Experience in adverse events detection in an emergency department: Nature of events

doi 101111j1742-6723200600897x Emergency Medicine Australasia (2007) 19 9ndash15

copy 2007 The AuthorsJournal compilation copy 2007 Australasian College for Emergency Medicine and Australasian Society for Emergency Medicine

Blackwell Publishing AsiaMelbourne AustraliaEMMEmergency Medicine Australasia1742-6731copy 2006 The Authors Journal compilation copy 2006 Australasian College for Emergency Medicine and Australasian Society forEmergency Medicine2007191915Original ArticleNature of eventsJ Hendrie

et al

Correspondence Dr James Hendrie Emergency Department Austin Health Heidelberg Vic 3084 Australia Email jameshendrie austinorgau

James Hendrie MBBS FACEM MD Emergency Physician Luke Sammartino MBBS FRACP MD Emergency Registrar Mervyn J SilvapullePhD Statistician George Braitberg MBBS FACEM FACMT Director of Emergency Medicine

ORIGINAL RESEARCH

Experience in adverse events detection in an emergency department Nature of eventsJames Hendrie1 Luke Sammartino1 Mervyn J Silvapulle2 and George Braitberg13

1Emergency Department Austin Health 2Department of Biometrics Monash University and 3Department of Medicine University of Melbourne Melbourne Victoria Australia

Abstract

Objective The study was performed to determine the nature of adverse events in an ED

Methods The methodology has been described in the accompanying paper Two by two tables wereanalysed using the two-tailed Fisherrsquos exact test A P-value of le005 was consideredsignificant Statistical analysis was performed using MINITAB

Results One hundred and ninety-four events were detected from a sample of 3222 patients Exceptwhere specified events with management causation le3 were excluded This excluded 24events (124) leaving 170 for analysis Errors of commission occurred in 55 andomission in 45 Errors of commission were significantly associated with prior eventserrors of omission with ED events (P le 00001 respectively) The most common cause ofevents was drug reactions 135 had a Naranjo score ge 1 054 ge 4 Prior events weresignificantly associated with adverse drug reactions (P le 00001) Drug reactions wereassociated with a lower preventability score (P le 00001) Diagnostic issues were presentin 12 All three categories that is diagnosis not considered diagnosis within thedifferential and seriousness not appreciated were associated preventability ge4 (P le 00001P le 002 and P le 0004 respectively) Diagnostic problems were significantly associatedwith ED events (P le 00001)

Conclusion In conclusion the data demonstrate that events fall into two sets prior events whichare associated with errors of commission drug reactions and lower preventability andED events which are associated with errors of omission diagnostic issues and highpreventability

Key words adverse event emergency department error

Introduction

In order to prevent adverse events (AE) an understand-ing of the nature of events in the ED would be advan-

tageous AE seldom happen in isolation but are theresult of many factors some of them patient-basedsome doctor or nurse-based and some system-based Inan earlier study critical incident monitoring was used

J Hendrie et al

10 copy 2007 The AuthorsJournal compilation copy 2007 Australasian College for Emergency Medicine and Australasian Society for Emergency Medicine

to provide insights into incidents and AE1 This tech-nique provides useful information regarding the causesof events such as communication problems and humanfactors such as haste distraction or fatigue Howeverbeing voluntary it is unable to provide objective quan-tification of the nature of events

The aviation industry provides a model for safetyimprovement The risk of death has fallen from 35 per100 000 flight hours in 1960 to 137 in 20052 In 1998 inthe USA there was not a single death from commercialaviation3 Each crash is analysed in detail before por-tioning responsibility for the crash and the train ofevents leading to a crash or near miss Recommenda-tions are then made to prevent recurrence In the carindustry the Toyota approach lsquoThe 5 Whysrsquo gives aclear indication of the depth of analysis applied4 InAustralia the mortality rate from car accidents hasbeen reduced from the peak in 1970 of 30 per 100 000ndash9 per 100 000 in 2001 through a combination of policedriver road and mechanical initiatives5 Closer to homelooking at the systems behind anaesthesia has helpedreduce a death rate of one in 20 000 two decades ago toone in 200 000 in 19936

The Harvard Medical Practice Study and the UtahColorado study both presented data on the nature ofAE7ndash9 These were derived from a hospital populationand not specifically from the ED

The accompanying paper reported the incidence andoutcome of AE The present paper reports the nature ofevents detected in the study We hypothesize that EDAE are associated with diagnostic error and are there-fore likely to be preventable

Materials and methods

The definitions and methodology were described in theaccompanying paper The term lsquoeventrsquo is used to includeboth incidents and AE

Diagnostic errors were subdivided into lsquodiagnosiscorrect but seriousness not appreciatedrsquo lsquodiagnosisconsidered within differential but alternative preferredrsquoand lsquodiagnosis not consideredrsquo Investigative issueswere subdivided into errors in interpreting laboratoryresults errors in X-ray and electrocardiogram (ECG)interpretation and lsquotest not requested when indicatedrsquo

The term adverse drug reaction (ADR) is limited toside-effects and allergic reactions whereas the termadverse drug event (ADE) also includes human factorssuch as delay in administration accidental overdose orincorrect medication The majority of drug reactions

was classified under a specific organ system (egcardiac arrhythmia under circulatory gastrointestinalhaemorrhage resulting from non-steroidal anti-inflammatory drugs under digestive) and thus do notappear in major diagnostic Category 21 in Table 1

Events were subdivided into errors of omission orcommission

Statistical analysis

Two by two tables were analysed using the two-tailedFisherrsquos exact test A P-value of le005 was consideredsignificant Two by two tables were created byamalgamating preventability categories into a non-preventable group (preventability lt5050 Categoriesone two and three) and a preventable group (prevent-ability gt5050 Categories four five and six) with respectto ADR side-effects allergic reactions and the threediagnostic issues Statistical analysis was performedusing MINITAB (Minitab State College PA USA)

Table 1 Major diagnostic categories (n = 170)

n

0 Pre MDCdagger 01 Nervous system 122 Eye 33 Ear nose and throat 54 Respiratory system 185 Circulatory system 266 Digestive system 177 Hepatobiliary and pancreas 58 Musculoskeletal and connective tissue 229 Skin subcutaneous tissue and breast 11

10 Endocrine nutrition and metabolic disease 011 Kidney and urinary tract 1212 Male reproductive system 213 Female reproductive system 114 Pregnancy childbirth puerperium 015 Newborns and neonates 016 Blood and blood-forming organs 217 Myeloproliferative disorders 018 Infectious and parasitic diseases 819 Mental diseases 220 Substance use and substance induced organic

mental disorder1

21 Injuries poisonings and toxic effects of drugs 2222 Burns 123 Factors influencing health status and other

contacts with health services0

daggerConsists of patients having tracheostomy bone marrow orliver transplantation MDC major diagnostic category

Nature of events


11

The Ethics Committee of the Austin and RepatriationMedical Centre approved the study

Results

The nature of events was classified according to majordiagnostic category (Table 1) Circulatory events werethe most common followed by lsquoinjuries poisonings andtoxic effect of drugsrsquo musculoskeletal and connectivetissue and then respiratory and digestive events

Errors of commission occurred in 55 and omissionin 45 Errors of commission were significantly asso-ciated with prior events (P le 00001 odds ratio 002995 confidence interval [CI] 0011ndash0075) and errors ofomission with ED events (P le 00001 odds ratio 3295 CI 1232ndash8557)

There were 45 drug reactions the most common sin-gle cause of events present in 265 of all events withmanagement causation This equates to a rate of 135Of the 265 with drug reactions 20 were caused byside-effects and 65 were judged to be allergic reac-tions Drug reactions were significantly associated withprior events (P le 00001 odds ratio 01 95 CI 003ndash0292)

Side-effects and allergic reactions revealed strongindependent associations with prior events (P le 00004odds ratio 0177 95 CI 005ndash051 and le0003 respec-tively odds ratio 0 95 CI 0ndash0467)

Table 2 documents drug-related events by class ofdrug (see below)

Antibiotic and cardiovascular medications were themost commonly implicated followed by analgesic andanti-tumour medications

Table 3 summarizes the types of drug-related compli-cations Rashes were the most common complicationdue to penicillin bactrim trimethoprim and ibuprofenCardiovascular complications ranged from posturalhypotension (from diltiazem) through to fatal bradycar-dia from amiodarone administered for rapid atrialfibrillation in the setting of myocardial infarction Therewere three cases of chemotherapy induced neutropeniatwo associated with fever Renal damage resulted fromgentamycin and angiotensin converting enzyme inhibi-tor Antipsychoticantiemetic agents including prochlo-rperazine caused various dyskinetic reactions Non-steroidal anti-inflammatory drugs such as aspirinand voltaren resulted in gastrointestinal bleeding Onepatient on dialysis developed hypercalcemia secondaryto calcium supplements

There is evidence that ADR were associated with alower preventability score (P le 00001 odds ratio 017595 CI 0073ndash04)

Side-effects as a subset of ADR were significantlyassociated with a lower preventability score (P le 001odds ratio 0329ndash078) as were allergic reactions(P le 00001 odds ratio 0 95 CI 0ndash0293) Eight lsquopriorrsquodrug reactions were judged to be preventable

Adverse drug reactions were classified according tothe Naranjo ADR probability scale (see Appendix I)1011

The Naranjo ADR probability score ranges from minus4to 13 with 0 or less considered doubtful 1ndash4 consideredpossible 5ndash8 considered probable and 9 or greater con-sidered definite (see Appendix I) Table 4 documents thescores for ADR in the study

On the basis of the Naranjo score 27 ADR are con-sidered lsquopossiblersquo and 18 considered lsquoprobablersquo Nonewas considered doubtful

Table 2 Medication-related events according to classof drug (n = 45)

Class

Cardiovascular 178Antibiotic 156Analgesic 133Anti-tumour 89Antipsychoticantiemetic 67Anticoagulant 44Antiseizure 44Anaestheticsedative 44Antiparkinsonian 44Antiasthmatic 22Antidepressant 22Steroid 22Sympathomimetic 22Mineral 22Other 88

Table 3 Types of drug-related complications (n = 45)

Type of complication

Allergycutaneous 244CVS 200Bleeding 133CNS 111GIT 111Marrow suppression 67Metabolic 44Renal 42Respiratory 42

CNS central nervous system CVS cardiovascular systemGIT gastrointestinal system

J Hendrie et al


Human factors in ADE were sought Delay in drugadministration (usually analgesia) was the most com-mon occurring in nine cases although one patient withpneumonia did not receive antibiotic therapy prior toleaving for the ward In three cases an excessive dosewas given and in five either no drug or the wrong drugwas given Three patients were given inadequate fluidand one was give excessive fluid during resuscitation

Diagnostic and investigative issues were assessed(see Table 5) Diagnostic issues were present in 40cases giving a rate of 12 In nine of the events thecorrect diagnosis was made but the seriousness wasnot appreciated in 10 cases the diagnosis was consid-ered within the differential but an alternative diagnosiswas preferred and in 21 events the diagnosis wasmissed entirely (Fig 1) All diagnostic issues bar twowere judged preventable lsquoDiagnosis correct serious-ness not appreciatedrsquo and lsquodiagnosis considered withindifferential alternative preferredrsquo were significantlyassociated with preventability (P le 0004 odds ratio 895 CI 1707ndashinfin and P le 002 odds ratio 804 95 CI106ndash3568 respectively) Failure to consider the diagno-sis was highly preventable (P le 00001 odds ratio 205595 CI 3087ndash8627)

Thirty-two diagnostic problems occurred in ED andsix prior to ED Thus diagnostic problems were signif-icantly associated with ED events (P le 00001 oddsratio 0086 95 CI 0028ndash023)

No AE resulted from laboratory misinterpretationError in X-ray interpretation occurred in 17 cases the

ECG was misinterpreted in four cases and in 12 casesit was judged that further investigation was indicated

Discussion

The most common cause of events was drug side-effectsand allergic reactions which together accounted for265 of events (135 of the study sample) Hafnerand colleagues11 reported a rate of 17 but onlycounted ADR with Naranjo scores of ge4 The compara-tive rate in the present study is 054

Several of the questions in Naranjo ADR probabilityscale are difficult to apply in an ED setting Most of thedrugs implicated in our study were given orally Givinga placebo (Appendix I Question 5) would entail aperiod of observation to mimic the real administrationof a possibly toxic drug Re-challenging is an ethicaldilemma and might need to be done with full monitor-ing and resuscitation facilities immediately availableAlthough most ED have the necessary facilities under-taking re-challenging is time consuming with protocolsextending to 12 h or more

In the present study ADR were strongly associatedwith prior events and lower preventability Incidentmonitoring systems are relatively poor at detectingevents resulting from drug reactions or allergy112 Thestrong association with prior events suggests that med-ications prescribed elsewhere result in drug reactionsthat force attendance at the ED

Diagnostic difficulties were detected in 40 cases(12 of the study sample) The diagnosis was correctbut the seriousness was not appreciated in a quarterand the diagnosis was within the differential but analternative was preferred in a further quarter In theremaining half the diagnosis was missed entirely Ofthe 253 patients with chest pain 24 were managedinappropriately in the ED Pope and colleaguesreported that 21 of patients with acute myocardialinfarction (AMI) and 23 with unstable angina (USA)were inappropriately discharged from the ED13 How-ever of all patients presenting with chest pain 038were mistakenly discharged from the ED with either

Table 4 Adverse drug reaction Naranjo score

Naranjo score n

1 02 83 144 55 66 37 68 39 0

Table 5 Diagnostic problems (n = 40)

Category

Diagnosis not considered 525Seriousness not appreciated 225Alternative diagnosis preferred 25

Figure 1 Diagnostic issues (n = 40)

ALTERNATIVE DIAGNOSISPREFERRED

SERIOUSNESS NOTAPPRECIATED

DIAGNOSIS NOT CONSIDERED

Nature of events


13

AMI or USA The range and overlap of symptoms inischaemic chest pain and other causes of chest pain isproblematic14 In up to 40 of patients who are laterproven to have myocardial infarction the initial ECG iseither normal or unchanged Enzyme testing (CKMB atthe time of the study) only reaches sufficient diagnosticsensitivity and specificity to allow confidence in diag-nosis more than eight hours after ictus although tropo-nin I estimation has improved this Exercise stresstesting has significant false positive and false negativerates Each step in the evaluation process howeveradds an incremental gain in diagnosis short of per-forming coronary angiography on every patient withchest pain Even a computer-based diagnostic aidcould only achieve a sensitivity of 95 in the detectionof AMI15 The validation of the Erlanger chest painevaluation protocol offers new precision in the evalua-tion of chest pain although physician judgement wasstill required to raise sensitivity for AMI from 976 to100 and acute coronary syndrome from 804 to99116

The present study supports the proposition put for-ward in HMPS that lsquothe high rate of negligence in EDAE reported may be due to diagnostic errorsrsquo7 In thepresent study events arising in the ED were stronglyassociated with errors of omission diagnostic issuesand high preventability Although preventability wasmeasured rather than negligence it is likely there is aclose correlation between the two In comparison withother specialties QAHCS reported that emergency wasthe site in 15 of cases of suffering AE9 Although theED AE rate is low relative to other specialties theemphasis on diagnosis in ED means that when eventsoccur they tend to have high preventability Thomasand colleagues report a negligence rate of 948 forAE involving emergency medicine providers8 Thomasattributes this to task complexity (multiple concurrenttasks uncertainty changing plans high workload)HMPS to time constraint and part-time physicians Theseverity of illness as measured by diagnostic-relatedgroups correlates with an increase in AE rate althoughthe negligence proportion remains constant17 Many ofthe sickest patients do enter the Emergency RoomThese are at increased risk of an AE because of theserious nature of their presentation and time constraintson diagnosis

Preventing diagnostic error will require a raft ofinterventions Assessment of focused educational acti-vities has shown improved performance in calculatingand executing drug ordering18 Lack of feedback can becountered to some extent by encouraging residents to

take lsquoa patient labelrsquo and follow up patients theyattended as well as the more traditional case presenta-tionmorbidity and mortality meetings19 Diagnosingrare conditions or unusual presentations is a challengeeven for senior faculty with a wide differential diagno-sis A balance has to be struck between squanderingtime and resources on the futile pursuit of unlikelydiagnoses and not missing rare but important condi-tions Quantitative reasoning such as the Wells criteriafor deep vein thrombosispulmonary embolism is animportant means to diagnostic precision2021 Sufficienttime to be thorough adequate supervision decreasingmultitasking and interruptions will all contribute2223

The aviation industry is often held up as a goodmodel for health safety reform its improvements arecommendable3 However there are clear differencesPilots are single tasked emergency physicians multi-tasked managing many patients simultaneously23

Patients present with a multitude of different prob-lems24 The large number of differential diagnosesmight leave an element of uncertainty in even the mostthorough work up Pilots in contrast perform almostidentical tasks day in day out

However clearly there are lessons to be learned Pilotsacknowledge that fatigue impairs decision-makingwhereas physicians do not22 Sixty-four per cent of phy-sicians felt they could handle crises effectively evenwhen fatigued compared with 27 of pilots Cognitivefunction declines with sleep deprivation After 24 h with-out sleep psychomotor function is impaired equivalentto a blood alcohol of measurement of 0125 The com-mercial aviation industry regulates work hours and restbreaks very judiciously Desynchronization of circadianrhythm by shift work decreases quantity and quality ofsleep26 Interruptions and distractions are one of the mostcommon causes of pilot error27 Chisholm reports a meaninterruption rate of 10 per hour and mean break-in-taskrate of seven per hour for emergency physicians23 Therewas a positive correlation between the number ofpatients being managed at one time and the number ofinterruptions This number of interruptions is unlikelyto be tolerated in the airline industry

In conclusion the data in the present study fallclearly into two sets (i) events occurring prior to EDstrongly associated with errors of commission drugside-effects and allergic reactions and lower prevent-ability and (ii) events arising in the ED strongly asso-ciated with errors of omission diagnostic error andhigher preventability Teaching diagnostic precisionand developing systems to prevent and detect diagnos-tic problems are clearly an important means of reducing

J Hendrie et al


preventable events in the ED Such systems includesenior staff with appropriate credentials to providesupervision and clinical decision support definedclinical pathways and funded risk managementprogrammes

Acknowledgements

Dr L Sammartino undertook the initial screening forevents Dr M Silvapulle provided statistical advice andanalysed the data using MINITAB aided by Mr AMatta Ms Fiona Nelson helped with data preparationAssociate professor Dr G Braitberg undertook theblinded review Dr Hendrie conceived the study judgedthose cases screening positive scored cases judged tobe incidents or AE created the Excel files and wrotethe paper

Competing interests

This study was funded in part by a grant from theDepartment of Human Services Victoria

Accepted 27 March 2006

References

1 Stella D Hendrie J Smythe J Graham I Experience with criticalincident monitoring in the emergency department Emerg Med1996 8 215ndash19

2 Aircraft Owners and Pilots Association Available fromURL httpwwwaopaorgspecialnewsroomstatssafetyhtml[Accessed 18 October 2006]

3 Schenkel S Promoting patient safety and preventing medicalerror in emergency departments Acad Emerg Med 2000 71204ndash22

4 Kohn LT Corrigan JM Donaldson MS To Err is Human Aus-tralian Transport Safety Bureau Available from URL httpatsbgovaupublications20051925-presentaspx [Accessed 18October 2006] Institute of Medicine National Academy Press1999

5 Australian Transport Safety Bureau Available from URL httpatsbgovaupublications20051925-presentaspx [Accessed 18October 2006]

6 Runciman WB Sellen A Webb RK et al Errors incidents andaccidents in anaesthetic practice Anaesth Intensive Care 199321 506ndash19

7 Leape LL Brennan TY Laird N et al The nature of adverseevents in hospitalised patients N Engl J Med 1991 324377ndash84

8 Thomas EJ Studdert DM Burstin HR et al Incidence and typesof adverse events and negligent care in Utah and Colorado MedCare 2000 38 261ndash71

9 Wilson RM Runciman WB Gibbert RW et al The quality inAustralian health care study Med J Aust 1995 163 458ndash71

10 Lanctocirct KL Naranjo CA Comparison of the Bayesian approachand a simple algorithm for assessment of adverse drug eventsClin Pharmacol Ther 1995 58 692ndash8

11 Hafner JW Belknap SM Sqillante MD et al Adverse drug eventsin the emergency department Ann Emerg Med 2002 39258ndash67

12 Cullen DJ Bates DW Small SD et al The incident reportingsystem does not detect adverse drug events a problem of qualityimprovement J Qual Improv 1995 21 541ndash8

13 Pope JH Anfderheide TP Ruthazer R et al Missed diagnosis ofacute cardiac ischaemia in the emergency department N EnglJ Med 2000 342 1163ndash70

14 Master AM Jaffe HL Pordy L Cardiac and non-cardiac chestpain a statistical study of lsquodiagnostic criteriarsquo Ann Intern Med1954 41 315ndash22

15 Baxt WG Shofer FS Sites FD et al A neural computational aidto the diagnosis of acute myocardial infarction Ann EmergMed 2002 39 366ndash73

16 Fesmire FM Hughes AD Fody EP et al The Erlanger Chest PainEvaluation Protocol a one-year experience with serial 12-leadECG monitoring two-hour delta serum markers measurementsand selective nuclear stress testing to identify and exclude acutecoronary syndromes Ann Emerg Med 2002 40 584ndash94

17 Brennan TA Leape LL Laird NM et al Incidence of adverseevents and negligence in hospitalised patients N Engl J Med1991 324 370ndash6

18 Nelson LS Gordon PE Simmons MD et al The benefit of houseofficer education on proper medication dose calculation andordering Acad Emerg Med 2000 7 1311ndash15

19 Croskerry P The feedback sanction Acad Emerg Med 2000 71232ndash8

20 Wells P Anderson DM Rodger M et al Excluding pulmonaryembolism at the bedside without diagnostic imaging manage-ment of patients with suspected pulmonary embolism presentingto the emergency department by using a simple clinical modeland D-dimer Ann Intern Med 2001 135 98ndash107

21 Wells PS et al Value of assessment of pretest probability of deepvenous thrombosis in clinical management Lancet 1997 3501795ndash8

22 Sexton JB Thomas EJ Helmreich RL Error stress and team-work in medicine and aviation cross sectional surveys BMJ2000 320 745ndash9

23 Chisholm CD Edgar K Collison BA et al Emergency depart-ment workplace interruptions are emergency physicianslsquointerrupt-drivenrsquo and lsquomultitaskingrsquo Acad Emerg Med 20007 1239ndash43

24 Biros MH Adams JG Errors in emergency a call to ActionAcad Emerg Med 2000 7 1173ndash4

25 Volpp KGM Grande D Residentsrsquo Suggestions for reducingerrors in teaching hospitals N Engl J Med 2003 348 851ndash5

26 Kuhn G Circadian rhythm shift work and emergency medicineAnn Emerg Med 2001 37 88ndash98

27 Dismukes K Young G Sumwait R Cockpit interruptions anddistractions effective management requires a careful balancingact ASRS Directline 1998 December 4ndash9

Nature of events


15

Appendix I

Naranjo adverse drug reaction probability scale

Yes No Unsure Score

1 Did the ADE appear after the drug was administered +2 minus1 02 Did the ADR improve when the drug was discontinued

or a specific antagonist was given+1 0 0

3 Did the ADR appear when the drug was readministered +2 minus1 04 Are the alternative causes that could have on their own

caused the reactionminus1 +2 0

5 Did the ADR appear when a placebo was given minus1 +1 06 Was the drug detected in the blood (or other fluid)

in concentrations known to be toxic+1 0 0

7 Was the reaction more severe when the dose was increased or less severe when the dose was decreased

+1 0 0

8 Did the patient have a similar reaction to the same or similar drugs in any previous exposure

+1 0 0

9 Was the ADE confirmed by any objective evidence +1 0 0

Total score

J Hendrie et al


to provide insights into incidents and AE1 This tech-nique provides useful information regarding the causesof events such as communication problems and humanfactors such as haste distraction or fatigue Howeverbeing voluntary it is unable to provide objective quan-tification of the nature of events

The aviation industry provides a model for safetyimprovement The risk of death has fallen from 35 per100 000 flight hours in 1960 to 137 in 20052 In 1998 inthe USA there was not a single death from commercialaviation3 Each crash is analysed in detail before por-tioning responsibility for the crash and the train ofevents leading to a crash or near miss Recommenda-tions are then made to prevent recurrence In the carindustry the Toyota approach lsquoThe 5 Whysrsquo gives aclear indication of the depth of analysis applied4 InAustralia the mortality rate from car accidents hasbeen reduced from the peak in 1970 of 30 per 100 000ndash9 per 100 000 in 2001 through a combination of policedriver road and mechanical initiatives5 Closer to homelooking at the systems behind anaesthesia has helpedreduce a death rate of one in 20 000 two decades ago toone in 200 000 in 19936

The Harvard Medical Practice Study and the UtahColorado study both presented data on the nature ofAE7ndash9 These were derived from a hospital populationand not specifically from the ED

The accompanying paper reported the incidence andoutcome of AE The present paper reports the nature ofevents detected in the study We hypothesize that EDAE are associated with diagnostic error and are there-fore likely to be preventable

Materials and methods

The definitions and methodology were described in theaccompanying paper The term lsquoeventrsquo is used to includeboth incidents and AE

Diagnostic errors were subdivided into lsquodiagnosiscorrect but seriousness not appreciatedrsquo lsquodiagnosisconsidered within differential but alternative preferredrsquoand lsquodiagnosis not consideredrsquo Investigative issueswere subdivided into errors in interpreting laboratoryresults errors in X-ray and electrocardiogram (ECG)interpretation and lsquotest not requested when indicatedrsquo

The term adverse drug reaction (ADR) is limited toside-effects and allergic reactions whereas the termadverse drug event (ADE) also includes human factorssuch as delay in administration accidental overdose orincorrect medication The majority of drug reactions

was classified under a specific organ system (egcardiac arrhythmia under circulatory gastrointestinalhaemorrhage resulting from non-steroidal anti-inflammatory drugs under digestive) and thus do notappear in major diagnostic Category 21 in Table 1

Events were subdivided into errors of omission orcommission

Statistical analysis

Two by two tables were analysed using the two-tailedFisherrsquos exact test A P-value of le005 was consideredsignificant Two by two tables were created byamalgamating preventability categories into a non-preventable group (preventability lt5050 Categoriesone two and three) and a preventable group (prevent-ability gt5050 Categories four five and six) with respectto ADR side-effects allergic reactions and the threediagnostic issues Statistical analysis was performedusing MINITAB (Minitab State College PA USA)

Table 1 Major diagnostic categories (n = 170)

n

0 Pre MDCdagger 01 Nervous system 122 Eye 33 Ear nose and throat 54 Respiratory system 185 Circulatory system 266 Digestive system 177 Hepatobiliary and pancreas 58 Musculoskeletal and connective tissue 229 Skin subcutaneous tissue and breast 11

10 Endocrine nutrition and metabolic disease 011 Kidney and urinary tract 1212 Male reproductive system 213 Female reproductive system 114 Pregnancy childbirth puerperium 015 Newborns and neonates 016 Blood and blood-forming organs 217 Myeloproliferative disorders 018 Infectious and parasitic diseases 819 Mental diseases 220 Substance use and substance induced organic

mental disorder1

21 Injuries poisonings and toxic effects of drugs 2222 Burns 123 Factors influencing health status and other

contacts with health services0

daggerConsists of patients having tracheostomy bone marrow orliver transplantation MDC major diagnostic category

Nature of events


11


Results














Class






J Hendrie et al







Discussion






Naranjo score n

1 02 83 144 55 66 37 68 39 0


Category






Nature of events


13









J Hendrie et al



Acknowledgements


Competing interests



References




























Nature of events


15

Appendix I


Yes No Unsure Score








+1 0 0


+1 0 0


Total score

Nature of events


11


Results














Class






J Hendrie et al







Discussion






Naranjo score n

1 02 83 144 55 66 37 68 39 0


Category






Nature of events


13









J Hendrie et al



Acknowledgements


Competing interests



References




























Nature of events


15

Appendix I


Yes No Unsure Score








+1 0 0


+1 0 0


Total score

J Hendrie et al







Discussion






Naranjo score n

1 02 83 144 55 66 37 68 39 0


Category






Nature of events


13









J Hendrie et al



Acknowledgements


Competing interests



References




























Nature of events


15

Appendix I


Yes No Unsure Score








+1 0 0


+1 0 0


Total score

Nature of events


13









J Hendrie et al



Acknowledgements


Competing interests



References




























Nature of events


15

Appendix I


Yes No Unsure Score








+1 0 0


+1 0 0


Total score

J Hendrie et al



Acknowledgements


Competing interests



References




























Nature of events


15

Appendix I


Yes No Unsure Score








+1 0 0


+1 0 0


Total score

Nature of events


15

Appendix I


Yes No Unsure Score








+1 0 0


+1 0 0


Total score

Experience in adverse events detection in an emergency department: Nature of events

Documents

Transcript of Experience in adverse events detection in an emergency department: Nature of events