Article Text

Impact of unacceptable behaviour between healthcare workers on clinical performance and patient outcomes: a systematic review
  1. Linda Guo1,2,
  2. Benjamin Ryan3,
  3. Isabel Anne Leditschke1,4,
  4. Kimberley J Haines5,6,
  5. Katrina Cook7,
  6. Lars Eriksson8,
  7. Olusegun Olusanya9,
  8. Tanya Selak10,11,
  9. Kiran Shekar2,12,
  10. Mahesh Ramanan2,7,12,13
  1. 1 Adult Intensive Care Services, Mater Health Services Brisbane, Brisbane, Queensland, Australia
  2. 2 School of Medicine, The University of Queensland, Brisbane, Queensland, Australia
  3. 3 Intensive Care Unit, Princess Alexandra Hospital, Brisbane, Queensland, Australia
  4. 4 Mater Research Institute, The University of Queensland, Brisbane, Queensland, Australia
  5. 5 Department of Physiotherapy, Western Health, Footscray, Victoria, Australia
  6. 6 Melbourne Medical School, The University of Melbourne, Melbourne, Victoria, Australia
  7. 7 Intensive Care Unit, Caboolture Hospital, Caboolture, Queensland, Australia
  8. 8 The University of Queensland Library, The University of Queensland, Brisbane, Queensland, Australia
  9. 9 Intensive Care Unit, The Royal London Hospital, London, UK
  10. 10 Department of Anaesthesia, Wollongong Hospital, Wollongong, New South Wales, Australia
  11. 11 School of Medicine, University of Wollongong, Wollongong, New South Wales, Australia
  12. 12 Adult Intensive Care Services, The Prince Charles Hospital, Brisbane, Queensland, Australia
  13. 13 Critical Care Division, The George Institute for Global Health, University of New South Wales, Sydney, New South Wales, Australia
  1. Correspondence to Dr Linda Guo, Adult Intensive Care Services, Mater Health Services Brisbane, South Brisbane 4101, Queensland, Australia; linda.guo{at}health.qld.gov.au

Abstract

Background Recent studies suggest that displays of unacceptable behaviour, including bullying, discrimination and harassment, between healthcare workers (HCWs) may impair job performance, and in turn, increase the frequency of medical errors, adverse events and healthcare-related complications. The objective of this systematic review was to summarise the current evidence of the impact of unacceptable behaviour occurring between HCWs on clinical performance and patient outcomes.

Methods We searched MEDLINE, Embase, PsycINFO and CINAHL from 1 January 1990 to 31 March 2021. The search results were screened by two independent reviewers and studies were included if they were original research that assessed the effects of unacceptable behaviour on clinical performance, quality of care, workplace productivity or patient outcomes. Risk of bias was assessed using tools relevant to the study design and the data were synthesised without meta-analysis.

Results From the 2559 screened studies, 36 studies were included: 22 survey-based studies, 4 qualitative studies, 3 mixed-methods studies, 4 simulation-based randomised controlled trials (RCTs) and 3 other study designs. Most survey-based studies were low quality and demonstrated that HCWs perceived a relationship between unacceptable behaviour and worse clinical performance and patient outcomes. This was supported by a smaller number of higher quality retrospective studies and RCTs. Two of four RCTs produced negative results, possibly reflecting inadequate power or study design limitations. No study demonstrated any beneficial effect of unacceptable behaviour on the study outcomes.

Conclusions Despite the mixed quality of evidence and some inconsistencies in the strengths of associations reported, the overall weight of evidence shows that unacceptable behaviour negatively affects the clinical performance of HCWs, quality of care, workplace productivity and patient outcomes. Future research should focus on the evaluation and implementation of interventions that reduce the frequency of these behaviours.

  • communication
  • diagnostic errors
  • human factors
  • patient safety
  • teamwork

Data availability statement

All data relevant to the study are included in the article or uploaded as supplemental information.

Statistics from Altmetric.com

Request Permissions

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.

Introduction

Unacceptable behaviour in the workplace is negative behaviour that violates the norms for mutual respect, and includes rude, dismissive, aggressive and discourteous behaviour.1 2 In a 2015 survey of 606 doctors across three teaching hospitals in England, 31% of doctors described being subject to rude, dismissive or aggressive behaviour a few times a week or more, with junior doctors and registrars being twice as affected as consultants.2 Similarly, in a recent systematic review,3 high rates of bullying and undermining behaviour were found to be experienced by nurses, allied health staff, medical students and doctors worldwide. Though the high prevalence of these behaviours within the healthcare sector has been demonstrated, their impact on clinical performance and patient outcomes is less well characterised.

The selection of appropriate terminology to describe these behaviours poses several challenges. While the word ‘incivility’ has been used extensively in the medical literature to describe unacceptable behaviour,4 the word ‘incivility’ has a difficult history, particularly for people from racial and ethnic minorities, as it evokes memories of their struggle for equity.5 The term ‘unacceptable behaviour’ encompasses bullying, discrimination and harassment as legislated or reflected in non-regulatory instruments in several countries internationally.6 Furthermore, while most definitions of workplace bullying require that the behaviour is frequent and has clear intention of harming the target,7 unacceptable behaviour also encompasses behaviour with (1) unclear or ambiguous intent to harm, (2) low intensity and (3) low frequency7 8 as the effects of these behaviours are independent of the frequency of exposure.9

Exposure to rudeness, one aspect of unacceptable behaviour, affects task performance8 10 through multiple mechanisms. Rudeness produces negative affect, which serves as a signal that there is a problem in the environment. As a result, cognitive resources are reallocated from task completion to situational processing and postulating methods of remediation.11 Exposure to rudeness also produces a state of physiological arousal, narrowing of perception and more selective processing. This produces deficits in creative problem-solving, comprehension and recall of prior knowledge.10 Furthermore, rudeness can incite retaliatory behaviour or conscious effort to not allocate effort towards the required tasks,12 resulting in reduced helpfulness13 and impaired teamwork performance. The impact on task performance is also experienced by those who are witnesses to, but not direct victims of, rude behaviour.8 10

Unacceptable behaviour also has effects on healthcare worker (HCW) health and well-being, including decreased job satisfaction,7 14 professional demotivation, and desire to leave a medical specialty or the medical profession entirely.2 7 It also impacts mental health, which in turn affects physical health,14 in part due to victims engaging in potentially harmful behaviour.2 Thus, unacceptable behaviour in healthcare appears to be common and is potentially associated with a range of negative consequences for HCWs, patients, families of patients and the broader healthcare system. In this context, our primary objective was to perform a systematic review of the published evidence on the impact of exposure to unacceptable behaviour between HCWs on clinical performance. Secondary objectives of our review included whether exposure to unacceptable behaviour affects quality of care, workplace productivity and patient outcomes.

Methods

This systematic review protocol, developed according to the Preferred Reporting Items for Systematic review and Meta-Analysis Protocols (PRISMA-P) statement,15 was prospectively registered on the International Prospective Register of Systematic Reviews (PROSPERO) on 17 January 2020 (PROSPERO ID: CRD42020165818). It was conducted and reported according to PRISMA guidelines.16

Inclusion/exclusion criteria

Studies were included if they were original research published in the English language that assessed the effects of unacceptable behaviour between HCWs. Randomised controlled trials (RCTs) (including simulation studies that did not involve patients), non-randomised trials, cohort studies, before and after, qualitative and mixed-methods studies were included. The primary outcome was the clinical performance of HCWs, including both technical (diagnostic, procedural)17 and non-technical (teamwork, communication)4 18 aspects of performance. The secondary outcomes were patient outcomes (including mortality, hospital length of stay, complications), quality of care and workplace productivity. Other secondary outcomes (patient satisfaction, HCW job satisfaction, burnout, stress and depression) were included in the search, but were removed post hoc following the registration of the protocol and initial search due to the large volume of primary literature pertaining to the other outcomes. Systematic reviews, editorials, opinion articles, conference proceedings and abstracts were excluded.

Literature search

We searched the online databases MEDLINE, Embase, PsycINFO and CINAHL using Medical Subject Headings and text words from 1 January 1990 to 31 March 2021. We started the search in 1990 both as an arbitrary marker of contemporaneity and as a marker of the commencement of the patient safety movement following the Bristol paediatric cardiac surgery incidents.19 This was supplemented by checking references of identified articles. The date of search was 29 May 2021. The search strategies for the various databases were formulated by the medical librarian in the author group (LE) and are detailed in our online supplemental material (page 11).

Supplemental material

Study selection and data management

Literature search results were exported and stored in Microsoft Excel. References were managed using EndNote. Two authors (LG and BR) independently screened titles and abstracts of the articles identified by the literature search. For those that were relevant, full texts were obtained, and compared with the predetermined inclusion criteria. Data from each relevant article were independently extracted by two authors (LG and BR) according to preprepared forms. We extracted the study characteristics, participants, methodology, type of intervention (if present) and all reported outcomes relevant to our stated objectives. Discrepancies in study screening and data extraction were resolved by discussion or through review by a third author (MR).

Risk of bias in individual studies

The quality of individual studies was appraised using the Cochrane Risk of Bias tool for RCTs, the Critical Appraisal Skills Programme (CASP) Qualitative Checklist for qualitative studies and a modified Newcastle–Ottawa Scale (NOS, see online supplemental material page 13) for other study types. The modified NOS was scored out of a total of seven stars. For the purposes of this review, we defined zero to two stars as low quality, three to five stars as moderate quality and six to seven stars as high quality. This was completed by two authors in duplicate (LG and BR). Disagreement was resolved by review by the senior author (MR).

Data synthesis

A meta-analysis using the DerSimonian-Laird random-effects estimator to pool the results was planned. However, due to the lack of common outcome measures between studies and lack of reporting in raw proportions, such pooling was not possible and hence abandoned. Instead, a descriptive data analysis was performed and we produced separate summaries of the data pertaining each of the following outcomes: clinical performance, quality of care, workplace productivity and patient outcomes.

Results

Search results

Our search strategy identified 2557 titles and abstracts for screening and two further articles were identified through reference searches of the studies included in this review (see figure 1). After the removal of duplicates, 2073 titles remained. Two hundred eleven were selected for full-text review, of which, 175 were excluded. One hundred twenty-one studies did not have outcomes related to clinical performance or patient outcomes. Forty-two studies assessed, at least in part, unacceptable behaviour displayed by patients or families, without delineation between the effects of HCW and patient behaviour on the outcomes measured. Five studies contained no original data, one study was not in English, four studies were conducted in a non-clinical setting and two studies examined a student population only. Following these exclusions, 36 studies remained for inclusion in the review. There were 4 qualitative studies, 22 survey-based studies, 3 mixed-methods studies, 1 retrospective cohort study, 1 prospective cohort study, 1 database study and 4 RCTs. Of the four RCTs, two were classroom based,18 20 one was an in situ simulation17 and one4 did not provide sufficient information for the setting to be determined. Twenty-one studies were conducted in North America, six in the Middle East, four in Europe, three in East Asia and two in Oceania. The majority of studies (29 of 36, 80.5%) involved nurses, with smaller proportions of studies including physicians (15 of 36, 41.6%), nursing assistants or care aides (4 of 36, 11.1%), and allied health (2 of 36, 5.5%). Detailed characteristics of the included studies are provided in online supplemental table 1.

Figure 1

Preferred Reporting Items for Systematic review and Meta-Analysis diagram for the impact of unacceptable behaviour between healthcare workers on clinical performance and patient outcomes.58

Methodological quality assessment

The four qualitative studies assessed using the CASP checklist were of high quality, scoring between 8 and 10 on the CASP checklist. The overall quality of studies assessed using a modified NOS was low to moderate. There were 22 low-quality studies (zero to two stars), 5 moderate-quality studies (three to five stars) and 1 high-quality study21 (six to seven stars). All four RCTs were judged to have a significant risk of bias. Three4 17 20 scored low risk in at least three of the six categories, and either unclear or high risk in the remaining categories. The fourth18 scored high risk in two categories, unclear risk in two categories and low risk in two categories.

Clinical performance

A total of 11 studies addressed clinical performance as an outcome. This consisted of four RCTs with significant risk of bias, two low-quality mixed-methods studies and five survey studies (four low quality and one moderate quality).

The four RCTs found in the search were all simulation based. HCWs were randomised to be exposed to unacceptable behaviour from a simulation lab manager,18 visiting expert,17 operating surgeon4 or surgeon handing over to intensive care unit (ICU) staff upon transfer.20 Riskin et al 17 found poorer diagnostic (p<0.00035) and procedural (p=0.0002) scores in those exposed to unacceptable behaviour. Katz et al 4 showed that anaesthesiology residents had significantly poorer performance (p=0.009) when the simulated surgeon was portrayed as impatient, even without being overtly intimidating or using inappropriate language. In this study, the participants’ sensitivity to criticism was measured and similar between both control and intervention groups. In contrast, Johnson et al 18 found no significant difference in cardiopulmonary resuscitation, cognitive or teamwork performance post-exposure. In Avesar et al,20 ICU doctors received a handover containing an incorrect diagnosis from a surgeon exhibiting either rude or neutral behaviour. The percentage of doctors that challenged the incorrect diagnosis was smaller in the group that received a rude handover (55%, 11 out of 20 doctors) compared with the group that received a neutral handover (71%, 15 out of 21 doctors), but this did not reach statistical significance (p=0.28). The percentage of clinicians challenging diagnostic error was higher in more experienced clinicians, but this also did not reach statistical significance.

The remaining seven survey and mixed-methods studies demonstrated some degree of negative impact of unacceptable behaviour on HCW clinical performance. In a large Australian survey of 5178 hospital staff including 4411 clinical HCWs, 54.7% of respondents described the perceived negative impact of unacceptable behaviour on teamwork as moderate or major; this effect was more prominent in nursing and allied health than medical staff.22 This was echoed in a large Saudi Arabian study,23 in which 84.6% of respondents either somewhat or strongly agreed that unacceptable behaviour negatively affects performance. Two studies24 25 have used multiple survey tools to demonstrate negative associations between unacceptable behaviour and the following outcomes: team creative performance, intent to share knowledge,24 quality and quantity of routine tasks performed and innovation in the workplace.25 Lastly, in one survey study26 and two mixed-methods studies,2 27 respondents described impaired clinical performance as a result of inefficient working practice and avoidant behaviour,2 being hindered from performing job duties to their fullest potential26 and impaired ability to innovate and improve service delivery27 following exposure to unacceptable behaviour.

Quality of care

Seven studies examined the effect of unacceptable behaviour on quality of nursing care. There was one moderate-quality prospective cohort study, five low-quality survey studies and one high-quality qualitative study. In the prospective cohort study,28 there was an association between unacceptable behaviour at the first time point and missed nursing care at the second time point 2 years later, but not quality of care. Five survey studies found significant associations between unacceptable behaviour and perceived quality of care,23 29–32 however in two of these,31 32 the finding was specific to some sources of unacceptable behaviour only. Coleman31 found an association with supervisor interactions, but not nurse or physician interactions. In Alshehry et al,32 this relationship was demonstrated with nurse interactions, but not supervisor or physician interactions. It also only applied to some domains of quality of care (prevention of complications, well-being and self-care, functional readaptation, responsibility and rigour) and overall quality of care, but not others (patient satisfaction, health promotion, nursing care organisation).

In the qualitative study,33 participants described impedance to the delivery of nursing care due to loss of culture of reciprocal help between colleagues and decreased quality of clinical handover following exposure to unacceptable behaviour. This resulted in self-reported delays or omissions in laboratory tests and medication administration, as well as delays in documentation.

Workplace productivity

Three low-quality survey studies34–36 found significant associations between the frequency of unacceptable behaviour and self-reported workplace productivity. Berry et al 34 found that 46.7% of respondents perceived that their productivity would decrease after a workplace bullying event, while 29.4% reported that their productivity would increase; however, there was an association between the frequency of exposure to unacceptable behaviour and perception of a negative impact on productivity. Respondents who reported no exposure to these behaviours reported minimal adverse impact, while respondents who reported more frequent exposure were more likely to report decreased productivity as a result. Using the Workplace Limitations Questionnaire, the two other studies showed a reduction in productivity of 20% overall35 or between 9.5% and 22.1% for each specific domain of the questionnaire (time management, physical demands, mental/interpersonal demands and output demands).36 Using mean percentage reductions in productivity and the average salaries of nurses in the USA, the estimated monetary loss due to unacceptable behaviour in one of these studies was $11 581 per nurse per year (20% loss of productivity, multiplied by hourly rate of $30.54, based on an annual salary of $63 523.00, with usual working hours and usual annual leave).35 Though identical methodology was reportedly applied in the second study36 with similar working values (9.5%–22.1% loss of productivity, annual salary for registered nurses of $50 481.60), the mean annual cost of lost productivity was $1484.03 per registered nurse per year.

Patient outcomes

Twenty studies assessed the effects of unacceptable behaviour on patient outcomes. There was 1 high-quality retrospective cohort study, 1 moderate-quality database study, 1 moderate-quality mixed-methods study, 14 survey studies (2 moderate quality and 12 low quality) and 3 high-quality qualitative studies.

The retrospective cohort study21 showed the rate of postoperative complications (medical and surgical) was 11.9%–14.3% higher for patients whose surgeon had one or more coworker reports of unprofessional behaviour in the preceding 3 years. Though the patients of surgeons with more coworker reports had a higher American Society of Anaesthesiology class at baseline, and therefore a higher burden of systemic disease, this effect remained significant in modelling that controlled for patient characteristics. In the database study,37 the Pennsylvania Patient Safety Reporting System was searched over a 2-year period for events as a result of workplace bullying. Of these, 44% resulted in either a medication error or procedure, treatment or test-related error or complication. In the mixed-methods study, Arnetz et al 38 demonstrated a significant association between the frequency of unacceptable behaviour of a hospital unit and its frequency of central line-associated bloodstream infections over a 6-month period, but not catheter-associated urinary tract infections, pressure ulcers or ventilator-associated events.

Five survey studies showed that respondents (a combination of nurses, physicians and other employees) felt that there was a negative relationship between unacceptable behaviour in the workplace and patient outcomes—the percentages of respondents that answered affirmatively to this question were 32.8%,39 66%,40 66.4%,23 88.6%41 and 94%,42 respectively. In another survey study,22 49.8% of respondents reported a moderate or major impact on patient care and frequency of errors and this impact was more frequently reported by nurses and administrative staff than medical staff.

In two other survey studies, respondents also reported that they were aware of either a near miss (27.3% of participants43) or adverse event (12.5%43 and 14%40) as a result of unacceptable behaviour.

In another four survey-based studies, a variety of statistical techniques were used to evaluate associations between nurses’ exposure to workplace unacceptable behaviour and perceived frequency of negative patient outcomes.29 30 44 45 These found significant associations between unacceptable behaviour and the perceived frequency of errors,30 adverse events (such as medication errors, nosocomial infections and falls29 44) and near misses or negative outcomes.45 Three studies found either direct23 46 or indirect47 associations between the frequency of unacceptable behaviour and patient safety culture of their workplace, while a fourth study48 demonstrated an association with individual HCW safety competence.

In three qualitative studies,49–51 respondents have attributed the relationship between unacceptable behaviour and poorer patient outcomes to impaired decision-making, decreased efficacy of communication and heightened anxiety resulting in increased frequency of errors,49 reluctance to help others50 and disengagement and inattention while performing clinical duties.51

The key findings of our review are presented in table 1.

Table 1

Current evidence of the effects of unacceptable behaviour between HCWs on clinical performance, quality of care, workplace productivity and patient outcomes

Discussion

Key findings

This systematic review has demonstrated that HCWs perceive important associations between unacceptable behaviour perpetrated by other HCWs, impaired clinical performance and negative patient healthcare outcomes. This is substantiated by a smaller number of recent higher quality studies with more rigorous methodology and more objective outcome measures. While there are some inconsistencies in the strength of the associations reported, the overall weight of evidence suggests that unacceptable behaviour has widespread negative consequences with considerable implications for quality in healthcare delivery and patient safety. Of note, no study demonstrated beneficial effects from unacceptable behaviour. Attention from healthcare organisations and their administrators is warranted as interventions that effectively reduce the frequency of these behaviours are likely to result in safer and higher quality care for patients.

Importance of findings

This is the first systematic review to examine the effects of unacceptable behaviour between HCWs on clinical performance and patient outcomes. The inclusion of several large databases, journals and reference checks was aimed at maximising the yield of our search and increasing the sensitivity of the search.

Overall, the studies identified in our systematic review support the concept that unacceptable behaviour between HCWs impairs clinical performance, workplace productivity and quality of care, and results in negative patient outcomes. Of note, however, only two of the four RCTs found in the search yielded a statistically significant result. The lack of significant effect in the other two RCTs may in part be explained by lower participant numbers20 resulting in the study being insufficiently powered to detect an effect, or less robust methodology,18 as evidenced by poorer scoring on the Cochrane Risk of Bias tool. In addition, simulation-based studies only approximate the behaviour of HCWs in their working environment; though, the ability to perform clinical RCTs is also limited. Effective randomisation is not possible as exposure to unacceptable behaviour cannot be controlled, hence data from the clinical setting will be necessarily limited to observational studies.

In agreement with the majority of the subjective survey-based studies, the higher quality retrospective studies confirm that the impact of these behaviours is substantial. This is congruent with other rigorous studies that have drawn associations between patient complaints, as a surrogate for the unacceptable behaviour of surgeons,52 53 and rate of postoperative complications which affirm both the substantive consequences of these behaviours and the utility of this study type in measuring them. A related retrospective study found a negative impact on HCW hand hygiene and medication protocol compliance as a result of unacceptable behaviour perpetrated by patients and families.54

The high-quality qualitative studies identified in our review demonstrate that the multiple mechanisms by which these behaviours affect HCWs are in keeping with the broader psychological literature, including heightened anxiety,49 reduced helpfulness50 and development of avoidance behaviours.2 This and the mounting evidence of the psychological and physical impact of these behaviours on HCWs22 27 provide another important motivation to address these behaviours.

Our review also suggests that unacceptable behaviour may have considerable health economic implications, which should serve as an additional driver for institutional change. This is supported by the increase in frequency of healthcare-related complications21 38 and quantification of the monetary cost secondary to reduced workplace productivity35 36 associated with these behaviours, though there is a large disparity between the estimates available and the calculation is ultimately based on the subjective assessment of percentage of productivity lost. The economic impact of unacceptable behaviour may in part be due to decreased job satisfaction, increased intention among HCWs to leave their current position2 7 14 and therefore higher rates of staff attrition and turnover, however these outcomes were outside the scope of this review.

Limitations

The major limitation of our review is the quality of the studies. The majority of studies were survey based, and findings are self-reported and therefore ultimately subjective in nature. Though RCTs typically represent a higher level of evidence than cohort studies, the RCTs used to answer this question were exclusively simulation based and therefore only an approximation of real working practice. Due to the wide range of terms used to describe unacceptable behaviour in the literature, including bullying, incivility, rude, dismissive and aggressive behaviour and horizontal violence, there is a possibility that not all relevant studies were found in our search. Other possible sources of bias include sampling bias of the studies by the authors or publication bias of negative studies. Measures of inter-rater agreement were not calculated. As our search targeted the effects of individuals’ behaviour on HCW performance, the effects of negative team, departmental or institutional culture were not thoroughly examined.

Future research

The impact of unacceptable behaviour in healthcare has the potential to be far-reaching with negative effects on patient outcomes, adverse events and HCW performance. The weight of the evidence, despite its inconsistencies, strongly suggests that unacceptable behaviour has negative consequences for HCWs and patients alike, and hence by extension, for the healthcare system as a whole. We believe that future research should be targeted at evaluating interventions that may reduce the frequency and impact of unacceptable behaviour, such as interprofessional education55 and feedback and follow-up with clinicians who display these behaviours as identified by patient56 or coworker reports.57 Some of the inconsistencies in the literature may be addressed by high-quality qualitative and prospective observational studies. However, we suggest that this is of lesser importance than interventions aimed at reducing the frequency of unacceptable behaviour.

Conclusions

Unacceptable behaviour between HCWs is associated with a suite of negative consequences including impaired HCW clinical performance, decreased workplace productivity, decreased quality of care, increased adverse events and worse patient outcomes. Though the overall effects of unacceptable behaviour were detrimental to patient care, there were inconsistencies in the strength of associations reported. There is a need for the development and implementation of strategies to reduce unacceptable behaviour in healthcare and mitigate its impacts on patient safety.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplemental information.

Ethics statements

Patient consent for publication

Ethics approval

This study does not involve human participants.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • Twitter @maheshramanan

  • Contributors The study was overseen by MR. LE performed the search. LG and BR performed the study screening, data extraction and quality assessment. All authors contributed to, read, provided feedback on and approved the final manuscript. LG is the study guarantor.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • 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.

Linked Articles