Identifying relevant studies

Several electronic databases were searched to identify relevant studies up to July 2019: Medline, CINAHL, PsychINFO, and PubMed. Keywords chosen in the search included ‘midwife’ (and its variations, e.g. midwifery, midwives), in combination with burnout-related terms including ‘burnout,’ ‘exhaustion,’ and ‘compassion fatigue.’ The search strategies were tailored for each of the databases’ thesaurus terms and field headings. A total of 1034 articles were identified (Medline=219; CINAHL=203; PsychINFO=151; PubMed=461). Eliminating all non-English articles and removing duplicates reduced this number to 598 articles eligible for title and abstract review. From these, 92 articles were deemed to be relevant and underwent closer review. Articles were evaluated on the following inclusion criteria: 1) Article must be on practicing midwives (e.g. studies involving nurse-midwives were included, but articles on student midwives or retired midwives were excluded); 2) Article must report on burnout among midwives; if other healthcare providers were included, results must be stratified, so that midwifery-specific results can be extracted; 3) Article must identify associated factor(s); and 4) Article must be written in English and have full-text available. A total of 27 articles met these criteria (Table 1).

Figure 1

Search strategy flowchart

Charting the data

Authors RS and BS collaborated extensively via a shared spreadsheet to review inclusion criteria and select articles, with any and all differences being settled through discussion or input from the supervising author, KS. Articles meeting the inclusion criteria were reviewed and data were extracted and charted pertaining to study setting, study design, burnout measures, study results, factors associated with burnout, and recommendations to address burnout. Factors that were significantly linked to burnout in quantitative studies were independently extracted by the first two authors and are summarized in Table 2. The supervising author reviewed all data points reported in Tables 1 and 2.

Measuring burnout

The most commonly used measures to assess burnout were the Maslach Burnout Inventory (MBI), utilized in 11 of the studies^{11,16,22-24,29-34}, and the Copenhagen Burnout Inventory (CBI), which was used in 12 of the 27 studies^{3,7,13-15,17-21,27,28}. See Table 1 for a full list of instruments that were used to assess burnout.

The Maslach Burnout Inventory includes 22 items that measure burnout along three dimensions: emotional exhaustion, depersonalization, and reduced personal accomplishment. Each item or statement is assessed on a 7-point Likert scale. The 9 items on the emotional exhaustion subscale measure feelings of being emotionally drained and exhausted by work. The 5 items on the depersonalization subscale assess the degree to which people are impersonal in their treatment of clients/patients³⁵. Personal accomplishment is measured with 8 items that assess feelings of competence and achievement with respect to work. Scores on the MBI subscales are always reported separately, whereas the CBI has both a full-scale score and three subscale scores: personal burnout, work-related burnout, and client-related burnout. Personal burnout is measured with 6 items that assess general burnout and can be completed by anyone, regardless of occupational status. Work-related burnout is measured with 7 items that ask respondents to rate the degree of physical and psychological fatigue related to work. The client-related burnout subscale includes 6 items that measure fatigue and exhaustion related to caring for others³⁶.

Prevalence of burnout

All authors reported on the prevalence of burnout in their study population, using either proportions, means or medians. Prevalence of work-related burnout (as measured with the Copenhagen Burnout Inventory) was highest among Australian midwives, especially those working in non-caseload/continuity models (median scores of 48.4 and 46.4)^14,21 and Western Canadian midwives (median score 46.4), and lowest among Norwegian midwives (mean score 19.1)⁷. Scores on the Maslach Burnout Inventory were lowest among midwives in Turkey (mean score on the emotional exhaustion (EE) subscale was 13.9) and the Netherlands (mean score 19.9)²⁹, and highest among midwives in Senegal (80% scored in the moderate to high range on the EE subscale)¹⁶. Prevalence of burnout varied by model of care, with case loading midwives consistently reporting less burnout than midwives who work in other models^{13,15,17,18,20,21}. Most of this evidence comes from Australia and New Zealand.

Studies utilizing burnout inventories were able to expand on their findings by describing the prevalence of burnout subdomains. Of those using the MBI, the ‘emotional exhaustion’ subscale emerged as the most frequently cited dimension of burnout, followed by ‘depersonalization’, and then ‘personal accomplishment’^11,31,32. Respondents with a high score in emotional exhaustion and depersonalization, and a low score in personal accomplishment are considered severely burnt-out. In their sample of 56 Australian midwives, Mollart et al.¹¹ found 60.7% to have moderate to high levels of emotional exhaustion, higher than in other studies from the US (41.8 %)³⁰ and Iran (41.9 %)³¹ but lower than those reported by midwives in Senegal (80%)³⁷. A similar pattern emerged amongst studies using the CBI: ‘personal burnout’ was the most prevalent dimension of burnout, followed closely by ‘work-related’ burnout, and then ‘client-related’ burnout in a distant third-place^3,7,19-21,27. Stoll and Gallagher³ reported the moderate-to-severe burnout prevalence of these subscales amongst Canadian midwives as 74.9%, 42.5%, and 20.3%, respectively, compared to Fenwick et al.²⁰ who reported these figures as 64.9%, 43.8%, and 10.4%, respectively, among Australian midwives.

Factors associated with burnout

After reviewing the articles, 26 factors that were significantly associated with burnout were identified (Table 2). These included both organizational factors (e.g. insufficient organizational support /distress related to organizational structure /poor or stressful work environment /poor perceptions of practice environment, not enough time off, poor pay) as well as individual factors unique to the midwife (e.g. young age, less work experience, marital status).

The most commonly reported factors linked to burnout were: insufficient organizational support for profession; poor or stressful work environment^{3,7,15,27,28,30,33}; working in non-case-load/non-continuity models of care (such as hospital shift work)^{13,15,17,18,20,21}; less work experience in maternity care^{11,19,27,29,30,34}; younger age^{7,15,19,27,30,31}; high workload/not enough time off^{3,15,23,27,30}; trauma/stress experienced at work^22,25,26,37; interpersonal conflict with colleagues/low recognition of midwives^15,27,28,34. For a full list of factors, see Tables 1 and 2.

Recommendations

Most authors offered recommendations for improving working conditions for midwives and reducing the prevalence of burnout. There was a diversity of suggestions offered to reduce midwife burnout, many of which overlap with and appear to target factors identified in Table 2.

The most widely reported recommendations were: offering more work-related education, improving organizational support, and working in a caseload-model of midwifery practice. Other recommendations included: better support after traumatic events, education for midwives to learn ways of coping with occupational stress and inter-professional education or programs to reduce inter-professional bullying and conflict. Less frequently reported suggestions included: offering part-time work options or career development opportunities, and promoting exercise and physical activity, as a way to reduce stress. See Table 1 for a full description of recommendations.

Burnout measures

We used different assessment tools to measure burnout. The most commonly used tools for measuring burnout were the MBI³⁵ and CBI³⁶. An important difference between these two scales is their theoretical underpinnings: the MBI describes burnout as a syndrome of depersonalization, reduced personal accomplishment, and emotional exhaustion as related to ‘people work’³⁵, whereas the CBI describes burnout as ‘fatigue and exhaustion’ resulting in personal, work-related and client-related burnout³⁶. While these differences may appear to be nuanced, they ultimately make it difficult to make direct comparisons between burnout studies. We recommend that future studies consider using the CBI in preference to the MBI for several reasons including: ‘depersonalization’ is perhaps better seen as a coping mechanism rather than a dimension of burnout, the MBI questions do not always adapt well to diverse cultures, and to reduce potential issues regarding distribution rights as some versions of the MBI are not in the public domain³⁵.

Burnout factors

In total, 26 factors were associated with burnout in the included literature and were reported in this scoping review. These factors could be broadly stratified into: 1) sociodemographic or lifestyle factors of the midwife (e.g. age, activity level, physical health, parental and relationship status); and 2) systemic and organization factors that affect the midwife (e.g. level of autonomy, inadequate facilities, low wage). The most widely supported factors for burnout included an approximately equal proportion of these two categories, suggesting that of the two there is no single domain that is disproportionately associated with burnout in the profession. ‘Less work experience in maternity care’ and ‘younger age’ are interrelated factors that emerged as the sociodemographic characteristics receiving support from the greatest number of articles. ‘Insufficient organizational support /poor or stressful work environment’, ‘practicing in non-caseload/non-continuity models of care’ (such as hospital shift work), and ‘high workload’, were the most prominent systemic and organizational-related factors found in the literature. Interventions for addressing burnout among midwives, therefore, may wish to consider avenues for reducing workload (e.g. hiring more midwives on staff, enabling case-load midwives to take fewer clients) – especially for midwives that are of young age or are early in their careers. Closely related to reducing work-load is remuneration of midwives. Higher pay enables midwives to take on fewer shifts or carry a smaller caseload.

The largest global survey of the midwifery work force to date revealed that midwives are deeply committed to their work, but experience many challenges. For example, midwives across low-, middle- and high-income countries reported loss of autonomy and power within the healthcare system as major barriers to providing high quality care. Disrespect from senior medical staff, low pay and lack of leadership opportunities were commonly reported and illustrate the uphill battle many midwives face⁴⁰. Findings from this report and the current scoping review demonstrate the need to improve working conditions for midwives, so they can continue providing high quality care while also enjoying better work-life balance, emotional well-being and mental health⁴⁰.

By their nature, factors characteristic of the midwife can be difficult and often impossible to change. Accordingly, hospital administrations and clinics should first look inward and address the policies and organizational factors that might contribute to burnout among midwives and other health professionals. A recent publication from the UK supports this point. Of the close to 2000 midwives who participated in an online survey, 83% reported personal burnout, and 67% work-related burnout⁴¹, placing UK midwives at the top of the list when compared to burnout scores from other high-income countries that utilized the CBI (Table 1). Perceived resource inadequacy was the strongest predictor of work-related burnout. Other factors associated with burnout included younger age (40 years or less), less work experience (<30 years), having a disability and reporting low levels of support from midwifery managers and low professional recognition⁴¹. These results are in line with the main findings of this scoping review.

Several additional contributing factors for burnout were identified in one of the studies. These included poor self-rated physical or mental health^27,38, low levels of physical activity¹¹, night shifts only compared to mixed shifts (day and night)¹¹, practice location (home versus hospital)¹³, and coping style¹³. Future research should consider including these factors in their investigations, to better understand the role they may have in contributing to burnout.

Additionally, researchers examining burnout in midwifery may also want to consider factors known to contribute to burnout in physicians, including fears of litigation, the increasing reliance on technology, and the growing uncertainties regarding the future of medicine; none of these factors was explicitly explored in the included articles. Research on physician burnout also delves into how the nature of a physician’s tasks influences and aggravates their symptoms of burnout. Similar to the findings from this review, a study on Canadian physicians showed that 64% feel that their workload is excessive, and that 48% reported that their workload had increased in the past year⁴². One possible explanation for this perceived increase in workload has been the increase in mundane or clerical tasks, which have been shown to compromise a physician’s sense of job satisfaction³. Perhaps the same can be applied to midwives: clerical tasks and other responsibilities, not directly related to healthcare, may cause them to experience a high workload and contribute to burnout. Two of the included studies^37,38 reported significant linkages between burnout and low task satisfaction. Further research on the impact of non-healthcare tasks have on the overall workload of midwives is warranted.

It is important to mention that the frequency of specific burnout factors across studies does not necessarily translate into how much significance (or ‘weight’) they should be assigned. A widely identified burnout factor may contribute only a small portion towards the prevalence or severity of burnout in midwifery (and vice versa for a seldom mentioned factor). This uncertainty on the weight each factor should be given is further compounded by the ambiguous or ‘broaden-compassing’ nature of many of the specified factors. For example, in the case of ‘high workload’ it is not always clear how or by what measure workload is being considered or interpreted by survey respondents. This area of uncertainty regarding how burnout factors interplay, specifically how to aggregate or separate factors, and to what degree individual factors contribute to burnout would certainly benefit from additional qualitative research.

Finally, this global scoping review included studies from many different high-resource countries and one low-resource country. The context for midwifery practice is very different across countries and affect the way midwives experience work and work-related burnout. It is of interest to note that several of the included studies were conducted as part of an international working group of midwives and researchers called WHELM – Work, Health and Emotional Lives of Midwives¹⁴. The WHELM group uses a standardized survey to collect data about burnout, occupational stress, intentions to leave, quality of life and other factors, from midwives in many high-resource countries, including Australia, New Zealand, the UK, Canada, Germany, and Scandinavia. In the future, data from midwives from different countries participating in the WHELM study might be pooled, to better understand similarities and differences in how burnout is experienced by midwives in different countries.

Burnout recommendations

‘Offering more work-related education’ emerged as a commonly cited recommendation. However, the content and purpose of suggested additional education varied across the studies. Offering more work-related education included: further training on increasing autonomy and clinical decision making¹⁹, educational workshops on preventing and addressing work-related burnout³¹, and ongoing education and clinical mentorship, to increase clinical competency and build confidence^19,37.

Offering midwives the option to practice a caseload model of midwifery care was found to be associated with reduced burnout. In a recent cross-sectional survey of 542 Australian midwives, a direct comparison between caseload and non-caseload midwifery revealed the latter group scored higher on all three CBI subscales of burnout (p<0.001)¹⁷. These results are supported by other studies demonstrating that midwives practicing caseload midwifery experience less burnout than those working in non-caseload models^{13,15,18,20,21}. These findings present a clear and unanimous recommendation for policy makers and healthcare administrators seeking to address burnout through changes in practice models by supporting caseload models. Studies about caseload midwives can be informative in terms of anticipating factors linked to burnout among midwives who practice this model of care. For example, Stoll and Gallagher³ studied case-load midwives in Western Canada. Their recommendations to reduce burnout were based on open-ended comments from midwives about strategies to reduce burnout and increase job satisfaction. Recommendations fell into four general areas: more time off /better pay, more flexible practice structures /change in model of care, more respect from profession /more support from colleagues and more support with professional issues (such as help with obtaining hospital privileges or more support after critical incidents). Specific recommendations within these four areas included: part-time work options, support for sick days/vacation coverage, more pay per course of care and more pay for complex clients, salaried rather than fee-for-service payment schemes in rural and remote areas, more off-call career opportunities, and initiatives to reduce intra-professional bullying and inter-professional conflict³.

Similar to the concerns raised in the previous section, the level of support for a particular burnout-reduction strategy across the included literature may not necessarily be representative of its success or effectiveness in reducing burnout, should it be implemented. These inquiries would be best addressed with additional investigation, in particular from studies utilizing longitudinal designs evaluating the efficacy of these recommendations.

Burnout is a complex issue that requires complex interventions. Because it is an occupational phenomenon, the onus for change is placed on organizations, such as hospitals, professional organizations, health policy makers and regulatory bodies. Finally, this review uncovered some inconclusive findings. For example, two studies identified that being married is linked to burnout^24,38 whereas two other studies identified being single or divorced as a risk factor^7,30. Similarly, high work-load and long hours were linked to burnout in some studies^3,15,23,30 but in another study midwives who worked more hours per week also reported higher scores on the personal accomplishment subscale of the MBI²⁹. Future studies might shed light on how these factors relate to burnout, ideally using qualitative study designs. Such designs can also elicit detailed responses about the kind of partner support that buffers against burnout and how work autonomy relates to burnout.

Limitations

This scoping review is not without its limitations. First, given the evolving nature of occupational burnout as a unique state distinct from other psycho-social constructs, the search terms used for this review may not have been fully inclusive of all the terminology used – presently and historically – to describe burnout. This may have resulted in certain relevant papers not being considered for inclusion. Our review was also limited by screening out articles that were not available in English, or those that were not accessible by database subscriptions held by the University of British Columbia. Further, there may be publication bias as a consequence of studies with significant findings being preferentially selected by journals for publication – and hence inclusion in this review. Finally, this scoping review did not include a quality assessment of articles.