Article Text

Retention of doctors in emergency medicine: a scoping review of the academic literature
  1. Daniel Darbyshire1,2,
  2. Liz Brewster1,
  3. Rachel Isba1,3,
  4. Richard Body4,5,
  5. Usama Basit6,
  6. Dawn Goodwin1
  1. 1Health Innovation One, Lancaster University Lancaster Medical School, Lancaster, UK
  2. 2Emergency Department, Salford Royal Hospitals NHS Trust, Salford, UK
  3. 3Paediatric Emergency Department, North Manchester General Hospital, Manchester, UK
  4. 4Division of Cardiovascular Sciences, The University of Manchester, Manchester, UK
  5. 5Emergency Department, Manchester University NHS Foundation Trust, Manchester, UK
  6. 6Department of Accident and Emergency, Ipswich Hospital NHS Trust, Colchester, Essex, UK
  1. Correspondence to Dr Daniel Darbyshire, Health Innovation One, Lancaster University Lancaster Medical School, Lancaster LA1 4YB, UK; dsdarbyshire{at}


Introduction Workforce issues prevail across healthcare; in emergency medicine (EM), previous work improved retention, but the staffing problem changed rather than improved. More experienced doctors provide higher quality and more cost-effective care, and turnover of these physicians is expensive. Research focusing on staff retention is an urgent priority.

Methods This study is a scoping review of the academic literature relating to the retention of doctors in EM and describes current evidence about sustainable careers (focusing on factors influencing retention), as well as interventions to improve retention. The established and rigorous JBI scoping review methodology was followed. The data sources searched were MEDLINE, Embase, Cochrane, HMIC and PsycINFO, with papers published up to April 2020 included. Broad eligibility criteria were used to identify papers about retention or related terms, including turnover, sustainability, exodus, intention to quit and attrition, whose population included emergency physicians within the setting of the ED. Papers which solely measured the rate of one of these concepts were excluded.

Results Eighteen papers met the inclusion criteria. Multiple factors were identified as linked with retention, including perceptions about teamwork, excessive workloads, working conditions, errors, teaching and education, portfolio careers, physical and emotional strain, stress, burnout, debt, income, work–life balance and antisocial working patterns. Definitions of key terms were used inconsistently. No factors clearly dominated; studies of correlation between factors were common. There were minimal research reporting interventions.

Conclusion Many factors have been linked to retention of doctors in EM, but the research lacks an appreciation of the complexity inherent in career decision-making. A broad approach, addressing multiple factors rather than focusing on single factors, may prove more informative.

  • emergency care systems
  • emergency departments
  • emergency department
  • management
  • HR management
  • training

Data availability statement

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

This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See:

Statistics from

Key messages

What is already known on this subject

  • Addressing the retention of emergency physicians has been identified as a high priority for research in emergency medicine (EM).

  • More experienced clinicians provide higher-quality and more cost-effective care. Keeping them in the specialty is vital now that efforts to improve recruitment have yielded positive results.

What this study adds

  • Dozens of factors that influence retention have been identified in the literature.

  • There is a real lack of studies looking at ways to improve retention in EM.

  • Future work should address complexity; understanding the multiple interacting factors associated with retention is more likely to be beneficial than replicating correlational studies.


In 2012, an editorial summarising the finding of an interim report from the emergency medicine (EM) task force (a multiprofessional group set-up by the UK government’s Department of Health to address workforce issue in EM) emphatically outlined the problem:

‘Speak it loudly and speak it clearly: the specialty of Emergency Medicine (EM) in the UK has a medical staffing crisis’.1

In the intervening years, many of the recommendations contained in the report have been instigated. Consultant numbers have increased across the UK.2 Recruitment to EM training has remained consistently above 85%.1 3 An alternative route into EM training has been developed.4 The number of clinical nurse specialists, advanced practitioners and physician associates within the ED has been expanded.4 Despite these successes, growth in ED attendances continues to outstrip that of the workforce.2 Problems with attrition from training programmes5 and exodus of established clinicians via early retirement6 mean that the workforce crisis may have changed, but it certainly has not been solved. Nor is it a problem unique to the UK. The landscape may be different, but staffing EDs is a problem worldwide.7–11 Similar stories can be told across other specialties: psychiatry12 13; paediatrics14 and general practice15 16 are examples. This issue also affects emergency nurses, the largest group working in EDs.17

Facilitating long and productive careers, be it for emergency physicians (EPs) or the other equally vital staff groups, is of paramount importance to sustainable long-term staffing of EDs. While perhaps self-evident, it is becoming increasingly apparent that more experienced clinicians provide higher quality and better value care for their patients.18–20 Much of the previous work related to EM careers has focused on reasons for leaving, with the literature on burnout, the challenging working environment and the impact of out-of-hours working continuing to develop. There is, therefore, a clear gap in the literature to view this problem from an alternative perspective: not to look at why people leave but to focus on why those who stay do so, despite the universal challenges. This review is part of a programme of work focused on addressing that gap,21 one of the key problems facing the specialty of EM.22

This review is framed in terms of retention, but the terms used in academic and policy documents are inconsistent and lack clarity. The different definitions of retention, expanded on in table 1, relate to efforts by,23–25 or the structure of,25 the employing organisation to keep staff, or the proportion of workers still with an organisation after a period of time.26–28 We use the term retention in reference to its dictionary definition. The Merriam-Webster dictionary has three descriptions which, when taken as a whole, clarify the meaning of the term ‘retention’ without positioning it too tightly within a specific academic domain.29 These definitions are ‘the act of retaining’, ‘the power of retaining’ and ‘something retained’.29 Retention therefore is something that can be done, can be done in a particular way and has been done. This brings us to ‘retain’, which is the transitive verb to the noun of retention. The definition ‘to keep in possession or use’ is helpful as it refers to both place and action.29 Our use of the term is not time specific, but we recognise that in certain contexts, where measurement is important, a more technical definition may be required.

Table 1

Definitions of retention from the limited number of sources which define the term

As well as retention, the literature contains a myriad of other terms which overlap in stated definition and usage with many being used interchangeably. For example, the word ‘attrition’ was frequently used interchangeably with the terms ‘dropouts’, ‘turnover’, ‘brain drain’, ‘losses’, ‘premature departure’ and ‘separation’.30 The commonly used terms for both staying in a role or leaving it are defined in table 2.

Table 2

Definitions for terms related to retention

Because of these definitional inconsistencies, the search included a wide selection of these terms. The included papers are those that address retention, as we have defined, regardless of the terminology used by the authors.


The protocol for this review was published in advance and is available open access.31 This paper focuses on the academic literature; the scoping review of the grey literature discussed in the protocol will be reported separately.

The aim of this study, aligning with the scoping review methodology,32 was to map the extent of the literature directly pertaining to retention of doctors in EM. More specifically, this involves identifying the types of evidence available, collating the key characteristics of papers, identifying the key definitions and concepts, and delineating and analysing the gaps in the literature. This is in keeping with the predetermined review question:

  • Primary question: What is known about retention of doctors in EM?

  • Subquestion 1: What factors have been studied relating to retention of doctors in EM?

  • Subquestion 2: What interventions have been implemented to improve retention of doctors in EM?

A search of MEDLINE, Embase, Cochrane, HMIC and PsycINFO was initially completed on 15 March 2019 by Helen Elwell, clinical librarian at the British Medical Association Library, and then updated for papers published in the interim, on 14 April 2020 (Cochrane and MEDLINE) and 21 April (Embase, HMIC and PsycINFO). This was supplemented by searches of Business Source Complete, Proquest Business Premium Collection and Emerald Insight. The search terms for Ovid MEDLINE are available in table 3, with the remainder in online supplemental appendix 1. Reflecting the nature of scoping reviews and the research questions, this search aimed for breadth of coverage. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) checklist is included in online supplemental appendix 2.

Table 3

Ovid MEDLINE search strategy

All searches were limited to English language. No date limitations were applied. Given the vast number of results, a team-based multistage approach was undertaken. Titles were reviewed by DD and clearly irrelevant items were excluded. Abstracts were then independently reviewed by DD and UB. To ensure consistency, this was piloted with tranches of 20 until complete adherence was achieved and reviewers were in frequent communication during the abstract screening process. Abstracts were reviewed against the inclusion criteria (see figure 1 and the protocol31), with those clearly not meeting the criteria excluded. Full-text articles were then accessed and again compared with the inclusion criteria; see figure 1 for inclusion and exclusion criteria and figure 2 for the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA) diagram.

Figure 1

Inclusion and exclusion criteria. *Type 1 EDs are ‘consultant led 24 hour service with full resuscitation facilities and designated accommodation for the reception of accident and emergency patients’.82

Figure 2

Preferred Reporting Items for Systematic Reviews and Meta-Analyses diagram.

The following data were extracted from the included articles: citation, methodology, factors influencing retention, efforts to improve retention and other findings relevant to retention of doctors in EM. Data are presented to summarise the different approaches to doctor retention in EM that are represented in the literature and to give a picture of where the gaps in the literature lie. Papers not pertaining to EPs and those that did not go beyond measuring a rate of retention (or attrition) were excluded. This is because defining the rate of retention in EM is not a research question for this study and would be best answered with a complementary methodology, such as systematic review and meta-analysis.


The result of database searching is presented in the PRISMA diagram (see figure 2). The studies excluded at the eligibility (full-text reading) stage, including rationale for exclusion, are summarised in online supplemental appendix 3. Methodological details and study characteristics of included papers are available in online supplemental appendix 4. A brief summary of each paper alongside the links to the research questions is in table 4.

Table 4

Summary of included papers

Factors influencing retention

The identified papers explored retention in different ways. Three papers analysed factors that correlated with intention to leave, five correlated with reasons for having quit or attrition rates,33–37 two with reasons for continuing EM work38 39 and in one case with reasons why EPs might leave or why they stay.40 Two of the papers found no practically applicable correlations: margin in life (a psychological theory of adult development) scale had no correlation with intention to quit,41 and while a low score on the Global Job Satisfaction instrument scale was correlated with leaving EM, its test characteristics meant that it was not a useful predictor of quitting.35 A pragmatic literature review described elements of EM that negatively impacted retention42; a case report and discussion explored incidents that may lead an EP to quit43; a letter44 commenting on an included study36 mainly reiterated points from the letter writers own study (also included in this review), which correlated burnout with intention to quit.45 The two remaining letters focused on sustainable careers in Ireland and Australia, respectively.46 47 The diversity of approaches used in the articles included in the review has led to a large number of different factors correlating with retention in EM; these are documented in table 5.

Table 5

Items related to retention, attrition or intention to leave EM

Efforts to influence retention

The majority of papers did not directly address efforts to improve retention.33 34 36–38 41 44 45 48 49 Of those that did, only three drew conclusions from empirical work.35 39 40 The participants in James and Gerrard’s study said that improving flow and staffing would improve retention,39 while those in the study from Fitzgerald et al. thought that the emergence of self-care and compassion dialogues may be beneficial.40 Lloyd et al. stated that work scheduling (rostering), as an extrinsic part of job satisfaction, is amenable to change and therefore has potential to improve retention.35 However, it should be noted that in this study, evaluating the predictive validity of two job satisfaction scales, while one of the scales they tested had a statistically significant correlation with attrition, we found that the test characteristics of this relationship mean it lacks predictive utility.

The other studies offered suggestions from a range of perspectives. One paper referenced documents on ageing and the EM workforce,50 while another offered an unreferenced statement that in some countries, ‘older doctors are not required to participate in after-hour rosters’.47 An editorial explained that streamlining training and a ministerial review of the broader medical workforce in Ireland, both ongoing when published in 2014, might help retention.46 The case report and discussion paper provided examples and references for stress management techniques that are relevant to EPs, which the authors postulated might help career sustainability.43 Smith and Dasan’s review paper highlighted measures to improve retention, reflecting some of the work of the previous section, specifically job planning, less than full time working, portfolio careers, appropriate remuneration and well-being.42 They then introduced sustainability work from the UK’s Royal College of Emergency Medicine (RCEM).

Other findings related to retention

The participants in the study by Fitzgerald et al. universally identified with the term ‘sustainability’ when discussing their emotional and physical status related to their work.40 This parallels the terminology used in the review by Smith and Dasan, also from the UK, along with the materials from RCEM that it references.42

In their study of residency trained US EPs, Hall and Wakeman found that clinicians tended to decrease clinical work and increase other types of work, such as teaching and administration, as their careers progressed.36 Takakuwa et al. found that policies related to ageing were inconsistent for the EPs approaching the final years of their career.50 While both studies were from the USA, this message is mirrored in Holmes’ opinion piece from Australia.47

In their large study of French EPs, Estryn-Behar et al. found that working conditions may be more important than pay.48 Related to money, Xu and Veloski38 and Xu et al34 found that having educational debt was associated with staying in EM in the USA.

A study from Brazil by Feitosa-Filho et al. found that 81.3% of EPs planned to stop working in the ED in the following 15 years.49 A letter by Mallon (from the USA),44 commenting on Hall and Wakeman (again from the USA),36 expressed concern about the possibility of overestimating attrition in the USA, leading to having too many trained EPs, leading in turn to job insecurity and a fall in the perceived value of EPs.


It has been over 20 years since the first paper on retention of EPs, identified by this review, was published,43 with a seeming trend of increased activity in this domain reflecting the growth of EM research globally.51 Despite this relative growth, the absolute number of papers is low, and those that have been produced display significant methodological heterogeneity. The most frequently used methodological approach has been measurement, using a pre-existing scale of a psychological construct and testing to see if it is correlated with retention (or a term related to it).39 41 45 48 Burnout is the most assessed construct,45 48 reflecting the prominence of burnout research in both the EM52 53 and wider medical literature.54–56 Again, reflecting the wider medical literature on burnout, problems arise with definitions and interpretations of the term, different cut-offs used for the threshold for defining burnout, different burnout inventories used and type I errors (false positives) when multiple tests for correlation are undertaken.57

Despite these issues, it is useful that two studies from different continents, using two different validated measures, have linked burnout with retention (both via intention to leave),45 48 a finding that is replicated in the nursing profession,58 teachers59 and volunteers.60 Margin in life (a psychological theory of adult development) was not correlated with intention to leave41—the measure is most often correlated with readiness for change such, as organisational restructuring or merger.61 62 While global job satisfaction was correlated with attrition, Xu et al. found that the correlation was not strong enough to use the scale predictively,34 a finding consistent with the broader human resources literature, which finds that intrinsic job satisfaction is negatively correlated to turnover, whereas extrinsic job satisfaction has no statistically convincing link.63

The second prominent group of studies measured aspects of work life and alongside either attrition33 36 or intention to leave,49 or described policies related to retention in the final third of an EP’s working life.50 Of the many aspects of work life that Feitosa-Filho et al. assessed, job satisfaction— measured as a single multiple choice question with the options ‘satisfied’, ‘neutral’ and ‘dissatisfied’—was the only one showing a statistically significant correlation with intention to leave.49 The study by Lloyd et al., discussed earlier, linked job satisfaction and quitting EM but not strongly enough to offer a predictive test.35 Feitosa-Filho et al. found that 64% of their EPs who were satisfied and 94% who were dissatisfied intended to quit in the next 15 years; however, the baseline characteristics of their study from Brazil make it equally difficult to apply a different practice setting.49 This does not mean that job satisfaction should be discounted—there is a long history from economics marking satisfaction as a ‘major determinant of labour market mobility’64—and it has been linked with concepts related to retention across several professional groups, including nurses,65 general practitioners,66 physician assistants and nurse practitioners.67 The second aspect of work life relates to training, with board certification (postgraduate specialty examinations in the USA) and fellowships (a period, generally a year, of subspecialty training related to the primary training specialty) correlating with lower attrition.33 36 What it is about fellowship or board certification that influences attrition is not clear, but other studies have linked high academic achievement while at medical school34 and the intellectual content, specifically diagnostic challenges, of the specialty as important.38 These features can be threatened by a lack of training or supervision, excessive workloads and poor working conditions.37

Most of the studies examined retention from a broad, though necessarily superficial perspective. However, two studies took the opposite approach, gaining in-depth accounts from a relatively smaller number of participants.39 40 Describing the physical and emotional strain of working in the ED as ‘unsustainable’ adds credence to the idea that psychological measures (such as burnout) may have utility in efforts to improve retention while simultaneously suggesting that such measures may be an oversimplification. The more social aspects of EM, such as the flattened hierarchy39 and peer social support,40 move the discussion away from the individual approach to retention to the idea that the interactions between the people involved in the work of EM might be key.

The papers in this study support the notion that pay is linked to retention26 with higher income correlating with lower attrition36 and with those who leave the specialty having had lower incomes than those who stayed,33 though this finding could be skewed by salaries generally rising with career length. Income was reported as a major factor in decisions to stay in EM.38 Educational debt is another factor, representing a strong correlate with staying in EM in one study34 and a minor factor in another.38 It should be noted that these studies are from the USA, where both income for doctors and educational debt are significantly higher than most other counties, with the study by Estryn-Behar et al. from France concluding that ‘working conditions may be more important than pay’.48 The relationship between pay and retention is more complex than a linear correlation, so that even with high pay, ‘pay dissatisfaction can lead to turnover’.26 Other factors, beyond the amount of remuneration received, make pay more complex with perceptions of fairness being the most important. This is described at two levels. This first, distributive justice, refers to the distribution of pay within an organisation,68 while the second, procedural justice, is about the process through which pay is administered,69 with both repeatedly linked to retention both within69 and outside of healthcare.26 68

While some of the studies presented here discussed aspects of work that may be amenable to change, in order to improve retention, none tested this as a hypothesis directly. The lack of interventions in the academic literature may be due to them being reported elsewhere. It is highly unlikely that a change to a single aspect of work influencing retention would lead to measurable change—the required number of participants and scale of impact would likely be too large to be feasible. Moving towards recognising, studying and implementing change with complexity,70 rather than imposing false notions of simplicity, will be key to any successful interventions, something that the review by Smith and Dasan alludes to.42

The concept of career change or evolution may be more closely aligned to careers in EM than the more linear concept of promotion, demotion and resignation. Hall and Wakeman found that clinicians decreased clinical work and increased other work though their career.36 Portfolio careers, here meaning role diversity within a profession rather than the definition more common outside of healthcare—‘individuals develop a portfolio of skills that they sell to a range of clients’,71 are gaining increasing prominence in discourses about health professions careers.42 72 The idea that using skills developed through professional training and experience in related roles helps prevent people getting bored or jaded has strong face validity and, while there is a small body of research supporting this, the findings are not conclusive.16

A recent body of work published by the UK medical regulator (General Medical Council (GMC)) started with the premise that patient safety is dependent on doctors’ well-being, integrated a summary of the existing academic literature with case studies and developed the ABCs of doctors’ core needs; the findings of this review can be mapped to the ABC structure (see figure 3). While not the primary aim of this work, it is clear that retention was within its broader remit, with the foreword from the chair of the GMC stating that ‘If we act together we will avoid losing good doctors and seize a golden opportunity to tackle the challenges the health service must meet now and in the future’.73

Figure 3

Review findings mapped to the ABC of doctors' core needs.

Referring to the table of definitions (table 2), we found that while the papers use terms related to retention, there is a universal lack of clear definitions; for example, Hall et al. use the term ‘career longevity’ without defining it.33 Estryn-Behar et al. do not define ‘intention to leave’, but they do state the question they use to measure it, and in other instances, the terms related to retention come from interview participants rather than the researcher.39 40 Lloyd et al. use an approach that avoids clearly defined terms, instead using short descriptive statements such as ‘left job and EM altogether’ consistently.35 However, the general lack of clarity in terms related to retention is a real weakness of this body of literature. Terms such as ‘attrition rate’, meaning different things in different papers, if indeed they are clearly defined at all, are a potential source of confusion and misinterpretation. The table of excluded papers (online supplemental appendix 3) reinforces this confusion, with papers relating to intention to leave referring to the workplace rather than the profession, and many papers presenting an estimate for levels of intention to leave or turnover, which, while useful in that specific context, does not help with developing understanding of retention.

The scoping review process has inherent limitations; we have described the factors that influence retention but not the scale of influence of each factor. The breadth of types of papers meant that several different quality appraisal tools would have been required to do this and a decision was therefore made that this would not have added significantly to the current study.

The literature related to retention of doctors in EM yielded a variety of factors with complicated and mostly unclear interactions. Interventions to improve retention have a very limited research base. Linked to the factors influencing retention, it is likely that programmes to address a single issue are unlikely to be effective; instead, holistic approaches cutting across the multiple domains of work life should be trialled. Future research needs to embrace this complexity rather than try to eradicate it.

Data availability statement

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


The authors acknowledge Helen Elwell, Clinical Librarian at the British Medical Association Library, for her expertise in contributing to the search development.


Supplementary materials


  • Handling editor Richard John Parris

  • Twitter @dsdarbyshire, @richardbody

  • Contributors DD conceived the project. DD, LB, RB and DG developed the initial proposals for funding. DD, LB, DG, RI and RB developed the protocol. DD and UB conducted the search and extraction. DD produced the initial draft of the manuscript. LB and DG are providing direct PhD supervision for DD. RB and RI are supporting the PhD with specialty and institutional expertise, respectively. All authors contributed to and revised the final manuscript.

  • Funding DD is funded by a National Institute for Health Research Doctoral Fellowship for this research project. Initiation of the study was supported by the BMA Foundation Kathleen Harper Award and the RCEM Young Investigator Award.

  • 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.

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.

Linked Articles