in this group has not been successful and remains a challenge for health care and society.

The between-group differences in favor of tailored behavioral medicine

treatment regarding pain-related disability, pain intensity and pain control and fear of movement/(re) injury, which were previously observed post-treatment and at the two-year follow-up, were not maintained 10 years after treatment. The study highlights the challenges associated with the control of threats to internal validity in the long-term perspective.

Our results suggest that physical therapists, although well-educated and

motivated to integrate behavioral medicine into physical therapy, mainly perform their analyses of patients’ pain conditions based on biomedical findings, indicating a need for innovative and evidence-based implemen-tation strategies.

Subgroup assignment according to the ÖMPSQ appears to be valid and

stable over time. It could be used as a basis for stratified treatment in pa-tients with musculoskeletal pain in primary health care but needs to be complemented with assessments of individual characteristics and behav-iors related to the outcomes of interest.

Clinical implications Prognostic factors such as chronic pain (particularly CWP) and female gender need to be considered when allocating health care resources and planning treatment to improve the long-term treatment outcomes.

Physical therapists are recommended to use early assessment of prognostic factors, and subgroup assignment based on the ÖMPSQ can be recommended to target treatment to different subgroups. Given the positive outcomes over

62

the first two years after the TBT intervention, treatment should also be tailored based on individual functional behavioral analyses of key behaviors and to the patient´s biomedical and psychosocial condition. To improve long-term out-comes, strategies for the maintenance of health behavioral changes are sug-gested to be included.

Methods for integrating behavioral medicine treatment into physical ther-apy need to be evaluated and improved. Integrating into physical therapy ed-ucation (both undergraduate and beyond), supervised skills-training, follow-ups of physical therapists’ clinical behavior, and health economic incentives to facilitate a change in clinical practice are suggested to succeed.

Future research Assignment to subgroups with low, moderate and high pain condition severity based on the ÖMPSQ should be further validated in different pain populations and in larger samples. Future studies are needed to evaluate the effects of strat-ified treatment by targeting prognostic factors based on the ÖMPSQ in com-bination with individual functional behavioral analyses of key behaviors and health outcomes. Strategies for the maintenance of behavioral change in indi-viduals with musculoskeletal pain should be explored in qualitative studies and subsequently evaluated in intervention studies.

A stepped care design would be used in future studies to differentiate treat-ments according to the patient’s need and pain condition severity to evaluate both treatment effects and the health-related cost-effectiveness.

To facilitate changes in physical therapists’ clinical reasoning and to inte-grate behavioral medicine into clinical practice, more extensive interviews in combination with video-observations are suggested. When a clinical behav-ioral change is integrated, the effect on treatment outcomes needs to be eval-uated.

63

Svensk sammanfattning

Muskuloskeletal smärta är vanligt förekommande och drabbar de flesta indi-vider någon gång under livet. För de flesta klingar smärtan av inom några veckor, men ca: 10 % utvecklar kronisk smärta (>3 månader) vilket ofta med-för ett stort lidande för individen med fysiska och psykosociala konsekvenser samt en ökad samhällskostnad i form av sjukskrivning och vårdkonsumtion. Trots framsteg inom smärtforskning, både inom biomedicinska och psykoso-ciala områden, finns det inga effektiva metoder för att behandla och förebygga kronisk smärta vilket är en utmaning både för sjukvården och samhället. Kro-nisk smärta är ett komplext tillstånd som involverar biologiska, psykosociala och beteendefaktorer. Under de senaste decennierna har intresset och kun-skapen om psykosociala faktorer och strategier för hälsofrämjande beteende-förändring ökat inom smärtforskning och är också huvudfokus i denna av-handling.

Syftet med avhandlingen var att studera prognostiska faktorer för vårdkon-sumtion relaterat till smärta i befolkningen, samt att undersöka långtidseffek-ter av individuell beteendemedicinsk behandling för patienter med musku-loskeletal smärta i primärvården. Ett annat syfte var att öka kunskapen om beteendemedicinsk kompetens hos fysioterapeuter avseende kliniskt resone-rande vid muskuloskeletala smärttillstånd, samt att identifiera förutsättningar och utmaningar för att integrera beteendemedicin i fysioterapi.

Studie 1 var en prospektiv kohortstudie. Ett representativ urval gjordes på 3928 personer mellan 20-74 år i Hallands län. Data samlades in genom fråge-formulär vid fem tillfällen under åren 1995-2016. Samband mellan bak-grundsvariabler, kronisk smärta, hälsorelaterad livskvalitet och vårdkonsumt-ion vid 21-årsuppföljningen undersöktes och analyserades. Resultatet visade att kronisk smärta vid baslinjemätningen 1995, hög ålder samt kvinnligt kön predicerade hög vårdkonsumtion år 2016. Ett mönster av hög vårdkonsumtion över hela tidsperioden kunde också identifieras hos en liten grupp, vilken ka-raktäriserades av kronisk generaliserad smärta och kvinnligt kön.

Studie Ⅱ var en 10-årsuppföljning av en randomiserad kontrollerad studie på patienter med muskuloskeletal smärta, där skräddarsydd beteendemedi-cinsk behandling jämfördes med fysisk träning. Båda behandlingsprotokollen utfördes av fysioterapeuter i primärvården. Det var ingen signifikant skillnad avseende funktionsnivå, smärta, smärtkontroll eller rörelserädsla mellan grup-perna vid 10-årsuppföljningen. Den statistiskt säkerställda skillnaden mellan grupperna, med större effekt för skräddarsydd beteendemedicinsk behandling

64

som rapporterades direkt efter avslutad behandling och efter 2-år, kvarstod därmed inte vid 10-årsuppfölningen.

I studie Ⅲ undersöktes hur 12 fysioterapeuter inom primärvården genom-förde patientbesök avseende undersökning av biomedicinska och psykolo-giska riskfaktorer, analys av patientens smärtproblem samt i vilken utsträck-ning beteendeförändringstekniker användes. Data samlades in genom video-inspelning vid ett initialt besök med patienter som sökte vård för muskuloske-letal smärta. Filmerna analyserades både kvantitativt och kvalitativt. Resulta-tet visade att en majoritet av fysioterapeuterna undersökte biomedicinska och psykologiska riskfaktorer, använde flera beteendeförändringstekniker och analyserade patientens smärtproblem utifrån biomedicinska undersöknings-fynd. Ingen av fysioterapeuterna integrerade psykosociala- och beteendefak-torer i sin analys.

I studie Ⅳ undersöktes samtidig validitet av indelning till tre subgrupper med olika svårighetsgrad av smärta baserat på Örebro Pain Screening Quest-ionnaire(ÖMPSQ), vilka jämfördes med referensinstrumenten Pain Disability Index(PDI), Tampa Scale for Kinesiophobia(TSK) och Pain Catastrophizing Scale(PCS). Stabilitet över tid av subgruppsindelningen undersöktes också vid två tillfällen med två till tre veckor mellan mätningarna.

Konklusionerna i denna avhandling är: 1) Man bör tidigt uppmärksamma individer med kronisk smärta, kvinnligt kön och hög ålder för att kunna er-bjuda en riktad behandling och minska vårdkonsumtionen på lång sikt; 2) Fy-sioterapeuter bör systematiskt undersöka riskfaktorer för sämre prognos. ÖMPSQ kan rekommenderas för indelning till subgrupper med olika svårig-hetsgrad av smärttillstånd, men behöver kompletteras med andra undersök-ningar utifrån patientens behov; 3) Smärtbehandling bör vara riktad mot pro-gnostiska subgrupper, men även utifrån individuell funktionell beteendeana-lys av nyckelbeteenden och undersökningsfynd för att kunna optimera be-handlingseffekterna. Strategier för bibehållande och återfallsprevention bör vara inkluderade i behandlingen; 4) Metoder för att integrera beteendemedicin i fysioterapi behöver utvärderas och förbättras. Förslag på åtgärder är imple-mentering i grundutbildning och vidareutbildning för fysioterapeuter samt kli-nisk träning och handledning.

65

Acknowledgements

Jag vill framföra ett varmt tack till alla personer som varit betydelsefulla för mig under min forskautbildning och i arbetet med denna avhandling. Jag vill särskilt tacka: Alla patienter och sjukgymnaster som deltagit i studierna. Min huvudhandledare Pernilla Åsenlöf, för att du gav mig möjligheten att börja och för att du på ett uppmuntrande, generöst och tålmodigt sätt har lotsat mig igenom hela forskarutbildningen. Jag är så glad över att fått tagit del av din vetenskapliga kompetens, klokhet och outtröttliga engagemang. Min handledare Ingrid Demmelmaier, för att du har varit så tydlig och struk-turerad i din handledning. Jag uppskattar också att du är positiv, noggrann, tillgänglig för frågor och funderingar och att du kan konsten att kombinera vetenskaplig stringens med en skopa humor. Min handledare Stefan Bergman, för att du på ett generöst sätt har bidragit med din kunskap och låtit mig få ta del av ditt forskningsprojekt. Tack också för att du på ett uppmuntrande och positivt sätt har kunnat lyfta och diskutera det som varit viktigt i vår forskning. Övriga medförfattare till delarbetena: Per Lindberg, Eva Denison, Cathrin Martin, Maria Sandborgh och Susanne Pettersson. Tack för ett gott samarbete och för att ni på olika sätt har delat med er av er kunskap och bidragit med värdefulla synpunkter. Doktorandgruppen, Institutionen för neurovetenskap, Fysioterapi, för intres-santa och givande diskussioner och trevlig samvaro både på och utanför job-bet: Susanna Tuvemo Johnson, Ylva Åkerblom, Regina Bendrik, Ann Essner, Hanna Ljungberg och Hedvig Zetterberg. Tack även till f.d. doktorander: Søren Spörndly-Nees, Carina Hagman, Åsa Revenäs, Henrik Johansson, Bir-git Wahlberg, Sara Holm, Mikael Andersson, Helena Igelström, Charlotte Urell och Annika Bring.

66

Special Interest Group (SIG) i beteendemedicin under ledning av Pernilla Åsenlöf: Per Lindberg, Ingrid Demmelmaier, Annika Bring, Sara Holm, He-lena Igelström, Åsa Revenäs, Søren Spörndly-Nees, Lena Zetterberg, Ylva Åkerblom, Hanna Ljungberg, Hedvig Zetterberg och Sofia Wagner. Mina rumskompisar under doktorandtiden: Susanna Tuvemo Johnson och Ylva Åkerblom, tidigare även Birgit Wahlberg och Åsa Revenäs. Vi har delat stort och smått, glädjeämnen och sorger, vilket har varit ovärderligt. Alla fantastiska och generösa kollegor på fysioterapeutprogrammet, Uppsala universitet. Ni alla bidrar verkligen till ett positivt arbetsklimat! Institutionen för neurovetenskap, Uppsala universitet, under ledning av tidi-gare biträdande prefekt Cathrin Martin och studierektor Lena Zetterberg, för att jag har fått möjligheten att genomföra min forskarutbildning i en stimule-rande miljö. Tack också till seniora forskare för uppmuntran och kloka syn-punkter längs vägen. VO Paramedicin, Akademiska sjukhuset, under ledning av Marianne Chris-tensen, som möjliggjort för mig att vara tjänstledig för forskarstudier. Min mamma Carin Jidenius och min bror Stefan Jidenius med familj, Elin, Isak och Anna. Ni representerar en stabil grund att stå på och påminner mig om vem jag är och var jag kommer ifrån. Axel, min man och livskamrat i nöd och lust sedan 1996, tack för du finns och för att du alltid har trott på mig. Våra barn, Louise och Sofie, för att ni är fan-tastiska människor och ger mig ett annat perspektiv på tillvaron. Det är ni tre som är de viktigaste och bästa i mitt liv! Denna avhandling har delfinansierats med anslag från: Reumatikerförbundet Initiativ smärta, FoU-medel landstinget i Uppsala, Vårdforskningsmedel Uppsala universitet, AFA försäkring, ALF-medel Aka-demiska sjukhuset, Svenskt reumaforum SveReFo och Minnesfonden Fysio-terapeuterna.

67

References

1. Merskey H, Bogduk N, International Association for the Study of Pain. Task Force on T. Classification of chronic pain: descriptions of chronic pain syndromes and definitions of pain terms. Vol 2. Seattle: IASP Press; 1994.

2. Treede R-D, Rief W, Barke A, et al. A classification of chronic pain for ICD-11. Pain. 2015;156(6):1003-1007.

3. World Health Organization. ICD-11 Beta Draft website. (cited, available: http://apps.who.int/classifications/icd11/browse/l-m/en. Accessed July 2016).

4. Kovacs FM, Abraira V, Zamora J, Fernandez C. The transition from acute to subacute and chronic low back pain: a study based on determinants of quality of life and prediction of chronic disability. Spine. 2005;30(15):1786-1792.

5. Mantel KE, Peterson CK, Humphreys BK. Exploring the Definition of Acute Low Back Pain: A Prospective Observational Cohort Study Comparing Outcomes of Chiropractic Patients With 0-2, 2-4, and 4-12 Weeks of Symptoms. Journal of manipulative and physiological therapeutics. 2016;39(3):141-149.

6. Stanton TR, Latimer J, Maher CG, Hancock M. Definitions of recurrence of an episode of low back pain: a systematic review. Spine. 2009;34(9):E316-322.

7. Wolfe F, Smythe HA, Yunus MB, et al. The American College of Rheumatology 1990 Criteria for the Classification of Fibromyalgia. Report of the Multicenter Criteria Committee. Arthritis and rheumatism. 1990;33.

8. Von Korff M, Miglioretti DL. A prognostic approach to defining chronic pain. Pain. 2005;117(3):304-313.

9. Dunn KM, Croft PR, Main CJ, Von Korff M. A prognostic approach to defining chronic pain: replication in a UK primary care low back pain population. Pain. 2008;135(1-2):48-54.

10. Bergman S, Herrstrom P, Hogstrom K, Petersson IF, Svensson B, Jacobsson LT. Chronic musculoskeletal pain, prevalence rates, and sociodemographic associations in a Swedish population study. The Journal of rheumatology. 2001;28(6):1369-1377.

11. Hagen K, Linde M, Heuch I, Stovner LJ, Zwart JA. Increasing prevalence of chronic musculoskeletal complaints. A large 11-year

68

follow-up in the general population (HUNT 2 and 3). Pain medicine (Malden, Mass). 2011;12(11):1657-1666.

12. Breivik H, Collett B, Ventafridda V, Cohen R, Gallacher D. Survey of chronic pain in Europe: prevalence, impact on daily life, and treatment. European journal of pain. 2006;10(4):287-333.

13. Kinge JM, Knudsen AK, Skirbekk V, Vollset SE. Musculoskeletal disorders in Norway: prevalence of chronicity and use of primary and specialist health care services. BMC musculoskeletal disorders. 2015;16:75.

14. Hauser W, Wolfe F, Henningsen P, Schmutzer G, Brahler E, Hinz A. Untying chronic pain: prevalence and societal burden of chronic pain stages in the general population - a cross-sectional survey. BMC public health. 2014;14:352.

15. Hoy D, Bain C, Williams G, et al. A systematic review of the global prevalence of low back pain. Arthritis and rheumatism. 2012;64(6):2028-2037.

16. Gerdle B, Bjork J, Henriksson C, Bengtsson A. Prevalence of current and chronic pain and their influences upon work and healthcare-seeking: a population study. The Journal of rheumatology. 2004;31(7):1399-1406.

17. Gerdle B, Bjork J, Coster L, Henriksson K, Henriksson C, Bengtsson A. Prevalence of widespread pain and associations with work status: a population study. BMC musculoskeletal disorders. 2008;9:102.

18. Kennedy J, Roll JM, Schraudner T, Murphy S, McPherson S. Prevalence of persistent pain in the U.S. adult population: new data from the 2010 national health interview survey. The journal of pain : official journal of the American Pain Society. 2014;15(10):979-984.

19. van Hecke O, Torrance N, Smith BH. Chronic pain epidemiology and its clinical relevance. British journal of anaesthesia. 2013;111(1):13-18.

20. Manchikanti L, Singh V, Falco FJ, Benyamin RM, Hirsch JA. Epidemiology of low back pain in adults. Neuromodulation : journal of the International Neuromodulation Society. 2014;17 Suppl 2:3-10.

21. Nordstoga AL, Nilsen TIL, Vasseljen O, Unsgaard-Tondel M, Mork PJ. The influence of multisite pain and psychological comorbidity on prognosis of chronic low back pain: longitudinal data from the Norwegian HUNT Study. BMJ open. 2017;7(5):e015312.

22. van Tulder M, Becker A, Bekkering T, et al. European guidelines for the management of acute nonspecific low back pain in primary care. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society. 2006;15 Suppl 2:S169-191.

23. Koes BW, van Tulder M, Lin CW, Macedo LG, McAuley J, Maher C. An updated overview of clinical guidelines for the management of

69

non-specific low back pain in primary care. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society. 2010;19(12):2075-2094.

24. Costa Lda C, Maher CG, McAuley JH, et al. Prognosis for patients with chronic low back pain: inception cohort study. Bmj. 2009;339:b3829.

25. Stanton TR, Henschke N, Maher CG, Refshauge KM, Latimer J, McAuley JH. After an episode of acute low back pain, recurrence is unpredictable and not as common as previously thought. Spine. 2008;33(26):2923-2928.

26. Henschke N, Maher CG, Refshauge KM, et al. Prognosis in patients with recent onset low back pain in Australian primary care: inception cohort study. Bmj. 2008;337:a171.

27. Marchand S. The physiology of pain mechanisms: from the periphery to the brain. Rheumatic diseases clinics of North America. 2008;34(2):285-309.

28. Fenton BW, Shih E, Zolton J. The neurobiology of pain perception in normal and persistent pain. Pain Manag. 2015;5(4):297-317.

29. Sanzarello I, Merlini L, Rosa MA, et al. Central sensitization in chronic low back pain: A narrative review. Journal of back and musculoskeletal rehabilitation. 2016;29(4):625-633.

30. Woolf CJ. Central sensitization: implications for the diagnosis and treatment of pain. Pain. 2011;152(3 Suppl):S2-15.

31. Zale EL, Lange KL, Fields SA, Ditre JW. The relation between pain-related fear and disability: a meta-analysis. The journal of pain : official journal of the American Pain Society. 2013;14(10):1019-1030.

32. Linton SJ, Shaw WS. Impact of psychological factors in the experience of pain. Physical therapy. 2011;91(5):700-711.

33. Pincus T, McCracken LM. Psychological factors and treatment opportunities in low back pain. Best practice & research Clinical rheumatology. 2013;27(5):625-635.

34. Sullivan MJ, Thorn B, Haythornthwaite JA, et al. Theoretical perspectives on the relation between catastrophizing and pain. The Clinical journal of pain. 2001;17(1):52-64.

35. Keefe FJ, Rumble ME, Scipio CD, Giordano LA, Perri LM. Psychological aspects of persistent pain: current state of the science. The journal of pain : official journal of the American Pain Society. 2004;5(4):195-211.

36. Pincus T, Burton AK, Vogel S, Field AP. A systematic review of psychological factors as predictors of chronicity/disability in prospective cohorts of low back pain. Spine. 2002;27(5):E109-120.

37. Wertli MM, Eugster R, Held U, Steurer J, Kofmehl R, Weiser S. Catastrophizing-a prognostic factor for outcome in patients with low

70

back pain: a systematic review. The spine journal : official journal of the North American Spine Society. 2014;14(11):2639-2657.

38. Wideman TH, Sullivan MJ. Differential predictors of the long-term levels of pain intensity, work disability, healthcare use, and medication use in a sample of workers' compensation claimants. Pain. 2011;152(2):376-383.

39. Fordyce W. Behavioral methods for chronic pain and illness. St Louis: CV Mosby Co; 1976.

40. Vlaeyen JW, Linton SJ. Fear-avoidance and its consequences in chronic musculoskeletal pain: a state of the art. Pain. 2000;85(3):317-332.

41. Leeuw M, Goossens ME, Linton SJ, Crombez G, Boersma K, Vlaeyen JW. The fear-avoidance model of musculoskeletal pain: current state of scientific evidence. Journal of behavioral medicine. 2007;30(1):77-94.

42. Vlaeyen JW, Linton SJ. Fear-avoidance model of chronic musculoskeletal pain: 12 years on. Pain. 2012.

43. Gheldof EL, Crombez G, Van den Bussche E, et al. Pain-related fear predicts disability, but not pain severity: a path analytic approach of the fear-avoidance model. European journal of pain. 2010;14(8):870.e871-879.

44. Grotle M, Foster NE, Dunn KM, Croft P. Are prognostic indicators for poor outcome different for acute and chronic low back pain consulters in primary care? Pain. 2010;151(3):790-797.

45. Valentin GH, Pilegaard MS, Vaegter HB, et al. Prognostic factors for disability and sick leave in patients with subacute non-malignant pain: a systematic review of cohort studies. BMJ open. 2016;6(1):e007616.

46. Campbell P, Foster NE, Thomas E, Dunn KM. Prognostic Indicators of Low Back Pain in Primary Care: Five-Year Prospective Study. The Journal of Pain. 2013;14(8):873-883.

47. Larsson B, Bjork J, Borsbo B, Gerdle B. A systematic review of risk factors associated with transitioning from regional musculoskeletal pain to chronic widespread pain. European journal of pain. 2012;16(8):1084-1093.

48. Downie A, Williams CM, Henschke N, et al. Red flags to screen for malignancy and fracture in patients with low back pain: systematic review. Bmj. 2013;347:f7095.

49. Nicholas MK, Linton SJ, Watson PJ, Main CJ, Decade of the Flags" Working G. Early identification and management of psychological risk factors ("yellow flags") in patients with low back pain: a reappraisal. Physical therapy. 2011;91(5):737-753.

50. Bergman S. Psychosocial aspects of chronic widespread pain and fibromyalgia. Disability and rehabilitation. 2005;27(12):675-683.

71

51. World Health Organization, International Classification of Diseases website. (cited, available: http://www.who.int/classifications/icd/en/. Accessed July 2016.).

52. Picavet HS, Struijs JN, Westert GP. Utilization of health resources due to low back pain: survey and registered data compared. Spine. 2008;33(4):436-444.

53. Ferreira ML, Machado G, Latimer J, Maher C, Ferreira PH, Smeets RJ. Factors defining care-seeking in low back pain--a meta-analysis of population based surveys. European journal of pain (London, England). 2010;14(7):747 e741-747.

54. Mannion AF, Wieser S, Elfering A. Association between beliefs and care-seeking behavior for low back pain. Spine. 2013;38(12):1016-1025.

55. Gore M, Sadosky A, Stacey BR, Tai KS, Leslie D. The burden of chronic low back pain: clinical comorbidities, treatment patterns, and health care costs in usual care settings. Spine. 2012;37(11):E668-677.

56. Sanchez RJ, Uribe C, Li H, et al. Longitudinal evaluation of health care utilization and costs during the first three years after a new diagnosis of fibromyalgia. Current medical research and opinion. 2011;27(3):663-671.

57. Stewart WF, Yan X, Boscarino JA, et al. Patterns of health care utilization for low back pain. Journal of pain research. 2015;8:523-535.

58. Henly SJ, Wyman JF, Gaugler JE. Health trajectory research: a call to action for nursing science. Nursing research. 2011;60(3 Suppl):S79-82.

59. Kongsted A, Kent P, Axen I, Downie AS, Dunn KM. What have we learned from ten years of trajectory research in low back pain? BMC musculoskeletal disorders. 2016;17(1):1-11.

60. Hartvigsen J, Davidsen M, Sogaard K, Roos EM, Hestbaek L. Self-reported musculoskeletal pain predicts long-term increase in general health care use: a population-based cohort study with 20-year follow-up. Scandinavian journal of public health. 2014;42(7):698-704.

61. Hansson T, Jensen I. Swedish Council on Technology Assessment in Health Care (SBU). Chapter 6. Sickness absence due to back and neck disorders. Scandinavian journal of public health Supplement. 2004;63:109-151.

62. Inspektionen för socialförsäkringen. Årlig rapport 2016: en rapport från Inspektionen för socialförsäkringen (Rapport 2017:2). 2017; http://www.inspsf.se/digitalAssets/6/6074_3981813_isf-2017-2_web.pdf. Accessed August 31, 2017.

63. Flor H, Turk DC. Chronic pain: an integrated biobehavioral approach. Lippincott Williams & Wilkins; 2015.

64. Bergman S, Jacobsson LT, Herrstrom P, Petersson IF. Health status as measured by SF-36 reflects changes and predicts outcome in

72

chronic musculoskeletal pain: a 3-year follow up study in the general population. Pain. 2004;108(1-2):115-123.

65. Jonsdottir T, Aspelund T, Jonsdottir H, Gunnarsdottir S. The Relationship between Chronic Pain Pattern, Interference with Life and Health-Related Quality of Life in a Nationwide Community Sample. Pain Management Nursing. 2014;15(3):641-651.

66. Engel GL. The need for a new medical model: a challenge for biomedicine. Science. 1977;196(4286):129-136.

67. Gatchel RJ, Peng YB, Peters ML, Fuchs PN, Turk DC. The biopsychosocial approach to chronic pain: scientific advances and future directions. Psychological bulletin. 2007;133(4):581-624.

68. Alvarez AS, Pagani M, Meucci P. The clinical application of the biopsychosocial model in mental health: a research critique. American journal of physical medicine & rehabilitation / Association of Academic Physiatrists. 2012;91(13 Suppl 1):S173-180.

69. McLaren N. A critical review of the biopsychosocial model. The Australian and New Zealand journal of psychiatry. 1998;32(1):86-92; discussion 93-86.

70. Michie SW, Robert; Campbell, Rona; Brown Jamie and Gainforth Heather ABC of behaviour change theories. Vol First edition. Great Britain: SIlverback Publishing; 2014.

71. Vlaeyen JW, de Jong J, Geilen M, Heuts PH, van Breukelen G. Graded exposure in vivo in the treatment of pain-related fear: a replicated single-case experimental design in four patients with chronic low back pain. Behaviour research and therapy. 2001;39(2):151-166.

72. Skinner BF. Science and human behavior. Journal of consulting psychology. 1953;17(3):233-233.

73. Fordyce WE. Behavioral methods for chronic pain and illness. Saint Louis, Mo: Mosby; 1976.

74. Bandura A. Health promotion by social cognitive means. Health education & behavior : the official publication of the Society for Public Health Education. 2004;31(2):143-164.

75. Bandura A. Social cognitive theory: an agentic perspective. Annual review of psychology. 2001;52:1-26.

76. Keogh A, Tully MA, Matthews J, Hurley DA. A review of behaviour change theories and techniques used in group based self-management programmes for chronic low back pain and arthritis. Manual therapy.

77. Bandura A. Social foundations of thought and action: a social cognitive theory. Englewood Cliffs, N.J: Prentice-Hall; 1986.

78. Bauman AE, Reis RS, Sallis JF, Wells JC, Loos RJ, Martin BW. Correlates of physical activity: why are some people physically active and others not? Lancet. 2012;380(9838):258-271.

79. Denison E, Asenlof P, Lindberg P. Self-efficacy, fear avoidance, and pain intensity as predictors of disability in subacute and chronic

73

musculoskeletal pain patients in primary health care. Pain. 2004;111(3):245-252.

80. Costa Lda C, Maher CG, McAuley JH, Hancock MJ, Smeets RJ. Self-efficacy is more important than fear of movement in mediating the relationship between pain and disability in chronic low back pain. European journal of pain. 2011;15(2):213-219.

81. van Middelkoop M, Rubinstein SM, Kuijpers T, et al. A systematic review on the effectiveness of physical and rehabilitation interventions for chronic non-specific low back pain. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society. 2011;20(1):19-39.

82. Turk DC, Adams LM. Using a biopsychosocial perspective in the treatment of fibromyalgia patients. Pain Manag. 2016;6(4):357-369.

83. Liddle SD, David Baxter G, Gracey JH. Physiotherapists' use of advice and exercise for the management of chronic low back pain: a national survey. Manual therapy. 2009;14(2):189-196.

84. Dahm KT, Brurberg KG, Jamtvedt G, Hagen KB. Advice to rest in bed versus advice to stay active for acute low-back pain and sciatica. Cochrane database of systematic reviews. 2010(6):CD007612.

85. Frost H, Lamb SE, Doll HA, Carver PT, Stewart-Brown S. Randomised controlled trial of physiotherapy compared with advice for low back pain. Bmj. 2004;329(7468):708.

86. Sizer PS, Jr., Brismee JM, Cook C. Medical screening for red flags in the diagnosis and management of musculoskeletal spine pain. Pain practice : the official journal of World Institute of Pain. 2007;7(1):53-71.

87. Scascighini L, Toma V, Dober-Spielmann S, Sprott H. Multidisciplinary treatment for chronic pain: a systematic review of interventions and outcomes. Rheumatology. 2008;47(5):670-678.

88. Kamper SJ, Apeldoorn AT, Chiarotto A, et al. Multidisciplinary biopsychosocial rehabilitation for chronic low back pain: Cochrane systematic review and meta-analysis. Bmj. 2015;350:h444.

89. Bileviciute-Ljungar I, Norrefalk JR. Beneficial long-term effects of multiprofessional assessment vs. rehabilitation program in patients with musculoskeletal pain. Pain practice : the official journal of World Institute of Pain. 2014;14(3):228-236.

90. Westman A, Linton SJ, Theorell T, Ohrvik J, Wahlen P, Leppert J. Quality of life and maintenance of improvements after early multimodal rehabilitation: a 5-year follow-up. Disability and rehabilitation. 2006;28(7):437-446.

91. Norrefalk J-R, Linder J, Ekholm J, Borg K. A 6-year follow-up study of 122 patients attending a multiprofessional rehabilitation programme for persistent musculoskeletal-related pain. International Journal of Rehabilitation Research. 2007;30(1):9-18.

74

92. Svanberg M, Stalnacke BM, Enthoven P, Brodda-Jansen G, Gerdle B, Boersma K. Impact of emotional distress and pain-related fear on patients with chronic pain: Subgroup analysis of patients referred to multimodal rehabilitation. Journal of rehabilitation medicine : official journal of the UEMS European Board of Physical and Rehabilitation Medicine. 2017;49(4):354-361.

93. Engel GL. A unified concept of health and disease. Perspectives in biology and medicine. 1960;3:459-485.

94. Engel GL. The clinical application of the biopsychosocial model. The American journal of psychiatry. 1980;137(5):535-544.

95. Haynes SN, Leisen MB, Blaine DD. Design of individualized behavioral treatment programs using functional analytic clinical case models. Psychological Assessment. 1997;9(4):334-348.

96. Nicholas MK, George SZ. Psychologically informed interventions for low back pain: an update for physical therapists. Physical therapy. 2011;91(5):765-776.

97. Booth J, Moseley GL, Schiltenwolf M, Cashin A, Davies M, Hubscher M. Exercise for chronic musculoskeletal pain: A biopsychosocial approach. Musculoskeletal care. 2017.

98. Caspersen CJ, Powell KE, Christenson GM. Physical activity, exercise, and physical fitness: definitions and distinctions for health-related research. Public health reports (Washington, DC : 1974). 1985;100(2):126-131.

99. Geneen LJ, Moore RA, Clarke C, Martin D, Colvin LA, Smith BH. Physical activity and exercise for chronic pain in adults: an overview of Cochrane Reviews. Cochrane database of systematic reviews. 2017;1:Cd011279.

100. van Middelkoop M, Rubinstein SM, Verhagen AP, Ostelo RW, Koes BW, van Tulder MW. Exercise therapy for chronic nonspecific low-back pain. Best practice & research Clinical rheumatology. 2010;24(2):193-204.

101. Macedo LG, Saragiotto BT, Yamato TP, et al. Motor control exercise for acute non-specific low back pain. Cochrane database of systematic reviews. 2016;2:Cd012085.

102. Henschke N, Ostelo RW, van Tulder MW, et al. Behavioural treatment for chronic low-back pain. Cochrane database of systematic reviews. 2010(7):CD002014.

103. Fordyce WE, Fowler RS, Jr., Lehmann JF, Delateur BJ, Sand PL, Trieschmann RB. Operant conditioning in the treatment of chronic pain. Archives of physical medicine and rehabilitation. 1973;54(9):399-408.

104. Åsenlöf P, Denison E, Lindberg P. Individually Tailored Treatment Targeting Activity, Motor Behavior, and Cognition Reduces Pain–Related Disability: A Randomized Controlled Trial in Patients With Musculoskeletal Pain. The Journal of Pain. 2005;6(9):588-603.

75

105. Pisters MF, Veenhof C, de Bakker DH, Schellevis FG, Dekker J. Behavioural graded activity results in better exercise adherence and more physical activity than usual care in people with osteoarthritis: a cluster-randomised trial. Journal of physiotherapy. 2010;56(1):41-47.

106. Smeets RJ, Vlaeyen JW, Hidding A, Kester AD, van der Heijden GJ, Knottnerus JA. Chronic low back pain: physical training, graded activity with problem solving training, or both? The one-year post-treatment results of a randomized controlled trial. Pain. 2008;134(3):263-276.

107. Macedo LG, Latimer J, Maher CG, et al. Effect of motor control exercises versus graded activity in patients with chronic nonspecific low back pain: a randomized controlled trial. Physical therapy. 2012;92(3):363-377.

108. George SZ, Zeppieri G. Physical therapy utilization of graded exposure for patients with low back pain. The Journal of orthopaedic and sports physical therapy. 2009;39(7):496-505.

109. Leeuw M, Goossens ME, van Breukelen GJ, et al. Exposure in vivo versus operant graded activity in chronic low back pain patients: results of a randomized controlled trial. Pain. 2008;138(1):192-207.

110. Macedo LG, Smeets RJ, Maher CG, Latimer J, McAuley JH. Graded activity and graded exposure for persistent nonspecific low back pain: a systematic review. Physical therapy. 2010;90(6):860-879.

111. George SZ, Wittmer VT, Fillingim RB, Robinson ME. Comparison of graded exercise and graded exposure clinical outcomes for patients with chronic low back pain. The Journal of orthopaedic and sports physical therapy. 2010;40(11):694-704.

112. George SZ, Zeppieri G, Jr., Cere AL, et al. A randomized trial of behavioral physical therapy interventions for acute and sub-acute low back pain (NCT00373867). Pain. 2008;140(1):145-157.

113. Lopez-de-Uralde-Villanueva I, Munoz-Garcia D, Gil-Martinez A, et al. A Systematic Review and Meta-Analysis on the Effectiveness of Graded Activity and Graded Exposure for Chronic Nonspecific Low Back Pain. Pain medicine (Malden, Mass). 2016;17(1):172-188.

114. International Society of Behavioral Medicine. About ISBM. https://www.isbm.info/about-isbm/. Accessed August 31, 2017.

115. Main CJ, George SZ. Psychologically informed practice for management of low back pain: future directions in practice and research. Physical therapy. 2011;91(5):820-824.

116. Foster NE, Delitto A. Embedding psychosocial perspectives within clinical management of low back pain: integration of psychosocially informed management principles into physical therapist practice--challenges and opportunities. Physical therapy. 2011;91(5):790-803.

117. Bryant C, Lewis P, Bennell KL, et al. Can Physical Therapists Deliver a Pain Coping Skills Program? An Examination of Training Processes and Outcomes. Physical therapy. 2014.

76

118. Nielsen M, Keefe FJ, Bennell K, Jull GA. Physical therapist-delivered cognitive-behavioral therapy: a qualitative study of physical therapists' perceptions and experiences. Physical therapy. 2014;94(2):197-209.

119. Main CJ, Sowden G, Hill JC, Watson PJ, Hay EM. Integrating physical and psychological approaches to treatment in low back pain: the development and content of the STarT Back trial's 'high-risk' intervention (StarT Back; ISRCTN 37113406). Physiotherapy. 2012;98(2):110-116.

120. van der Wees PJ, Jamtvedt G, Rebbeck T, de Bie RA, Dekker J, Hendriks EJ. Multifaceted strategies may increase implementation of physiotherapy clinical guidelines: a systematic review. The Australian journal of physiotherapy. 2008;54(4):233-241.

121. Overmeer T, Boersma K, Denison E, Linton SJ. Does teaching physical therapists to deliver a biopsychosocial treatment program result in better patient outcomes? A randomized controlled trial. Physical therapy. 2011;91(5):804-819.

122. Asenlof P, Denison E, Lindberg P. Individually tailored treatment targeting motor behavior, cognition, and disability: 2 experimental single-case studies of patients with recurrent and persistent musculoskeletal pain in primary health care. Physical therapy. 2005;85(10):1061-1077.

123. Åsenlöf P, Denison E, Lindberg P. Long-term follow-up of tailored behavioural treatment and exercise based physical therapy in persistent musculoskeletal pain: A randomized controlled trial in primary care. European journal of pain. 2009;13(10):1080-1088.

124. Overmeer T, Peterson G, Landen Ludvigsson M, Peolsson A. The effect of neck-specific exercise with or without a behavioral approach on psychological factors in chronic whiplash-associated disorders: A randomized controlled trial with a 2-year follow-up. Medicine. 2016;95(34):e4430.

125. Bennell KL, Ahamed Y, Jull G, et al. Physical Therapist-Delivered Pain Coping Skills Training and Exercise for Knee Osteoarthritis: Randomized Controlled Trial. Arthritis care & research. 2016;68(5):590-602.

126. Michie S, Fixsen D, Grimshaw JM, Eccles MP. Specifying and reporting complex behaviour change interventions: the need for a scientific method. Implementation science : IS. 2009;4:40.

127. Achterberg T, Huisman-de Waal GG, Ketelaar NA, Oostendorp RA, Jacobs JE, Wollersheim HC. How to promote healthy behaviours in patients? An overview of evidence for behaviour change techniques. Health promotion international. 2011;26.

128. Michie S, Johnston M, Francis J, Hardeman W, Eccles M. From Theory to Intervention: Mapping Theoretically Derived Behavioural

77

Determinants to Behaviour Change Techniques. Applied Psychology: An International Review. 2008;57(4):660-680.

129. Michie S, Richardson M, Johnston M, et al. The behavior change technique taxonomy (v1) of 93 hierarchically clustered techniques: building an international consensus for the reporting of behavior change interventions. Annals of behavioral medicine : a publication of the Society of Behavioral Medicine. 2013;46(1):81-95.

130. van Achterberg T, Huisman-de Waal GG, Ketelaar NA, Oostendorp RA, Jacobs JE, Wollersheim HC. How to promote healthy behaviours in patients? An overview of evidence for behaviour change techniques. Health promotion international. 2011;26(2):148-162.

131. Fjeldsoe B, Neuhaus M, Winkler E, Eakin E. Systematic review of maintenance of behavior change following physical activity and dietary interventions. Health psychology : official journal of the Division of Health Psychology, American Psychological Association. 2011;30(1):99-109.

132. Greaves CJ, Sheppard KE, Abraham C, Hardeman W, Roden M, Evans PH. Systematic review of reviews of intervention components associated with increased effectiveness in dietary and physical activity interventions. BMC public health. 2011;11.

133. Samdal GB, Eide GE, Barth T, Williams G, Meland E. Effective behaviour change techniques for physical activity and healthy eating in overweight and obese adults; systematic review and meta-regression analyses. The international journal of behavioral nutrition and physical activity. 2017;14(1):42.

134. McDermott MS, Oliver M, Iverson D, Sharma R. Effective techniques for changing physical activity and healthy eating intentions and behaviour: A systematic review and meta-analysis. British journal of health psychology. 2016;21(4):827-841.

135. Dombrowski SU, Sniehotta FF, Avenell A, Johnston M, MacLennan G, Araújo-Soares V. Identifying active ingredients in complex behavioural interventions for obese adults with obesity-related co-morbidities or additional risk factors for co-morbidities: a systematic review. Health psychology review. 2012;6(1):7-32.

136. Olander EK, Fletcher H, Williams S, Atkinson L, Turner A, French DP. What are the most effective techniques in changing obese individuals' physical activity self-efficacy and behaviour: a systematic review and meta-analysis. The international journal of behavioral nutrition and physical activity. 2013;10:29.

137. Williams SL, French DP. What are the most effective intervention techniques for changing physical activity self-efficacy and physical activity behaviour--and are they the same? Health education research. 2011;26(2):308-322.

138. Dusseldorp E, van Genugten L, van Buuren S, Verheijden MW, van Empelen P. Combinations of techniques that effectively change health

78

behavior: evidence from Meta-CART analysis. Health psychology : official journal of the Division of Health Psychology, American Psychological Association. 2014;33(12):1530-1540.

139. Zangoni G, Thomson OP. 'I need to do another course' - Italian physiotherapists' knowledge and beliefs when assessing psychosocial factors in patients presenting with chronic low back pain. Musculoskeletal science & practice. 2017;27:71-77.

140. Sanders T, Foster NE, Bishop A, Ong BN. Biopsychosocial care and the physiotherapy encounter: physiotherapists' accounts of back pain consultations. BMC musculoskeletal disorders. 2013;14:65.

141. Sandborgh M, Åsenlöf P, Lindberg P, Denison E. Implementing behavioural medicine in physiotherapy treatment. Part II: Adherence to treatment protocol. Advances in Physiotherapy. 2010;12(1):13-23.

142. Green AJ, Jackson DA, Klaber Moffett JA. An observational study of physiotherapists’ use of cognitive-behavioural principles in the management of patients with back pain and neck pain. Physiotherapy. 2008;94(4):306-313.

143. Synnott A, O'Keeffe M, Bunzli S, Dankaerts W, O'Sullivan P, O'Sullivan K. Physiotherapists may stigmatise or feel unprepared to treat people with low back pain and psychosocial factors that influence recovery: a systematic review. Journal of physiotherapy. 2015;61(2):68-76.

144. Overmeer T, Boersma K, Main CJ, Linton SJ. Do physical therapists change their beliefs, attitudes, knowledge, skills and behaviour after a biopsychosocially orientated university course? Journal of evaluation in clinical practice. 2009;15(4):724-732.

145. Ferguson FC, Morison S, Ryan CG. Physiotherapists' Understanding of Red Flags for Back Pain. Musculoskeletal care. 2014.

146. Demmelmaier I, Denison E, Lindberg P, Asenlof P. Physiotherapists' telephone consultations regarding back pain: a method to analyze screening of risk factors. Physiotherapy theory and practice. 2010;26(7):468-475.

147. Alexanders J, Anderson A, Henderson S. Musculoskeletal physiotherapists’ use of psychological interventions: a systematic review of therapists’ perceptions and practice. Physiotherapy. 2015;101(2):95-102.

148. Singla M, Jones M, Edwards I, Kumar S. Physiotherapists' assessment of patients' psychosocial status: are we standing on thin ice? A qualitative descriptive study. Manual therapy. 2015;20(2):328-334.

149. Boersma K, Linton SJ. Screening to identify patients at risk: profiles of psychological risk factors for early intervention. The Clinical journal of pain. 2005;21(1):38-43; discussion 69-72.

150. Foster NE, Hill JC, O'Sullivan P, Hancock M. Stratified models of care. Best practice & research Clinical rheumatology. 2013;27(5):649-661.

79

151. Bergbom S, Boersma K, Linton SJ. When Matching Fails: Understanding the Process of Matching Pain-Disability Treatment to Risk Profile. Journal of occupational rehabilitation. 2015;25(3):518-526.

152. Turk DC. The potential of treatment matching for subgroups of patients with chronic pain: lumping versus splitting. The Clinical journal of pain. 2005;21(1):44-55; discussion 69-72.

153. Linton SJ, Boersma K. Early identification of patients at risk of developing a persistent back problem: the predictive validity of the Orebro Musculoskeletal Pain Questionnaire. The Clinical journal of pain. 2003;19(2):80-86.

154. Hill JC, Dunn KM, Main CJ, Hay EM. Subgrouping low back pain: a comparison of the STarT Back Tool with the Orebro Musculoskeletal Pain Screening Questionnaire. European journal of pain. 2010;14(1):83-89.

155. Von Korff M, Moore JC. Stepped care for back pain: activating approaches for primary care. Annals of internal medicine. 2001;134(9 Pt 2):911-917.

156. Bower P, Gilbody S. Stepped care in psychological therapies: access, effectiveness and efficiency. Narrative literature review. The British journal of psychiatry : the journal of mental science. 2005;186:11-17.

157. Åsenlof P, Demmelmaier I, Emilson C, Pettersson S, Bergman S. STEP-UP: an innovative stepped-care protocol for tailored behavioral medicine treatment in the management of musculoskeletal pain in primary care. Annals of the rheumatic diseases. 2013;72.

158. Tait RC, Pollard CA, Margolis RB, Duckro PN, Krause SJ. The Pain Disability Index: psychometric and validity data. Archives of physical medicine and rehabilitation. 1987;68(7):438-441.

159. Tait RC, Chibnall JT, Krause S. The Pain Disability Index: psychometric properties. Pain. 1990;40(2):171-182.

160. Soer R, Koke AJ, Vroomen PC, et al. Extensive validation of the pain disability index in 3 groups of patients with musculoskeletal pain. Spine. 2013;38(9):E562-568.

161. McGrath P. Handbook of pain assessment, second edition. Vol 43. OTTAWA: Canadian Psychological Association; 2002:61-61.

162. Ware JE, Jr., Gandek B. Overview of the SF-36 Health Survey and the International Quality of Life Assessment (IQOLA) Project. Journal of clinical epidemiology. 1998;51(11):903-912.

163. Sullivan M, Karlsson J. The Swedish SF-36 Health Survey III. Evaluation of criterion-based validity: results from normative population. Journal of clinical epidemiology. 1998;51(11):1105-1113.

164. Sullivan M, Karlsson J, Ware Jr JE. The Swedish SF-36 Health Survey—I. Evaluation of data quality, scaling assumptions, reliability

80

and construct validity across general populations in Sweden. Social science & medicine. 1995;41(10):1349-1358.

165. Persson L-O, Karlsson J, Bengtsson C, Steen B, Sullivan M. The Swedish SF-36 Health Survey II. Evaluation of Clinical Validity: Results from Population Studies of Elderly and Women in Gothenborg. Journal of clinical epidemiology. 1998;51(11):1095-1103.

166. Kori SH, Miller RP, Todd DD. Kinisophobia: A new view of chronic pain behavior. Pain Management. 1990;3(1):35-43.

167. Lundberg MKE, Styf J, Carlsson SG. A psychometric evaluation of the Tampa Scale for Kinesiophobia — from a physiotherapeutic perspective. Physiotherapy theory and practice. 2004;20(2):121-133.

168. Sullivan MJL, Bishop SR, Pivik J. The Pain Catastrophizing Scale: Development and validation. Psychological Assessment. 1995;7(4):524-532.

169. Scott W, Wideman TH, Sullivan MJ. Clinically meaningful scores on pain catastrophizing before and after multidisciplinary rehabilitation: a prospective study of individuals with subacute pain after whiplash injury. The Clinical journal of pain. 2014;30(3):183-190.

170. Sullivan MJ, Adams H, Rhodenizer T, Stanish WD. A psychosocial risk factor--targeted intervention for the prevention of chronic pain and disability following whiplash injury. Physical therapy. 2006;86(1):8-18.

171. Linton SJ, Hallden K. Can we screen for problematic back pain? A screening questionnaire for predicting outcome in acute and subacute back pain. The Clinical journal of pain. 1998;14(3):209-215.

172. Hockings RL, McAuley JH, Maher CG. A systematic review of the predictive ability of the Orebro Musculoskeletal Pain Questionnaire. Spine. 2008;33(15):E494-500.

173. Westman AE, Boersma K, Leppert J, Linton SJ. Fear-avoidance beliefs, catastrophizing, and distress: a longitudinal subgroup analysis on patients with musculoskeletal pain. The Clinical journal of pain. 2011;27(7):567-577.

174. Nordeman L, Nilsson B, Moller M, Gunnarsson R. Early access to physical therapy treatment for subacute low back pain in primary health care: a prospective randomized clinical trial. The Clinical journal of pain. 2006;22(6):505-511.

175. Hurley DA, Dusoir TE, McDonough SM, Moore AP, Baxter GD. How effective is the acute low back pain screening questionnaire for predicting 1-year follow-up in patients with low back pain? The Clinical journal of pain. 2001;17(3):256-263.

176. Forsbrand M, Grahn B, Hill JC, Petersson IF, Sennehed CP, Stigmar K. Comparison of the Swedish STarT Back Screening Tool and the Short Form of the Orebro Musculoskeletal Pain Screening

81

Questionnaire in patients with acute or subacute back and neck pain. BMC musculoskeletal disorders. 2017;18(1):89.

177. Abraham C, Wood CE, Johnston M, et al. Reliability of Identification of Behavior Change Techniques in Intervention Descriptions. Annals of behavioral medicine : a publication of the Society of Behavioral Medicine. 2015.

178. Abraham C, Wood CE, Johnston M, Francis J, Hardeman W, Richardson M. Reliability of identification of behavior change techniques in intervention descriptions. Annals of behavioral medicine : a publication of the Society of Behavioral Medicine. 2015;49.

179. Jakobsson U, Westergren A. Statistical methods for assessing agreement for ordinal data. Scandinavian journal of caring sciences. 2005;19(4):427-431.

180. http://vassarstats.net/kappa.html. Accessed 0410, 2017. 181. Joud A, Petersson IF, Englund M. Low back pain: epidemiology of

consultations. Arthritis care & research. 2012;64(7):1084-1088. 182. Andersson HI, Ejlertsson G, Leden I, Schersten B. Impact of chronic

pain on health care seeking, self care, and medication. Results from a population-based Swedish study. Journal of epidemiology and community health. 1999;53(8):503-509.

183. Jensen JC, Haahr JP, Frost P, Andersen JH. The significance of health anxiety and somatization in care-seeking for back and upper extremity pain. Family practice. 2012;29(1):86-95.

184. Hicks GE, Gaines JM, Shardell M, Simonsick EM. Associations of back and leg pain with health status and functional capacity of older adults: findings from the retirement community back pain study. Arthritis and rheumatism. 2008;59(9):1306-1313.

185. Bosmans JE, Pool JJ, de Vet HC, van Tulder MW, Ostelo RW. Is behavioral graded activity cost-effective in comparison with manual therapy for patients with subacute neck pain? An economic evaluation alongside a randomized clinical trial. Spine. 2011;36(18):E1179-1186.

186. Magalhaes MO, Muzi LH, Comachio J, et al. The short-term effects of graded activity versus physiotherapy in patients with chronic low back pain: A randomized controlled trial. Manual therapy. 2015;20(4):603-609.

187. George SZ, Fritz JM, Bialosky JE, Donald DA. The effect of a fear-avoidance-based physical therapy intervention for patients with acute low back pain: results of a randomized clinical trial. Spine. 2003;28(23):2551-2560.

188. Friedrich M, Gittler G, Arendasy M, Friedrich KM. Long-Term Effect of a Combined Exercise and Motivational Program on the Level of Disability of Patients With Chronic Low Back Pain. Spine.

82

2005;30(9):995-1000 1010.1097/1001.brs.0000160844.0000171551.af.

189. Carroll LJ, Ferrari R, Cassidy JD, Cote P. Coping and recovery in whiplash-associated disorders: early use of passive coping strategies is associated with slower recovery of neck pain and pain-related disability. The Clinical journal of pain. 2014;30(1):1-8.

190. Hill JC, Dunn KM, Lewis M, et al. A primary care back pain screening tool: identifying patient subgroups for initial treatment. Arthritis and rheumatism. 2008;59(5):632-641.

191. Karran EL, McAuley JH, Traeger AC, et al. Can screening instruments accurately determine poor outcome risk in adults with recent onset low back pain? A systematic review and meta-analysis. BMC Med. 2017;15(1):13.

192. Hill JC, Whitehurst DG, Lewis M, et al. Comparison of stratified primary care management for low back pain with current best practice (STarT Back): a randomised controlled trial. Lancet. 2011;378(9802):1560-1571.

193. Murphy SE, Blake C, Power CK, Fullen BM. Comparison of a Stratified Group Intervention (STarT Back) With Usual Group Care in Patients With Low Back Pain: A Nonrandomized Controlled Trial. Spine. 2016;41(8):645-652.

194. Foster NE, Mullis R, Hill JC, et al. Effect of stratified care for low back pain in family practice (IMPaCT Back): a prospective population-based sequential comparison. Annals of family medicine. 2014;12(2):102-111.

195. Bier JD, Sandee-Geurts JJW, Ostelo R, Koes BW, Verhagen AP. Can primary care for back and/or neck pain in the Netherlands benefit from stratification for risk groups according to the STarT Back Tool-classification? Archives of physical medicine and rehabilitation. 2017.

196. Bergbom S, Flink IK, Boersma K, Linton SJ. Early psychologically informed interventions for workers at risk for pain-related disability: does matching treatment to profile improve outcome? Journal of occupational rehabilitation. 2014;24(3):446-457.

197. Soer R, Koke AJ, Speijer BL, et al. Reference Values of the Pain Disability Index in Patients With Painful Musculoskeletal and Spinal Disorders: A Cross-national Study. Spine. 2015;40(9):E545-551.

198. Prochaska JO, Velicer WF. The transtheoretical model of health behavior change. American journal of health promotion : AJHP. 1997;12(1):38-48.

199. Synnott A, O'Keeffe M, Bunzli S, et al. Physiotherapists report improved understanding of and attitude toward the cognitive, psychological and social dimensions of chronic low back pain after Cognitive Functional Therapy training: a qualitative study. Journal of physiotherapy. 2016;62(4):215-221.

83

200. Kazdin AE. Research design in clinical psychology. Vol 4. uppl. Boston, MA: Allyn and Bacon; 2014.

201. Fewtrell MS, Kennedy K, Singhal A, et al. How much loss to follow-up is acceptable in long-term randomised trials and prospective studies? Archives of disease in childhood. 2008;93(6):458-461.

202. Kristman V, Manno M, Cote P. Loss to follow-up in cohort studies: how much is too much? European journal of epidemiology. 2004;19(8):751-760.

203. Daykin AR, Richardson B. Physiotherapists’ Pain Beliefs and Their Influence on the Management of Patients With Chronic Low Back Pain. Spine. 2004;29(7):783-795.

204. Parry R, Pino M, Faull C, Feathers L. Acceptability and design of video-based research on healthcare communication: Evidence and recommendations. Patient Education and Counseling. 2016.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1360

Editor: The Dean of the Faculty of Medicine

A doctoral dissertation from the Faculty of Medicine, UppsalaUniversity, is usually a summary of a number of papers. A fewcopies of the complete dissertation are kept at major Swedishresearch libraries, while the summary alone is distributedinternationally through the series Digital ComprehensiveSummaries of Uppsala Dissertations from the Faculty ofMedicine. (Prior to January, 2005, the series was publishedunder the title “Comprehensive Summaries of UppsalaDissertations from the Faculty of Medicine”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-328363

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2017