Objective To identify the risk factors associated with complaints, malpractice claims and impaired performance in medical practitioners.
Design Systematic review.
Data sources Ovid-Medline, Ovid Embase, Scopus and Cochrane Central Register of Controlled Trials were searched from 2011 until March 2020. Reference lists and Google were also handsearched.
Results Sixty-seven peer-reviewed papers and three grey literature publications from 2011 to March 2020 were reviewed by pairs of independent reviewers. Twenty-three key factors identified, which were categorised as demographic or workplace related. Gender, age, years spent in practice and greater number of patient lists were associated with higher risk of malpractice claim or complaint. Risk factors associated with physician impaired performance included substance abuse and burn-out.
Conclusions It is likely that risk factors are interdependent with no single factor as a strong predictor of a doctor’s risk to the public. Risk factors for malpractice claim or complaint are likely to be country specific due to differences in governance structures, processes and funding. Risk factors for impaired performance are likely to be specialty specific due to differences in work culture and access to substances. New ways of supporting doctors might be developed, using risk factor data to reduce adverse events and patient harm.
PROSPERO registration number PROSPERO registration number: CRD42020182045.
- clinical governance
- risk management
- health policy
- human resource management
- health & safety
Data availability statement
All data relevant to the study are included in the article or uploaded as online supplemental information.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
Strengths and limitations of this study
Comprehensive search terminology and rigorous review methodology to identify studies from a range of academic databases and grey literature sources.
A large number of factors associated with doctors at risk of malpractice claims, complaints or impaired performance were identified.
High levels of heterogeneity precluded conducting pooled analyses.
Differences in healthcare culture and governance between countries may limit the generalisability of the findings.
Medical practitioners have a responsibility to ensure the delivery of high quality and safe patient care. At times, the care provided may not be considered satisfactory or result in poor patient outcomes and a complaint may be lodged against a practitioner. Practitioners may receive complaints directly from patients, or through employers or organisations that do not have regulatory or licensing powers. Complaints about a clinician’s conduct can also be lodged with practitioner licensing boards.1 Some practitioner licensing boards carry out investigations into practitioner conduct, largely relying on complaints and claims from patients, peers and employers to identify potential instances of misconduct and impaired performance (ie, impairment in ability to practice).2 In contrast, claims are lawsuits or statements that have been filed for compensation for injuries caused by alleged negligence or omission.3 4 However, not all impaired performance results in, or justifies a complaint or claim. Impaired performance, therefore, may provide insights into quality and safety-related problems before they result in poor patient outcomes and subsequent formal complaints or malpractice claims. Factors that interfere with a doctor’s ability to function pose a risk to patient safety, whether they result in complaints, claims or impaired performance.
The identification of practitioners at higher risk of complaints and subsequent claims has been examined in previous studies.5–7 Several predictive factors have been identified, and these can be categorised as system or personal risk factors. Examples of system factors include country of initial training, clinical workload and practice setting or sector (eg, solo vs group practice, private vs public health sector).8 Personal, or ‘demographic’, risk factors are specific to the individual; characteristics that have been reported to affect risk of complaints include age, sex, mental state, medical specialty and number of prior complaints.7 Given the increasing volume and complexity of patient care needs, as well as technical changes in clinical practice, it is necessary to gain a better insight into the factors which may lead to complaints, malpractice claims or impaired performance.
Understanding the factors contributing to risk will allow practitioner boards to make more objective assessments of doctors on receipt of a complaint and allow for better targeted monitoring of higher risk medical practitioners with imposed conditions or restrictions on their registration.9 Therefore, the purpose of this systematic review was to examine the research evidence provided in peer-reviewed and grey literature, to identify the risk factors associated with complaints, malpractice claims and impaired performance in medical practitioners.
The published protocol (Prospero registration number: CRD42020182045) guided the review in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement (PRISMA).10
A comprehensive search strategy was developed in consultation with a research librarian to search Ovid-Medline, Ovid Embase, Scopus and the Cochrane Central Register of Controlled Trials for peer-reviewed literature and Google for grey literature. Databases were searched on 27 March 2020 for English language articles published between 2011 and 2020. This was accompanied by handsearching the reference lists of relevant review articles. The full search strategy for all databases is shown in online supplemental appendix A.
An example, illustrating the search strategy for Scopus, is as follows:
TITLE-ABS-KEY(“Doctor*” OR “physician*” OR “medical officer*”) AND TITLE-ABS-KEY(“malpractice” OR “negligen*” OR “impair*“) AND TITLE-ABS-KEY(“risk*")
Original peer-reviewed research studies and grey literature articles were included in the systematic review if they met the following additional criteria: (1) involving doctors; (2) identifying risk factors for impaired performance and malpractice claims or complaint; (3) providing disaggregated demographic characteristics about the doctors in the sample. For this review, the terms ‘physician’, ‘doctor’ and ‘medical practitioner’ are used interchangeably to refer to registered medical doctors. Articles were excluded if they were: (1) systematic reviews, reviews of reviews (umbrella reviews) and other research syntheses; (2) grey literature opinion, letters to the editor, commentary or case report; involved student doctors or patients; (3) reported aggregated sample characteristics; (4) published before 2011; (5) published in a language other than English and (6) focused on system breakdown factors (eg, informed consent processes, organisational use of medical device or treatment, management processes) defensive medicine or litigation.
Screening and data extraction
The results of the searches were entered into EndNote citation management software (V.8.2; Thompson Reuters, New York, New York, USA), and duplicates were removed. The study titles and abstracts were entered into Rayyan, a free web and mobile application screening tool.11 For each study, title and abstract were independently screened by pairs of reviewers for inclusion according to the prespecified criteria. Disagreements were resolved via discussion. Abstracts flagged as potentially relevant by reviewers underwent full-text review, again by independent pairs of reviewers.
The data were extracted independently by pairs of reviewers into a form specifically designed for the review and piloted for usability prior to data extraction. The extraction form included author(s) name, year of publication, country where the study was conducted, study design, characteristics and risk factors of doctors and related data (eg, measures of relative risk such as OR, rate ratios, HRs), study limitations and study results. Where disaggregated data were reported for multiple professions, only data reported for doctors were extracted.
Risk of bias
Methodological quality of the included peer-reviewed studies was assessed using the following The Joanna Briggs Institute critical appraisal tools: Checklist for Cohort Studies, Checklist for Analytical Cross-Sectional Studies, Checklist for Randomised Controlled Trials, Checklist for Quasi-Experimental Studies and the Checklist for Case–Control Studies.12 13 Tools were selected based on study design and piloted on a sample of six research papers. Study quality was appraised by pairs of independent reviewers, with disagreements resolved via discussion.
Data processing and analysis
A narrative synthesis was performed for this review. Synthesis included numerical statistical summaries, textual commentaries, and tabular and graphical representations.
Patient and public involvement
Patients and the public were not involved in the design and conduct of this review.
The combined searches yielded 38 884 articles. Following removal of 13 794 duplicates, 25 090 abstracts and 329 full-texts were reviewed, with 67 articles meeting inclusion criteria. Figure 1 presents the PRISMA diagram for the identification, screening, and inclusion processes. The grey literature search identified 11 reports, with three meeting the inclusion criteria.
Description of the included studies
Characteristics of the included studies are in table 1. Publication years of included studies (2011–2020) are provided in figure 2. Most studies were conducted in the USA (60%, n=40), followed by Australia (12%, n=8). Two studies were conducted each in Canada, Denmark, Japan, Taiwan and the UK. One study was conducted in each of Chile, Columbia, Egypt, France, India, Iran, Italy, The Netherlands and Spain.
Most studies were assessed as having considerable or potential flaws or limitations in their design, conduct or analysis that could distort the results (see table 2).
The included 67 articles identified 23 factors associated with doctors at risk of malpractice claims, complaints and/or impaired performance. The 23 factors, categorised as either demographic or workplace related, are shown in figure 3. Due to the variety of different settings and methods, it was deemed that the methodological heterogeneity was too broad to permit the calculation of pooled results. The results from studies are therefore reported in the form of a narrative synthesis.
Older doctors were generally found to be at 1.1–1.4 times greater risk of malpractice claims or complaints when compared with younger doctors15 ,16 17. (One study investigating the predictive impact of age on claims reported age to have no effect.18 The extent to which age is confounded with patient volumes is unclear and potentially important, as detailed examination has shown that the number of claims per 1000 patient encounters is higher for clinicians with lower numbers of patient encounters.19
Younger doctors at risk of impaired performance were more likely to be affected by environmental factors (eg, work induced fatigue),20 while older doctors (eg, 59–90 years of age) were at greater risk of impaired performance due to cognitive disorders.21 22 According to two studies, age was not a good predictor of impaired performance23 or burn-out.24 Three studies reported the average age of doctors enrolled in physician health programmes (PHPs) was 40–49 years of age25–27 and doctors referred for assessment of substance use or psychiatric were also more likely to be 45–64 years old.28 Finally, alcohol abuse or dependence was more likely in those who were younger.29 30
Compared with females, males made up 91%–92% of disciplined doctors,16 31 were 1.3 times more likely to be the subject of a complaint,32 and have a history of complaint.33 34 Similarly, doctors presenting to PHP were predominantly male,25 27 making up 85% of the misconduct doctor population.35
Compared with females, males were 1.4–2.3 times more likely to be the primary defendant in a malpractice case.33 36 Gómez-Durán et al18 reported no gender difference concerning recurrent paid malpractice claims. Similar to age, there is potential for confounding association with differences in the number of complaints or claims per 1000 patient encounters, and due to differences in the gender composition of different clinical specialties.19
Two studies reported no gender differences in self-reported sleep-related impaired performance20 or burn-out.24 While males represented over half (57%–86%) the doctors presenting to PHPs,26 37 one study reported female surgeons had almost twice the rates of alcohol abuse or dependence as male surgeons.30 Doctors with evidence of neurocognitive disorders or referred for psychiatric examination were predominantly male (90% vs 70%).21 22 28
One Australian study reported an 18%, 31% and 46% higher risk of complaint if a doctor was born in Asia, Africa or the Middle East, respectively.38 In contrast, two studies from the USA reported no statistical differences between Caucasian and non-Caucasian doctors.1 39
One study reported referrals for assessment of Caucasian doctors were proportionately higher (84% vs 71%) when compared with general doctor characteristics group published by the American Medical Association.28 40 Whereas a second study did not find an association between ethnicity and burn-out, as a predictor of impaired performance.24
No association was found between doctor malpractice and marital status.25 33
Partnered doctors were 1.3–2 times more likely to suffer alcohol abuse or dependence, and doctors whose spouse was also employed were 1.43 times more likely to suffer burn-out-related impaired performance.29 30 41
Mental health, personality and burn-out
One study reported that surgeons were 1.9 times more likely than physicians to have complaints attributed to interpersonal behaviour problems (eg, complaints about bullying, discrimination).32
One study reported that doctors with malpractice claims had a variety of mental health diagnoses with the most common being mood disorders (12%) and substance abuse (12%).25
A small number of studies addressed mental health,23 27 28 37 42 43 personality23 and burn-out.30 41 44 A higher proportion of doctors with impaired performance had mental health issues including bipolar disorder, depression and anxiety23 27 37 and problems with mental health increased with age.43 Doctors and their peers were often aware of the presence of an underlying issue with one-third of doctors self-referring for cognitive assessment.23 Higher levels of burn-out (eg, depersonalisation) related impaired performance was reported by three of four studies30 41 44 and associated with more hours of direct patient care, more hours on-call and more shifts per month.24 Doctors with emotional exhaustion also showed greater alcohol use or dependence.30
Physical health and diet
While physical health declined with age,43 one study found that, for doctors self-referring to a PHP, mental health problems were more common than physical (26% vs 1%).37 Doctors with diets that were high in plant-based foods and low in saturated fat and added sugars had less sleep-related impaired performance.20
In Australia, compared with physicians, surgeons were at 2.1 times higher risk of complaints regarding substance use.32 In Indiana, USA, drug or alcohol abuse and drug diversion were the most common reasons for a disciplinary complaint.45
One study found no relationship between either malpractice or maladaptive behaviours for a sample of emergency medicine doctors, reporting that doctors who attended worship services had an odds of 0.4 of maladaptive behaviours including smoking (10%), marijuana or cocaine use (10%) and alcohol use.24
Data from PHPs in multiple jurisdictions found alcohol was the most commonly used substance, with reports between 8% and 15% of medical practitioners screened positive for alcohol abuse.37 The secondary consequences of alcohol use were reported to include addiction problems,37 46 mental and physical impaired performance (eg, reduced alertness, impaired adaptive tracking)30 46 and recent major medical errors.30
In general, community doctors have a slightly higher sleep-related impairment score compared with reference populations (ie, the 2000 General US Census).20 Surgeons’ performance post-on-call shift (ie, when a clinician is able to be contacted to provide care if necessary, but not formally on duty) was similar to, or worse than, the performance of intoxicated surgeons to a degree where dose-dependent ethanol-induced performance closely resembled the decrease of performance over a 14-hour night shift.46 Similarly, a study from the USA47 reported that doctors remained impaired (eg, attentional failures) even after over 4 hours of sleep. This study also reported an inverse relationship between fatigue and technical skill.48
One study found doctors with neurocognitive disorders (defined as having a deficit in recent memory, executive functioning, social cognition, global functioning or visuospatial functioning during the study period) had more unsolicited patient complaints compared with age-matched and sex-matched comparators.21
Cognitive performance is affected by age43 and deteriorates throughout the course of successive shifts, being significantly worse in the fifth and sixth shift, and/or longer shift length.49 Doctors presented to PHPs with four principal reasons for impaired performance: (1) brain disease (48%); (2) mood/anxiety disorders or treatment side effects (28%); (3) substance use (9%) and (4) traumatic brain injury (7%).23 Of the doctors clinically referred for neuropsychological assessment, impaired doctors were referred predominantly with known neurological disease, known psychiatric (eg, Parkinson’s disease, major depression) or suspected cognitive diagnoses.22
Medical training location
Evidence was mixed on whether locally trained doctors or internationally trained doctors were more subject to complaints.50 A Canadian study found that 33% of doctors subject to claims were trained internationally, while these doctors make up 23% of the doctor population.16 An Australian study found doctors trained internationally had lower odds of being complaint-prone than those trained in Australia.51
Through the complaints process, one study identified a greater percentage of internationally trained doctors with a neurocognitive impairment (33% vs 23%),21 while another study found that relatively fewer internationally trained doctors (74% vs 83%) were referred for fitness-to-practice concerns.28
General practitioners (GPs) and family medicine specialists comprise a large proportion of total complaints (49%–65%).9 16 50 53 Other high-risk specialties include psychiatry, surgery, obstetrics and gynaecology.16 31 51 53 For example, surgeons were noted to be 2.3 times as likely to be subject to complaint compared with GPs.32
Overall, general surgery had higher risk of malpractice when compared with other surgical subspecialties,18 54 however, agreement was not universal.55 Grey literature and published academic papers reported GPs and family medicine specialists comprise a large proportion of total claims (34%–62%) and were up to three times as likely to be subject to claims.33 50 55 56 Other high-risk specialties include psychiatry, obstetrics and gynaecology, which were noted to be between four and 17 times as likely to be subject to malpractice claims compared with GPs.33 57
According to the location of the study, specialty as a risk factor varied between countries. In Australia, impaired performance was less of a risk for specialists,27 while in the USA two studies found a higher relative risk for surgeons, psychiatrists and family medicine doctors.28 41 A further study conducted in the USA found that surgeons were 1.9 times more likely than non-surgeons to enrol in a PHP because of alcohol-related problems26 and 0.5 times less likely because of opioid use.
Certification status and clinical workload
Licensure loss was associated with lapsed or missed certification.58 Greater number of hours worked per week,3 54 greater number of patients seen19 59–61 and more years in practice60 were associated with a 1.0–1.1 times higher risk of malpractice claims.
Doctors with evidence of neurocognitive disorders were more likely than age-matched and sex-matched doctors to lack board recertification.21 The highest percentage of doctors with impaired performance was found in the GP group (39%), with specialists comprising 26%.27 Doctors working longer hours and additional nights on call had lower prevalence of alcohol abuse,29 30 but 2.0–2.4 times higher risk of burn-out.24 41
Poor performance on medical knowledge and licensing exams1 and language assessments62 were predictors of complaints and fitness to practice issues. There was a 3.6 times greater risk of complaints during patient monitoring and follow-up,32 however, a 2019 Canadian report found that the majority of complaints were related to the clinical aspects of care, rather than issues with communication and/or unethical or improper behaviour.43
The use of procedure-specific patient education brochures reduced the likelihood of surgeons being sued.63
Time in practice
Frequency of disciplinary cases increased with years in practice43 50; it has been postulated that those with more years in practice had less factual knowledge and were less likely to adhere to standards.43
Frequency of malpractice claims increased with years in practice.39 51 60 61 64 In contrast, three studies reported practitioners with fewer years in practice had higher odds of a malpractice claim or medical litigation.55 65 66
Junior doctors who spent less time in practice than their senior counterparts were 1.3 times more likely to suffer impaired performance.41 However, there was no association between duration of practice and impaired performance.24 Sixteen per cent of doctors clinically referred for neuropsychological assessment and subsequently categorised as impaired were currently working full time or at reduced capacity.22
Claims history, insurance and doctor spending
Between 14% and 28% of doctors who receive complaints have experienced complaints previously.16 53 67
Doctors who were a recipient of a previous claim were reported to be at 1.9 times greater risk of a repeated claim,18 even when controlling for age, gender and specialty.18 52 US doctors whose insurance providers required participation in educational courses were less likely to be sued.63 Greater average doctor spending (eg, visits, tests, procedures) was associated with reduced risk of malpractice claim.68 69 The likelihood of multiple claims increased with the duration of study coverage, the calendar year and the country.1 18 34 52 70 The potential confounding effect of doctor activity level was not accounted for in the aforementioned studies.
Previous malpractice experience was associated with greater risk of impaired performance.44
Clinical practice setting or sector
One study reported complaints were more frequent for inpatient settings,15 while two studies reported that complaints were more frequent for doctors working in outpatient settings and independent practice.31 67 Doctors in solo/independent practice were more likely to be disciplined than those in group practices.16 One study reported a higher percentage of complaints against the private sector compared with the public system.15
Seven studies reported claims were more frequent for inpatient settings,15 17 55 71–74 while two studies reported that claims were more frequent for doctors working in outpatient settings and independent practice.3 54 One study reported a similar number of claims regarding inpatient and outpatient settings.75 Doctors in solo/independent practice were more likely to be sued or have recent claims than those in group practices,3 54 they were also more likely to do less well on assessments and recertification exams.43 Conversely, those with five or more claims were more likely to move into solo practice than doctors with no claims.34
While three studies reported that doctors practising in metropolitan areas were at greater risk of complaint than doctors practising in regional and remote areas,9 15 32 one study reported no relationship between geographical location and frequency of claims.51 Conversely, doctors (ophthalmologists) in regional medical centres had a longer time to the first complaint than those practising at academic medical centres.14 In the USA, complaints were more likely to be upheld in the Midwest than in the South.77 In Australia, doctors practising in Queensland or Victoria had a higher risk of a complaint than in other states.31 38
While two studies reported that doctors practising in metropolitan areas were at greater risk of malpractice claim than doctors practising in regional and remote areas,17 52 two studies reported no relationship between geographical location and frequency of claims.34 54
No relationship was found between the size of a doctor’s practice community and the rate of alcohol abuse/dependence.29
We examined the research evidence provided in peer-reviewed and grey literature identifying and characterising high-risk practitioners. This systematic review found older age, males, longer time in practice, having a history of claims, specialty and heavier clinical workload to be most frequently cited as risk factors for malpractice claims or complaints. The issue of risk factors and predicting complaints and claims is far more complex and context specific than a list of factors suggests. For example, in some countries, it may not be possible to pursue lawsuits for compensation, rather compensation for poor outcomes maybe settled through non-legal means. Differences in settlement practices may explain differences in the rate of complaints and claims between countries. The evidence examining risk factors for impaired performance was limited and likely subject to bias due to analysis primarily based on self-reported measures.24 30 41 44 For example, studies that used objective measures to record burn-out found fatigue and substance use to impair clinical practice,30 46 whereas studies using self-reported measures of burn-out found impaired performance stemming from burn-out manifested in exhaustion, depersonalisation of patients, and reduced sense of personal accomplishment.24 30 41 44
Factors associated with higher risk of malpractice claims or complaint are multifactorial and likely interdependent. For example, one contributing factor to a doctor experiencing burn-out may be higher clinical workload.41 High patient load or workload was associated with complaints, medical errors and poor performance on assessment.43 50 High workload and fatigue go hand in hand, with fatigue consistently linked to impaired performance (eg, diminished technical skills).20 46–48 For example, cognitive performance was found to deteriorate throughout a work shift, and cumulatively worsen over the course of successive shifts.49 Impaired performance such as attentional failures can remain even after a period of rest.47 Crucially, the decrease in performance/skill due to fatigue is similar to, or greater than performance impairments seen with alcohol intoxication.46 While the included studies did not directly link impaired performance with malpractice claims or complaint, it is likely that a diminishment or loss of ability plays a role in a doctor’s risk for malpractice claim or complaint.
Medical governance and regulatory bodies (eg, Medical Board of Australia, Federation of State Medical Boards—America) typically collect standard information on doctors (eg, age, gender, specialty).78 It is not practical or ethical for governance or regulatory bodies to collect and apply all 23 factors identified by the review. Instead, factors that are not currently collected by medical governance and regulatory bodies could perhaps be collected by employers, medical defence insurers and doctor health programmes which target high risk members. The need to collect information needs to be balanced and justified against registrants’ right to privacy. While no single factor was a strong predictor of risk of claims, complaint or impaired performance, collection of data on all factors would establish the extent to which each factor contributes to a doctor’s risk within the local context.
The findings highlight the need to include medical practitioner voices in the discourse on risk and prevention and the contribution of a context-informed perspective that includes awareness of demographic, workplace, social and legal effects. New ways of supporting doctors might be developed, using risk factor data to reduce adverse events and patient harm. For example, widespread education, peer support, mentoring and early intervention for factors associated with complaints, malpractice and impaired performance may facilitate early identification of at-risk doctors, including identification by colleagues, supervisors and patients. It may also help doctors recognise the factors in their own practice that elevates their risk of complaints, malpractice or impaired performance and encourage behavioural change. For instance, understanding that high patient load is a risk factor, medical practitioners might make strategic decisions around the volume of work they choose to take on. Similarly, employers might use the understanding that high patient volume puts medical practitioners and patients at risk to establish safe and effective workload policies and organisational culture practices. To provide a clearer picture on a doctor’s risk, baseline data for the normal population should also be reported alongside doctor population data. Reporting data from both populations would enable absolute rather than relative risk to be determined and facilitate pooling of data for analysis.
Strengths and limitations
A comprehensive search and rigorous review processes was used to search and appraise studies from a range of academic databases and grey literature sources. We were unable to pool data for analyses due to the heterogeneity of data collection and analysis methods. The majority of studies investigating risk of malpractice claim or complaint were descriptive and involved no comparator groups, limiting the ability to generalise study findings to the wider population of doctors. Furthermore, few studies adjusted statistical analyses to account for factors (eg, workload) likely to impact the risk of malpractice claim or impaired performance. Finally, the research regarding characteristics and risk factors for malpractice and/or impaired performance predominantly comes from the USA. It is possible that differences in healthcare culture (eg, attitudes towards healthcare, spending and responsibility) and governance (eg, malpractice legislation) are country specific, thereby, limiting the generalisability of the review findings.
We identified 23 risk factors for doctors at risk of malpractice claims, complaint and/or impaired performance. The risk factors include demographic factors such as age and gender, as well as work-based characteristics such as clinical workload and the geographical location in which a doctor practices. No single factor was a strong predictor of risk of claims, complaint or impaired performance, therefore, a range of workplace and demographic information should be collected to establish risk profiles.
Data availability statement
All data relevant to the study are included in the article or uploaded as online supplemental information.
Mary Simons contributed to the development of the search strategy and Catalin Tufanaru provided guidance on the selection of risk of bias tools for critical appraisal.
Twitter @DrLilAustin, @DFajardoPulido, @peter_hibbert, @JBraithwaite1, @wileslouise79, @RLystad, @JanetCLong, @frapport, @RClaywilliams
Contributors EEA, VD, RN, DFP, AP and RC-W conceived the review, RN conducted the search. EEA, VD, RN, DFP, RC-W, GA, YT, RPL, TT, SH, JCL and FR screened titles and abstracts. EEA, VD, RN, DFP, RC-W, GA, PDH, LKW and TT screened full texts. EEA, VD, RN, DFP and RC-W extracted data and undertook the critical appraisal. EEA, VD, DFP and RC-W undertook the qualitative analysis. EEA and VD wrote the introduction, EA and RC-W wrote the discussion. RC-W, PDH and JB revised the first draft of the paper. All members of the team revised the final draft paper.
Funding This work was supported by The Medical Council of New South Wales.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.