Physical activity, burnout and quality of life in medical students: A systematic review

Abstract Background Medical students are at risk of burnout and reduced quality of life (QoL). The risk of burnout doubles from third to sixth year of medical school, and medical students have an 8%–11% lower QoL than nonmedical students. It is imperative to prevent this, as burnout and reduced QoL is independently associated with errors in practice. This systematic review aims to examine whether physical activity/exercise is associated with burnout and/or QoL in medical students. Methods Articles were identified through database searches of Embase, Medline, PsycINFO, Scopus and Web of Science. Studies were included if both physical activity/exercise and burnout or QoL were measured and limited to those focussing on medical students. Risk of bias was assessed using accredited cohort and cross‐sectional checklists. A narrative synthesis was conducted due to heterogeneity in the dataset. Findings Eighteen studies were included, comprising 11,500 medical students across 13 countries. Physical activity was negatively associated with burnout and positively associated with QoL. Furthermore, the findings were suggestive of a dose–response effect of physical activity on both burnout and QoL; higher intensities and frequencies precipitated greater improvements in outcomes. Conclusions This multinational review demonstrates that physical activity is associated with reduced burnout and improved QoL in medical students. It also identifies a paucity of research into the optimal intensity, frequency, volume and mode of physical activity. Further research, building on this review, is likely to inform the long overdue development of evidence‐based, well‐being curricula. This could involve incorporating physical activity into medical education which may improve well‐being and better prepare students for the demands of medical practice.


| BACKGROUND
Burnout and quality of life (QoL) are two important concepts which affect well-being.
Burnout is 'a syndrome conceptualized as resulting from chronic workplace stress that has not been successfully managed'. 1 There are three founding domains of burnout: emotional exhaustion (feelings of exhaustion due to stress), depersonalisation (distancing and impersonalising of one's work) and personal accomplishment (feelings of achievement and competence in work). 2 The World Health Organisation defines QoL as 'an individual's perception of their position in life in the context of the culture and value systems in which they live and in relation to their goals, expectations, standards and concerns'. 3 Burnout and QoL are interrelated, with evidence of significant associations between QoL and all three domains of burnout. 4 Recent findings show that 31.5% of UK doctors have 'high burnout'. 5 Furthermore, both burnout and reduced QoL in physicians have been independently associated with errors in practice and thus patient safety. [6][7][8] Burnout has also been associated with reduced professional work effort in doctors. 9 Medical students are particularly at risk of burnout and/or reduced QoL due to stressors experienced during training, including time pressure, coping with death and suffering, workload and maintaining work-life balance. 10 Medical students show an 8%-11% decreased QoL compared with non-medical students of the same age. 11 Additionally, the risk of burnout has been shown to double from third year to sixth year of medical school. 12 Even small changes in burnout domains have been linked to a 7% increase in 'serious thoughts' of dropping out in the next year. 13 This provides a strong impetus for improving well-being strategies which students can use both in medical school and throughout their careers as doctors. There has been increasing interest in the integration of well-being curricula to undergraduate medical education and more widely in health education 14 with varying approaches such as mindfulness, 15 lifestyle education 16 and resilience training. 17 Medical students are particularly at risk of burnout and/or reduced QoL due to stressors experienced during training.
One possible strategy is increasing medical students' levels of physical activity ('any bodily movement produced by skeletal muscles that requires energy expenditure' 18 ) and/or exercise ('a subcategory of physical activity that is planned, structured, repetitive and purposeful in the sense that the improvement or maintenance of one or more components of physical fitness is the objective' 19 ).
Indeed, physical activity has been linked to improvement in all domains of QoL, 20 and an exercise programme in medical residents and fellows has been shown to significantly raise QoL. 21 Furthermore, a systematic review of multiple professions demonstrated a negative relationship between physical activity and the key component of burnout, emotional exhaustion. 22 To date, there has been no systematic review demonstrating the extent to which, if any, physical activity (including exercise) has an effect on burnout/QoL in medical students. Understanding this is an important step in informing future development of evidence-based well-being curricula.
Understanding this is an important step in informing future development of evidence-based well-being curricula.

| Aim and objectives
This systematic review aims to provide clarity on whether physical activity/exercise has a role to play in the well-being of medical students. Specifically, 1. Is physical activity/exercise associated with burnout in medical students?
2. Is physical activity/exercise associated with QoL in medical students?

| Protocol
The systematic review protocol was registered on PROSPERO (https://www.crd.york.ac.uk/prospero/display_record.php?ID= CRD42020182616). The PRISMA checklist has been followed in the review process. 23

| Eligibility criteria
Inclusion criteria are as follows: • Medical students, defined as individuals enrolled in an undergraduate medicine degree course at a higher education institution

| Study selection
Search results were exported into Endnote, reference management software. Titles and abstracts were screened to produce a shortlist according to the inclusion and exclusion criteria. Full texts were then obtained and reviewed to achieve a final list of studies for inclusion.

| Data collection
Data extraction was completed using a pre-designed extraction form.
Author and date, country, number of participants, age, gender/sex, year of study, physical activity/exercise measure, burnout and/or QoL measure, other outcome measures and a summary of results were collected.

| Risk of bias assessment
Cross-sectional studies were critically appraised using the Center for Evidence-Based Management (CEBMa) checklist. 24 Cohort studies were critically appraised using the Critical Appraisal Skills Programme (CASP) cohort study checklist. 25 Questions 7 and 12 referring to the results and implications for practice were removed from the CASP checklist; these could not be categorised for inclusion in the table however are considered in detail in the results and discussion of this review. An additional question referring to response rate was added to allow comparison to the cross-sectional studies. No studies were excluded based on critical appraisal.
All stages of the review were undertaken by two authors (CT & ES) working in parallel. Any disagreements were resolved by discussion.

| Synthesis of results
A narrative analysis was conducted for the main outcomes of burnout and QoL. Subgroups within the data were identified and analysed: pre-clinical versus clinical students, physical activity/exercise intensity/volume/frequency/mode and questionnaire measures. Due to heterogeneity of study measures used and reported statistics, a metaanalysis was not possible.

| Risk of bias assessment
Supporting information Tables S1 and S2 display the results of quality assessment including response rates. A 50% response rate was considered 'good'. Overall quality of the studies was reasonable. An area of weakness for almost all cross-sectional studies (n = 15) was the lack of a sample size based on considerations of statistical power. Several studies (n = 8) did not use confidence intervals for the results.
The average response rate was 57%. Three studies had response rates of less than 30%. 26-28

| Measures
Physical activity/exercise was commonly measured with a questionnaire developed individually for the study (n = 11). Three used the Godin Leisure Time Exercise Questionnaire, 29 three the International Physical Activity Questionnaire 30 and one the Simple Lifestyle Indicator questionnaire. 31 For transparency, the terminology used in each study has been transferred to this review (physical activity or exercise) in order to acknowledge the difference between the two terms.
The Maslach Burnout Inventory (MBI) 32 or an adapted form of this, was the most common burnout measure (n = 9). Three used the MBI student survey, 33 one the MBI general survey 32 and one the MBI human services version. 32 One used the Oldenburg Burnout Inventory student version 34 and one used the Burnout Measure short version. 35 For QoL, three used the SF-36. 36 Three used the WHO-QOL-Bref version. 37 Two studies developed a questionnaire for their study. The VERAS-Q 38 was also used as an additional measure in one study.

| Burnout and physical activity/exercise
Burnout findings are summarised in Table 1. Six studies 26,39-43 found a negative association between physical activity/exercise and all burnout components reported. Correlations ranged from À0.20 to À0.44; these are small-to-medium effect sizes. 44 Three studies 27,28,45 found that physical activity/exercise was associated with some burnout components but had no effect on others. Two studies 45,47 found no relationship between exercise and burnout.
Nine of the 11 burnout studies directly measured emotional exhaustion, the key component of burnout. 2 Of these, three 26,39,40 found negative correlations between exercise and emotional exhaustion ranging from À0.272 to À0.44. Two 27,42 found that low levels of physical activity significantly predicted higher emotional exhaustion scores. However, four studies 28,45-47 did not find a significant relationship between physical activity/exercise and emotional exhaustion.

| QoL and physical activity
QoL findings are summarised in Table 2. Six studies 40,[48][49][50][51][52] found that physical activity/exercise improved all QoL domains measured, while two studies found that physical activity improved some, but not all, of the domains. 53 42 Cross-sectional n = 731 Godin leisure-time questionnaire MBI Participants who had burnout had lower exercise scores (p = 0.034). Participants who reported they 'often' did exercise were less likely to experience burnout than participants who reported 'sometimes' or 'rarely' (P < 0.001).
Additionally, those who reported performing regular exercise were less likely to experience burnout than those who reported they 'rarely' High emotional exhaustion was predicted by low reported exercise frequency (never/rarely versus often), p < 0.001.
Macilwraith et al. No association was found between leisure-time exercise and cynicism. There was also no significant correlation between exercise and burnout.
Youssef et al. 95% CI, p < 0.0001) Mean emotional exhaustion, depersonalisation and burnout were lower in students who met the aerobic exercise guidelines (P < 0.01). Mean emotional exhaustion score and prevalence of high emotional exhaustion and burnout were lower in those who met the CDC strength guidelines (P < 0.0001). These parameters were also lower for students who met both aerobic and strength guidelines.
Mean QoL scores were higher for those following the CDC guidelines:

| Pre-clinical versus all med students
Four studies recruited pre-clinical students only. 39 found that a low intensity of physical activity in comparison to high was a significant predictor of higher emotional exhaustion score.
Additionally, moderate or low intensities of physical activity were significant predictors of a lower personal accomplishment score, whereas high levels were not. 27 A third study 50 found a significant association between both moderate and high levels of physical activity and better QoL for all components. Low activity level was also significantly associated with most QoL components, except physical health and social relationships. 50 Six studies used a continuous variable for exercise. 26,28,39,40,45,53 Three of these studies 26,39,40 found a negative association between increasing exercise and either exhaustion 40 or total burnout. 26,39 Additionally, a low exercise level was a significant unique predictor of burnout, 26 exhaustion and low professional efficacy. 28 For QoL, significant correlations between total physical activity level and QoL (including all subdomains) were found. 53 Additionally, moderate physical activity was a significant predictor of QoL. 53 A second study found a weak positive correlation between professional efficacy and increasing physical activity levels. However, they found no correlation between increasing physical activity and emotional exhaustion or cynicism. 45 Four studies recorded the frequency of physical activity/exercise without differentiating for intensity. 42,48,49,52 It was found that physical and mental component scores had a significant dose-response effect with scores increasing as frequency of physical activity increased. 49 Additionally, greater physical activity frequencies (3-

4/4 + per week) had significantly higher scores in all four domains of
QoL. 52 This was supported by another study which found that midterm QoL was predicted by greater exercise frequency. 48 Students with burnout had lower exercise levels. 42 Additionally, those who 'often' exercised were less likely to experience burnout than those who 'rarely' or 'sometimes' exercised. 41 Only one study 41  Four studies 43,46,47,54 recorded physical activity/exercise with a 'yes/no' or 'agree/disagree' method and therefore did not provide information on intensity, volume, frequency or mode.

| Key findings
This systematic review aimed to discern whether physical activity/ exercise is associated with burnout and/or QoL in medical students.
The findings suggest that physical activity/exercise is associated with reduced burnout and increased QoL. Additionally, the data indicated that although all levels of physical activity/exercise can precipitate improvements in QoL and burnout, higher intensities and frequencies may be required for the greatest effect.
The findings suggest that physical activity/exercise is associated with reduced burnout and increased QoL.
Despite most studies supporting these findings, two studies 46,47 drew contradictory conclusions. This may have been due to method of exercise measurement; the studies used a binary 'yes/no' measure.
Wide variation may have been present within the 'yes' category, ranging from walking to athlete training. This may explain the lack of a difference between yes and no groups. Although two other studies 43,54 also used dichotomous reporting styles, these were supportive of the consistent finding that physical activity/exercise reduced burnout. This may have been because they specified regular physical activity/exercise in their questions, diminishing this issue to some extent.
A wide range of tools were used to measure the outcomes of interest in this review. For example, the MBI is designed to assess the three dimensions of emotional exhaustion, depersonalisation and personal accomplishment. 55 Adaptations have been developed; the MBI-HSS is most applicable to human services jobs, such as the medical profession. Additionally, the MBI-GS is applicable to roles without a large human service element, and the MBI-SS is for students who are not in full-time employment. 33 The studies investigating burnout which found mixed results or no effect of physical activity used either the MBI-GS or MBI-SS. 28,[45][46][47] It is plausible that this observation is due to the questionnaires being less applicable to the medical student population. Despite 'student' status, medical students spend a large proportion of time with patients in the future workplace.
Non-MBI measures were also used, complicating results interpretation. The Burnout Measure is highly correlated with the emotional exhaustion subscale of the MBI, 56 while the OBI has been validated 57 as a measure of two dimensions: disengagement and exhaustion. The studies using these measures supported the majority findings so are unlikely to have been influenced by questionnaire choice.
Differences have also been identified in QoL questionnaires; the SF-36 measures health-related QoL, while the WHOQOL-BREF measures global QoL. 58 There was less discrepancy in results between studies using different questionnaires than for burnout studies suggesting that questionnaire choice was less of an issue for QoL studies.
This is the first review to specifically examine the effect of physical activity/exercise on burnout and/or QoL in medical students.
Reviews in other populations found a negative correlation between physical activity and emotional exhaustion in employees from a mixture of professions. 22 Additionally, individual studies in teachers 59 and medical residents/fellows 21 have found negative correlations between physical activity and burnout.
This is the first review to specifically examine the effect of physical activity/ exercise on burnout and/or QoL in medical students.
There are contrasting reviews; a recent meta-analysis of RCTs showed no significant difference in burnout between physical activity intervention and control groups in employees of various professions. 60 However, this meta-analysis only included four studies with a wide range of exercise modalities. 60 Previous research of QoL is more limited. Surgeons who completed aerobic and strengthening exercises according to US guidelines had high QoL scores compared with those who did not meet the guidelines. 61 Additionally, a study of University students (18-30 years; some from physical health/physiotherapy) found that physical activity was positively correlated to QoL. 62 However, this was only for certain types of physical activity such as household tasks rather than leisure-time physical activity, for which a relationship was not established. These studies support our findings; however, they must be treated with caution. More detail is required on the type of physical activity which precipitates improvements.

| Strengths
This is the first review to examine the effect of physical activity/ exercise on either burnout and/or QoL in medical students. It therefore provides a unique synthesis of information drawn from around the globe which has implications for medical schools worldwide. The consideration of both burnout and QoL allows for the opportunity to evaluate whether physical activity influenced both, one or neither of these interlinked concepts. Additionally, subgroups, such as physical activity/exercise intensity, have been identified and analysed in detail to draw further conclusions from the findings.
A further strength is the demographics of studies included. Over 11,500 students from all stages of medical school over 13 countries were included. This suggests that the improvement of QoL and reduction of burnout via physical activity demonstrated are not limited to a single medical curriculum or stage of study.
Over 11,500 students from all stages of medical school over 13 countries were included.

| Future research and implications
There is a paucity of high methodological quality research in this area, demonstrated by the critical appraisal results. Future research should focus on longitudinal cohort and interventional studies to allow exploration of causality. More detail on physical activity/exercise is necessary; the intensity, frequency, volume and mode must be investigated.
Additionally, a clearer distinction between physical activity and exercise is required, as this will influence recommendations made to medical schools. Sample sizes based on considerations of statistical power, homogeneity in burnout/QoL questionnaires, greater consideration of confounding factors and longer follow-up in cohort studies will produce higher quality studies. Greater homogeneity of measures used will also allow for a future meta-analysis. Importantly, future research must identify the minimum level of physical activity/exercise required to reduce burnout and increase QoL significantly. This would allow for maximal inclusivity when implemented into medical education.
Future research must identify the minimum level of physical activity/exercise required to reduce burnout and increase QoL significantly. This would allow for maximal inclusivity.
Medical admissions processes are being refined to identify students with 'grit' 63,64 and the resilience to be able to complete a challenging programme of study. 65 Evidence of sporting activity in applications is unlikely to be impactful compared with academic performance, admission tests and interviews, which is rightly supportive of applicants from widening participation backgrounds. 66 National stakeholders 67 emphasise the importance of facilitating access to medicine to students with disabilities and using physical activity as a factor in selection may be discriminatory. The authors therefore do not suggest that physical activity should be used as a selection criterion for medical school admissions.
Despite this opinion, the evidence presented in this systematic review leads the authors to propose that physical activity should be used as part of a spectrum of well-being activities on offer during medical school. In this way, physical activity would not be compulsory for all medical students, rather, it could form part of a compulsory well-being programme alongside other options, such as mindfulness.
Medical curricula are subject to repeated pressure to increase content on several fronts including sub-specialisation of medical disciplines, medical advances, external stakeholder pressures and highprofile patient safety events. 68,69 While there is increasing interest in well-being curricula within medical schools, engagement from students is variable. 70 As a course with high cognitive load, focus is maintained on credit-bearing modules which are likely to be assessed in formal examinations. With crowded timetables students may be impeded from accessing their previous levels of activity; it is possible that a debate around required curriculum content for medical students is overdue. There is evidence that student engagement is higher with peer-led or self-directed well-being activities 71 and strategies which support this (in addition to freeing up time) such as reduced gym membership and support for student sports societies could be encouraged. Several US schools offer credit-bearing modules for weight loss in obese students, 72 and this approach could be applied with credits available for physical activity engagement.
If we are serious about facilitating learning in our students, we need to create a culture in which self-care is valued and facilitated. 73 Protected time for well-being is a model which already exists in some educational activities in the UK. For example, many colleges and universities do not timetable learning on Wednesday afternoon to allow students to pursue interests of their choice, often sport.
This could be successfully extended to medical education. Importantly, the authors suggest that well-being activities should be given specific time in the timetable rather than expecting students to find time outside of medical education. It would also be interesting to investigate how these well-being activities are best delivered; the COVID-19 pandemic has mandated the use of online technologycould a physical activity intervention be successfully delivered in this way?
Well-being activities should be given specific time in the timetable rather than expecting students to find time outside of medical education.
Additionally, it is imperative that medical students are educated about the importance of physical activity for both physical and mental health. By doing so, they will be equipped to offer advice to their patients and are more likely to understand the benefits for themselves.
It is pivotal that strategies are taken to improve well-being in medical school, especially at a time where medical students are facing the unique challenges that the COVID-19 pandemic has imposed. The authors believe that this systematic review provides an exciting basis for the inclusion of physical activity into well-being measures during medical school.

| CONCLUSION
This multinational systematic review demonstrates that physical activity is associated with lower burnout and increased QoL in medical students. Following further research addressing the limitations identified in this review, medical schools could prioritise well-being of medical students via the implementation of tailored physical activity. This is of great significance for medical education and beyond, as the next generation of doctors may be better prepared to manage the demands of medical practice and offer optimal care to their patients.