BMJ Open Is Committed to Open Peer Review. As Part of This Commitment We Make the Peer Review History of Every Article We Publish Publicly Available

BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

BMJ Open is committed to open peer review. As part of this commitment we make the peer review history of every article we publish publicly available.

When an article is published we post the peer reviewers’ comments and the authors’ responses online. We also post the versions of the paper that were used during peer review. These are the versions that the peer review comments apply to.

The versions of the paper that follow are the versions that were submitted during the peer review process. They are not the versions of record or the final published versions. They should not be cited or distributed as the published version of this manuscript.

BMJ Open is an open access journal and the full, final, typeset and author-corrected version of record of the manuscript is available on our site with no access controls, subscription charges or pay-per-view fees (http://bmjopen.bmj.com).

If you have any questions on BMJ Open’s open peer review process please email [email protected] http://bmjopen.bmj.com/ on September 26, 2021 by guest. Protected copyright. BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

The effect of interventions for the wellbeing, satisfaction and flourishing of general practitioners – A systematic review ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2020-046599

Article Type: Original research

Date Submitted by the 09-Nov-2020 Author:

Complete List of Authors: Naehrig, Diana; The University of Sydney, The Faculty of Health and Medicine Schokman, Aaron; The University of Sydney, The Faculty of Health and Medicine Hughes, Jessica; The University of Sydney Library Epstein, Ronald; University of Rochester School of Medicine, Family Medicine Research Programs Hickie, Ian; The University of Sydney, Brain and Mind Centre Glozier, Nick; The University of Sydney, The Faculty of Health and Medicine

PRIMARY CARE, GENERAL MEDICINE (see Internal Medicine), MENTAL Keywords: HEALTH, Organisational development < HEALTH SERVICES http://bmjopen.bmj.com/ ADMINISTRATION & MANAGEMENT

on September 26, 2021 by guest. Protected copyright.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 40 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 5 6 7 8 9 I, the Submitting Author has the right to grant and does grant on behalf of all authors of the Work (as defined 10 in the below author licence), an exclusive licence and/or a non-exclusive licence for contributions from authors 11 who are: i) UK Crown employees; ii) where BMJ has agreed a CC-BY licence shall apply, and/or iii) in accordance 12 with the terms applicable for US Federal Government officers or employees acting as part of their official 13 duties; on a worldwide, perpetual, irrevocable, royalty-free basis to BMJ Publishing Group Ltd (“BMJ”) its 14 licensees and where the relevant Journal is co-owned by BMJ to the co-owners of the Journal, to publish the 15 Work in this journal and any other BMJ products and to exploit all rights, as set out in our licence. 16 17 The Submitting Author accepts and understands that any supply made under these terms is made by BMJ to 18 the Submitting Author Forunless you peer are acting as review an employee on behalf only of your employer or a postgraduate 19 student of an affiliated institution which is paying any applicable article publishing charge (“APC”) for Open 20 Access articles. Where the Submitting Author wishes to make the Work available on an Open Access basis (and 21 intends to pay the relevant APC), the terms of reuse of such Open Access shall be governed by a Creative 22 Commons licence – details of these licences and which Creative Commons licence will apply to this Work are set 23 out in our licence referred to above. 24 25 Other than as permitted in any relevant BMJ Author’s Self Archiving Policies, I confirm this Work has not been 26 accepted for publication elsewhere, is not being considered for publication elsewhere and does not duplicate 27 material already published. I confirm all authors consent to publication of this Work and authorise the granting 28 of this licence. 29 30 31 32 33 34 35 36 37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44 45 on September 26, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 40

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 TITLE PAGE 5 6 Title The effect of interventions for the wellbeing, satisfaction, and flourishing 7 of general physicians – A systematic review 8 Authors Diana Naehrig, ORCID (0000-0002-1162-4092), 9 10 Dr.med., FMH Radioonkologie, MSc Coach Psych, 11 PhD candidate, The University of Sydney, Faculty of Health and Medicine, 12 Level 5, Professor Marie Bashir Centre, Missenden Road, Camperdown, 13 NSW 2050, Australia, [email protected]; 14 15 Aaron Schokman, ORCID (0000-0003-0419-9347), BSc, MPhil, 16 17 PhD candidate, The University of Sydney, Faculty of Health and Medicine, 18 ForLevel 5,peer Professor Mariereview Bashir Centre, only Missenden Road, Camperdown, 19 NSW 2050, Australia, [email protected]; 20 21 Jessica Kate Hughes, ORCID (0000-0003-0046-1686), Master of 22 Information Studies, Assistant Librarian, The University of Sydney 23 Library, Rm 202, Fisher Library F03, The University of Sydney, NSW 24 25 2006, Australia, [email protected]; 26 27 Ron Mark Epstein, ORCID (0000-0002-3564-9163), MD 28 Co-Director, Center for Communication and Disparities Research, 29 Professor of Family Medicine, Oncology and Medicine (Palliative Care), 30 American Cancer Society of Clinical Research Professor, University of 31 Rochester School of Medicine and Dentistry, 1381 South Avenue, 32 33 Rochester, NY 14620, USA, [email protected]; 34 35 Ian Hickie, ORCID (0000-0001-8832-9895), AM MD FRANZCP FASSA 36 FAHM 37 Co-Director, Health and Policy and Professor of Psychiatry, NHMRC http://bmjopen.bmj.com/ 38 Senior Principal Research Fellow, The University of Sydney, Faculty of 39 40 Health and Medicine, Brain and Mind Centre, Level 4, Building F, 94 41 Mallett Street, Camperdown, NSW 2050, Australia, 42 [email protected]; 43 44 Nick Glozier, ORCID (0000-0002-0476-9146), 45 Professor of Psychological Medicine, The University of Sydney, Faculty of on September 26, 2021 by guest. Protected copyright. 46 Health and Medicine, Central Clinical School, Level 5, Professor Marie 47 48 Bashir Centre, Missenden Road, Camperdown, NSW 2050, Australia, 49 [email protected] 50 Corresponding Diana Naehrig, The University of Sydney, Faculty of Health and Medicine, 51 author Level 5, Professor Marie Bashir Centre, Missenden Road, Camperdown, 52 NSW 2050, Australia, [email protected] 53 54 55 Word count 3262 56 manuscript 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 40 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 ABSTRACT 5 6 7 Objectives 8 Clinician wellbeing has been recognised as an important pillar of health care. However, 9 research mainly addresses mitigating the negative aspects of stress or burnout, rather 10 than enabling positive aspects. With the added strain of a pandemic, identifying how 11 best to maintain and support the wellbeing, satisfaction and flourishing of frontline 12 doctors is now more important than ever. 13 14 15 Design 16 We systematically reviewed the effect of any type of intervention on wellbeing, and 17 related positive outcomes in General Practitioners (GPs). 18 For peer review only 19 Data sources 20 We searched MEDLINE, PsycINFO, Embase, CINAHL, and Scopus from 2000 to 21 22 2020. 23 24 Study selection 25 Studies with more than 50% GPs, trialling any type of intervention, looking at 26 wellbeing, satisfaction, flourishing and related positive outcomes were included. The 27 Cochrane risk of bias 2 tool was applied. 28 29 30 Results 31 Data Extraction 32 We retrieved 14,792 records, 94 studies underwent full text review. We included 19 33 studies in total. Six randomised controlled trials, three non-randomised, controlled 34 trials, and eight non-controlled studies. There were a total of 1141 participants, with 35 interventions targeting individuals or organisations, plus two quasi-experimental articles 36 37 evaluating health system policy change. http://bmjopen.bmj.com/ 38 39 Data synthesis 40 Individual mindfulness interventions were the most common (k=9) with medium to 41 large within (0.37-1.05) and between group (0.5-1.5) effect sizes for mindfulness 42 outcomes, and small to medium effect sizes for a range of other positive outcomes 43 including resilience, compassion, and empathy. Studies assessing other interventions or 44 45 positive outcomes (including wellbeing, satisfaction) were of limited size and quality. on September 26, 2021 by guest. Protected copyright. 46 47 Conclusions 48 There is remarkably little evidence on how to improve PCP wellbeing beyond using 49 mindfulness interventions, particularly interventions addressing organisational or 50 system factors. This was further undermined by inconsistent reporting, and overall high 51 52 risk of bias. We need to conduct research in this space with the same rigour with which 53 we approach clinical intervention studies in patients. 54 55 Registration 56 We submitted the systematic review protocol for registration on PROSPERO. 57 58 59 Funding source 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 40

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Dr Diana Naehrig is funded through the Raymond Seidler PhD scholarship. 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44 45 on September 26, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 40 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 ARTICLE SUMMARY 5 6 Strenghts and limitations of this study 7 8  Whilst burnout has been a general focus of research, interventions to improve 9 10 the wellbeing of general practitioners (primary care physicians) appear sparse. 11  The strength of this study is the extensive and systematic approach taken to 12 13 evaluate interventions aimed at improving wellbeing, satisfaction, flourishing 14 15 and other positive outcomes in general practitioners 16 17  The systematic review was registered on PROSPERO a priori before 18 commencingFor the peerdata selection review and extraction process.only 19 20  The limitations of this systematic review is the dependency on the number of 21 22 retrieved and included publications, and their quality of methodology and 23 24 reporting. 25 26 27 28 29 INTRODUCTION 30 31 Mental ill-health, burnout and stress amongst health care practitioners are a huge 32 concern internationally with impacts on individual doctors and their families, patient 33 34 care, and the sustainability of the health care system (1). Hence, improving clinician 35 36 wellbeing has been added to the more commonly shared health system goals of: 37 http://bmjopen.bmj.com/ 38 improved care for individuals, better population health, and reduced health care costs 39 (2-5). Despite this stated aim, few studies have evaluated interventions to improve 40 41 wellbeing, satisfaction, and flourishing in general practitioners (GPs) (6-12), rather than 42 43 taking a problem-focused context, such as reduction of burnout and stress. 44 45 on September 26, 2021 by guest. Protected copyright. 46 a) Factors associated with satisfaction and wellbeing in GPs 47 48 Primary care physicians are less satisfied with their job than other specialists in the 49 50 USA (13, 14), a ten-year decline in job satisfaction for British GPs has been reported 51 (15), and a Norwegian longitudinal study reported dwindling GP satisfaction over seven 52 53 years (16). In primary care, job satisfaction correlates with practice resources, an 54 55 ordered atmosphere, a practice culture that enables communication, and ease of 56 providing quality care (13, 17, 18), and is inversely related to isolation and low sense of 57 58 community (19, 20). 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 40

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 For wellbeing, Murray et al. conducted a cross-sectional study exclusively exploring 5 6 positive mental health and psychological resources (wellbeing, resilience, self-efficacy 7 8 and hope) of GPs (21). GPs positive mental health was comparable to the general 9 10 population and older and female GPs fared best, suggesting interventions for younger 11 and male GPs may be most useful. 12 13 14 15 b) Systematic reviews of interventions aimed at improving satisfaction and 16 wellbeing in GPs 17 18 A European collaborationFor peerconducted areview systematic review only and qualitative study looking 19 20 at positive determinants of satisfaction in GPs. They identified general work related 21 22 factors (i.e. workload, responsibility, recognition and income), and GP specific factors 23 (i.e. competence, intellectual stimulus and work-life balance) (22, 23). However, there 24 25 does not appear to be a systematic review looking at interventions to improve 26 27 satisfaction exclusively in GPs. 28 29 A systematic review of interventions to improve the psychological wellbeing of GPs 30 identified only four articles; two cognitive-behavioural, one mindfulness-based 31 32 intervention and one self-help information approach (24). They applied a dual model of 33 34 languishing/flourishing and the presence of mental illness/absence of mental illness 35 matrix (25-28). 36 37 http://bmjopen.bmj.com/ 38 39 Overall, little seems known about which interventions are efficacious in promoting GPs 40 41 wellbeing and satisfaction. Currently, COVID-19 places enormous additional strain on 42 health professionals which impacts their physical, mental and social wellbeing (29, 30). 43 44 GPs as the first port of call are particularly affected, whilst playing a crucial role in 45 on September 26, 2021 by guest. Protected copyright. 46 supporting population health (31-33). Now more than ever, efforts must be made to 47 48 ensure GPs remain well. 49 50 51 OBJECTIVES 52 53 We systematically reviewed studies exploring the effect of any type of intervention on 54 the wellbeing, satisfaction and wellbeing of GPs. 55 56 57 58 59 METHODS 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 40 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Data sources and search strategy 5 6 We followed PRISMA guidelines (34), and consulted a specialist librarian (JKH). 7 8 MEDLINE, PsycINFO, Embase, CINAHL and Scopus were searched from 01/2000 to 9 10 01/2020. 11 PICO search terms included GPs and synonyms; interventions and implementations 12 13 directed at the individual and profession at large; outcomes included wellbeing, 14 15 satisfaction, flourishing, synonyms and antonyms (search strategy, supplement 1). 16 Titles, abstracts, text, key terms and subject headings were searched for English 17 18 publications. EligibleFor articles peer were hand-searched review for further only relevant references. 19 20 21 22 Study selection and data extraction 23 Sample screening of titles was performed by two reviewers (DNA, AS) and showed 24 25 excellent agreement. DNA screened all titles. DNA and AS independently examined all 26 27 abstracts and full-text records for inclusion using Covidence (35). Any disagreement 28 29 was resolved through discussion, or third reviewer adjudication (NG). 30 Studies with more than 50% GPs (family doctors, primary care physicians) working in a 31 32 practice setting or medical centre, reporting on wellbeing, satisfaction, flourishing, 33 34 mindfulness, resilience, empathy, engagement, balance, empowerment, compassion, 35 work-related morale and control measures were included. We excluded studies 36 37 exclusively reporting on burnout, distress or mental ill-health. http://bmjopen.bmj.com/ 38 39 Data including author, year, type of study, participants, intervention, pre- and post- 40 41 outcome measures, and results were extracted. (Table 1). 42 43 44 Data synthesis and analysis 45 on September 26, 2021 by guest. Protected copyright. 46 We calculated within- and between-group absolute change and effect sizes (Hedges’ g) 47 48 (see table 2, 3 and supplement 2, 3) (36). We compared mean outcome scores and 49 standard deviations (SDs) at baseline with post-intervention scores. Where several post- 50 51 intervention measures were reported, the primary outcome point nominated by the 52 53 authors was selected. We utilised SD*, which takes different sample sizes into account 54 (formulae in supplement 2). For within-group we calculated the pooled SD* based on 55 56 pre- and post- interventions SDs, for between-group analysis the effect size was 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 40

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 calculated based on the pooled SD* of control and intervention group at baseline (37) 5 6 (supplement 2). 7 8 Positive effect sizes indicated an effect for the intervention. Effect sizes of 0.2, 0.5 and 9 10 0.8 were considered small, moderate and large, respectively (38). 11 12 Risk of bias 13 14 Two reviewers (AS, DNA) independently applied the Cochrane RoB2 (39) to RCTs. 15 16 Any discrepancies were discussed, and a consensus was achieved. The other studies 17 18 were assessed byFor DNA. peer review only 19 20 21 Patient and public involvement 22 23 No patient are involved. 24 25 26 RESULTS 27 28 The database searches rendered 14,792 records in total. After removing duplicates, 29 30 studies conducted before the year 2000, and adding 12 studies through hand search, 31 10,759 studies were screened. We eliminated 9,682 records by title, and 983 by abstract, 32 33 leaving 94 studies for full-text assessment. 34 35 36 37 Study characteristics & design http://bmjopen.bmj.com/ 38 We included nineteen studies in the systematic review (2, 10, 40-56) (Table 1 and 2, 39 40 and PRISMA-Flowchart Figure 1). Six randomised controlled trials (RCT), three non- 41 42 randomised controlled trials (CBA), eight non-controlled interventions (NCBA), and 43 two reports from a longitudinal cohort during which a health policy change was 44 45 introduced, which we considered as ‘naturalistic’ interventions (41, 56). Five studies on September 26, 2021 by guest. Protected copyright. 46 47 included a qualitative component. Only one RCT (50) and two CBAs (44, 47) utilised 48 49 active controls. Five RCTs and one CBA (40) had a waitlist control arm. Publications 50 from the United States (8/19, 42%), Europe (8/19, 42%), Australia (2/19) and Israel 51 52 (1/19) were included (Table 1). 53 54 55 56 Participants 57 The total population enrolled was 1,141 for the 17 intervention studies. The two studies 58 59 reporting on the same panel survey population (41, 56) included approximately 2,000 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 40 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 participants each year. Mean age ranged from 40 – 54.9 years, and sex from 8% - 76.9% 5 6 male participants (Table 1). Attrition for intervention groups varied from 0% to 20%, 7 8 for controls from 0% to 24%. One outlier had a total attrition rate of 80% (51). 9 10 11 Intervention type 12 13 We found considerable variation in intervention type, length and dose-intensity. Three 14 15 groups were distinguished based upon the focus of the intervention: individual/personal 16 (13/19, 68% of studies, N=930), organisational (4/19, 21%, N=211) and naturalistic 17 18 interventions on Fora systemic peer level (policy review change) (2/19, only11%). 19 20 Individual mindfulness-based interventions were most common (9/13, 69%), followed 21 22 by educational training or experiential workshops (3/13, 23%) with one coaching 23 intervention (46). Two organisational interventions trialled the addition of clerical 24 25 support or scribes, and two explored an organisational improvement programme. Two 26 27 studies from the United Kingdom examined the effects of the introduction of a pay for 28 29 performance scheme (41, 56) (Table 2). 30 31 32 Outcomes and their measures 33 34 The definitions of outcomes and measurement tools varied considerably. Only one 35 study clearly stated one a priori primary outcome (51), with most using a battery of self- 36 37 reported outcome measures (supplement 4). These included a range of validated and http://bmjopen.bmj.com/ 38 39 non-validated job satisfaction metrics in eight studies, mindfulness tools in six, 40 41 resilience measures in four, compassion and empathy tools were each used thrice, the 42 positive and negative affect scale was used twice. The WHO-5 wellbeing index was 43 44 used once (supplement 4). Not one study evaluated flourishing. Negative outcome 45 on September 26, 2021 by guest. Protected copyright. 46 measures were often concurrently reported. Sixteen studies employed the Maslach 47 48 Burnout Inventory or other stress-related measures. 49 50 51 Intervention effects 52 53 The between-group change for controlled studies and within-group change for 54 55 intervention arms are presented in Tables 2, 3 and supplement 3. 56 57 58 a. Individual focussed interventions 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 40

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 i) Mindfulness (k=9) 5 6 Six mindfulness interventions (3 RCTs, 1 CBA, 2 NCBA) evaluated mindfulness 7 8 outcomes (FFMQ, MAAS, BAER) and reported moderate to high between-group effect 9 10 sizes (k=4) ranging from 0.5 – 0.88 for mindfulness with an outlier at 1.5 (42). Within- 11 group ES (k=6) showed moderate effect sizes (range 0.47 – 0.78) with one study outlier 12 13 at 0.37 and one at 1.05 (Tables 2 and 3). 14 15 Studies frequently utilised resilience, compassion and empathy measures with overall 16 low to moderate effect sizes. One RCT and two NCBAs measured resilience (BRS, RS- 17 18 14, CD-RISC), wherebyFor between-grouppeer review ES (k=1) was moderateonly at 0.61, while within- 19 20 group (k=3) effect sizes were low to moderate (range 0.17 – 0.51). Compassion 21 22 measures (SCBC, SCS) were reported in three studies (1 RCT, 2 NCBAs). Between- 23 group ES (k=1) was 0.73, whilst within-group ES (k=3) varied considerably (-0.04 to 24 25 0.77). Three studies reported on empathy (JSPE) (1 RCT, 1 CBA, 1 NCBA) with very 26 27 low 0.02, respectively moderate between-group 0.44 ES (k=2), and within-group ES 28 29 ranging from 0.2 - 0.44 (k=3) (Tables 2 and 3). 30 Two mindfulness studies (NCBAs) measured positive affect (PANAS), only one 31 32 reported a within-group ES (0.52). One NCBA reported a within-group ES (0.52) for 33 34 wellbeing (WHO-5), another NCBA reported an ES of 0.46 for self-reflection. 35 These effect sizes are generally supported by the results reported in the individual 36 37 studies (Table 1). Several interventions included repeated measures at later time-points , http://bmjopen.bmj.com/ 38 39 i.e. during maintenance phase (48), showing an ongoing impact of mindfulness practice. 40 41 Qualitative results suggested increased wellbeing and compassion towards self and 42 others (40), respectively, improved awareness, acceptance, peacefulness and openness 43 44 (55) after the intervention. 45 on September 26, 2021 by guest. Protected copyright. 46 47 48 ii. Training, workshops and coaching (k=4) 49 For training, workshops and coaching interventions, we were only able to obtain data to 50 51 calculate the ES of one RCT (46) and one CBA (47). Low between-group effect sizes 52 53 for work-related morale (0.3), quality of work-life (0.27), and low ES for both measures 54 within-group (0.43 and 0.45 respectively) were found for Gardiner’s CBA (47). Very 55 56 low effect sizes for job satisfaction and resilience both for between-group (0.06, 0.13), 57 58 and within-group (0.13, 0.24) change were observed in Dyrbye’s RCT (46). 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 40 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 These results are reflected in the individual study results (Table 1). Barcons (44) did not 5 6 detect any significant between-group differences for overall job satisfaction, whilst 7 8 Margalit (50) demonstrated significant improvement in self-esteem between groups. 9 10 11 b. Organisational interventions (k=4) 12 13 One RCT and three NCBAs trialled organisational interventions. Means and standard 14 15 deviations were not provided; therefore, we were not able to calculate effect sizes. 16 Linzer et al. (RCT, n=166) demonstrated that workflow interventions, communication 17 18 and overall qualityFor improvements peer benefited review satisfaction onlyin the intervention group (49). 19 20 Whilst Dunn et al. (NCBA, n=32), showed that quality improvement projects in the 21 22 workplace showed significant improvement in quality work competence ratings but 23 fluctuating satisfaction levels (2). Two smaller uncontrolled trials (n=13 in total) 24 25 investigated the addition of clerical staff to the practice. Pozdnyakova et al. showed that 26 27 the addition of clerical staff led to an improvement in satisfaction with the clinic 28 29 workflow from 2/6 to all 6 GPs in a single practice, but did not report on any other 30 measures of wellbeing (52). Similarly, Contratto et al. reported improved quality of life 31 32 and professional balance for seven general medicine physicians in a mixed-methods 33 34 approach (45). 35 36 37 c. Systemic interventions (k=2) http://bmjopen.bmj.com/ 38 39 The introduction of a new contract with pay for performance scheme showed a 40 41 significant improvement in job satisfaction (56) with an effect size of 0.44 between 42 2004 and 2005. Allen et al. used the same data and included a 2008 survey to look at 43 44 satisfaction as a function of the exposure of GPs to the pay per performance scheme. 45 on September 26, 2021 by guest. Protected copyright. 46 Whilst job satisfaction declined again in 2008, the exposure to the scheme didn’t affect 47 48 satisfaction (41). 49 50 51 Risk of bias 52 53 The types of intervention and study settings precluded blinding for randomised 54 controlled studies (no allocation concealment for waitlist control groups), and the 55 56 outcome measures were participant reported throughout, and as such all studies were 57 58 rated as high risk of bias by the Cochrane RoB2. 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 40

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 5 6 DISCUSSION 7 8 Strengths, limitations and interpretation of evidence 9 10 We identified nineteen studies, half of which were published in the last four years, 11 demonstrating an increased interest in the improvement of wellbeing and satisfaction of 12 13 GPs. In comparison, a systematic review from 2016 looking at interventions to reduce 14 15 burnout in physicians included fifteen RCTs and 37 cohort studies with 20 studies 16 conducted before 2010 (57), suggesting burnout has been a research focus for longer. 17 18 The considerableFor heterogeneity peer in the definitionreview and measurement only of constructs, study 19 20 design, participant numbers, intervention types, intervention dose (ranging from six to 21 22 53 hours), follow-up periods, quality and reporting precluded a meta-analysis. It is 23 challenging to draw robust conclusions regarding the (comparative) effectiveness of the 24 25 different types of interventions reviewed. 26 27 28 29 Mindfulness interventions provided the most comprehensive and robust data with 30 moderate to large effect sizes on mindfulness outcomes, and low to moderate effect 31 32 sizes on compassion, resilience and empathy. Our results are supported by two reviews 33 34 looking at the effects of mindfulness-based interventions on wellbeing (58, 59), more 35 generally. Lomas et al. conducted a systematic review and meta-analysis and assessed 36 37 41 studies with approximately 2,100 participants. They found a wide range of self- http://bmjopen.bmj.com/ 38 39 reported outcomes (with both positive and negative measures of wellbeing). Reported 40 41 within-group effect sizes for mindfulness, positive wellbeing (or life satisfaction), and 42 compassion hovered around a moderate 0.5 mark, ES for empathy was 0.31; whilst for 43 44 RCTs, the between-group ES for mindfulness, life satisfaction, and compassion were 45 on September 26, 2021 by guest. Protected copyright. 46 around 0.3 (58). 47 48 Scheepers et al. contributed a narrative review of 23 studies looking at mindfulness- 49 based interventions for wellbeing in doctors of all ages and specialities. Review authors 50 51 noted some caveats; considerable variation in type and dose-intensity of practice, sparse 52 53 long-term outcome data, and methodological limitations. They cautioned that 54 mindfulness practice involves time and dedication, which is not always feasible for busy 55 56 health care professionals. In sum, the conclusions they drew are tentatively positive. 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 40 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 In contrast to Lomas et al. our ES for mindfulness are higher between-group than 5 6 within-group, which is somewhat unusual. This may be explained by a small study 7 8 effect, and it may be driven by one study (42) with a decline in mindfulness over time 9 10 for their control group. 11 12 13 We identified four studies evaluating coaching and experiential workshops for GPs, 14 15 which showed low effect sizes for satisfaction measures and moderate ES for work- 16 related morale and quality of work life. There does not appear to be much literature on 17 18 coaching for healthFor care professionals.peer review One quasi-experimental only study by Gardiner et al. 19 20 looked at ‘cognitive behavioural coaching’ in rural Australian GPs and demonstrated a 21 22 significant within-group reduction in distress levels for the coachees (60). Resilience 23 training for a range of different physicians that had completed training was investigated 24 25 in a recent systematic review. Four RCTs and five observational studies were included. 26 27 The authors flag heterogenous study design and use of outcome measures, as well as 28 29 quality issues with weak evidence for the interventions, whilst indicating potential for 30 improvement of resilience (61). 31 32 33 34 We found four small-scale organisational interventions that suggested improved (job) 35 satisfaction, as did one large-scale health policy intervention of performance-related pay 36 37 in the UK. For burnout, a paucity of interventions trials delivered at organisational and http://bmjopen.bmj.com/ 38 39 systemic levels has been previously reported (62, 63), the authors suggest to actively 40 41 design such trials. Similarly, Dyrbye et al., concluded that whilst useful, an individual 42 intervention such as coaching is no replacement for organizational improvement (46). 43 44 Shanafelt and colleagues have collated their vast research into burnout and put forward 45 on September 26, 2021 by guest. Protected copyright. 46 nine organisational strategies to address burnout and physician wellbeing through 47 48 leadership (64). Despite calls for action, these avenues have not been adequately 49 addressed or reported to date, at least not for GPs, and may warrant further exploration. 50 51 Considering the time it takes to gather and report data, it is understandable that 52 53 organizations might feel pressure to implement programs based on preliminary data. 54 55 56 Suggestions for future research 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 40

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Based on our findings, we provide some suggestions which may be useful for future 5 6 research into wellbeing and satisfaction for GPs. Stronger collaboration amongst 7 8 researchers in this space may also lead to improved results. 9 10 11 a) Selection of outcomes & outcome measures 12 13 The reduction in burnout and stress is often equated with an improvement in wellbeing 14 15 or satisfaction. We argue that the improvement of negative outcomes doesn’t 16 necessarily indicate a presence of satisfaction or wellbeing. This aligns with the dual 17 18 continuum modelFor of mental peer health / mental review ill-health and only flourishing / languishing (25- 19 20 28). Good mental health isn’t automatically linked to flourishing, nor is mental ill- 21 22 health an indicator of languishing. Other authors have made similar statements (10, 24, 23 42, 58). We did not find a single study about flourishing in GPs. 24 25 Clearly defining the constructs ‘wellbeing’ and ‘satisfaction’, whilst utilising validated 26 27 wellbeing and satisfaction measures, will enhance clarity, consistency and 28 29 comparability of study design and reporting. Brady et al., who conducted a systematic 30 review looking at the definition and measurement of ‘physician wellness’ stated that 31 32 there needs to be consensus and clarity of definition, if we want to improve the quality 33 34 and comparability of research in this space (65). Whilst this would improve the next 35 phase of studies, the urgency in calls for actions may need to be balanced against the 36 37 calls for consistency among studies. http://bmjopen.bmj.com/ 38 39 40 41 b) Organisational and systemic interventions 42 When looking at known drivers of burnout (organisational culture, workplace 43 44 conditions, lack of control and autonomy), it’s not surprising that individual 45 on September 26, 2021 by guest. Protected copyright. 46 interventions are not as effective as desired (49, 62, 66-68). Hence, more combined 47 48 approaches targeting both individuals and organisations have been proposed. The same 49 might be true for wellbeing interventions. 50 51 A 2017 British meta-analysis contrasted different types of interventions for burnout on 52 53 the individual doctor and on the systemic level, whereby systemic interventions appear 54 more effective (69). Similarly, groups in the USA state that the approach must be 55 56 combined and include organisational interventions (1, 70, 71), mostly focusing on time 57 58 management, rostering, workflow management, staffing and use of information 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 40 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 technology solutions. Overall, there is a scarcity of organisational interventions aimed at 5 6 reducing burnout (72), and conclusions from the two meta-analyses of interventions to 7 8 reduce burnout should be considered provisional. We endorse an intensified effort to 9 10 explore organisational interventions to improve wellbeing and satisfaction. 11 12 13 c) Physical interventions 14 15 We didn’t find any physical interventions (i.e. exercise, nutrition) geared towards 16 improving GPs’ wellbeing. Sparse research on exercise or diet interventions for doctors 17 18 exists. A PakistaniFor cross-sectional peer survey review revealed that 76%only of nearly 1200 doctors, 19 20 nurses and dentists did not exercise at all, and only one participant ate according to 21 22 USDA dietary guidelines (73). While a US cross-sectional survey of 303 physicians 23 found that less than 25% knew the American Heart Association (AHA) dietary 24 25 recommendations, whilst around half knew and followed their physical activity 26 27 guidelines (74). Two systematic reviews looked at exercise and burnout in the general 28 29 population, one was inconclusive (75), the other stated that physical activity effectively 30 reduces burnout (76). Both identified methodological issues and no long term follow- 31 32 up. Seeing the paucity of data, this might provide an avenue for further investigation. 33 34 35 d) Quality and risk of bias 36 37 Areas for risk of bias are inherent in this type of research. However, measures can be http://bmjopen.bmj.com/ 38 39 taken to reduce bias for example by using active controls in randomized studies as 40 41 suggested by other review authors (59), by consistently publishing study protocols a 42 priori, and controlling for participant attributes, such as prior engagement in 43 44 mindfulness practice. Ideally, the same rigourous approach should be applied to 45 on September 26, 2021 by guest. Protected copyright. 46 intervention studies for clincians, as to clinical interventions studies for patients. 47 48 49 CONCLUSION 50 51 Despite increasing interest in the improvement of wellbeing and satisfaction, there is 52 53 remarkably little evidence beyond mindfulness interventions aimed at individual GPs. 54 Few studies utilize validated measures of wellbeing and satisfaction, and favour burnout 55 56 tools. Studies looking into organisational and systemic interventions remain sparse, and 57 58 conclusions about their effectiveness may be permature. 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 40

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 In light of the COVID-19 pandemic and the added strain to primary care, programmes 5 6 to support and research GP wellbeing should be prioritised by policymakers and 7 8 governments worldwide. 9 10 11 ACKNOWLEDGEMENTS 12 13 Registration 14 15 The protocol was submitted for registration on PROSPERO, due to high workload a 16 registration is still pending. 17 18 For peer review only 19 20 Funding statement 21 22 Dr Diana Naehrig was supported through the Raymond Seidler PhD scholarship. The 23 funding source had no influence on the study design, collection, analysis or 24 25 interpretation of data, the writing of the manuscript nor the decision to submit the article 26 27 for publication. Award/Grant number is not applicable. 28 29 30 Competing interest statement 31 32 All authors have completed the Unified Competing Interest Form. 33 34 Beyond the Raymond Seidler PhD scholarship for Diana Naehrig there was no support 35 from any organisation for the submitted work. The funding source had no influence on 36 37 the study design, collection, analysis or interpretation of data, the writing of the report http://bmjopen.bmj.com/ 38 39 nor the decision to submit the article for publication. 40 41 Ian Hickie has declared financial relationships outside of the submitted work in the 42 previous three years and other relationships or activities. 43 44 45 on September 26, 2021 by guest. Protected copyright. 46 Authors’ Contributions 47 48 DNA is the guarantor and corresponding author and attests that all listed authors meet 49 authorship criteria and that no others meeting the criteria have been omitted. 50 51 Authors contributed to the study conception and design (DNA, JKH, RE, IH, NG), the 52 53 acquisition (DNA, AS, NG), analysis (DNA, AS, NG) and interpretation (DNA, AS, 54 RE, IH, NG) and the drafting or critical revision of important intellectual content and 55 56 final approval of the version to be published and agree to be accountable for all aspects 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 40 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 of the work in ensuring that questions related to the accuracy or integrity of any part of 5 6 the work are appropriately investigated and resolved. 7 8 9 10 Transparency declaration 11 Dr Diana Naehrig (the manuscript’s guarantor) affirms that the manuscript is an honest, 12 13 accurate, and transparent account of the study being reported; that no important aspects 14 15 of the study have been omitted; and that any discrepancies from the study as planned 16 and registered have been explained. 17 18 For peer review only 19 20 Data sharing statement 21 22 All included studies are published. We will consider sharing data upon reasonable 23 request. 24 25 26 27 Copyright 28 29 Dr Diana Naehrig (corresponding author) has the right to grant on behalf of all authors 30 and does grant on behalf of all authors, an exclusive licence on a worldwide basis to the 31 32 BMJ Publishing Group Ltd to permit this article to be published in BMJ editions and 33 34 any other BMJPGL products and sublicences such use and explit all subsidiary rights, 35 as set out in our licence. 36 37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44 45 on September 26, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 40

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 REFERENCES 5 6 1. West CP, Dyrbye LN, Shanafelt TD. Physician burnout: contributors, 7 consequences and solutions. Journal of Internal Medicine. 2018;283(6):516-29. 8 2. Dunn PM, Arnetz BB, Christensen JF, Homer L. Meeting the imperative to 9 10 improve physician well-being: Assessment of an innovative program. Journal of 11 General Internal Medicine. 2007;22(11):1544-52. 12 3. Luchterhand C, Rakel D, Haq C, Grant L, Byars-Winston A, Tyska S, et al. 13 Creating a Culture of Mindfulness in Medicine. Wmj. 2015;114(3):105-9. 14 4. Spinelli WM. The phantom limb of the triple aim. Mayo Clin Proc. 15 2013;88(12):1356-7. 16 5. Wallace JE, Lemaire JB, Ghali WA. Physician wellness: a missing quality 17 18 indicator. The Lancet.For 2009;374(9702):1714-21. peer review only 19 6. West CP, Dyrbye LN, Rabatin JT, Call TG, Davidson JH, Multari A, et al. 20 Intervention to promote physician well-being, job satisfaction, and professionalism: a 21 randomized clinical trial. JAMA Intern Med. 2014;174(4):527-33. 22 7. Pollak KI, Nagy P, Bigger J, Bilheimer A, Lyna P, Gao X, et al. Effect of 23 teaching motivational interviewing via communication coaching on clinician and patient 24 25 satisfaction in primary care and pediatric obesity-focused offices. Patient Educ Couns. 26 2016;99(2):300-3. 27 8. Gidwani R, Nguyen C, Kofoed A, Carragee C, Rydel T, Nelligan I, et al. Impact 28 of Scribes on Physician Satisfaction, Patient Satisfaction, and Charting Efficiency: A 29 Randomized Controlled Trial. Ann Fam Med. 2017;15(5):427-33. 30 9. Dyrbye LN, West CP, Richards ML, Ross HJ, Satele D, Shanafelt TD. A 31 randomized, controlled study of an online intervention to promote job satisfaction and 32 33 well-being among physicians. Burnout Research. 2016;3(3):69-75. 34 10. Fortney L, Luchterhand C, Zakletskaia L, Zgierska A, Rakel D. Abbreviated 35 mindfulness intervention for job satisfaction, quality of life, and compassion in primary 36 care clinicians: a pilot study. Annals of Family Medicine. 2013;11(5):412-20. 37 11. Dyrbye LN, Shanafelt TD, Gill PR, Satele DV, West CP. Effect of a http://bmjopen.bmj.com/ 38 Professional Coaching Intervention on the Well-being and Distress of Physicians: A 39 Pilot Randomized Clinical Trial. JAMA Intern Med. 2019. 40 41 12. Barcons C, Garc√≠a B, Sarri C, Rodr√≠guez E, Cunillera O, Parellada N, et al. 42 Effectiveness of a multimodal training programme to improve general practitioners' 43 burnout, job satisfaction and psychological well-being. BMC family practice. 44 2019;20(1):155. 45 13. Alidina S, Rosenthal MB, Schneider EC, Singer SJ, Friedberg MW. Practice on September 26, 2021 by guest. Protected copyright. 46 environments and job satisfaction in patient-centered medical homes. Ann Fam Med. 47 48 2014;12(4):331-7. 49 14. Leigh JP, Tancredi DJ, Kravitz RL. Physician career satisfaction within 50 specialties. BMC Health Serv Res. 2009;9:166. 51 15. Sherlock C, John C. Adaptation practice: teaching doctors how to cope with 52 stress,anxiety and depression by developing resilience. British Journal of Medical 53 Practitioners. 2016;9(2). 54 16. Rosta J, Aasland OG, Nylenna M. Changes in job satisfaction among doctors in 55 56 Norway from 2010 to 2017: a study based on repeated surveys. BMJ Open. 57 2019;9(9):e027891. 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 40 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 17. Linzer M, Manwell LB, Williams ES, Bobula JA, Brown RL, Varkey AB, et al. 5 6 Working conditions in primary care: physician reactions and care quality. Annals of 7 internal medicine. 2009;151(1):28-36. 8 18. Friedberg MW, Chen PG, Van Busum KR, Aunon F, Pham C, Caloyeras J, et al. 9 Factors affecting physician professional satisfaction and their implications for patient 10 care, health systems, and health policy. Rand health quarterly. 2014;3(4). 11 19. Quinn MA, Wilcox A, Orav J, Bates DW, Simon SR. The relationship between 12 perceived practice quality and quality improvement activities and physician practice 13 14 dissatisfaction, professional isolation, and work-life stress. Medical care. 2009:924-8. 15 20. Frey JJ, 3rd. Professional Loneliness and the Loss of the Doctors' Dining Room. 16 Ann Fam Med. 2018;16(5):461-3. 17 21. Murray MA, Cardwell C, Donnelly M. GPs' mental wellbeing and psychological 18 resources: a cross-sectionalFor peer survey. Br review J Gen Pract. 2017;67(661):e547-e54. only 19 22. Le Floch B, Bastiaens H, Le Reste JY, Lingner H, Hoffman R, Czachowski S, et 20 al. Which positive factors give general practitioners job satisfaction and make general 21 22 practice a rewarding career? A European multicentric qualitative research by the 23 European general practice research network. BMC Fam Pract. 2019;20(1):96. 24 23. Le Floch B, Bastiaens H, Le Reste JY, Lingner H, Hoffman RD, Czachowski S, 25 et al. Which positive factors determine the GP satisfaction in clinical practice? A 26 systematic literature review. BMC Fam Pract. 2016;17(1):133. 27 24. Murray MA, Murray L, Donnelly M. Systematic review of interventions to 28 29 improve the psychological well-being of general practitioners. BMC Family Practice. 30 2016;17(1). 31 25. Keyes CL. Mental illness and/or mental health? Investigating axioms of the 32 complete state model of health. J Consult Clin Psychol. 2005;73(3):539-48. 33 26. Keyes CL. Promoting and protecting mental health as flourishing: a 34 complementary strategy for improving national mental health. The American 35 psychologist. 2007;62(2):95-108. 36 37 27. Keyes CL. The mental health continuum: From languishing to flourishing in life. http://bmjopen.bmj.com/ 38 Journal of health and social behavior. 2002:207-22. 39 28. Westerhof GJ, Keyes CL. Mental Illness and Mental Health: The Two Continua 40 Model Across the Lifespan. J Adult Dev. 2010;17(2):110-9. 41 29. Sethi BA, Sethi A, Ali S, Aamir HS. Impact of Coronavirus disease (COVID- 42 19) pandemic on health professionals. Pak J Med Sci. 2020;36(COVID19-S4):S6-S11. 43 30. Liu Q, Luo D, Haase JE, Guo Q, Wang XQ, Liu S, et al. The experiences of 44 45 health-care providers during the COVID-19 crisis in China: a qualitative study. The on September 26, 2021 by guest. Protected copyright. 46 Lancet Global Health. 2020;8(6):e790-e8. 47 31. Kohlbacher S. GPs heading for burnout as COVID stresses pile up 48 www.medicalrepublic.com.au: The Medical Republic; 2020 [18/09/2020]. Available 49 from: http://medicalrepublic.com.au/gps-heading-for-burnout-as-covid-stresses-pile- 50 up/34342. 51 52 32. RACGP. Mental health and wellbeing for GPs and patients - Caring for yourself: 53 Royal Australian College of General Practitioners; 2020 [Available from: 54 https://www.racgp.org.au/coronavirus. 55 33. Scott A. The impact of COVID-19 on GPs and non-GP specialists in private 56 practice. Melbourne Institute: Applied Economic & Social Research, The University of 57 Melbourne; 2020. 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 40

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 34. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, Ioannidis JP, et al. 5 6 The PRISMA statement for reporting systematic reviews and meta-analyses of studies 7 that evaluate health care interventions: explanation and elaboration. PLoS Med. 8 2009;6(7):e1000100. 9 35. Innovation VH. Covidence systematic review software. Veritas Health 10 Innovation Melbourne, Australia; 2016. 11 36. Borenstein M, Hedges LV, Higgins JPT, Rothstein HR. Effect Sizes Based on 12 Means. Introduction to Meta-Analysis. Chichester, UK: John Wiley & Sons Ltd.; 2009. 13 14 p. 21-32. 15 37. Morris SB. Estimating effect sizes from pretest-posttest-control group designs. 16 Organizational research methods. 2008;11(2):364-86. 17 38. Higgins JP, Green S. Cochrane handbook for systematic reviews of 18 interventions. Chichester,For peerEngland; Hoboken. review NJ: Wiley-Blackwell; only 2008. 19 39. Sterne JA, Savović J, Page MJ, Elbers RG, Blencowe NS, Boutron I, et al. RoB 20 2: a revised tool for assessing risk of bias in randomised trials. Bmj. 2019;366. 21 22 40. Verweij H, Waumans RC, Smeijers D, Speckens AEM, Lucassen PLBJ, 23 Donders ART, et al. Mindfulness-based stress reduction for GPs: results of a controlled 24 mixed methods pilot study in Dutch primary care. British Journal of General Practice. 25 2016;66(643):e99-e105. 26 41. Allen T, Whittaker W, Sutton M. Does the proportion of pay linked to 27 performance affect the job satisfaction of general practitioners? Social Science & 28 29 Medicine. 2017;173:9-17. 30 42. Amutio A, Martinez-Taboada C, Hermosilla D, Delgado LC. Enhancing 31 relaxation states and positive emotions in physicians through a mindfulness training 32 program: A one-year study. Psychology, Health & Medicine. 2015;20(6):720-31. 33 43. Asuero AM, Queralto JM, Pujol-Ribera E, Berenguera A, Rodriguez-Blanco T, 34 Epstein RM. Effectiveness of a mindfulness education program in primary health care 35 professionals: a pragmatic controlled trial. J Contin Educ Health Prof. 2014;34(1):4-12. 36 37 44. Barcons C, Garcia B, Sarri C, Rodriguez E, Cunillera O, Parellada N, et al. http://bmjopen.bmj.com/ 38 Effectiveness of a multimodal training programme to improve general practitioners' 39 burnout, job satisfaction and psychological well-being. BMC family practice. 40 2019;20(1):155. 41 45. Contratto E, Romp K, Estrada CA, Agne A, Willett LL. Physician Order Entry 42 Clerical Support Improves Physician Satisfaction and Productivity. Southern Medical 43 Journal. 2017;110(5):363-8. 44 45 46. Dyrbye LN, Shanafelt TD, Gill PR, Satele DV, West CP. Effect of a on September 26, 2021 by guest. Protected copyright. 46 Professional Coaching Intervention on the Well-being and Distress of Physicians: A 47 Pilot Randomized Clinical Trial. JAMA Internal Medicine. 2019;05:05. 48 47. Gardiner M, Lovell G, Williamson P. Physician you can heal yourself! 49 Cognitive behavioural training reduces stress in GPs. Family Practice. 2004;21(5):545- 50 51. 51 52 48. Krasner MS, Epstein RM, Beckman H, Suchman AL, Chapman B, Mooney CJ, 53 et al. Association of an educational program in mindful communication with burnout, 54 empathy, and attitudes among primary care physicians. JAMA. 2009;302(12):1284-93. 55 49. Linzer M, Poplau S, Grossman E, Varkey A, Yale S, Williams E, et al. A 56 Cluster Randomized Trial of Interventions to Improve Work Conditions and Clinician 57 Burnout in Primary Care: Results from the Healthy Work Place (HWP) Study. Journal 58 of General Internal Medicine. 2015;30(8):1105-11. 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 40 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 50. Margalit AP, Glick SM, Benbassat J, Cohen A, Katz M. Promoting a 5 6 biopsychosocial orientation in family practice: effect of two teaching programs on the 7 knowledge and attitudes of practising primary care physicians. Med Teach. 8 2005;27(7):613-8. 9 51. Montero-Marin J, Gaete J, Araya R, Demarzo M, Manzanera R, Alvarez de Mon 10 M, et al. Impact of a blended web-based mindfulness programme for general 11 practitioners: A pilot study. Mindfulness. 2018;9(1):129-39. 12 52. Pozdnyakova A, Laiteerapong N, Volerman A, Feld LD, Wan W, Burnet DL, et 13 14 al. Impact of Medical Scribes on Physician and Patient Satisfaction in Primary Care. 15 Journal of General Internal Medicine. 2018;33(7):1109-15. 16 53. Rees CS, Craigie MA, Slatyer S, Crawford C, Bishop M, McPhee E, et al. Pilot 17 study of the effectiveness of a Mindful Self-Care and Resiliency program for rural 18 doctors in Australia.For Aust peerJ Rural Health. review 2020;28(1):22-31. only 19 54. Schroeder DA, Stephens E, Colgan D, Hunsinger M, Rubin D, Christopher MS. 20 A Brief Mindfulness-Based Intervention for Primary Care Physicians: A Pilot 21 22 Randomized Controlled Trial. American Journal of Lifestyle Medicine. 2018;12(1):83- 23 91. 24 55. van Wietmarschen H, Tjaden B, van Vliet M, Battjes-Fries M, Jong M. Effects 25 of mindfulness training on perceived stress, self-compassion, and self-reflection of 26 primary care physicians: a mixed-methods study. Bjgp Open. 27 2018;2(4):bjgpopen18X101621. 28 29 56. Whalley D, Gravelle H, Sibbald B. Effect of the new contract on GPs' working 30 lives and perceptions of quality of care: A longitudinal survey. British Journal of 31 General Practice. 2008;58(546):8-14. 32 57. West CP, Dyrbye LN, Erwin PJ, Shanafelt TD. Interventions to prevent and 33 reduce physician burnout: a systematic review and meta-analysis. The Lancet. 34 2016;388(10057):2272-81. 35 58. Lomas T, Medina JC, Ivtzan I, Rupprecht S, Eiroa-Orosa FJ. A systematic 36 37 review of the impact of mindfulness on the well-being of healthcare professionals. J http://bmjopen.bmj.com/ 38 Clin Psychol. 2018;74(3):319-55. 39 59. Scheepers RA, Emke H, Epstein RM, Lombarts K. The impact of mindfulness- 40 based interventions on doctors' well-being and performance: A systematic review. Med 41 Educ. 2020;54(2):138-49. 42 60. Gardiner M, Kearns H, Tiggemann M. Effectiveness of cognitive behavioural 43 coaching in improving the well-being and retention of rural general practitioners. Aust J 44 45 Rural Health. 2013;21(3):183-9. on September 26, 2021 by guest. Protected copyright. 46 61. Venegas CL, Nkangu MN, Duffy MC, Fergusson DA, Spilg EG. Interventions 47 to improve resilience in physicians who have completed training: A systematic review. 48 PLoS ONE. 2019;14(1):e0210512. 49 62. Maslach C, Leiter MP. Early predictors of job burnout and engagement. Journal 50 of applied psychology. 2008;93(3):498. 51 52 63. Awa WL, Plaumann M, Walter U. Burnout prevention: a review of intervention 53 programs. Patient Educ Couns. 2010;78(2):184-90. 54 64. Shanafelt TD, Noseworthy JH. Executive Leadership and Physician Well-being: 55 Nine Organizational Strategies to Promote Engagement and Reduce Burnout. Mayo 56 Clin Proc. 2017;92(1):129-46. 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 40

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 65. Brady KJS, Trockel MT, Khan CT, Raj KS, Murphy ML, Bohman B, et al. 5 6 What Do We Mean by Physician Wellness? A Systematic Review of Its Definition and 7 Measurement. Acad Psychiatry. 2018;42(1):94-108. 8 66. Linzer M, Poplau S, Brown R, Grossman E, Varkey A, Yale S, et al. Do Work 9 Condition Interventions Affect Quality and Errors in Primary Care? Results from the 10 Healthy Work Place Study. Journal of General Internal Medicine. 2017;32(1):56-61. 11 67. Gregory ST, Menser T. Burnout Among Primary Care Physicians: A Test of the 12 Areas of Worklife Model. Journal of Healthcare Management. 2015;60(2):133-48. 13 14 68. Cossman JS, Street D. Policy, autonomy, and physician satisfaction. Journal of 15 Health Care for the Poor & Underserved. 2010;21(3):898-912. 16 69. Panagioti M, Panagopoulou E, Bower P, Lewith G, Kontopantelis E, Chew- 17 Graham C, et al. Controlled interventions to reduce burnout in physicians: a systematic 18 review and meta-analysis.For peer JAMA internal review medicine. 2017;177(2):195-205. only 19 70. West C, Dyrbye L, Erwin P, Shanafelt T. Interventions to prevent and reduce 20 physician burnout: a systematic review and meta-analysis. The Lancet. 21 22 2016;388(10057):2272-81. 23 71. Swensen SJ, Shanafelt T. An Organizational Framework to Reduce Professional 24 Burnout and Bring Back Joy in Practice. The Joint Commission Journal on Quality and 25 Patient Safety. 2017;43(6):308-13. 26 72. Gregory ST, Menser T, Gregory BT. An Organizational Intervention to Reduce 27 Physician Burnout. Journal of Healthcare Management. 2018;63(5):338-52. 28 29 73. Ahmad W, Taggart F, Shafique MS, Muzafar Y, Abidi S, Ghani N, et al. Diet, 30 exercise and mental-wellbeing of healthcare professionals (doctors, dentists and nurses) 31 in Pakistan. PeerJ. 2015;3:e1250. 32 74. Aggarwal M, Singh Ospina N, Kazory A, Joseph I, Zaidi Z, Ataya A, et al. The 33 Mismatch of Nutrition and Lifestyle Beliefs and Actions Among Physicians: A Wake- 34 Up Call. American Journal of Lifestyle Medicine. 2020;14(3):304-15. 35 75. Ochentel O, Humphrey C, Pfeifer K. Efficacy of exercise therapy in persons 36 37 with burnout. A systematic review and meta-analysis. Journal of sports science & http://bmjopen.bmj.com/ 38 medicine. 2018;17(3):475. 39 76. Naczenski LM, Vries JD, Hooff M, Kompier MAJ. Systematic review of the 40 association between physical activity and burnout. J Occup Health. 2017;59(6):477-94. 41 42 43 44 45 on September 26, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 40 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 5 Table 1. Evidence table 6 7 Partici- 8 pants (n, Control 9Author, mean Interven- (n, Positive Other 10Date, Study age, % tion (n, mean outcome outcome Post-intervention Follow-up score 11design male) mean age) age) measure measure Baseline scores score (mean, SD, n) Results - Summary 12 Between 2004 and Time at 12 months. Time at 48 months. 2005 overall job 13 Mean. 2004: Life Mean. 2005: Life Mean. 2008: Life satisfaction 14 satisfaction (4.649). satisfaction (5.095). satisfaction (5.008). increased (also see 15 GPs in Overall JS (4.567), Overall JS (5.201), Overall JS (4.728), Whalley 2008), then 16 the UK. Physical working Physical working Physical working fell from 2005 to 2004: conditions (4.862), conditions (5.044), conditions (5.129), 2008. There is a 17 n=1950, Choose method of Choose method of Choose method of positive effect 18 47.0yrs, For peer reviewworking (4.636), workingonly (4.892), working (4.640), (though not 19 66.2% Colleagues (5.515), Colleagues (5.599), Colleagues (5.602), statistically 20Allen 2017 male. Pay for Recognition for Recognition for Recognition for significant) between 21 2005: perform- good work (4.224), good work (4.726), good work (4.495), QOF/P4P income Panel n=2000, ance (P4P) Responsibility Responsibility Responsibility exposure and job 22survey 47.9yrs, scheme; (4.976), (5.406), (5.276), satisfaction in 2005 23linked with 63.6 % Quality and Job Remuneration Remuneration Remuneration (t-ratio 1.74), 24QOF data male. Outcomes satisfacti (4.376), Opportunity (5.387), Opportunity (4.849), Opportunity though not so in 252004/5 2008: Framework on (JSS to use abilities to use abilities to use abilities 2008 (t-ratio 0.14). (first year) n=1986, (QOF) WCW). (4.787), Hours of (5.147), Hours of (5.074), Hours of The P4P exposure 26and 2007/8 48.7yrs, introduced Life work (3.914), work (4.802), work (4.205), shows now 27(fourth 63.3% in NHS in satisfacti P4P Variety in job Variety in job Variety in job significant effect on 28year) male. 2004 NA on exposure (5.011). (5.269). (5.276). GPs job satisfaction. 29 Mean and SD. Time at 8 weeks. 30 INTERVENTION Mean and SD. (n=21) FFMQ Total INTERVENTION 31 (3.34, 0.44), FFMQ (n=21) FFMQ Total 32 Observing (3.33, (3.71, 0.51), FFMQ 33 0.60), FFMQ Observing (3.98, 34 Describing (3.58, 0.64), FFMQ 0.72), FFMQ act Describing (3.83, 35 with awareness 0.62), FFMQ act 36 (3.16, 0.87), FFMQ with awareness 37 Nonjudging (3.42, (3.48, 0.65), FFMQ http://bmjopen.bmj.com/ 38 0.64), FFMQ Nonjudging (3.78, 39 Nonreactivity (3.17, 0.66), FFMQ 0.51). SRSI Nonreactivity (3.46, 40 relaxation (2.54, 0.63). SRSI 41 0.53), SRSI positive relaxation (3.08, 42 energy (3.09, 0.64), 0.61), SRSI positive Time at 12 months. 43 SRSI mindfulness energy (3.80, 0.82), Mean and SD. (3.74, 0.89), SRSI SRSI mindfulness INTERVENTION 44 Transcedence (2.67, (4.45, 0.71), SRSI FFMQ Total (3.85, 45 8 week 0.82). CONTROL Transcedence (3.27, 0.49), FFMQ on September 26, 2021 by guest. Protected copyright. 46 MBSR (n=21) FFMQ Total 1.02). CONTROL Observing (4.09, 47 programme (3.51, 0.25), FFMQ (n=21) FFMQ Total 0.62), FFMQ The MBSR (group Observing (3.02, (3.34, 0.33), FFMQ Describing (4.01, programme 48 course 0.88), FFMQ Observing (2.83, 0.58), FFMQ act (statistically) 49 Physic- 8x2.5h/wk, Describing (3.88, 0.98), FFMQ with awareness significantly 50 ians 1x8h 0.53), FFMQ Act Describing (3.82, (3.55, 0.69), FFMQ improves 51 (approx retreat plus with awareness 0.58), FFMQ Act Nonjudging (3.96, mindfulness and 52 70% homework) (3.93, 0.70), FFMQ with awareness 0.66), FFMQ relaxation measures primary , 10 month Nonjudging (4.19, (3.91, 0.61), FFMQ Nonreactivity (3.58, (particularly positive 53 care) in mainten- 0.52), FFMQ Nonjudging (4.16, 0.55). SRSI energy and 54 public or ance period Nonreactivity (3.23, 0.52), FFMQ relaxation (3.16, transcendence) at 8 55 private (group Mindful- 0.57). SRSI Nonreactivity (3.26, 0.78), SRSI positive weeks after 56 practice course ness relaxation (2.52, 0.67). SRSI energy (3.99, 0.81), baseline. Sustained Amutio in Spain 10x2.5h/mt (FFMQ), 0.5), SRSI positive relaxation (2.60, SRSI mindfulness and even improved 572015 (n=42, plus Relax- energy (3.01, 0.62), 0.51), SRSI positive (4.60, 1.01), SRSI longterm effects at 58 47.3 yrs, homework) WLC ation Heart SRSI mindfulness energy (3.01, 0.6), Transcedence (3.65, 12 months follow- 59RCT 42.9%) (n=21) (n=21) (SRSI-3) rate (4.29, 0.92), SRSI SRSI mindfulness 1.36). up are shown. 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 24 of 40

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Transcedence (2.64, (4.24, 0.95), SRSI 5 0.93). Transcedence (2.40, 6 0.92). 7 Time at 8 weeks. Mean and SD. Mean and SD. 8 INTERVENTION INTERVENTION 9 (n=43). FFMQ Total (n=43) FFMQ Total 10 (129.6, 22.2), FFMQ (141.6, 16.5), FFMQ 11 Observing (26.1, Observing (30.4, 8.6), FFMQ 5.1), FFMQ 12 Describing (28.2, Describing (28.9, 13 5.7), FFMQ act with 5.3), FFMQ act with 14 awareness (25.3, awareness (27.4, 15 5.9), FFMQ 4.7), FFMQ 16 Nonjudging (27.1, Nonjudging (30.6, 8.0), FFMQ 6.2), FFMQ 17 Nonreactivity (21.9, Nonreactivity (24.1, 18 For peer review4.3). JSPE Total 3.0).only JSPE Total 19 (119.5, 13.1), JSPE (123.0, 9.2), JSPE 20 Perspective taking Perspective taking (54.8, 6.5), JSPE (56.1, 5.3), JSPE 21 Compassionate care Compassionate care 22 8 week MB (47.2, 5.1), JSPE (49.0, 3.9), JSPE 23 psycho- standing in pts standing in pts For mindfulness 24 educational shoes (13.1, 1.8). shoes (13.5, 1.1). total there is program: CONTROL (n=25). CONTROL (n=25) significant 25 Didactic FFMQ Total (120.5, FFMQ Total (121.5, improvement for 26 material, 14.4), FFMQ 16.0), FFMQ the intervention 27 mindful- Observing (24.5, Observing (24.1, (moderate change). 28 ness 4.3), FFMQ 4.5), FFMQ For empathy total 29 Primary meditation, Describing (26.4, Describing (26.5, there is a non- health narrative/a 5.2), FFMQ Act with 5.5), FFMQ Act with significant increase 30 care ppreciative awareness (23.6, awareness (23.0, from pre to post 31 profes- enquiry, 6.8), FFMQ 5.9), FFMQ measure in the 32 sionals discussion. Mindfuln Nonjudging (26.3, Nonjudging (27.4, intervention group, 33 (41 Group ess Burnout 6.5), FFMQ 6.8), FFMQ only an increase in physician course (FFMQ), (MBI), Nonreactivity (19.8, Nonreactivity (20.2, compassionate care 34 s) in 8x2.5h/wk, Empathy mood 2.9). JSPE Total 3.5). JSPE Total was statistically 35 Spain, 1x8h WLC (JSPE), disturban (120.8, 10.1), JSPE (119.0, 10.7), JSPE significant. Overall 36 public retreat plus (n=25 also self- ce Perspective taking Perspective taking mindfulness and 37Asuero system homework. total, report on (POMS), (59.8, 7.0), JSPE (59.6, 6.3), JSPE empathy http://bmjopen.bmj.com/ 382014 (n=68 (n=43 total, 18 energy, intervent Compassionate care Compassionate care (compassionate total, 47 23 physi- wellbeing ion (47.4, 4.9), JSPE (46.9, 4.6), JSPE care) were 39RCT (2010 - yrs, 8% physicians, cians, and evaluatio standing in pts standing in pts improved by the 402012) male) 48.8yrs) 46.9yr) activity n shoes (13.6, 0.9). shoes (12.5, 3.0). NA programme. 41 42 43 44 45 on September 26, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 25 of 40 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Median and IQR. 5 INTERVENTION 6 (n=20) FR 7 satisfaction at work (15, 13, 16.25), FR Time at 10 months. 8 work tension (12, Median and IQR. 9 11, 14), FR INTERVENTION 10 Professional (n=20) FR 11 competence (5, 4, satisfaction at work 6), FR work pressure (16, 15, 17), FR work 12 (12, 10, 12.25), FR tension (13, 11, 16), 13 professional FR Professional 14 promotion (12, 9, competence (5, 4.5, 15 13), FR relationship 6), FR work pressure 16 superiors (4, 4, 6.5), (12, 10, 12.50), FR FR relationship professional 17 peers (6, 5.75, 7), FR promotion (12, 10, 18 For peer reviewextrinsic status (5.5, 13),only FR relationship 19 4 ,6), FR monotony superiors (4, 4, 6.5), 20 (4.5, 4, 6), FR total FR relationship (75, 72, 77.5). BPRS peers (5, 5, 6), FR 21 total (23.50, 22, extrinsic status (5, 5, 22 24.25). CONTROL 6.5), FR monotony 23 (n=18), FR (5, 4, 5), FR total 24 Burnout satisfaction at work (76, 73, 83). BPRS (MBI), (14, 10.5, 16), FR total (20.50, 19, 22). 25 Brief work tension (13.5, CONTROL (n=18), FR 26 Psychiatri 12.25, 16.75), FR satisfaction at work 27 MTP and c Rating Professional (12, 9, 16), FR work 28 IBST group Scale competence (4.5, 3, tension (15, 1, 16), No statistically 29 programme (BPRS), 6), FR work pressure FR Professional significant changes plus Rout- Psychoph (12, 12, 13), FR competence (5, 4, in job satisfaction 30 routine ine armacolo professional 6), FR work pressure (FR). For wellbeing, 31 programme mental gy use, promotion (11, 10, (11, 11, 12), FR a statistically 32 . 9 x 1hr per -health opinions 13.75), FR professional significant decrease 33 Primary week, (6h sup- about relationship promotion (12, 11, was seen from pre health Training on port mental superiors (6, 4, 15), FR relationship to post measure in 34 care psychology, pro- illness, 6.75), FR superiors (6, 4, 8), the intervention 35Barcons profess- 3h on gram administr relationship peers FR relationship group, as the BPRS 362019 ionals in psychiatry for ative and (5, 5, 6), FR extrinsic peers 6, 5, 7), FR measures negative 37 Spain, and 1h prim- Job health status (5, 5, 6), FR extrinsic status (6, 4, constructs, this is http://bmjopen.bmj.com/ CBA, mixed public social ary Satisfacti care monotony (5, 4, 6), FR monotony (4, interpreted as an 38methods, system work). care on (FR- indicator 5.75), FR total (76, 3, 5), FR total (77, improvement of 392016 - 2017 (n=38). (n=20) (n=18) JS) s 73, 80.5). 75, 78). NA wellbeing. 40 Qol, personal 41 Burnout balance, burnout 42 (MBI), improved. Personal- physician Productivity 43 General professio product- increased (work 44 medicine 1 clerical nal ivity, relative value unit) 45Contratto physician staff was balance, EMR Time at 4 months. per session on September 26, 2021 by guest. Protected copyright. 462017 s in the hired in a physician documen QoL: 14% bad, 14% QoL: 0% bad, 0% increased. Reports USA, GP practice satisfacti tation. neutral, 71% good. neutral, 100% good. of feeling more 47NCBA, working to enter on Qual Balance: 14% Balance: 0% supported, focused 48mixed part-time orders defined inter- Dissatisfied, 43% Dissatisfied, 29% on pt care, less 49methods (n=7) (n=7). NA as QoL. views. neutral, 43% good. neutral, 71% good. stress, less fatigue 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 26 of 40

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Organisat- 5 ional inter- 6 vention Intervention 7 programme ongoing from 2000 to improve onward. Emotional 8 physician and work-related 9 wellbeing exhaustion 10 (quality decreased 11 improve- significantly over ment the study period. 12 project, QWC measures of 13 participant organizational 14 Primary data- health significantly 15 care in guided improved initially 16 urban interv- Burnout and remained USA ention) (n= (MBI), acceptable and 17 (n=32, 25 22 to 32 Physician Quality 2001: 55% of Time at 24 months. Time at 24 months. stable during the 18Dunn 2007 internists range over Forsatisfacti peerWork reviewphysicians were 2003:only 84% were 2005: 74% were rest of the study. 19 6 family the inter- on Compete somewhat or very somewhat or very somewhat or very Satisfaction 20NCBA, medicine vention (ACP/ASI nce satisfied with their satisfied with their satisfied with their fluctuated 2000-2005 1 NP) period). NA M) (QWC) practice practice practice. throughout. 21 Means and SD. Time at 5 months. 22 INTERVENTION Means. 23 (n=44), Resilience INTERVENTION 24 (31.0, 6.3), Job (n=44), Resilience satisfaction (43.4, (32.3), Job 25 Resilien- 10.7), UWES vigor satisfaction (44.4), Statistically 26 ce (CD- (3.9, 1.0), UWES UWES vigor (4.1), significant 27 RISC), Dedication (4.5, UWES Dedication improvement for 28 Global 1.1), UWES (4.6), UWES resilience from pre 29 Job Absorption (4.2, Absorption (4.1), to post intervention, Satisfact- 1.0), Empowerment Empowerment at no change in job 30 ion-12 at work (55.5, 11.9). work (58.2). satisfaction. 31 Physician 6 coaching (subscale CONTROL (n=44), CONTROL (n=44), Burnout, emotional 32 s in USA sessions of PJSS), Resilience (30.6, Resilience (31.2), exhaustion 33 (67 (3.5h) work 5.7), Job satisfaction Job satisfaction decreased. QoL generalis facilitated engage- (42.8, 10.6), UWES (43.2), UWES vigor improved. No 34 ts, 21 by a ment vigor (4.0, 1.2), (4.2), UWES statistically 35Dyrbye subspec- profess- (UWES), UWES Dedication Dedication (4.7), significant 362019 ialists). sional empower (4.6, 1.0), UWES UWES Absorption differences in 37 (n=88, coach over ment at Absorption (4.1, (4.2), depersonalization, http://bmjopen.bmj.com/ 38RCT, 45.5% 5mts WLC work Burnout 1.1), Empowerment Empowerment at engagement, or 2017 - 2018 male). (n=44). (n=44) scale. (MBI). at work (57.3, 14.0). work (60.3). meaning in work. 39 No significant 40 improvement in 41 Family compassion and 42 medicine Burnout resilience over time. practit- (MBI), Participants had 43 ioners in Shortened Depressi Time at 4 weeks. Time at 9 months. improvements 44 USA MBSR on & Mean score and Mean score and Mean score and compared with 45 (n=30, course, 18h Compas- Anxiety Confidence Interval. Confidence Interval. Confidence Interval baseline at all on September 26, 2021 by guest. Protected copyright. 46 87% total, 14h sion (DASS), n=30. Resilience RS- n=28. Resilience RS- at 9mts. n=23. follow-up time 47Fortney family weekend, (SCBC), Perceive 14 (79.9, CI 75.2- 14 (82, CI 77.1- Resilience RS-14 points for Burnout, 2013 medicine 2x 2h plus Resili- d Stress 84.6), Compassion 86.8), Compassion (81.4, CI 76.2-86.6), depression, anxiety 48 physic- homework ence (RS- scale SCBC (27.6, CI 25.9- SCBC (27.4, CI 25.6- Compassion SCBC and perceived 49NCBA ians) (n=30). NA 14). (PSS), 29.3). 29.1). (28.3, CI 26.5-30.1). stress. 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 27 of 40 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Overall the 5 intervention group 6 showed higher 7 scores post intervention than 8 the control group 9 (positive trend). 10 However, no 11 statistically significant change 12 seen for work 13 related morale or 14 QoWL for the 15 intervention group. 16 Work Means and SD. Total Only physicological related measure without Time at 4 weeks. distress significantly 17 distress sub-measures Means and SD. lower for 18 ControlFor peermeasure reviewgiven.INTERVENTIO INTERVENTIONonly intervention group. 19 Cognitive group Work (WRD-7), N (n=86), Work (n=77), Work- When looking at 20 behavioural had related General related morale total related morale total those GPs that stress other morale psycholo 31.83 (6.75). QOWL 34.62 (6.11). QoWL scored low for 21 manage- devel- measure gical total 25.32 (6.64). total 28.24 (6.35). Time at 12 weeks. morale, there was a 22 GPs in ment opmen (WRM-7), distress CONTROL (n=24), CONTROL (n=19), Means and SD. 56% reduction pre 23Gardiner metro- course, 1x tal Quality (GHQ- Work related Work related INTERVENTION to post intervention, 242004 politan 3h, 15h course of work 12), morale total 31.42 morale total (n=62), Work compared with a Australia over 5 wks. s. life coping (6.19). QoWL total 32.21 (6.73). QoWL related morale total 29% reduction in 25CBA (n=110) (n=85). (n=25) (QoWL-6) styles. 23.16 (5.86). total 24.32 (6.36). 35.70 (6.01). the control group. 26 Time at 8 weeks. Time at 12 months. 27 Mean and CI. Baer Mean and CI. Baer 28 Mean and CI. Baer mindfulness total mindfulness total 29 mindfulness total (52.9, 95% CI 51 to (55, 95% CI 53.0 to CME (45.2, 95% CI 43.3- 54.8, n=59), Baer 56.9, n=56), Baer 30 programme 47.1, n=60), Baer mindfulness mindfulness 31 : Didactic mindfulness observe (30.6, 95% observe (31.1, 95% 32 material, observe (25.6, 95% CI 29.4 to 31.8, CI 29.8 to 32.3, 33 mindful- CI 24.4-26.8, n=60), n=59), Baer n=56), Baer Baseline scores and ness Baer mindfulness mindfulness mindfulness follow-up scores at 34 meditation, nonreact (19.7, CI nonreact (22.9, CI nonreact (23.9, CI 15 months are not 35 narrative/a 95% 18.7-20.7, 95% 21.8 to 23.9, 95% 22.9 to 24.9, reported for the 36 ppreciative n=60). JSPE total n=59). JSPE total n=56). JSPE total purpose of this SR. 37 enquiry, Burnout (116.2, 95% CI (120.6, 95% CI 118.2 (121.4, 95% CI 119.0 Over time, all http://bmjopen.bmj.com/ 38 discussion. (MBI), 114.2-118.9, n=60), to 123.0, n=59), to 123.8, n=56), measures for 28h total. 8 Mood JSPE compassionate JSPE compassionate JSPE compassionate mindfulness, 39 Primary x 2.5hr per (POMS), care (48.6, 95% CI care (49.8, 95% CI care (50.4, 95% CI burnout, physician 40 care week, 1x Big 5 47.5-49.7, n=60), 48.7 to 50.9, n=59), 49.3 to 51.5, n=56), belief, mood and, 41 physici- 7h, 10- personali JSPE perspective JSPE perspective JSPE perspective personality 42 ans in the month Mindfuln ty taking (57.1, 95% CI taking (59.1, 95% CI taking (59.7, 95% CI improved, the Krasner USA mainten- ess factors, 55.6 to 58.6, n=60), 57.6 to 60.6, n=59), 58.2 to 61.2, n=56), largest effect size 432009 (n=70, ance phase (Baer), Physician JSPE standing in pts JSPE standing in pts JSPE standing in pts was observed for 44 54% (2.5 h/mo). Empathy belief shoes (10.8, 95% CI shoes (11.7, 95% CI shoes (11.4, 95% CI mindfulness at 15 45NCBA male) (n=70) NA (JSPE) (PBS) 10.4 to 11.5, n=60). 11.1 to 12.2, n=59). 10.9 to 11.9, n=56). mts. on September 26, 2021 by guest. Protected copyright. 46 Time at 12 to 47 18months. Number Primary Number given in %. given in %. 48 care Intervent- INTERVENTION INTERVENTION 49 clinicians ions to (n=83) High work (n=67) High work 50 in the improve control (score >2) control (score >2) 51 USA communica 96%, High 4.6%, High Satisfaction (n=166, > tion; Work Stress, satisfaction (>3) satisfaction (>3) improved with 52 80% workflow, Control, burnout, 38.5%. CONTROL 40.0%. CONTROL workflow 53 physic- quality Satisfacti chaos, (n=83) High work (n=72) High work interventions, 54Linzer 2015 ians, 47.3 improve- on (from intent to control 13.2%. High control 11.4%, High targeted QI 55 yrs, 48% ment (QI). WLC PWS and leave, satisfaction (>3) satisfaction (>3) projects, RCT male) n=83 n=83 MEMO). varia 51.8%. 45.7%. communication. 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 28 of 40

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Significant 5 improvement on 6 CPD self-esteem post 7 teach- intervention. The ing interactive teaching 8 progr- approach improved 9 CPD am Knowled Time at 6 months. self-esteem more 10 teaching (n=22 Attitude ge, Mean, SD. Self- Mean, SD. Self- than the didactic 11Margalit programs didac- to intention esteem (72.1, 14.5), esteem (76.5, 12.9), teaching did. No 2005 GPs in (n=22 for tic patient , Attitude to patient Attitude to patient improvement on 12 Israel interactive meth- care, self- attitude, care (34.5, 12.2) care (36.5, 19.9) attitude to patient 13RCT (n=44) program) od) esteem burnout (n=44) (n=44) NA care. 14 Brief 15 blended 16 web-based mindful- Time at 4 weeks. 17 ness Mean, SD. 1 Mean, SD. 1 Benefits in pos 18 intervent- For peer reviewsession/wk (n=28): session/wkonly (n=28): PANAS and MAAS 19 ion,1x 4h PANAS-pos (32.19, PANAS-pos (33.44, for two or more 20 face to 6.72). MAAS (64, 5.42). MAAS (66.67, weekly meditation face, 8 Positive 12.07). CD-RISC 10.88). CD-RISC session. No benefits 21 GPs in online affect Negative (38.96, 8.96). 2 (40.19, 5.17). 2 for one weekly 22Montero- Spain sessions (2 (PANAS), affect sessions/wk (n=30). sessions/wk (n=30). practice. While face 23Martin (n=290, weekly awarenes (PANAS), PANAS-pos (32.03, PANAS-pos (35.00, to face attendence 242018 49 yrs, sessions s (MAAS), Burnout 6.38). MAAS (61.77, 4.91). MAAS (66.37, was good, very high 22.5% over 4 wks) resilience subtypes( 13.41). CD-RISC 11.03). CD-RISC attrition rate for 25NCBA males) (n=58). NA (CD-RISC) BCSQ) (38.80, 8.58). (41.28, 4.32). online component. 26 Of six physicians, 27 100% were satisfied 28 Burnout, Time at 3 months. with clinic workflow 29 Time n=6. Number of post-pilot (vs. 33% spent on n=6. Number of responses 'agree' or pre-pilot), and 83% 30 EHR responses 'agree' or 'strongly agree'. were satisfied with 31 documen 'strongly agree'. Satisfied with clinic EHR use post-pilot 32 tation, Satisfied with clinic workflow 6/6 (vs. 17% pre- pilot). 33 patient workflow 2/6 (33%). (100%). Calm Physician burnout Academic satisfacti Calm atmosphere in atmosphere in work was low at baseline 34 general on with work area 0/6 (0%). area 2/6 (33%). and did not change 35 internal doctor– Satisfied with Satisfied with post-pilot. Mean 36 medicine patient quality of patient quality of patient time spent on post- 37 clinic Clinic Physician relations interactions 5/6 interactions 6/6 clinic EHR http://bmjopen.bmj.com/ 38Pozdnya- (n=6 sessions workplac hip, (83%). Satisfied with (100%). Satisfied documentation kova 2018 faculty, with and e attitudes quality of with quality of decreased from 1.65 39 n=325 without a satisfacti towards communication with communication with to 0.76 h per clinic 40NCBA, 2007 patients). scribe. NA on scribes. patient 4/6 (67%). patient 5/6 (83%). session (p = 0.02). 41 Rural 42 medical practiti- 43 oners in MSCR, 44 Australia Mindful 45 (57% Self-Care on September 26, 2021 by guest. Protected copyright. 46 GPs). and For the WHO-5 47 (n=13 Resiliency wellbeing scale total, n=4 program Burnout there was no 48 qual (7h face-to- measure change pre-post for 49 research, face (short one doctor, 50 n=7 session) Wellbein version deterioration for 51Rees 2020 quant and 3x 1h g (WHO- 10-item), Time at 4 weeks. two doctors with no research, video- 5), psycholo n=7. Mean. n=7. Mean. reliable change, 52NCBA, 40 yrs, conference Positive gical Wellbeing (61.1). Wellbeing (71.4). Improvement for 53mixed 76.9% (follow up affect strain Positive affect not Positive affect not four doctors with no 54methods male) sessions) NA (PANAS). (GHQ-12) reported. reported. reliable change. 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 29 of 40 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Participants in the 5 MMC group 6 reported significant 7 improvements over Time at 4 weeks. Time at 3 monthrs. time for MAAS (also 8 Mindful Mean, SD. Mean, SD. Mean, SD. PSS and MBI), 9 Primary Attention INTERVENTION INTERVENTION INTERVENTION whereas in the 10 care Mindful Awarene Stress (n=16) MAAS (3.42, (n=15) MAAS (3.62, (n=13) MAAS (4.04, control group, there 11 physici- Medicine ss (PPS), 0.96), BRS (21.62, 0.89), BRS (22.33, 1.02), BRS (24.15, were no ans in the Curriculum (MAAS), Burnout 4.45), SCBCS (26.31, 4.74), SCBCS (27.66, 5.47), SCBCS (27.84, improvements. 12Schroeder USA = modified Resilienc (MBI), 4.51). CONTROL 3.22). CONTROL 4.09). CONTROL There were no 132018 (n=33, version of e (BRS), Meditati (n=17) MAAS (3.32, (n=14) MAAS (3.08, (n=13) MAAS (3.18, significant 14 42.76 yrs, MBSR, 1x Compassi on 0.76), BRS (18.70, 0.76), BRS (19.42, 0.58), BRS (18.82, improvement for 15RCT, 27% 13h & 2x WLC on Practice 5.13), SCBCS (27.00, 4.21), SCBCS (26.07, 5.32), SCBCS (25.07, resilience or 162014 - 2015 male) 2h (n=16) (n=17) (SCBC), (MPQ) 4.97). 4.73). 5.85). compassion. Significant 17 improvement of 18 For peer review only self-compassion and 19 Primary self-reflection. Six 20 care months after PSS physic- and SCS were still 21 ians in improved. PSS 22Van the Adapted significantly 23Wietmar- Nether- MBSR reduced. Qual: 24schen 2018 lands programme Self- awareness, (n=54, , weekly Compassi Cohen acceptance, 25NCBA, 87% GPs, group on (SCS), Perceive Time at 8 weeks. Time at 6 months. peacefulness and 26mixed 40 yrs, sessions for Self- d Stress Mean, SD. SCS (2.9, Mean, SD. SCS (3.4, Mean, SD. SCS (3.7, openness improved 27methods, 22% 8 weeks, 26 reflection Scale 0.7, n=50), GRAS 0.6, n=50), GRAS 0.7, n=21), GRAS through 282015 - 2016 male) hours total. NA (GRAS). (PSS) (87.6, 7.7, n=44). (90.9, 6.7, n=44). (90.2, 10.9, n=17). intervention. 29 Time at 8 weeks. Mean, SE. 30 Mean, SE. INTERVENTION 31 INTERVENTION (n=43) JSPE (119.35, 32 (n=43) JSPE (117.4, 1.49), FFMQ Total 33 1.53), FFMQ Total (143.08, 2.19), (136.21, 2.23), FFMQ Observing 34 FFMQ Observing (28.4, 0.68), FFMQ 35 (26.36, 0.69), FFMQ Describing (29.77, 36 Describing (28.26, 0.71), FFMQ Acting 37 0.9), FFMQ Acting with awareness http://bmjopen.bmj.com/ 38 with awareness (28.1, 0.69), FFMQ (27.12, 0.71), FFMQ Non-judging (32.36, 39 Non-judging (31.16, 0.79), FFMQ Non- 40 0.81), FFMQ Non- reacting (24.47, 41 reacting (23.34, 0.56, 0.57). Mindfulness skills 42 0.57). CONTROL CONTROL (n=20) increased (n=20) JSPE (116.18, JSPE (117.93, 1.98). significantly in the 43 1.92). FFMQ Total FFMQ Total (135.45, MBSR group. 44 General (135.48, 2.65), 2.67), FFMQ Empathy remained 45 practiti- FFMQ Observing Observing (26.33, the same. The on September 26, 2021 by guest. Protected copyright. 46 oner (25.85, 0.82), FFMQ 0.83), FFMQ qualitative data 47 trainers Describing (28.33, Describing (28.2, indicated that the in the 0.9), FFMQ Acting 0.91), FFMQ Acting MBSR course 48 Nether- Empathy with awareness (28, with awareness increased their 49Verweij lands MBSR (JSPE-20), Work 0.9), FFMQ Non- (27.41, .91), FFMQ wellbeing and 502016 (n=50, training 8x Mindfuln engagem judging (30.49, Non-judging (30.32, compassion towards 51 54.9 yrs, 2.5h, 1x 8h ess ent, 0.98), FFMQ Non- 0.99), FFMQ Non- themselves and CBA, mixed 66% retreat WLC (FFMQ- Burnout reacting (23.12, reacting (23.49, others, including 52methods male) (n=30) (n=20) 39) (UBOS-C) 0.7). 0.7). their patients. 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 30 of 40

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Statistically 5 significant 6 GPs in improvement in job 7 the UK. satisfaction. Job 2004: pressure and work 8 n=2105, hours significantly 9 mean New declined. Most GPs 10 age introduct- reported that the 11 46.9yrs, ion of pay new contract had 66% for increased their 12 male. perform- income (88%), but 13 2005: ance decreased their 14Whalley n=1349, system Job professional 152008 mean happened pressure, autonomy (71%), 16 age in 2004 Job Time at one year. and increased their Panel 48.6yrs, (after the design Mean, SD. 2004: JSS Mean, SD. 2005: JSS administrative 17survey, 65 % 2004 and time total (4.58, 1.39, total (5.17, 1.28, (94%) and clinical 182004, 2005 male. survey) NA ForJSS WCW peerpressures reviewn=2081) n=1345)only (86%) workloads. 19 20 Key: RCT Randomised Controlled Trial; CBA Controlled Before and After Trial; NCBA Non-Controlled 21 Before and After Trial; QOF Quality and Outcomes Framework; P4P Pay for Performance; NP Nurse 22 Practitioner; WCW Warr Cook Wall; JS Job satisfaction; JSS Job Satisfaction Scale; MBI Maslach Burnout 23 Inventory; JSPE Jefferson Scale of Physician Empathy); POMS Profile Of Mood States; FFMQ Five Facet 24 Mindfulness Questionnaire; SRSI Smith Relaxation States Inventory; NA Not applicable; PBS Physician 25 Belief Scale; SCS Self Compassion Scale (Neff); MPQ Meditation Practice Questionnaire, FR-JS Font Roja 26 Job Satisfaction Questionnaire; BPRS Brief Psychiatric Rating Scale; IQR Inter Quartile Range; ACP College 27 of Physicians; ASIM American Society of Internal Medicine; DASS Depression and Anxiety Scale; PJSS 28 Physician Job Satisfaction Scale (3 dimensions JS, career satisfaction, and specialty satisfaction), 12-item 29 Global Job Satisfaction subscale used; UWES Utrecht Work Engagement Score; CD-RISC Connor- 30 Davidson Resilience Scale; PANAS Positive And Negative Affect Scale, MAAS Mindful attention 31 awareness scale; GHQ General Health Questionnaire; AWS Areas of Work Life Scale (positive measures 32 include control, reward, community, values); BRS Brief Resilience Scale; SCBCS Santa Clara Brief 33 Compassion Scale (an abbreviation of the Sprecher and Fehr’s Compassionate Love Scale); MPQ 34 Meditation Practice Questionnaire; UBOS-C Utrecht Burnout Scale for Contactual Occupations (this is 35 the validated Dutch version of the Maslach Burnout Inventory); GRAS Groningen Reflection Ability Scale; 36 EHR Electronic Health Record; MTP Multimodal training programme; IBST integrated brief systemic 37 therapy, PPS Perceived Stress Scale. http://bmjopen.bmj.com/ 38 39 40 41 42 43 44 45 on September 26, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 31 of 40 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Table 2. Overview of included studies 5 6 7 Within group Between group 8 Study effect sizes effect sizes (exp 9 Intervention level Intervention type Study ID (Author, Date) design Outcome Measure (experiment) vs control group) 10 11 Intervention targeted 12 at the individual Mindfulness Amutio 2015 RCT FFMQ 0.78 1.5 13 Asuero 2014 RCT FFMQ, JSPE 0.64, 0.31 0.57, 0.44 14 Schroeder 2018 RCT MAAS, BRS, SCBC 0.63, 0.51, 0.35 0.88, 0.61, 0.73 15 Verweij 2016 CBA FFMQ, JSPE 0.47, 0.20 0.5, 0.02 16 17 Fortney 2013 NCBA SCBC, RS-14 -0.04, 0.17 - 18 ForMontero-Marin peer 2018 reviewNCBA PANAS, only MAAS, CD-RISC 0.52, 0.37, 0.37 - 19 Rees 2020 NCBA PANAS, WHO-5 -, 0.52 - 20 21 Van Wietmarschen 2018 NCBA SCS, GRAS 0.77, 0.46 - 22 Krasner 2009 NCBA FFMQ-2 (BAER), JSPE 1.05, 0.44 - 23 Educational training / 24 experiential workshop Barcons 2019 CBA FR-JS - - 25 Gardiner 2004 CBA WRM-7, QoWL-6 0.43, 0.45 0.3, 0.27 26 27 Margalit 2005 RCT Self-esteem - - 28 Coaching Dyrbye 2019 RCT CD-RISC, PJSS 0.24, 0.10 0.13, 0.06 29 Professional balance, 30 Intervention at Clerical support / physician satisfaction 31 organisational level Scribes Contratto 2017 NCBA (QoL) - - 32 Physician workplace Pozdnyakova 2018 NCBA satisfaction - - 33 34 Organisational improvement 35 programme / QA Dunn 2007 NCBA Physician satisfaction - - 36 Work control, 37 Linzer 2015 RCT satisfaction - - http://bmjopen.bmj.com/ 38 Intervention at 39 Introduction of pay for NCBA Job satisfaction (JSS systemic level / policy 40 performance scheme Allen 2017 (naturalistic) WCW), Life satisfaction - - 41 42 NCBA 43 Whalley 2008 (naturalistic) JSS WCW 0.44 - 44 45 Key: Study design: RCT Randomised Controlled Trial; CBA Controlled Before and After trial; NCBA Non- on September 26, 2021 by guest. Protected copyright. 46 Controlled Before and After trial. 47 Outcome measures: JSS WCW Warr Cook Wall Job satisfaction Scale; JSPE Jefferson Scale of Physician 48 Empathy; FFMQ Five Facet Mindfulness Questionnaire; SRSI Smith Relaxation States Inventory; SCS Self 49 Compassion Scale (Neff); FR-JS Font Roja Job Satisfaction Questionnaire; SCBC Santa Clara brief 50 compassion scale (an abbreviation of the Sprecher and Fehr’s Compassionate Love Scale); PJSS Physician 51 Job Satisfaction Scale (3 dimensions JS, career satisfaction, and specialty satisfaction), 12-item Global 52 Job Satisfaction subscale used; CD-RISC Connor-Davidson Resilience Scale; PANAS Positive And Negative 53 Affect Scale, MAAS Mindful attention awareness scale; BRS Brief Resilience Scale; GRAS Groningen 54 Reflection Ability Scale; RS-14 Resilience, WHO-5 Wellbeing Index, WRM Work Related Morale Measure. 55 QoWL-6 Quality of Work Life. 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 32 of 40

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Table 3. Overview of within group and between group Effect Sizes (ES) for several 5 6 positive outcomes of mindfulness interventions 7 8 9 Outcome Mindfulness Resilience Compassion Empathy 10 11 12 13 Within 14 15 group ES 0.37 – 1.05 0.17 – 0.51 -0.04 - 0.77 0.2 – 0.44 16 17 Mindfulness (k=9) 18 interventions For peer review only 19 20 (k=9) 21 22 Between Only one Only one 0.02 - 0.44 23 group ES 0.5 – 1.5 study study 24 25 (k=4) 0.61 0.73 26 27 28 29 30 There were 9 studies that trialled mindfulness interventions (k=9). These studies utilized a range of 31 different outcomes measures and included tools for assessing mindfulness, resilience, compassion and 32 empathy. Within group effect sizes (ES) are shown, comparing before and after measures for the 33 interventi 34 35 36 37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44 45 on September 26, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 33 of 40 BMJ Open

1 2 3 Figure 1. PRISMA diagram BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 5 6 14792 studies imported for screening 1330 studies before yr 2000 removed 7 1861 Medline 8 3871 CINAHL 9 2746 Embase 10 5514 Scopus 800 PsycINFO 11 12 13 14 15 13462 studies 2715 duplicates removed 16 17 18 For peer review only 19 10747 studies 12 studies added through hand search 20 21 22 23 24 10759 study titles screened 25 9682 studies irrelevant 26 27 28 29 1077 study abstracts screened 983 studies excluded 30 351 wrong study design 31 179 wrong participants 32 172 wrong publication type 33 148 no intervention 147 wrong outcome 34 86 wrong setting 35 36 37 http://bmjopen.bmj.com/ 38 94 full-text studies assessed for 75 studies excluded 39 eligibility 20 wrong publication type 40 19 wrong participants 41 13 wrong outcome 42 11 wrong study design 43 5 wrong language 3 unavailable 44 2 no intervention 45 1 wrong setting on September 26, 2021 by guest. Protected copyright. 46 47 48 49 50 19 studies included 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 34 of 40 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

1 2 3 Supplements 4 5 6  Supplement 1. Medline OVID search strategy 7 8  Supplement 2. Formulae used for Effect Size calculations in Excel for within and between group ES. 9 10  Supplement 3. Effect sizes for outcome measures and sub-measures 11  Supplement 4. Outcome measures utilized in the included studies. 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40 41 on September 26, 2021 by guest. Protected copyright. 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 1 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 35 of 40 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

1 2 3 Supplement 1. Medline OVID search strategy 4 5 1. general practitioners/ 6 7 2. "general practitioner*".mp. 8 9 3. (GP or GPs).mp. 10 11 4. exp General Practice/ 12 5. Family Practice/ or Primary Health Care/ 13 14 6. "family practitioner*".mp. 15 16 7. "primary care practitioner*".mp.For peer review only 17 18 8. physicians, family/ or physicians, primary care/ 19 9. "family medicine".mp. 20 21 10. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 22 23 11. ("psychological well being" or "psychological wellbeing").mp. 24 25 12. (wellbeing or "well being").mp. 26 13. ("cognitive well being" or "cognitive wellbeing").mp. 27 28 14. flourishing.mp. 29 30 15. ((job or work) adj3 (satisf* or engage* or motivat*)).mp. 31 32 16. life satisfaction*.mp. 33 http://bmjopen.bmj.com/ 17. Job Satisfaction/ 34 35 18. resilien*.mp. 36 37 19. Resilience, Psychological/ 38 39 20. burnout.mp. 40

21. Burnout, Professional/ on September 26, 2021 by guest. Protected copyright. 41 42 22. Compassion fatigue/ 43 44 23. compassion fatigue.mp. 45 46 24. joy.mp. 47 25. "joy in practice".mp. 48 49 26. 11 or 12 or 13 or 14 or 15 or 16 or 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 50 51 27. Motivational interviewing/ 52 53 28. Mindfulness/ 54 29. Mediation/ 55 56 57 58 59 2 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 36 of 40 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

1 2 3 4 30. mindful*.mp. 5 31. meditat*.mp. 6 7 32. Self-Care/ 8 9 33. ("self care" or "self-care").mp. 10 11 34. Motivation/ 12 35. Professional autonomy/ 13 14 36. ("self determination" or "self-determination").mp. 15 16 37. exp Randomised ControlledFor Trial/ or Clinicalpeer Trial/ review only 17 18 38. trial.mp. 19 39. Early Medical Intervention/ or intervention.mp. 20 21 40. Counseling/ 22 23 41. Course?ling.mp. 24 25 42. Mentoring/ 26 43. mentor*.mp. 27 28 44. Education/ 29 30 45. education.mp. 31 32 46. workshop*.mp. http://bmjopen.bmj.com/ 33 47. training.mp. 34 35 48. coaching.mp. 36 37 49. facilitat*.mp. 38 39 50. 27 or 28 or 29 or 30 or 31 or 32 or 33 or 34 or 35 or 36 or 37 or 38 or 39 or 40 or 41 or 42 or 43 or 44 or 45 or 46 or 47 or 48 or 49 40 on September 26, 2021 by guest. Protected copyright. 41 51. 10 and 26 and 50 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 3 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 37 of 40 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

1 2 3 Supplement 2. Formulae used for Effect Size calculations in Excel for within and between group ES. 4 5 6 7 Within-group: 8 9 Mean_change_WG = Mean_Exp_post – Mean_Exp_pre 10 11 12 SD* pooled _WG = SQRT(((N_Exp_post-1) * SD_Exp_post^2) + (N_Exp_pre-1) * SD_Exp_pre^2)) / (N_Exp_post + N_Exp_pre – 2)) 13 14 15 ES_WG = Mean_change_WG / SD* pooled_WG 16 For peer review only 17 18 Between-group: 19 20 Mean_change_BG = (Mean_Exp_post – Mean_Exp_pre) – (Mean_Ctrl_post – Mean_Ctrl_pre) 21 22 23 SD* pooled_BG = SQRT(((N_Ctrl_pre-1) * SD_Ctrl_pre^2) + (N_Exp_pre-1) * SD_Exp_pre^2)) / (N_Ctrl_pre + N_Exp_pre – 2)) 24 25 26 ES_BG = Mean change_BG / SD* pooled_BG 27 28 29 30 Legend. ES = Effect size, WG = Within group, BG = Between group, Exp = experimental group, Ctrl = 31 control group, N = sample size, pre = pre-intervention measure, post = post-intervention measure, SQRT = 32

square root, ^2 = raised to the power of two. http://bmjopen.bmj.com/ 33 34 35 36 37 38 39 40 41 on September 26, 2021 by guest. Protected copyright. 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 4 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 38 of 40 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

1 2 3 Supplement 3. Effect sizes for outcome measures and sub-measures 4 5 Between group 6 Within group ES Hedges' g absolute mean 7 absolute mean Exp (Delta change SD* ES Hedges' g Name, change (post - SD* Exp mean/ SD* (experiment – Baseline (Delta Mean / 8 n pre/post Measure pre experiment) pooled pool) control group) pooled SD* pool) 9 10 With control 11 group 12 AMUTIO* FFMQ 0.37 0.48 0.78 0.54 0.36 1.51 13 n=21/21 Observe 0.65 0.62 1.05 0.84 0.75 1.12 14 Describe 0.25 0.67 0.37 0.31 0.63 0.49 15 Awareness 0.32 0.77 0.42 0.34 0.79 0.43 16 Non-judgement 0.36 0.65 0.55 0.39 0.58 0.67 Non-reactiveFor peer 0.29review0.57 only0.51 0.26 0.54 0.48 17 SRSI-3 Relaxation 0.54 0.57 0.95 0.46 0.52 0.89 18 SRSI-3 Pos energy 0.71 0.74 0.97 0.71 0.63 1.13 19 SRSI-3 Mindfulness 0.71 0.81 0.88 0.76 0.91 0.84 20 SRSI-3 Transcend 0.6 0.93 0.65 0.84 0.88 0.96 21 VERWEIJ** FFMQ 6.87 14.49 0.47 6.9 13.82 0.50 22 n=43/43 exp, 23 n=20/20 control Observe 2.04 4.49 0.45 1.56 4.28 0.36 24 Describe 1.51 4.69 0.32 1.64 4.52 0.36 25 Awareness 0.98 4.59 0.21 1.57 4.47 0.35 Non-judgement 1.2 5.25 0.23 1.37 5.04 0.27 26 Non-reactive 1.13 3.71 0.30 0.76 3.56 0.21 27 JSPE total 1.95 9.90 0.20 0.2 9.61 0.02 28 29 ASSUERO FFMQ 12.40 19.50 0.64 11.3 19.65 0.57 30 n=43/43 exp 31 n=25/25 control Observe 4.30 7.07 0.61 4.7 7.33 0.64 32 Describe 0.70 5.50 0.13 0.6 5.52 0.11

Awareness 2.10 5.33 0.39 2.7 6.24 0.43 http://bmjopen.bmj.com/ 33 Non-judgement 3.50 7.05 0.50 2.4 7.35 0.33 34 Non-reactive 2.20 3.71 0.59 1.8 3.85 0.47 35 JSPE total 3.50 11.32 0.31 5.3 12.10 0.44 36 Perspective 1.30 5.88 0.22 1.5 6.62 0.23 37 Compassion 1.80 4.54 0.40 2.3 5.03 0.46 In patients shoes 0.40 1.49 0.27 1.5 1.54 0.98 38 39 DYRBYE CD-RISC-10 1.30 5.44 0.24 0.7 5.46 0.13 40 n=40/40 exp 41 n= 39/39 control PJSS 1.00 9.94 0.10 0.6 10.11 0.06 on September 26, 2021 by guest. Protected copyright. UWES vigor 0.30 0.94 0.32 0.2 1.04 0.19 42 UWES dedication 0.10 0.87 0.11 -0.1 0.96 -0.10 43 UWES absorption -0.10 0.88 -0.11 -0.2 1.00 -0.20 44 Empowerment at work 2.7 11.29 0.24 -0.3 12.56 -0.02 45 46 GARDINER WRM-7 total 2.79 6.46 0.43 2 6.63 0.30 47 n=86/77 exp, 48 n=24/19 control QoWL-6 total 2.92 6.50 0.45 1.76 6.48 0.27 49 50 SCHROEDER MAAS 0.62 0.99 0.63 0.76 0.86 0.88 51 N=16/13 exp, 52 n=17/13 control BRS 2.53 4.93 0.51 2.95 4.81 0.61 SCBC 1.53 4.33 0.35 3.46 4.75 0.73 53 54 MARGALIT*** 55 n=44/44 Self-esteem 4.4 13.72 0.32 56 57 58 59 5 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 39 of 40 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

1 2 3 4 5 6 7 Without control 8 group 9 10 MONTERO- MARIN 11 n=30/30 PANAS 1 session/wk 1.25 6.10 0.20 12 13 PANAS 2 session/wk 2.97 5.69 0.52 MAAS 1 session/wk 2.67 11.49 0.23 14 MAAS 2 session/wk 4.59 12.28 0.37 15 CD-RISC total 1 16 session/wkFor peer 1.23review7.31 only0.17 17 CD-RISC total 2 18 session/wk 2.48 6.79 0.37

19 VON 20 WIETMARSCHEN SCS 0.5 0.65 0.77 - 21 n=50/50 GRAS 3.3 7.22 0.46 - 22 FORTNEY RS-14 2.1 12.28 0.17 - 23 n=30/28 SCBC -0.2 4.54 -0.04 - 24 25 KRASNER JSPE total 4 9.16 0.44 - 26 n=60/59 Perspective 2 5.79 0.35 - 27 Compassion 1.6 4.24 0.38 - In patients’ shoes 0.8 2.12 0.38 - 28 Mindfulness 2 factor 29 FFMQ (BAER) total 7.7 7.33 1.05 - 30 BAER observe 5 4.63 1.08 - 31 BAER non-react 3.2 3.95 0.81 -

32 REES WHO-5 (WB) 10.3 19.79 0.52 - http://bmjopen.bmj.com/ 33 n=7/7 PANAS-pos - - 34 35 WHALLEY JSS WCW 0.59 1.35 0.44 - n=2081/1345 36 37 38 39 Notes: Amutio*, no information given on pre-post sample size, assumption made that there was no attrition. 40 Verweij**, wait list control participants were included in both intervention and control groups. *** Margalit only 41 pooled group (experimental and control) analysis provided. For the FFMQ Verweij and Assuero have much larger on September 26, 2021 by guest. Protected copyright. 42 numbers than Amutio, this is presumably due to reporting of the total score count as opposed to the total mean 43 scores. 44 45 Outcome measures: JSS WCW Warr Cook Wall Job satisfaction Scale; JSPE Jefferson Scale of Physician Empathy; 46 FFMQ Five Facet Mindfulness Questionnaire; Baer (2 item version of FFMQ); SRSI Smith Relaxation States 47 Inventory; SCS Self Compassion Scale (Neff); FR-JS Font Roja Job Satisfaction Questionnaire; SCBC Santa Clara brief 48 compassion scale (an abbreviation of the Sprecher and Fehr’s Compassionate Love Scale); PJSS Physician Job 49 Satisfaction Scale (3 dimensions JS, career satisfaction, and specialty satisfaction), 12-item Global Job Satisfaction 50 subscale used; CD-RISC Connor-Davidson Resilience Scale; PANAS Positive And Negative Affect Scale, MAAS 51 52 Mindful attention awareness scale; BRS Brief Resilience Scale; GRAS Groningen Reflection Ability Scale; RS-14 53 Resilience, WHO-5 Wellbeing Index, WRM Work Related Morale Measure. QoWL-6 Quality of Work Life. 54 55 56 57 58 59 6 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 40 of 40 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

1 2 3 Supplement 4. Outcome measures utilized in the included studies. 4 5 No of 6 studies 7 Acronym Outcome measure used in 8 9 ACP/ASIM Physician satisfaction (Am College of Phys / Am Soc of Int Med) 1 10 BAER Baer mindfulness scale (2 item version of FFMQ) 1 11 BRS Brief resilience scale 1 12 13 CD-RSIC-10 Connor-Davidson Resilience Scale 2 14 FFMQ Five Facet Mindfulness Questionnaire 3 15 FR-JS Font Roja Job Satisfaction scale 1 16 For peer review only 17 GRAS Groningen Reflection Ability Scale 1 18 JSPE Jefferson Scale of Physician Empathy 3 19 JSS WCW Job Satisfaction Scale Warr-Cook-Wall 2 20 21 LS Life satisfaction 1 22 MAAS-15 Mindful attention awareness scale 2 23 PANAS Positive and negative affect scale 2 24 25 PJSS Physician Job satisfaction scale 1 26 PWS (MEMO) Physician Workplace Satisfaction 2 27 QoWL-6 Quality of work life 1 28 29 RS-14 Resilience 1 30 SCBC Santa Clara brief compassion scale 2 31 SCS Self compassion scale (Neff) 1 32

SRSI-3 Smith Relaxation states 1 http://bmjopen.bmj.com/ 33 34 UWES Utrecht work engagement score 1 35 WHO-5 Wellbeing Index 1 36 WRM-7 Work related morale measure 1 37 38 Self-esteem 1 39 40 41 on September 26, 2021 by guest. Protected copyright. 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 7 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

Page 41 of 40 BMJ Open

1 2 The PRISMA for Abstracts Checklist 3 4 5 TITLE CHECKLIST ITEM REPORTED 6 ON PAGE # 7 1. Title: Identify the report as a systematic review, meta-analysis, or both. 1 8 9 BACKGROUND 10 11 2. Objectives: The research question including components such as participants, interventions, comparators, and outcomes. 2 12 METHODS For peer review only 13 14 3. Eligibility criteria: Study and report characteristics used as criteria for inclusion. 2-4 15

16 4. Information sources: Key databases searched and search dates. http://bmjopen.bmj.com/ 17 18 5. Risk of bias: Methods of assessing risk of bias. 19 RESULTS 20 21 6. Included studies: Number and type of included studies and participants and relevant characteristics of studies. 4-8 22 23 7. Synthesis of results: Results for main outcomes (benefits and harms), preferably indicating the number of studies and participants for

24 each. If meta-analysis was done, include summary measures and confidence intervals. on September 26, 2021 by guest. Protected copyright. 25 26 8. Description of the effect: Direction of the effect (i.e. which group is favoured) and size of the effect in terms meaningful to clinicians and 27 patients. 28 29 DISCUSSION 30 9. Strengths and Limitations Brief summary of strengths and limitations of evidence (e.g. inconsistency, imprecision, indirectness, or risk of 8-12 31 of evidence: bias, other supporting or conflicting evidence) 32 33 10. Interpretation: General interpretation of the results and important implications 34 35 OTHER 36 37 11. Funding: Primary source of funding for the review. 13 38 12. Registration: Registration number and registry name. 39 40 41 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

The effect of interventions for the wellbeing, satisfaction and flourishing of general practitioners – A systematic review ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2020-046599.R1

Article Type: Original research

Date Submitted by the 26-May-2021 Author:

Primary Subject

General practice / Family practice http://bmjopen.bmj.com/ Heading:

Secondary Subject Heading: General practice / Family practice

PRIMARY CARE, GENERAL MEDICINE (see Internal Medicine), MENTAL Keywords: HEALTH, Organisational development < HEALTH SERVICES ADMINISTRATION & MANAGEMENT

on September 26, 2021 by guest. Protected copyright.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 44 BMJ Open

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 44

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 TITLE PAGE 5 6 Title The effect of interventions for the wellbeing, satisfaction, and flourishing 7 of general practitioners – A systematic review 8 Authors Diana Naehrig, ORCID (0000-0002-1162-4092), 9 10 Dr.med., FMH Radioonkologie, MSc Coach Psych, 11 PhD candidate, The University of Sydney, Faculty of Health and Medicine, 12 Level 5, Professor Marie Bashir Centre, Missenden Road, Camperdown, 13 NSW 2050, Australia, [email protected]; 14 15 Aaron Schokman, ORCID (0000-0003-0419-9347), BSc, MPhil, 16 17 PhD candidate, The University of Sydney, Faculty of Health and Medicine, 18 ForLevel 5,peer Professor Mariereview Bashir Centre, only Missenden Road, Camperdown, 19 NSW 2050, Australia, [email protected]; 20 21 Jessica Kate Hughes, ORCID (0000-0003-0046-1686), BA (Hons), 22 Grad Dip Arts (Art History and Curatorship), Master of Information 23 Studies (Librarianship), Assistant Librarian, The University of Sydney 24 25 Library, Rm 202, Fisher Library F03, The University of Sydney, NSW 26 2006, Australia, [email protected]; 27 28 Ron Mark Epstein, ORCID (0000-0002-3564-9163), MD 29 Co-Director, Center for Communication and Disparities Research, 30 Professor of Family Medicine, Oncology and Medicine (Palliative Care), 31 American Cancer Society of Clinical Research Professor, University of 32 33 Rochester School of Medicine and Dentistry, 1381 South Avenue, 34 Rochester, NY 14620, USA, [email protected]; 35 36 Ian Hickie, ORCID (0000-0001-8832-9895), AM MD FRANZCP FASSA 37 FAHM http://bmjopen.bmj.com/ 38 Co-Director, Health and Policy and Professor of Psychiatry, NHMRC 39 40 Senior Principal Research Fellow, The University of Sydney, Faculty of 41 Health and Medicine, Brain and Mind Centre, Level 4, Building F, 94 42 Mallett Street, Camperdown, NSW 2050, Australia, 43 [email protected]; 44 45 Nick Glozier, ORCID (0000-0002-0476-9146), on September 26, 2021 by guest. Protected copyright. 46 Professor of Psychological Medicine, The University of Sydney, Faculty of 47 48 Health and Medicine, Central Clinical School, Level 5, Professor Marie 49 Bashir Centre, Missenden Road, Camperdown, NSW 2050, Australia, 50 [email protected] 51 Corresponding Diana Naehrig, The University of Sydney, Faculty of Health and Medicine, 52 author Level 5, Professor Marie Bashir Centre, Missenden Road, Camperdown, 53 NSW 2050, Australia, [email protected] 54 55 56 Word count 3995 57 manuscript 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 44 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44 45 on September 26, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 44

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 ABSTRACT 5 6 7 Objectives 8 Clinician wellbeing has been recognised as an important pillar of health care. However, 9 research mainly addresses mitigating the negative aspects of stress or burnout, rather 10 than enabling positive aspects. With the added strain of a pandemic, identifying how 11 best to maintain and support the wellbeing, satisfaction and flourishing of general 12 practitioners (GPs) is now more important than ever. 13 14 15 Design 16 Systematic review. 17 18 Data sources For peer review only 19 We searched MEDLINE, PsycINFO, Embase, CINAHL, and Scopus from 2000 to 20 2020. 21 22 23 Study selection 24 Intervention studies with more than 50% GPs in the sample evaluating self-reported 25 wellbeing, satisfaction and related positive outcomes were included. The Cochrane Risk 26 of Bias 2 tool was applied. 27 28 29 Results 30 We retrieved 14,792 records, 94 studies underwent full text review. We included 19 31 studies in total. Six randomised controlled trials, three non-randomised, controlled 32 trials, eight non-controlled studies of individual or organisational interventions with a 33 total of 1141 participants. There were two quasi-experimental articles evaluating health 34 system policy change. 35 36 37 Quantitative and qualitative positive outcomes were extracted and analysed. Individual http://bmjopen.bmj.com/ 38 mindfulness interventions were the most common (k=9) with medium to large within 39 (0.37-1.05) and between group (0.5-1.5) effect sizes for mindfulness outcomes, and 40 small to medium effect sizes for other positive outcomes including resilience, 41 compassion, and empathy. Studies assessing other intervention foci or other positive 42 outcomes (including wellbeing, satisfaction) were of limited size and quality. 43 44 45 Conclusions on September 26, 2021 by guest. Protected copyright. 46 There is remarkably little evidence on how to improve GPs wellbeing beyond using 47 mindfulness interventions, particularly for interventions addressing organisational or 48 system factors. This was further undermined by inconsistent reporting, and overall high 49 risk of bias. We need to conduct research in this space with the same rigour with which 50 we approach clinical intervention studies in patients. 51 52 53 Registration 54 The systematic review protocol is registered on PROSPERO CRD42020164699. 55 . 56 57 58 Funding source 59 Dr Diana Naehrig is funded through the Raymond Seidler PhD scholarship. 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 44 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 ARTICLE SUMMARY 5 6 Strengths and limitations of this study 7 8  Whilst burnout has been a general focus of research, interventions to improve 9 10 the wellbeing of general practitioners (primary care physicians) appear sparse. 11  The strength of this study is the extensive and systematic approach taken to 12 13 evaluate interventions aimed at improving wellbeing, satisfaction, flourishing, 14 15 and other positive outcomes in general practitioners. 16 17  The systematic review was registered on PROSPERO a priori before 18 commencingFor the peerdata selection review and extraction process.only 19 20  English only articles were included. 21 22  The limitation of this systematic review is the dependency on the number of 23 24 retrieved and included publications, and their quality of methodology and 25 reporting. 26 27 28 29 30 31 INTRODUCTION 32 Mental ill-health, burnout and stress amongst health care practitioners are a huge 33 34 concern internationally with impacts on individual doctors and their families, patient 35 36 care, and the sustainability of the health care system (1). The wellbeing of clinicians has 37 http://bmjopen.bmj.com/ 38 been recognised as crucial, and has been added to the more commonly shared health 39 system goals of: improved care for individuals, better population health, and reduced 40 41 health care costs (2-5). Despite this stated aim, few studies have evaluated interventions 42 43 to improve wellbeing, satisfaction, and flourishing in general practitioners (GPs) (6-12), 44

typically taking a more traditional, problem-focused approach, such as investigating on September 26, 2021 by guest. Protected copyright. 45 46 causes and reduction of burnout and stress. 47 48 49 50 Medical doctors that provide primary care to patients are the backbone of health care 51 provision. In Australia, the UK and Europe, typically the term ‘general practitioner’ 52 53 (GP) is used, whereas North American articles generally refer to ‘primary care 54 55 physicians’ or ‘family doctors’. Terms are used interchangeably. 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 44

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Satisfaction and wellbeing in GPs are known to be associated with several factors. In 5 6 the USA, primary care physicians are less satisfied with their job than specialists (13, 7 8 14). A ten-year decline in job satisfaction for British GPs has been reported (15), and a 9 10 Norwegian longitudinal study reported dwindling GP satisfaction over seven years (16). 11 In primary care, job satisfaction correlates with practice resources, an ordered 12 13 atmosphere, a practice culture that enables communication, and ease of providing 14 15 quality care (13, 17, 18), and is inversely related to isolation and low sense of 16 community (19, 20). 17 18 For wellbeing, MurrayFor et al.peer conducted review a cross-sectional only study exclusively exploring 19 20 positive mental health and psychological resources (wellbeing, resilience, self-efficacy, 21 22 and hope) of GPs (21). GPs positive mental health was comparable to the general 23 population and older and female GPs fared best, suggesting interventions for younger 24 25 and male GPs may be most useful. 26 27 28 29 Systematic reviews and interventions aimed at improving satisfaction and wellbeing in 30 GPs appear sparse. A European collaboration conducted a systematic review and 31 32 qualitative study looking at positive determinants of satisfaction in GPs. They identified 33 34 general work-related factors (i.e., workload, responsibility, recognition, and income), 35 and GP specific factors (i.e., competence, intellectual stimulus, and work-life balance) 36 37 (22, 23). However, there does not appear to be a systematic review looking at http://bmjopen.bmj.com/ 38 39 interventions to improve satisfaction exclusively in GPs. 40 41 A systematic review of interventions to improve the psychological wellbeing of GPs 42 identified only four articles; two cognitive-behavioural, one mindfulness-based 43 44 intervention and one self-help information approach (24). They applied a dual model of 45 on September 26, 2021 by guest. Protected copyright. 46 languishing/flourishing and the presence of mental illness/absence of mental illness 47 48 matrix (25-28). 49 50 51 Overall, little seems known about which interventions are efficacious in promoting GPs 52 53 wellbeing and satisfaction. In contrast to more extensive research on burnout, distress, 54 and mental ill health with a view to treat, avoid or mitigate negative outcomes, we 55 56 explicitly aimed to apply a positive lens and focus on interventions that enhance GPs’ 57 58 satisfaction and wellbeing, or promote environments and individual behaviours that may 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 44 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 better enable wellbeing. We purposely included interventions on any level; directed at 5 6 the individual (i.e., training, workshops), the organisation (i.e., work-flow 7 8 improvements in the practice) and on a systemic level (i.e., policy change). It is useful 9 10 to bear in mind that GPs are typically high functioning individuals and are not a priori a 11 clinical population, which is why we believe a positive framework is most preproperate. 12 13 Currently, reports indicate that COVID-19 places enormous additional strain on health 14 15 professionals which impacts their physical, mental, and social wellbeing (29, 30). GPs 16 as the first port of call may be particularly affected, whilst playing a crucial role in 17 18 supporting populationFor health peer (31-33). Effortsreview must be made only to understand how GPs 19 20 remain well, and if necessary, put measures in place to assist this. 21 22 23 OBJECTIVES 24 25 We systematically reviewed studies exploring the effect of any type of intervention on 26 27 the wellbeing, satisfaction, and wellbeing of GPs. We broaden and expand upon the 28 29 existing literature, deliberately including any type of intervention, and a range of 30 positive outcomes, and explore if there have been more recent intervention studies 31 32 conducted in this field. 33 34 35 METHODS 36 37 Data sources and search strategy http://bmjopen.bmj.com/ 38 39 We followed PRISMA guidelines (34), and consulted a specialist librarian (JKH). 40 MEDLINE, PsycINFO, Embase, CINAHL and Scopus were searched on 13th January 41 42 2020. 43 44 PICO search terms included GPs and synonyms; interventions and implementations 45 on September 26, 2021 by guest. Protected copyright. 46 directed at the individual, the level of the organisation or practice and the health care 47 system; outcomes included wellbeing, satisfaction, flourishing, synonyms and antonyms 48 49 (search strategy, supplement 1). Titles, abstracts, text, key terms, and subject headings 50 51 were searched for English publications. Eligible articles and related systematic reviews 52 53 were hand-searched for further relevant references, and authors were asked to supply 54 full text articles where relevant conference abstracts only were available. 55 56 57 58 Study selection and data extraction 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 44

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Due to the large number of records, sample screening of 1% of titles (n=107) was 5 6 performed and discussed by two reviewers (DNA, AS) together, and excellent 7 8 agreement was reached. DNA screened all titles. DNA and AS independently and 9 10 separately examined all abstracts and full-text records for inclusion using Covidence 11 (35). Initial inter-rater reliability showed a proportionate agreement of 0.94 with 12 13 Cohen’s Kappa of 0.68 for abstracts. For full text screening there was a proportionate 14 15 agreement of 0.84 with Cohen’s Kappa 0.56, whereby both reviewers agreed to include 16 15/94 full text articles and exclude 64/94 articles. Initial disagreements on 15 articles 17 18 were resolved throughFor joint peer discussion, review or third reviewer only adjudication (NG). 19 20 Studies with more than 50% GPs (family doctors, primary care physicians) working in a 21 22 practice setting or medical centre, reporting on wellbeing, satisfaction, flourishing, 23 mindfulness, resilience, empathy, engagement, balance, empowerment, compassion, 24 25 work-related morale, and control measures were included. We excluded studies 26 27 exclusively reporting on burnout, distress, or mental ill-health. 28 29 Data including author, year, type of study, participants, intervention, pre- and post- 30 outcome measures, and quantitative and qualitative results were extracted. (Table 1). 31 32 33 34 Data synthesis and analysis 35 We calculated within- and between-group absolute change and effect sizes (Hedges’ g) 36 37 (see table 2, 3 and supplement 2, 3) (36). We compared mean outcome scores and http://bmjopen.bmj.com/ 38 39 standard deviations (SDs) at baseline with post-intervention scores. Where several post- 40 41 intervention measures were reported, the primary outcome point nominated by the 42 authors was selected. We utilised SD*, which takes different sample sizes into account 43 44 (formulae in supplement 2). For within-group we calculated the pooled SD* based on 45 on September 26, 2021 by guest. Protected copyright. 46 pre- and post- interventions SDs, for between-group analysis the effect size was 47 48 calculated based on the pooled SD* of control and intervention group at baseline (37) 49 (supplement 2). 50 51 Positive effect sizes indicated an effect for the intervention. Effect sizes of 0.2, 0.5 and 52 53 0.8 were considered small, moderate and large, respectively (38). 54 55 56 Risk of bias 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 44 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Two reviewers (AS, DNA) independently applied the Cochrane RoB2 (39) to RCTs. 5 6 Total RoB2 scores showed 100% agreement. Any discrepancies of sub-scores were 7 8 discussed, and consensus was achieved. The other studies were assessed by DNA. 9 10 11 Patient and public involvement 12 13 No patients are involved. 14 15 16 RESULTS 17 18 The database searchesFor rendered peer 14,792 review records in total. onlyAfter removing duplicates, 19 20 studies conducted before the year 2000, and adding 12 studies through hand search - 21 22 which included contacting authors for full text papers of relevant conference abstracts - 23 10,759 studies were screened. We eliminated 9,682 records by title, and 983 by abstract, 24 25 leaving 94 studies for full-text assessment. 26 27 28 29 Study characteristics & design 30 We included nineteen studies in the systematic review (2, 10, 40-56) (Table 1 and 2, 31 32 and PRISMA-Flowchart Figure 1). Six randomised controlled trials (RCT), three non- 33 34 randomised controlled trials (CBA), eight non-controlled interventions (NCBA), and 35 two reports from a longitudinal cohort during which a health policy change was 36 37 introduced, which we considered as ‘naturalistic’ interventions (41, 56). Five studies http://bmjopen.bmj.com/ 38 39 included a qualitative component. Only one RCT (50) and two CBAs (44, 47) utilised 40 41 active controls. Five RCTs and one CBA (40) had a waitlist control arm. Publications 42 from the United States (8/19, 42%), Europe (8/19, 42%), Australia (2/19) and Israel 43 44 (1/19) were included (Table 1). 45 on September 26, 2021 by guest. Protected copyright. 46 47 48 Participants 49 The total population enrolled was 1,141 for the 17 intervention studies (average 50 51 participants per study 67.1, range 6 – 290). The two studies reporting on the same panel 52 53 survey population (41, 56) included approximately 2,000 participants each year. Mean 54 age overall ranged from 40 – 54.9 years, and sex from 8% - 76.9% male participants 55 56 (Table 1). Attrition for intervention groups varied from 0% to 20%, for controls from 57 58 0% to 24%. One outlier had a total attrition rate of 80% (51). Eight studies reported 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 44

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 follow-up measures, timepoints ranged from 3 to 48 months post baseline (mean 14.6 5 6 months). 7 8 9 10 Intervention type 11 We found considerable variation in intervention type, length, and dose-intensity. Three 12 13 groups were distinguished based upon the focus of the intervention: individual/personal 14 15 (13/19, 68% of studies, N=930), organisational (4/19, 21%, N=211) and naturalistic 16 interventions on a systemic level (policy change in the UK) (2/19, 11%). 17 18 Individual mindfulness-basedFor peer interventions review were most commononly (9/13, 69%), followed 19 20 by educational training or experiential workshops (3/13, 23%) with one coaching 21 22 intervention (46). Two organisational interventions trialled the addition of clerical 23 support or scribes, and two explored an organisational improvement programme. Two 24 25 studies from the United Kingdom examined the effects of the introduction of a pay for 26 27 performance scheme (41, 56) (Table 2). 28 29 30 Outcomes and their measures 31 32 The definitions of outcomes and measurement tools varied considerably. Only one 33 34 study clearly stated one a priori primary outcome (51), with most using a battery of self- 35 reported outcome measures (supplement 4). These included a range of twelve validated 36 37 tools (BAER, BRS, CD-RISC-10, FFMQ, GRAS, JSPE, JSS-WCW, MAAS, PANAS, http://bmjopen.bmj.com/ 38 39 SCS, UWES, WHO-5) (57-75) as well as thirteen measures where no validation 40 41 information was obtainable. A range of job satisfaction measures were applied in eight 42 studies, mindfulness in six, resilience in four, compassion and empathy tools were each 43 44 used thrice, the positive and negative affect scale was used twice. The WHO-5 45 on September 26, 2021 by guest. Protected copyright. 46 wellbeing index was used once (supplement 4). Not one study evaluated flourishing. 47 48 Negative outcome measures were often concurrently reported. Sixteen studies employed 49 the Maslach Burnout Inventory or other stress-related measures. As the a priori aim of 50 51 the study was to explore effects of interventions on wellbeing, satisfaction, and other 52 53 positive outcomes, we did not extract and report results for negative outcome measures, 54 nor examine possible inter-relationships between positive and negative outcome 55 56 measures. 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 44 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Intervention effects 5 6 7 The between-group change for controlled studies and within-group change for 8 intervention arms are presented in Tables 2, 3 and supplement 3. 9 10 11 a. Individual focussed interventions 12 13 i) Mindfulness (k=9) 14 15 Six mindfulness interventions (3 RCTs, 1 CBA, 2 NCBA) evaluated mindfulness 16 outcomes (FFMQ, MAAS, BAER) and reported moderate to high between-group effect 17 18 sizes (k=4) rangingFor from 0.5peer – 0.88 for review mindfulness with only an outlier at 1.5 (42). Within- 19 20 group ES (k=6) showed moderate effect sizes (range 0.47 – 0.78) with one study outlier 21 22 at 0.37 and one at 1.05 (Tables 2 and 3). 23 Studies frequently utilised resilience, compassion, and empathy measures with overall 24 25 low to moderate effect sizes. One RCT and two NCBAs measured resilience (BRS, RS- 26 27 14, CD-RISC), whereby between-group ES (k=1) was moderate at 0.61, while within- 28 29 group (k=3) effect sizes were low to moderate (range 0.17 – 0.51). Compassion 30 measures (SCBC, SCS) were reported in three studies (1 RCT, 2 NCBAs). Between- 31 32 group ES (k=1) was 0.73, whilst within-group ES (k=3) varied considerably (-0.04 to 33 34 0.77). Three studies reported on empathy (JSPE) (1 RCT, 1 CBA, 1 NCBA) with very 35 low 0.02, respectively moderate between-group 0.44 ES (k=2), and within-group ES 36 37 ranging from 0.2 - 0.44 (k=3) (Tables 2 and 3). http://bmjopen.bmj.com/ 38 39 Two mindfulness studies (NCBAs) measured positive affect (PANAS), only one 40 41 reported a within-group ES (0.52). One NCBA reported a within-group ES (0.52) for 42 wellbeing (WHO-5), another NCBA reported an ES of 0.46 for self-reflection. 43 44 These effect sizes are generally supported by the results reported in the individual 45 on September 26, 2021 by guest. Protected copyright. 46 studies (Table 1). Several interventions included repeated measures at later time-points, 47 48 i.e., during maintenance phase (48), showing an ongoing impact of mindfulness 49 practice. Qualitative results suggested increased wellbeing and compassion towards self 50 51 and others (40), respectively, improved awareness, acceptance, peacefulness, and 52 53 openness (55) after the intervention. 54 55 56 ii. Training, workshops, and coaching (k=4) 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 44

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 For training, workshops and coaching interventions, we were only able to obtain data to 5 6 calculate the ES of one RCT (46) and one CBA (47). Low between-group effect sizes 7 8 for work-related morale (0.3), quality of work-life (0.27), and low ES for both measures 9 10 within-group (0.43 and 0.45 respectively) were found for Gardiner’s CBA (47). Very 11 low effect sizes for job satisfaction and resilience both for between-group (0.06, 0.13), 12 13 and within-group (0.13, 0.24) change were observed in Dyrbye’s RCT (46). 14 15 These results are reflected in the individual study results (Table 1). Barcons (44) did not 16 detect any significant between-group differences for overall job satisfaction, whilst 17 18 Margalit (50) demonstratedFor peer significant review improvement in self-esteemonly between groups. 19 20 21 22 b. Organisational interventions (k=4) 23 One RCT and three NCBAs trialled organisational interventions. Means and standard 24 25 deviations were not provided; therefore, we were not able to calculate effect sizes. 26 27 Linzer et al. (RCT, n=166) demonstrated that workflow interventions, communication 28 29 and overall quality improvements benefited satisfaction in the intervention group (49). 30 Whilst Dunn et al. (NCBA, n=32), showed that quality improvement projects in the 31 32 workplace showed significant improvement in quality work competence ratings but 33 34 fluctuating satisfaction levels (2). Two smaller uncontrolled trials (n=13 in total) 35 investigated the addition of clerical staff to the practice. Pozdnyakova et al. showed that 36 37 the addition of clerical staff led to an improvement in satisfaction with the clinic http://bmjopen.bmj.com/ 38 39 workflow from 2/6 to all 6 GPs in a single practice but did not report on any other 40 41 measures of wellbeing (52). Similarly, Contratto et al. reported improved quality of life 42 and professional balance for seven general medicine physicians in a mixed-methods 43 44 approach (45). 45 on September 26, 2021 by guest. Protected copyright. 46 47 48 c. Systemic interventions (k=2) 49 The introduction of a new contract with pay for performance scheme showed a 50 51 significant improvement in job satisfaction (56) with an effect size of 0.44 between 52 53 2004 and 2005. Allen et al. used the same data and included a 2008 survey to look at 54 satisfaction as a function of the exposure of GPs to the pay per performance scheme. 55 56 Whilst job satisfaction declined again in 2008, the exposure to the scheme didn’t affect 57 58 satisfaction (41). 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 44 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 5 6 Risk of bias 7 8 The types of intervention and study settings precluded blinding for randomised 9 10 controlled studies (no allocation concealment for waitlist control groups), and the 11 outcome measures were participant reported throughout, and as such all studies were 12 13 rated as high risk of bias by the Cochrane RoB2. 14 15 16 DISCUSSION 17 18 Strengths, limitations,For and peer interpretation review of evidence only 19 20 We identified nineteen studies, half of which were published in the last four years, 21 22 demonstrating an increased interest in the improvement of wellbeing and satisfaction of 23 GPs. In comparison, a systematic review from 2016 looking at interventions to reduce 24 25 burnout in physicians included fifteen RCTs and 37 cohort studies with 20 studies 26 27 conducted before 2010 (76), suggesting burnout has been a research focus for longer. Or 28 29 this may possibly indicate the focus is more generally shifting from a disease and ‘dis- 30 abled’ to an ‘en-abled’ approach when trying to design interventions for health care 31 32 professionals. 33 34 The considerable heterogeneity in the definition and measurement of constructs, study 35 design, participant numbers, intervention types, intervention dose (ranging from six to 36 37 53 hours), follow-up periods, quality and reporting precluded a meta-analysis. It is http://bmjopen.bmj.com/ 38 39 challenging to draw robust conclusions regarding the (comparative) effectiveness of the 40 41 different types of interventions reviewed. 42 43 44 Mindfulness interventions provided the most comprehensive and robust data with 45 on September 26, 2021 by guest. Protected copyright. 46 moderate to large effect sizes on mindfulness outcomes, and low to moderate effect 47 48 sizes on compassion, resilience, and empathy. Our results are supported by two reviews 49 looking at the effects of mindfulness-based interventions on wellbeing (77, 78), in 50 51 health care professionals more generally. Lomas et al. conducted a systematic review 52 53 and meta-analysis and assessed 41 studies with approximately 2,100 participants. They 54 found a wide range of self-reported outcomes (with both positive and negative measures 55 56 of wellbeing). Reported within-group effect sizes for mindfulness, positive wellbeing 57 58 (or life satisfaction), and compassion hovered around a moderate 0.5 mark, ES for 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 44

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 empathy was 0.31; whilst for RCTs, the between-group ES for mindfulness, life 5 6 satisfaction, and compassion were around 0.3 (77). 7 8 Scheepers et al. contributed a narrative review of 23 studies looking at mindfulness- 9 10 based interventions for wellbeing in doctors of all ages and specialities. Review authors 11 noted some caveats; considerable variation in type and dose-intensity of practice, sparse 12 13 long-term outcome data, and methodological limitations. They cautioned that 14 15 mindfulness practice involves time and dedication, which is not always feasible for busy 16 health care professionals. In sum, the conclusions they drew are tentatively positive. 17 18 In contrast to LomasFor et al. peer our ES for mindfulnessreview is higher only between-group than within- 19 20 group, which is somewhat unusual. This may be explained by one study (42) whose 21 22 positive outcome appeared determined by the decline in mindfulness over time in their 23 control group, rather than the intervention being effective. 24 25 26 27 We identified four studies evaluating coaching and experiential workshops for GPs, 28 29 which showed low effect sizes for satisfaction measures and moderate ES for work- 30 related morale and quality of work life. There does not appear to be much literature on 31 32 coaching for health care professionals. One quasi-experimental study by Gardiner et al. 33 34 looked at ‘cognitive behavioural coaching’ in rural Australian GPs and demonstrated a 35 significant within-group reduction in distress levels for the coachees (79). Resilience 36 37 training for a range of different physicians that had completed training was investigated http://bmjopen.bmj.com/ 38 39 in a recent systematic review. Four RCTs and five observational studies were included. 40 41 The authors flag heterogenous study design and use of outcome measures, as well as 42 quality issues with weak evidence for the interventions, whilst indicating potential for 43 44 improvement of resilience (80). 45 on September 26, 2021 by guest. Protected copyright. 46 47 48 We found four small-scale organisational interventions that suggested improved (job) 49 satisfaction, as did one large-scale health policy intervention of performance-related pay 50 51 in the UK. For burnout, a paucity of interventions trials delivered at organisational and 52 53 systemic levels has been previously reported (81, 82), the authors suggest to actively 54 design such trials. Similarly, Dyrbye et al., concluded that whilst useful, an individual 55 56 intervention such as coaching is no replacement for organisational improvement (46). 57 58 Shanafelt and colleagues have collated their vast research into burnout and put forward 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 44 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 nine organisational strategies to address burnout and physician wellbeing through 5 6 leadership (83). Despite calls for action, these avenues have not been adequately 7 8 addressed or reported to date, at least not for GPs, and warrant further exploration. 9 10 Considering the time it takes to gather and report data, it is understandable that 11 organizations might feel pressure to implement programs based on preliminary data. 12 13 Commendably, Dutch researchers recently investigated the effects of a mindful 14 15 leadership course in hospital-based medical specialists (84, 85). Both a qualitative 16 interview and a pre- and post- self-evaluation suggested an overall benefit of the 17 18 intervention withFor improved peer mindfulness, review life satisfaction only and leadership, reduced 19 20 burnout, and positive change in attitudes and behaviours towards self and others. Not all 21 22 participants benefited equally, suggesting a need to provide a range of interventions to 23 meet defend participants’ needs. Future investigation will need to explore what the 24 25 impact on individuals’ leadership style and on their teams is. 26 27 28 29 Limitations 30 We included English publications only, although purposely extending our search 31 32 globally. We excluded studies before the year 2000, because wellbeing literature in 33 34 medicine is a more recent development, and general practice is now likely quite 35 different than two decades ago. Self-reported outcome measures are typically subject to 36 37 bias, particularly considering studies included GPs from different settings and cultures, http://bmjopen.bmj.com/ 38 39 potentially introducing cultural bias, rendering comparisons challenging. 40 41 42 Suggestions for future research 43 44 Based on our findings, we provide some suggestions which may be useful for future 45 on September 26, 2021 by guest. Protected copyright. 46 research into wellbeing and satisfaction for GPs. Stronger collaboration amongst 47 48 researchers in this space may also lead to improved results. 49 50 51 a) Selection of outcomes & outcome measures 52 53 The reduction in burnout and stress is often equated with an improvement in wellbeing 54 or satisfaction. We argue that the improvement of negative outcomes doesn’t 55 56 necessarily indicate a presence of satisfaction or wellbeing. This aligns with the dual 57 58 continuum model of mental health / mental ill-health and flourishing / languishing (25- 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 44

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 28). Good mental health isn’t automatically linked to flourishing, nor is mental ill- 5 6 health an indicator of languishing. Other authors have made similar statements (10, 24, 7 8 42, 77). We did not find a single study about flourishing in GPs. 9 10 Clearly defining the constructs ‘wellbeing’ and ‘satisfaction’, whilst utilising validated 11 wellbeing and satisfaction measures, will enhance clarity, consistency and 12 13 comparability of study design and reporting. We suggest drawing on existing 14 15 frameworks, models, and definitions in the psychological literature (for different types 16 of wellbeing, satisfaction or flourishing) (86-89). To measure wellbeing we suggest the 17 18 Warwick EdinburghFor Mental peer Wellbeing review Scale (WEMWBS) only and for Job satisfaction the 19 20 Warr-Cook-Wall scale (WCW-JS), both of which have been validated in medical 21 22 populations (90, 91). Brady et al., who conducted a systematic review looking at the 23 definition and measurement of ‘physician wellness’ similarly stated that there needs to 24 25 be consensus and clarity of definition, if we want to improve the quality and 26 27 comparability of research in this space (92). Whilst this would improve the next phase 28 29 of studies, the urgency in calls for actions may need to be balanced against the calls for 30 consistency among studies. 31 32 33 34 b) Organisational and systemic interventions 35 With the dearth of research in this space, and the relatively small effects for individual 36 37 interventions, we believe it worthwhile to explore system and organisational http://bmjopen.bmj.com/ 38 39 interventions (i.e. mindful leadership training describe above) in the context of 40 41 wellbeing. 42 Considering what is known about burnout (drivers being organisational culture, 43 44 workplace conditions, lack of control and autonomy), it’s not surprising that individual 45 on September 26, 2021 by guest. Protected copyright. 46 interventions are not as effective as desired (49, 81, 93-95). Hence, more combined 47 48 approaches targeting both individuals and organisations have been proposed. 49 A 2017 British meta-analysis contrasted different types of interventions for burnout on 50 51 the individual doctor and on the systemic level, whereby systemic interventions appear 52 53 more effective (96). Similarly, groups in the USA state that the approach must be 54 combined and include organisational interventions (1, 97, 98), mostly focusing on time 55 56 management, rostering, workflow management, staffing and use of information 57 58 technology solutions. Overall, there is a scarcity of organisational interventions aimed at 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 44 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 reducing burnout (99), and conclusions from the two meta-analyses of interventions to 5 6 reduce burnout should be considered provisional. 7 8 In sum, we endorse an intensified effort to explore organisational interventions to 9 10 improve wellbeing and satisfaction, and believe a focus on leadership and improving the 11 culture at work is a good place to start. 12 13 14 15 c) Physical interventions 16 We didn’t find any physical interventions (i.e. exercise, nutrition) geared towards 17 18 improving GPs’ Forwellbeing. peer Sparse research review on exercise onlyor diet interventions for doctors 19 20 exists. A Pakistani cross-sectional survey revealed that 76% of nearly 1200 doctors, 21 22 nurses and dentists did not exercise at all, and only one participant ate according to 23 USDA dietary guidelines (100). While a US cross-sectional survey of 303 physicians 24 25 found that less than 25% knew the American Heart Association (AHA) dietary 26 27 recommendations, whilst around half knew and followed their physical activity 28 29 guidelines (101). Two systematic reviews looked at exercise and burnout in the general 30 population, one was inconclusive (102), the other stated that physical activity 31 32 effectively reduces burnout (103). Both identified methodological issues and no long- 33 34 term follow-up. Seeing the paucity of data, this might provide an avenue for further 35 investigation. 36 37 http://bmjopen.bmj.com/ 38 39 d) Quality and risk of bias 40 41 Areas for risk of bias are inherent in this type of research. However, measures can be 42 taken to reduce bias for example by using active controls in randomized studies as 43 44 suggested by other review authors (78), by consistently publishing study protocols a 45 on September 26, 2021 by guest. Protected copyright. 46 priori, and controlling for participant attributes, such as prior engagement in 47 48 mindfulness practice. Ideally, the same rigorous approach should be applied to 49 intervention studies for clinicians, as to clinical interventions studies for patients. 50 51 52 53 CONCLUSION 54 Despite increasing interest in the improvement of wellbeing and satisfaction, there is 55 56 remarkably little evidence beyond mindfulness interventions aimed at individual GPs. 57 58 Few studies utilize validated measures of wellbeing and satisfaction, and favour burnout 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 44

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 tools. Studies looking into organisational and systemic interventions remain sparse, and 5 6 conclusions about their effectiveness may be premature. 7 8 Considering the COVID-19 pandemic and the added strain to primary care, programmes 9 10 to support and research GP wellbeing should be prioritised by policymakers and 11 governments worldwide. 12 13 14 15 ACKNOWLEDGEMENTS 16 Registration 17 18 The protocol is registeredFor peeron PROSPERO review CRD42020164699. only 19 20 Ethics approval statement 21 22 We did not seek ethics approval, as this is a systematic review of previously published 23 24 data, that is available in the public domain. 25 26 27 Funding statement 28 29 Dr Diana Naehrig was supported through the Raymond Seidler PhD scholarship. The 30 31 funding source had no influence on the study design, collection, analysis or 32 33 interpretation of data, the writing of the manuscript nor the decision to submit the article 34 for publication. Award/Grant number is not applicable. 35 36 37 http://bmjopen.bmj.com/ 38 Competing interest statement 39 All authors have completed the Unified Competing Interest Form. 40 41 Beyond the Raymond Seidler PhD scholarship for Diana Naehrig there was no support 42 43 from any organisation for the submitted work. The funding source had no influence on 44 45 the study design, collection, analysis or interpretation of data, the writing of the report on September 26, 2021 by guest. Protected copyright. 46 nor the decision to submit the article for publication. 47 48 Ian Hickie has declared financial relationships outside of the submitted work in the 49 50 previous three years and other relationships or activities. 51 52 53 Authors’ Contributions 54 55 DNA is the guarantor and corresponding author and attests that all listed authors meet 56 57 authorship criteria and that no others meeting the criteria have been omitted. 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 44 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Authors contributed to the study conception and design (DNA, JKH, RE, IH, NG), the 5 6 acquisition (DNA, AS, NG), analysis (DNA, AS, NG) and interpretation (DNA, AS, 7 8 RE, IH, NG) and the drafting or critical revision of important intellectual content and 9 10 final approval of the version to be published and agree to be accountable for all aspects 11 of the work in ensuring that questions related to the accuracy or integrity of any part of 12 13 the work are appropriately investigated and resolved. 14 15 16 Transparency declaration 17 18 Dr Diana NaehrigFor (the manuscript’s peer guarantor) review affirms thatonly the manuscript is an honest, 19 20 accurate, and transparent account of the study being reported; that no important aspects 21 22 of the study have been omitted; and that any discrepancies from the study as planned 23 and registered have been explained. 24 25 26 27 Data sharing statement 28 29 All included studies are published. We will consider sharing data upon reasonable 30 request. 31 32 33 34 Copyright 35 Dr Diana Naehrig (corresponding author) has the right to grant on behalf of all authors 36 37 and does grant on behalf of all authors, an exclusive licence on a worldwide basis to the http://bmjopen.bmj.com/ 38 39 BMJ Publishing Group Ltd to permit this article to be published in BMJ editions and 40 41 any other BMJPGL products and sublicences such use and explit all subsidiary rights, 42 as set out in our licence. 43 44 45 on September 26, 2021 by guest. Protected copyright. 46 Figure 1. Prisma diagram 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 44

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 44 BMJ Open

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 44

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 44 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 50. Margalit AP, Glick SM, Benbassat J, Cohen A, Katz M. Promoting a 5 6 biopsychosocial orientation in family practice: effect of two teaching programs on the 7 knowledge and attitudes of practising primary care physicians. Med Teach. 8 2005;27(7):613-8. 9 51. Montero-Marin J, Gaete J, Araya R, Demarzo M, Manzanera R, Alvarez de Mon 10 M, et al. Impact of a blended web-based mindfulness programme for general 11 practitioners: A pilot study. Mindfulness. 2018;9(1):129-39. 12 52. Pozdnyakova A, Laiteerapong N, Volerman A, Feld LD, Wan W, Burnet DL, et 13 14 al. Impact of Medical Scribes on Physician and Patient Satisfaction in Primary Care. 15 Journal of General Internal Medicine. 2018;33(7):1109-15. 16 53. Rees CS, Craigie MA, Slatyer S, Crawford C, Bishop M, McPhee E, et al. Pilot 17 study of the effectiveness of a Mindful Self-Care and Resiliency program for rural 18 doctors in Australia.For Aust peerJ Rural Health. review 2020;28(1):22-31. only 19 54. Schroeder DA, Stephens E, Colgan D, Hunsinger M, Rubin D, Christopher MS. 20 A Brief Mindfulness-Based Intervention for Primary Care Physicians: A Pilot 21 22 Randomized Controlled Trial. American Journal of Lifestyle Medicine. 2018;12(1):83- 23 91. 24 55. van Wietmarschen H, Tjaden B, van Vliet M, Battjes-Fries M, Jong M. Effects 25 of mindfulness training on perceived stress, self-compassion, and self-reflection of 26 primary care physicians: a mixed-methods study. Bjgp Open. 27 2018;2(4):bjgpopen18X101621. 28 29 56. Whalley D, Gravelle H, Sibbald B. Effect of the new contract on GPs' working 30 lives and perceptions of quality of care: A longitudinal survey. British Journal of 31 General Practice. 2008;58(546):8-14. 32 57. Heritage B, Pollock C, Roberts LD. Confirmatory factor analysis of Warr, Cook, 33 and Wall's (1979) Job satisfaction scale. Australian Psychologist. 2015;50(2):122-9. 34 58. Warr P, Cook J, Wall T. Scales for the measurement of some work attitudes and 35 aspects of psychological well‐being. Journal of occupational Psychology. 36 37 1979;52(2):129-48. http://bmjopen.bmj.com/ 38 59. Baer RA, Smith G. T., Allen KB. Assessment of mindfulness by self-report: The 39 Kentucky Inventory of Mindfulness Skills. Assessment. 2004;11:191-206. 40 60. Baer RA, Smith GT, Hopkins J, Krietemeyer J, Toney L. Using self-report 41 assessment methods to explore facets of mindfulness. Assessment. 2006;13:27-45. 42 61. Baer RA, Smith GT, Lykins E, Button D, Krietemeyer J, Sauer S, et al. 43 Construct validity of the five facet mindfulness questionnaire in meditating and 44 45 nonmeditating samples. Assessment. 2008;15(3):329-42. on September 26, 2021 by guest. Protected copyright. 46 62. Smith BW, Dalen J, Wiggins K, Tooley E, Christopher P, Bernard J. The brief 47 resilience scale: assessing the ability to bounce back. International journal of behavioral 48 medicine. 2008;15(3):194-200. 49 63. Campbell‐Sills L, Stein MB. Psychometric analysis and refinement of the 50 connor–davidson resilience scale (CD‐RISC): Validation of a 10‐item measure of 51 52 resilience. Journal of Traumatic Stress: Official Publication of The International Society 53 for Traumatic Stress Studies. 2007;20(6):1019-28. 54 64. Cebolla A, Garcia-Palacios A, Soler J, Guillén V, Baños R, Botella C. 55 Psychometric properties of the Spanish validation of the Five Facets of Mindfulness 56 Questionnaire (FFMQ). The European Journal of Psychiatry. 2012;26(2):118-26. 57 65. Rostami A, Keshmiri F, Askari R, Jambarsang S, Shafiei M. Validation of 58 Groningen Reflection Ability Scale Questionnaire and Evaluation of Reflection Ability 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 24 of 44

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Level of Health Care Management Students. Evidence Based Health Policy, 5 6 Management and Economics. 2019;3(4):293-302. 7 66. Hojat M, Mangione S, Nasca TJ, Cohen MJM, Gonnella JS, Erdmann JB, et al. 8 The Jefferson scale of physician empathy: Development and primary psychometrics. 9 Educ, Psych Measur. 2001;61:349-65. 10 67. Hojat M, Gonnella JS, Nasca TJ, Mangione S, Veloksi JJ, Magee M. The 11 Jefferson Scale of Physician Empathy: further psychometric data and differences by 12 gender and specialty at item level. Academic Medicine. 2002;77(10):S58-S60. 13 14 68. Hojat M, Gonnella JS, Nasca TJ, Mangione S, Vergare M, Magee M. Physician 15 empathy: definition, components, measurement, and relationship to gender and 16 specialty. American Journal of Psychiatry. 2002;159(9):1563-9. 17 69. Brown KW, Ryan RM. The benefits of being present: mindfulness and its role in 18 psychological well-being.For peerJournal of personalityreview and social only psychology. 2003;84(4):822. 19 70. Watson D, Clark LA, Tellegen A. Development and validation of brief measures 20 of positive and negative affect: the PANAS scales. Journal of personality and social 21 22 psychology. 1988;54(6):1063. 23 71. Wagnild G. Development and use of the Resilience Scale (RS) with middle-aged 24 and older adults. Resilience in Children, Adolescents, and Adults: Springer; 2013. p. 25 151-60. 26 72. Neff KD. The development and validation of a scale to measure self- 27 compassion. Self and identity. 2003;2(3):223-50. 28 29 73. Seppälä P, Mauno S, Feldt T, Hakanen J, Kinnunen U, Tolvanen A, et al. The 30 Construct Validity of the Utrecht Work Engagement Scale: Multisample and 31 Longitudinal Evidence. Journal of Happiness Studies. 2008;10(4):459-81. 32 74. Bech P, Olsen LR, Kjoller M, Rasmussen NK. Measuring well‐being rather than 33 the absence of distress symptoms: a comparison of the SF‐36 Mental Health subscale 34 and the WHO‐Five well‐being scale. International journal of methods in psychiatric 35 research. 2003;12(2):85-91. 36 37 75. Topp CW, Ostergaard SD, Sondergaard S, Bech P. The WHO-5 Well-Being http://bmjopen.bmj.com/ 38 Index: a systematic review of the literature. Psychother Psychosom. 2015;84(3):167-76. 39 76. West CP, Dyrbye LN, Erwin PJ, Shanafelt TD. Interventions to prevent and 40 reduce physician burnout: a systematic review and meta-analysis. The Lancet. 41 2016;388(10057):2272-81. 42 77. Lomas T, Medina JC, Ivtzan I, Rupprecht S, Eiroa-Orosa FJ. A systematic 43 review of the impact of mindfulness on the well-being of healthcare professionals. J 44 45 Clin Psychol. 2018;74(3):319-55. on September 26, 2021 by guest. Protected copyright. 46 78. Scheepers RA, Emke H, Epstein RM, Lombarts K. The impact of mindfulness- 47 based interventions on doctors' well-being and performance: A systematic review. Med 48 Educ. 2020;54(2):138-49. 49 79. Gardiner M, Kearns H, Tiggemann M. Effectiveness of cognitive behavioural 50 coaching in improving the well-being and retention of rural general practitioners. Aust J 51 52 Rural Health. 2013;21(3):183-9. 53 80. Venegas CL, Nkangu MN, Duffy MC, Fergusson DA, Spilg EG. Interventions 54 to improve resilience in physicians who have completed training: A systematic review. 55 PLoS ONE. 2019;14(1):e0210512. 56 81. Maslach C, Leiter MP. Early predictors of job burnout and engagement. Journal 57 of applied psychology. 2008;93(3):498. 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 25 of 44 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 82. Awa WL, Plaumann M, Walter U. Burnout prevention: a review of intervention 5 6 programs. Patient Educ Couns. 2010;78(2):184-90. 7 83. Shanafelt TD, Noseworthy JH. Executive Leadership and Physician Well-being: 8 Nine Organizational Strategies to Promote Engagement and Reduce Burnout. Mayo 9 Clin Proc. 2017;92(1):129-46. 10 84. Kersemaekers WM, Vreeling K, Verweij H, van der Drift M, Cillessen L, van 11 Dierendonck D, et al. Effectiveness and feasibility of a mindful leadership course for 12 medical specialists: a pilot study. BMC Med Educ. 2020;20(1):34. 13 14 85. Vreeling K, Kersemaekers W, Cillessen L, van Dierendonck D, Speckens A. 15 How medical specialists experience the effects of a mindful leadership course on their 16 leadership capabilities: a qualitative interview study in the Netherlands. BMJ Open. 17 2019;9(12):e031643. 18 86. Keyes CL.For The Mental peer Health reviewContinuum: From only Languishing to Flourishing in 19 Life. Journal of Health and Social Behavior. 2002;43(2):207-22. 20 87. Ryff CD. Happiness is Everything, or is it. Explorations, Meaning of 21 22 Psychological Well-Being. Journal of Personality and Social Psychology. 23 1989;57(6):1069-81. 24 88. Seligman M. PERMA and the building blocks of well-being. The Journal of 25 Positive Psychology. 2018;13(4):333-5. 26 89. Seligman ME. Flourish: A visionary new understanding of happiness and well- 27 being.: Simon and Schuster.; 2012. 28 29 90. Tennant R, Hiller L, Fishwick R, Platt S, Joseph S, Weich S, et al. The 30 Warwick-Edinburgh Mental Well-being Scale (WEMWBS): development and UK 31 validation. Health Qual Life Outcomes. 2007;5:63. 32 91. Hills D, Joyce C, Humphreys J. Validation of a job satisfaction scale in the 33 Australian clinical medical workforce. Eval Health Prof. 2012;35(1):47-76. 34 92. Brady KJS, Trockel MT, Khan CT, Raj KS, Murphy ML, Bohman B, et al. 35 What Do We Mean by Physician Wellness? A Systematic Review of Its Definition and 36 37 Measurement. Acad Psychiatry. 2018;42(1):94-108. http://bmjopen.bmj.com/ 38 93. Linzer M, Poplau S, Brown R, Grossman E, Varkey A, Yale S, et al. Do Work 39 Condition Interventions Affect Quality and Errors in Primary Care? Results from the 40 Healthy Work Place Study. Journal of General Internal Medicine. 2017;32(1):56-61. 41 94. Gregory ST, Menser T. Burnout Among Primary Care Physicians: A Test of the 42 Areas of Worklife Model. Journal of Healthcare Management. 2015;60(2):133-48. 43 95. Cossman JS, Street D. Policy, autonomy, and physician satisfaction. Journal of 44 45 Health Care for the Poor & Underserved. 2010;21(3):898-912. on September 26, 2021 by guest. Protected copyright. 46 96. Panagioti M, Panagopoulou E, Bower P, Lewith G, Kontopantelis E, Chew- 47 Graham C, et al. Controlled interventions to reduce burnout in physicians: a systematic 48 review and meta-analysis. JAMA internal medicine. 2017;177(2):195-205. 49 97. West C, Dyrbye L, Erwin P, Shanafelt T. Interventions to prevent and reduce 50 physician burnout: a systematic review and meta-analysis. The Lancet. 51 52 2016;388(10057):2272-81. 53 98. Swensen SJ, Shanafelt T. An Organizational Framework to Reduce Professional 54 Burnout and Bring Back Joy in Practice. The Joint Commission Journal on Quality and 55 Patient Safety. 2017;43(6):308-13. 56 99. Gregory ST, Menser T, Gregory BT. An Organizational Intervention to Reduce 57 Physician Burnout. Journal of Healthcare Management. 2018;63(5):338-52. 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 26 of 44

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 100. Ahmad W, Taggart F, Shafique MS, Muzafar Y, Abidi S, Ghani N, et al. Diet, 5 6 exercise and mental-wellbeing of healthcare professionals (doctors, dentists and nurses) 7 in Pakistan. PeerJ. 2015;3:e1250. 8 101. Aggarwal M, Singh Ospina N, Kazory A, Joseph I, Zaidi Z, Ataya A, et al. The 9 Mismatch of Nutrition and Lifestyle Beliefs and Actions Among Physicians: A Wake- 10 Up Call. American Journal of Lifestyle Medicine. 2020;14(3):304-15. 11 102. Ochentel O, Humphrey C, Pfeifer K. Efficacy of exercise therapy in persons 12 with burnout. A systematic review and meta-analysis. Journal of sports science & 13 14 medicine. 2018;17(3):475. 15 103. Naczenski LM, Vries JD, Hooff M, Kompier MAJ. Systematic review of the 16 association between physical activity and burnout. J Occup Health. 2017;59(6):477-94. 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44 45 on September 26, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 27 of 44 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 5 Table 1. Evidence table 6 7 Partici- 8 pants (n, Control 9Author, mean Interven- (n, Positive Other 10Date, Study age, % tion (n, mean outcome outcome Post-intervention Follow-up score 11design male) mean age) age) measure measure Baseline scores score (mean, SD, n) Results - Summary 12 Between 2004 and Time at 12 months. Time at 48 months. 2005 overall job 13 Mean. 2004: Life Mean. 2005: Life Mean. 2008: Life satisfaction 14 satisfaction (4.649). satisfaction (5.095). satisfaction (5.008). increased (also see 15 GPs in Overall JS (4.567), Overall JS (5.201), Overall JS (4.728), Whalley 2008), then 16 the UK. Physical working Physical working Physical working fell from 2005 to 2004: conditions (4.862), conditions (5.044), conditions (5.129), 2008. There is a 17 n=1950, Choose method of Choose method of Choose method of positive effect 18 47.0yrs, For peer reviewworking (4.636), workingonly (4.892), working (4.640), (though not 19 66.2% Colleagues (5.515), Colleagues (5.599), Colleagues (5.602), statistically 20Allen 2017 male. Pay for Recognition for Recognition for Recognition for significant) between 21 2005: perform- good work (4.224), good work (4.726), good work (4.495), QOF/P4P income Panel n=2000, ance (P4P) Responsibility Responsibility Responsibility exposure and job 22survey 47.9yrs, scheme; (4.976), (5.406), (5.276), satisfaction in 2005 23linked with 63.6 % Quality and Job Remuneration Remuneration Remuneration (t-ratio 1.74), 24QOF data male. Outcomes satisfacti (4.376), Opportunity (5.387), Opportunity (4.849), Opportunity though not so in 252004/5 2008: Framework on (JSS to use abilities to use abilities to use abilities 2008 (t-ratio 0.14). (first year) n=1986, (QOF) WCW). (4.787), Hours of (5.147), Hours of (5.074), Hours of The P4P exposure 26and 2007/8 48.7yrs, introduced Life work (3.914), work (4.802), work (4.205), shows now 27(fourth 63.3% in NHS in satisfacti P4P Variety in job Variety in job Variety in job significant effect on 28year) male. 2004 NA on exposure (5.011). (5.269). (5.276). GPs job satisfaction. 29 Mean and SD. Time at 8 weeks. 30 INTERVENTION Mean and SD. (n=21) FFMQ Total INTERVENTION 31 (3.34, 0.44), FFMQ (n=21) FFMQ Total 32 Observing (3.33, (3.71, 0.51), FFMQ 33 0.60), FFMQ Observing (3.98, 34 Describing (3.58, 0.64), FFMQ 0.72), FFMQ act Describing (3.83, 35 with awareness 0.62), FFMQ act 36 (3.16, 0.87), FFMQ with awareness 37 Nonjudging (3.42, (3.48, 0.65), FFMQ http://bmjopen.bmj.com/ 38 0.64), FFMQ Nonjudging (3.78, 39 Nonreactivity (3.17, 0.66), FFMQ 0.51). SRSI Nonreactivity (3.46, 40 relaxation (2.54, 0.63). SRSI 41 0.53), SRSI positive relaxation (3.08, 42 energy (3.09, 0.64), 0.61), SRSI positive Time at 12 months. 43 SRSI mindfulness energy (3.80, 0.82), Mean and SD. (3.74, 0.89), SRSI SRSI mindfulness INTERVENTION 44 Transcendence (4.45, 0.71), SRSI FFMQ Total (3.85, 45 8-week (2.67, 0.82). Transcendence 0.49), FFMQ on September 26, 2021 by guest. Protected copyright. 46 MBSR CONTROL (n=21) (3.27, 1.02). Observing (4.09, 47 programme FFMQ Total (3.51, CONTROL (n=21) 0.62), FFMQ The MBSR (group 0.25), FFMQ FFMQ Total (3.34, Describing (4.01, programme 48 course Observing (3.02, 0.33), FFMQ 0.58), FFMQ act (statistically) 49 Physic- 8x2.5h/wk, 0.88), FFMQ Observing (2.83, with awareness significantly 50 ians 1x8h Describing (3.88, 0.98), FFMQ (3.55, 0.69), FFMQ improves 51 (approx retreat plus 0.53), FFMQ Act Describing (3.82, Nonjudging (3.96, mindfulness and 52 70% homework) with awareness 0.58), FFMQ Act 0.66), FFMQ relaxation measures primary , 10-month (3.93, 0.70), FFMQ with awareness Nonreactivity (3.58, (particularly positive 53 care) in mainten- Nonjudging (4.19, (3.91, 0.61), FFMQ 0.55). SRSI energy and 54 public or ance period 0.52), FFMQ Nonjudging (4.16, relaxation (3.16, transcendence) at 8 55 private (group Mindful- Nonreactivity (3.23, 0.52), FFMQ 0.78), SRSI positive weeks after 56 practice course ness 0.57). SRSI Nonreactivity (3.26, energy (3.99, 0.81), baseline. Sustained Amutio in Spain 10x2.5h/mt (FFMQ), relaxation (2.52, 0.67). SRSI SRSI mindfulness and even improved 572015 (n=42, plus Relax- 0.5), SRSI positive relaxation (2.60, (4.60, 1.01), SRSI long-term effects at 58 47.3 yrs, homework) WLC ation Heart energy (3.01, 0.62), 0.51), SRSI positive Transcendence 12 months follow- 59RCT 42.9%) (n=21) (n=21) (SRSI-3) rate SRSI mindfulness energy (3.01, 0.6), (3.65, 1.36). up are shown. 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 28 of 44

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 (4.29, 0.92), SRSI SRSI mindfulness 5 Transcendence (4.24, 0.95), SRSI 6 (2.64, 0.93). Transcendence 7 (2.40, 0.92). Time at 8 weeks. 8 Mean and SD. Mean and SD. 9 INTERVENTION INTERVENTION 10 (n=43). FFMQ Total (n=43) FFMQ Total 11 (129.6, 22.2), FFMQ (141.6, 16.5), FFMQ Observing (26.1, Observing (30.4, 12 8.6), FFMQ 5.1), FFMQ 13 Describing (28.2, Describing (28.9, 14 5.7), FFMQ act with 5.3), FFMQ act with 15 awareness (25.3, awareness (27.4, 16 5.9), FFMQ 4.7), FFMQ Nonjudging (27.1, Nonjudging (30.6, 17 8.0), FFMQ 6.2), FFMQ 18 For peer reviewNonreactivity (21.9, Nonreactivityonly (24.1, 19 4.3). JSPE Total 3.0). JSPE Total 20 (119.5, 13.1), JSPE (123.0, 9.2), JSPE Perspective taking Perspective taking 21 (54.8, 6.5), JSPE (56.1, 5.3), JSPE 22 Compassionate care Compassionate care 23 8-week MB (47.2, 5.1), JSPE (49.0, 3.9), JSPE 24 psycho- standing in pts standing in pts For mindfulness educational shoes (13.1, 1.8). shoes (13.5, 1.1). total there is 25 program: CONTROL (n=25). CONTROL (n=25) significant 26 Didactic FFMQ Total (120.5, FFMQ Total (121.5, improvement for 27 material, 14.4), FFMQ 16.0), FFMQ the intervention 28 mindful- Observing (24.5, Observing (24.1, (moderate change). 29 ness 4.3), FFMQ 4.5), FFMQ For empathy total Primary meditation, Describing (26.4, Describing (26.5, there is a non- 30 health narrative/a 5.2), FFMQ Act with 5.5), FFMQ Act with significant increase 31 care ppreciative awareness (23.6, awareness (23.0, from pre to post 32 profes- enquiry, 6.8), FFMQ 5.9), FFMQ measure in the 33 sionals discussion. Mindfuln Nonjudging (26.3, Nonjudging (27.4, intervention group, (41 Group ess Burnout 6.5), FFMQ 6.8), FFMQ only an increase in 34 physician course (FFMQ), (MBI), Nonreactivity (19.8, Nonreactivity (20.2, compassionate care 35 s) in 8x2.5h/wk, Empathy mood 2.9). JSPE Total 3.5). JSPE Total was statistically 36 Spain, 1x8h WLC (JSPE), disturban (120.8, 10.1), JSPE (119.0, 10.7), JSPE significant. Overall 37 public retreat plus (n=25 also self- ce Perspective taking Perspective taking mindfulness and http://bmjopen.bmj.com/ 38Asuero system homework. total, report on (POMS), (59.8, 7.0), JSPE (59.6, 6.3), JSPE empathy 2014 (n=68 (n=43 total, 18 energy, intervent Compassionate care Compassionate care (compassionate 39 total, 47 23 physi- wellbeing ion (47.4, 4.9), JSPE (46.9, 4.6), JSPE care) were 40RCT (2010 - yrs, 8% physicians, cians, , and evaluatio standing in pts standing in pts improved by the 412012) male) 48.8yrs) 46.9yr) activity n shoes (13.6, 0.9). shoes (12.5, 3.0). NA programme. 42 43 44 45 on September 26, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 29 of 44 BMJ Open

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 30 of 44

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Organisat- 5 ional inter- 6 vention Intervention 7 programme ongoing from 2000 to improve onward. Emotional 8 physician and work-related 9 wellbeing exhaustion 10 (quality decreased 11 improve- significantly over ment the study period. 12 project, QWC measures of 13 participant organizational 14 Primary data- health significantly 15 care in guided improved initially 16 urban interv- Burnout and remained USA ention) (n= (MBI), acceptable and 17 (n=32, 25 22 to 32 Physician Quality 2001: 55% of Time at 24 months. Time at 24 months. stable during the 18Dunn 2007 internists range over Forsatisfacti peerWork reviewphysicians were 2003:only 84% were 2005: 74% were rest of the study. 19 6 family the inter- on Compete somewhat or very somewhat or very somewhat or very Satisfaction 20NCBA, medicine vention (ACP/ASI nce satisfied with their satisfied with their satisfied with their fluctuated 2000-2005 1 NP) period). NA M) (QWC) practice practice practice. throughout. 21 Means and SD. Time at 5 months. 22 INTERVENTION Means. 23 (n=44), Resilience INTERVENTION 24 (31.0, 6.3), Job (n=44), Resilience satisfaction (43.4, (32.3), Job 25 Resilien- 10.7), UWES vigor satisfaction (44.4), Statistically 26 ce (CD- (3.9, 1.0), UWES UWES vigor (4.1), significant 27 RISC), Dedication (4.5, UWES Dedication improvement for 28 Global 1.1), UWES (4.6), UWES resilience from pre 29 Job Absorption (4.2, Absorption (4.1), to post intervention, Satisfact- 1.0), Empowerment Empowerment at no change in job 30 ion-12 at work (55.5, 11.9). work (58.2). satisfaction. 31 Physician 6 coaching (subscale CONTROL (n=44), CONTROL (n=44), Burnout, emotional 32 s in USA sessions of PJSS), Resilience (30.6, Resilience (31.2), exhaustion 33 (67 (3.5h) work 5.7), Job satisfaction Job satisfaction decreased. QoL generalis facilitated engage- (42.8, 10.6), UWES (43.2), UWES vigor improved. No 34 ts, 21 by a ment vigor (4.0, 1.2), (4.2), UWES statistically 35Dyrbye subspec- profess- (UWES), UWES Dedication Dedication (4.7), significant 362019 ialists). sional empower (4.6, 1.0), UWES UWES Absorption differences in 37 (n=88, coach over ment at Absorption (4.1, (4.2), depersonalization, http://bmjopen.bmj.com/ 38RCT, 45.5% 5mts WLC work Burnout 1.1), Empowerment Empowerment at engagement, or 2017 - 2018 male). (n=44). (n=44) scale. (MBI). at work (57.3, 14.0). work (60.3). meaning in work. 39 No significant 40 improvement in 41 Family compassion and 42 medicine Burnout resilience over time. practit- (MBI), Participants had 43 ioners in Shortened Depressi Time at 4 weeks. Time at 9 months. improvements 44 USA MBSR on & Mean score and Mean score and Mean score and compared with 45 (n=30, course, 18h Compas- Anxiety Confidence Interval. Confidence Interval. Confidence Interval baseline at all on September 26, 2021 by guest. Protected copyright. 46 87% total, 14h sion (DASS), n=30. Resilience RS- n=28. Resilience RS- at 9mts. n=23. follow-up time 47Fortney family weekend, (SCBC), Perceive 14 (79.9, CI 75.2- 14 (82, CI 77.1- Resilience RS-14 points for Burnout, 2013 medicine 2x 2h plus Resili- d Stress 84.6), Compassion 86.8), Compassion (81.4, CI 76.2-86.6), depression, anxiety, 48 physic- homework ence (RS- scale SCBC (27.6, CI 25.9- SCBC (27.4, CI 25.6- Compassion SCBC and perceived 49NCBA ians) (n=30). NA 14). (PSS), 29.3). 29.1). (28.3, CI 26.5-30.1). stress. 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 31 of 44 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Overall the 5 intervention group 6 showed higher 7 scores post intervention than 8 the control group 9 (positive trend). 10 However, no 11 statistically significant change 12 seen for work 13 related morale or 14 QoWL for the 15 intervention group. 16 Work Means and SD. Total Only physiological related measure without Time at 4 weeks. distress significantly 17 distress sub-measures given. Means and SD. lower for 18 ControlFor peermeasure reviewINTERVENTION INTERVENTIONonly intervention group. 19 Cognitive group Work (WRD-7), (n=86), Work (n=77), Work- When looking at 20 behavioural had related General related morale total related morale total those GPs that stress other morale psycholo 31.83 (6.75). QOWL 34.62 (6.11). QoWL scored low for 21 manage- devel- measure gical total 25.32 (6.64). total 28.24 (6.35). Time at 12 weeks. morale, there was a 22 GPs in ment opmen (WRM-7), distress CONTROL (n=24), CONTROL (n=19), Means and SD. 56% reduction pre 23Gardiner metro- course, 1x tal Quality (GHQ- Work related Work related INTERVENTION to post intervention, 242004 politan 3h, 15h course of work 12), morale total 31.42 morale total (n=62), Work compared with a Australia over 5 wks. s. life coping (6.19). QoWL total 32.21 (6.73). QoWL related morale total 29% reduction in 25CBA (n=110) (n=85). (n=25) (QoWL-6) styles. 23.16 (5.86). total 24.32 (6.36). 35.70 (6.01). the control group. 26 Time at 8 weeks. Time at 12 months. 27 Mean and CI. Baer Mean and CI. Baer 28 Mean and CI. Baer mindfulness total mindfulness total 29 mindfulness total (52.9, 95% CI 51 to (55, 95% CI 53.0 to CME (45.2, 95% CI 43.3- 54.8, n=59), Baer 56.9, n=56), Baer 30 programme 47.1, n=60), Baer mindfulness mindfulness 31 : Didactic mindfulness observe (30.6, 95% observe (31.1, 95% 32 material, observe (25.6, 95% CI 29.4 to 31.8, CI 29.8 to 32.3, 33 mindful- CI 24.4-26.8, n=60), n=59), Baer n=56), Baer Baseline scores and ness Baer mindfulness mindfulness mindfulness follow-up scores at 34 meditation, nonreact (19.7, CI nonreact (22.9, CI nonreact (23.9, CI 15 months are not 35 narrative/a 95% 18.7-20.7, 95% 21.8 to 23.9, 95% 22.9 to 24.9, reported for the 36 ppreciative n=60). JSPE total n=59). JSPE total n=56). JSPE total purpose of this SR. 37 enquiry, Burnout (116.2, 95% CI (120.6, 95% CI 118.2 (121.4, 95% CI 119.0 Over time, all http://bmjopen.bmj.com/ 38 discussion. (MBI), 114.2-118.9, n=60), to 123.0, n=59), to 123.8, n=56), measures for 28h total. 8 Mood JSPE compassionate JSPE compassionate JSPE compassionate mindfulness, 39 Primary x 2.5hr per (POMS), care (48.6, 95% CI care (49.8, 95% CI care (50.4, 95% CI burnout, physician 40 care week, 1x Big 5 47.5-49.7, n=60), 48.7 to 50.9, n=59), 49.3 to 51.5, n=56), belief, mood and 41 physici- 7h, 10- personali JSPE perspective JSPE perspective JSPE perspective personality 42 ans in the month Mindfuln ty taking (57.1, 95% CI taking (59.1, 95% CI taking (59.7, 95% CI improved, the Krasner USA mainten- ess factors, 55.6 to 58.6, n=60), 57.6 to 60.6, n=59), 58.2 to 61.2, n=56), largest effect size 432009 (n=70, ance phase (Baer), Physician JSPE standing in pts JSPE standing in pts JSPE standing in pts was observed for 44 54% (2.5 h/mo). Empathy belief shoes (10.8, 95% CI shoes (11.7, 95% CI shoes (11.4, 95% CI mindfulness at 15 45NCBA male) (n=70) NA (JSPE) (PBS) 10.4 to 11.5, n=60). 11.1 to 12.2, n=59). 10.9 to 11.9, n=56). mts. on September 26, 2021 by guest. Protected copyright. 46 Time at 12 to 47 18months. Number Primary Number given in %. given in %. 48 care Intervent- INTERVENTION INTERVENTION 49 clinicians ions to (n=83) High work (n=67) High work 50 in the improve control (score >2) control (score >2) 51 USA communica 96%, High 4.6%, High Satisfaction (n=166, > tion; Work Stress, satisfaction (>3) satisfaction (>3) improved with 52 80% workflow, Control, burnout, 38.5%. CONTROL 40.0%. CONTROL workflow 53 physic- quality Satisfacti chaos, (n=83) High work (n=72) High work interventions, 54Linzer 2015 ians, 47.3 improve- on (from intent to control 13.2%. High control 11.4%, High targeted QI 55 yrs, 48% ment (QI). WLC PWS and leave, satisfaction (>3) satisfaction (>3) projects, RCT male) n=83 n=83 MEMO). varia 51.8%. 45.7%. communication. 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 32 of 44

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Significant 5 improvement on 6 CPD self-esteem post 7 teach- intervention. The ing interactive teaching 8 progr- approach improved 9 CPD am Knowled Time at 6 months. self-esteem more 10 teaching (n=22 Attitude ge, Mean, SD. Self- Mean, SD. Self- than the didactic 11Margalit programs didac- to intention esteem (72.1, 14.5), esteem (76.5, 12.9), teaching did. No 2005 GPs in (n=22 for tic patient , Attitude to patient Attitude to patient improvement on 12 Israel interactive meth- care, self- attitude, care (34.5, 12.2) care (36.5, 19.9) attitude to patient 13RCT (n=44) program) od) esteem burnout (n=44) (n=44) NA care. 14 Brief 15 blended 16 web-based mindful- Time at 4 weeks. 17 ness Mean, SD. 1 Mean, SD. 1 Benefits in pos 18 intervent- For peer reviewsession/wk (n=28): session/wkonly (n=28): PANAS and MAAS 19 ion,1x 4h PANAS-pos (32.19, PANAS-pos (33.44, for two or more 20 face to 6.72). MAAS (64, 5.42). MAAS (66.67, weekly meditation face, 8 Positive 12.07). CD-RISC 10.88). CD-RISC session. No benefits 21 GPs in online affect Negative (38.96, 8.96). 2 (40.19, 5.17). 2 for one weekly 22Montero- Spain sessions (2 (PANAS), affect sessions/wk (n=30). sessions/wk (n=30). practice. While face 23Martin (n=290, weekly awarenes (PANAS), PANAS-pos (32.03, PANAS-pos (35.00, to face attendance 242018 49 yrs, sessions s (MAAS), Burnout 6.38). MAAS (61.77, 4.91). MAAS (66.37, was good, very high 22.5% over 4 wks) resilience subtypes( 13.41). CD-RISC 11.03). CD-RISC attrition rate for 25NCBA males) (n=58). NA (CD-RISC) BCSQ) (38.80, 8.58). (41.28, 4.32). online component. 26 Of six physicians, 27 100% were satisfied 28 Burnout, Time at 3 months. with clinic workflow 29 Time n=6. Number of post-pilot (vs. 33% spent on n=6. Number of responses 'agree' or pre-pilot), and 83% 30 EHR responses 'agree' or 'strongly agree'. were satisfied with 31 documen 'strongly agree'. Satisfied with clinic EHR use post-pilot 32 tation, Satisfied with clinic workflow 6/6 (vs. 17% pre- pilot). 33 patient workflow 2/6 (33%). (100%). Calm Physician burnout Academic satisfacti Calm atmosphere in atmosphere in work was low at baseline 34 general on with work area 0/6 (0%). area 2/6 (33%). and did not change 35 internal doctor– Satisfied with Satisfied with post-pilot. Mean 36 medicine patient quality of patient quality of patient time spent on post- 37 clinic Clinic Physician relations interactions 5/6 interactions 6/6 clinic EHR http://bmjopen.bmj.com/ 38Pozdnya- (n=6 sessions workplac hip, (83%). Satisfied with (100%). Satisfied documentation kova 2018 faculty, with and e attitudes quality of with quality of decreased from 1.65 39 n=325 without a satisfacti towards communication with communication with to 0.76 h per clinic 40NCBA, 2007 patients). scribe. NA on scribes. patient 4/6 (67%). patient 5/6 (83%). session (p = 0.02). 41 Rural 42 medical practiti- 43 oners in MSCR, 44 Australia Mindful 45 (57% Self-Care on September 26, 2021 by guest. Protected copyright. 46 GPs). and For the WHO-5 47 (n=13 Resiliency wellbeing scale total, n=4 program Burnout there was no 48 qual (7h face-to- measure change pre-post for 49 research, face (short one doctor, 50 n=7 session) Wellbein version deterioration for 51Rees 2020 quant and 3x 1h g (WHO- 10-item), Time at 4 weeks. two doctors with no research, video- 5), psycholo n=7. Mean. n=7. Mean. reliable change, 52NCBA, 40 yrs, conference Positive gical Wellbeing (61.1). Wellbeing (71.4). Improvement for 53mixed 76.9% (follow up affect strain Positive affect not Positive affect not four doctors with no 54methods male) sessions) NA (PANAS). (GHQ-12) reported. reported. reliable change. 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 33 of 44 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Participants in the 5 MMC group 6 reported significant 7 improvements over Time at 4 weeks. Time at 3 months. time for MAAS (also 8 Mindful Mean, SD. Mean, SD. Mean, SD. PSS and MBI), 9 Primary Attention INTERVENTION INTERVENTION INTERVENTION whereas in the 10 care Mindful Awarene Stress (n=16) MAAS (3.42, (n=15) MAAS (3.62, (n=13) MAAS (4.04, control group, there 11 physici- Medicine ss (PPS), 0.96), BRS (21.62, 0.89), BRS (22.33, 1.02), BRS (24.15, were no ans in the Curriculum (MAAS), Burnout 4.45), SCBCS (26.31, 4.74), SCBCS (27.66, 5.47), SCBCS (27.84, improvements. 12Schroeder USA = modified Resilienc (MBI), 4.51). CONTROL 3.22). CONTROL 4.09). CONTROL There was no 132018 (n=33, version of e (BRS), Meditati (n=17) MAAS (3.32, (n=14) MAAS (3.08, (n=13) MAAS (3.18, significant 14 42.76 yrs, MBSR, 1x Compassi on 0.76), BRS (18.70, 0.76), BRS (19.42, 0.58), BRS (18.82, improvement for 15RCT, 27% 13h & 2x WLC on Practice 5.13), SCBCS (27.00, 4.21), SCBCS (26.07, 5.32), SCBCS (25.07, resilience or 162014 - 2015 male) 2h (n=16) (n=17) (SCBC), (MPQ) 4.97). 4.73). 5.85). compassion. Significant 17 improvement of 18 For peer review only self-compassion and 19 Primary self-reflection. Six 20 care months after PSS physic- and SCS were still 21 ians in improved. PSS 22Van the Adapted significantly 23Wietmar- Nether- MBSR reduced. Qual: 24schen 2018 lands programme Self- awareness, (n=54, , weekly Compassi Cohen acceptance, 25NCBA, 87% GPs, group on (SCS), Perceive Time at 8 weeks. Time at 6 months. peacefulness, and 26mixed 40 yrs, sessions for Self- d Stress Mean, SD. SCS (2.9, Mean, SD. SCS (3.4, Mean, SD. SCS (3.7, openness improved 27methods, 22% 8 weeks, 26 reflection Scale 0.7, n=50), GRAS 0.6, n=50), GRAS 0.7, n=21), GRAS through 282015 - 2016 male) hours total. NA (GRAS). (PSS) (87.6, 7.7, n=44). (90.9, 6.7, n=44). (90.2, 10.9, n=17). intervention. 29 Time at 8 weeks. Mean, SE. 30 Mean, SE. INTERVENTION 31 INTERVENTION (n=43) JSPE (119.35, 32 (n=43) JSPE (117.4, 1.49), FFMQ Total 33 1.53), FFMQ Total (143.08, 2.19), (136.21, 2.23), FFMQ Observing 34 FFMQ Observing (28.4, 0.68), FFMQ 35 (26.36, 0.69), FFMQ Describing (29.77, 36 Describing (28.26, 0.71), FFMQ Acting 37 0.9), FFMQ Acting with awareness http://bmjopen.bmj.com/ 38 with awareness (28.1, 0.69), FFMQ (27.12, 0.71), FFMQ Non-judging (32.36, 39 Non-judging (31.16, 0.79), FFMQ Non- 40 0.81), FFMQ Non- reacting (24.47, 41 reacting (23.34, 0.56, 0.57). Mindfulness skills 42 0.57). CONTROL CONTROL (n=20) increased (n=20) JSPE (116.18, JSPE (117.93, 1.98). significantly in the 43 1.92). FFMQ Total FFMQ Total (135.45, MBSR group. 44 General (135.48, 2.65), 2.67), FFMQ Empathy remained 45 practiti- FFMQ Observing Observing (26.33, the same. The on September 26, 2021 by guest. Protected copyright. 46 oner (25.85, 0.82), FFMQ 0.83), FFMQ qualitative data 47 trainers Describing (28.33, Describing (28.2, indicated that the in the 0.9), FFMQ Acting 0.91), FFMQ Acting MBSR course 48 Nether- Empathy with awareness (28, with awareness increased their 49Verweij lands MBSR (JSPE-20), Work 0.9), FFMQ Non- (27.41, .91), FFMQ wellbeing and 502016 (n=50, training 8x Mindfuln engagem judging (30.49, Non-judging (30.32, compassion towards 51 54.9 yrs, 2.5h, 1x 8h ess ent, 0.98), FFMQ Non- 0.99), FFMQ Non- themselves and CBA, mixed 66% retreat WLC (FFMQ- Burnout reacting (23.12, reacting (23.49, others, including 52methods male) (n=30) (n=20) 39) (UBOS-C) 0.7). 0.7). their patients. 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 34 of 44

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Statistically 5 significant 6 GPs in improvement in job 7 the UK. satisfaction. Job 2004: pressure and work 8 n=2105, hours significantly 9 mean New declined. Most GPs 10 age introduct- reported that the 11 46.9yrs, ion of pay new contract had 66% for increased their 12 male. perform- income (88%), but 13 2005: ance decreased their 14Whalley n=1349, system Job professional 152008 mean happened pressure, autonomy (71%), 16 age in 2004 Job Time at one year. and increased their Panel 48.6yrs, (after the design Mean, SD. 2004: JSS Mean, SD. 2005: JSS administrative 17survey, 65 % 2004 and time total (4.58, 1.39, total (5.17, 1.28, (94%) and clinical 182004, 2005 male. survey) NA ForJSS WCW peerpressures reviewn=2081) n=1345)only (86%) workloads. 19 20 Key: RCT Randomised Controlled Trial; CBA Controlled Before and After Trial; NCBA Non-Controlled 21 Before and After Trial; QOF Quality and Outcomes Framework; P4P Pay for Performance; NP Nurse 22 Practitioner; WCW Warr Cook Wall; JS Job satisfaction; JSS Job Satisfaction Scale; MBI Maslach Burnout 23 Inventory; JSPE Jefferson Scale of Physician Empathy); POMS Profile Of Mood States; FFMQ Five Facet 24 Mindfulness Questionnaire; SRSI Smith Relaxation States Inventory; NA Not applicable; PBS Physician 25 Belief Scale; SCS Self Compassion Scale (Neff); MPQ Meditation Practice Questionnaire, FR-JS Font Roja 26 Job Satisfaction Questionnaire; BPRS Brief Psychiatric Rating Scale; IQR Inter Quartile Range; ACP College 27 of Physicians; ASIM American Society of Internal Medicine; DASS Depression and Anxiety Scale; PJSS 28 Physician Job Satisfaction Scale (3 dimensions JS, career satisfaction, and specialty satisfaction), 12-item 29 Global Job Satisfaction subscale used; UWES Utrecht Work Engagement Score; CD-RISC Connor- 30 Davidson Resilience Scale; PANAS Positive And Negative Affect Scale, MAAS Mindful attention 31 awareness scale; GHQ General Health Questionnaire; AWS Areas of Work Life Scale (positive measures 32 include control, reward, community, values); BRS Brief Resilience Scale; SCBCS Santa Clara Brief 33 Compassion Scale (an abbreviation of the Sprecher and Fehr’s Compassionate Love Scale); MPQ 34 Meditation Practice Questionnaire; UBOS-C Utrecht Burnout Scale for Contractual Occupations (this is 35 the validated Dutch version of the Maslach Burnout Inventory); GRAS Groningen Reflection Ability Scale; 36 EHR Electronic Health Record; MTP Multimodal training programme; IBST integrated brief systemic 37 therapy, PPS Perceived Stress Scale. http://bmjopen.bmj.com/ 38 39 40 41 42 43 44 45 on September 26, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 35 of 44 BMJ Open

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Table 2. Overview of included studies 5 6 Within group Between group 7 Study effect sizes effect sizes (exp 8 Intervention level Intervention type Study ID (Author, Date) design Outcome Measure (experiment) vs control group) 9 10 Intervention targeted 11 at the individual Mindfulness Amutio 2015 RCT FFMQ 0.78 1.5 12 Asuero 2014 RCT FFMQ, JSPE 0.64, 0.31 0.57, 0.44 13 14 Schroeder 2018 RCT MAAS, BRS, SCBC 0.63, 0.51, 0.35 0.88, 0.61, 0.73 15 Verweij 2016 CBA FFMQ, JSPE 0.47, 0.20 0.5, 0.02 16 Fortney 2013 NCBA SCBC, RS-14 -0.04, 0.17 - 17 18 ForMontero-Marin peer 2018 reviewNCBA PANAS, only MAAS, CD-RISC 0.52, 0.37, 0.37 - 19 Rees 2020 NCBA PANAS, WHO-5 -, 0.52 - 20 Van Wietmarschen 2018 NCBA SCS, GRAS 0.77, 0.46 - 21 22 Krasner 2009 NCBA FFMQ-2 (BAER), JSPE 1.05, 0.44 - 23 Educational training / experiential workshop Barcons 2019 CBA FR-JS - - 24 25 Gardiner 2004 CBA WRM-7, QoWL-6 0.43, 0.45 0.3, 0.27 26 Margalit 2005 RCT Self-esteem - - 27 Coaching Dyrbye 2019 RCT CD-RISC, PJSS 0.24, 0.10 0.13, 0.06 28 29 Professional balance, Intervention at Clerical support / physician satisfaction 30 organisational level Scribes Contratto 2017 NCBA (QoL) - - 31 Physician workplace 32 Pozdnyakova 2018 NCBA satisfaction - - 33 Organisational 34 improvement 35 programme / QA Dunn 2007 NCBA Physician satisfaction - - 36 Work control, Linzer 2015 RCT satisfaction - - 37 http://bmjopen.bmj.com/ 38 Intervention at Introduction of pay for NCBA Job satisfaction (JSS systemic level / policy 39 performance scheme Allen 2017 (naturalistic) WCW), Life satisfaction - - 40 41 NCBA 42 Whalley 2008 (naturalistic) JSS WCW 0.44 - 43 44

Key: Study design: RCT Randomised Controlled Trial; CBA Controlled Before and After trial; NCBA Non- on September 26, 2021 by guest. Protected copyright. 45 Controlled Before and After trial. 46 Outcome measures: JSS WCW Warr Cook Wall Job satisfaction Scale; JSPE Jefferson Scale of Physician 47 Empathy; FFMQ Five Facet Mindfulness Questionnaire; SRSI Smith Relaxation States Inventory; SCS Self 48 Compassion Scale (Neff); FR-JS Font Roja Job Satisfaction Questionnaire; SCBC Santa Clara brief 49 compassion scale (an abbreviation of the Sprecher and Fehr’s Compassionate Love Scale); PJSS Physician 50 Job Satisfaction Scale (3 dimensions JS, career satisfaction, and specialty satisfaction), 12-item Global 51 Job Satisfaction subscale used; CD-RISC Connor-Davidson Resilience Scale; PANAS Positive And Negative 52 53 Affect Scale, MAAS Mindful attention awareness scale; BRS Brief Resilience Scale; GRAS Groningen 54 Reflection Ability Scale; RS-14 Resilience, WHO-5 Wellbeing Index, WRM Work Related Morale Measure. 55 QoWL-6 Quality of Work Life. 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 36 of 44

1 2 3 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from 4 Table 3. Overview of within group and between group Effect Sizes (ES) for several 5 6 positive outcomes of mindfulness interventions 7 8 9 Outcome Mindfulness Resilience Compassion Empathy 10 11 12 13 Within 14 15 group ES 0.37 – 1.05 0.17 – 0.51 -0.04 - 0.77 0.2 – 0.44 16 17 Mindfulness (k=9) 18 interventions For peer review only 19 20 (k=9) 21 22 Between Only one Only one 0.02 - 0.44 23 group ES 0.5 – 1.5 study study 24 25 (k=4) 0.61 0.73 26 27 28 29 30 There were 9 studies that trialed mindfulness interventions (k=9). These studies utilized a range of 31 different outcomes measures and included tools for assessing mindfulness, resilience, compassion, and 32 empathy. Within group effect sizes (ES) are shown, comparing before and after measures for the 33 intervention. 34 35 36 37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44 45 on September 26, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 37 of 44 BMJ Open

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 38 of 44 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

1 2 3 Supplements 4 5 6 • Supplement 1. Medline OVID search strategy 7 8 • Supplement 2. Formulae used for Effect Size calculations in Excel for within and between group ES. 9 10 • Supplement 3. Effect sizes for outcome measures and sub-measures 11 • Supplement 4. Outcome measures utilized in the included studies. 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40 41 on September 26, 2021 by guest. Protected copyright. 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 1 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 39 of 44 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

21. Burnout, Professional/ on September 26, 2021 by guest. Protected copyright. 41 42 22. Compassion fatigue/ 43 44 23. compassion fatigue.mp. 45 46 24. joy.mp. 47 25. "joy in practice".mp. 48 49 26. 11 or 12 or 13 or 14 or 15 or 16 or 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 50 51 27. Motivational interviewing/ 52 53 28. Mindfulness/ 54 29. Mediation/ 55 56 57 58 59 2 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 40 of 44 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

1 2 3 4 30. mindful*.mp. 5 31. meditat*.mp. 6 7 32. Self-Care/ 8 9 33. ("self care" or "self-care").mp. 10 11 34. Motivation/ 12 35. Professional autonomy/ 13 14 36. ("self determination" or "self-determination").mp. 15 16 37. exp Randomised ControlledFor Trial/ or Clinicalpeer Trial/ review only 17 18 38. trial.mp. 19 39. Early Medical Intervention/ or intervention.mp. 20 21 40. Counseling/ 22 23 41. Course?ling.mp. 24 25 42. Mentoring/ 26 43. mentor*.mp. 27 28 44. Education/ 29 30 45. education.mp. 31 32 46. workshop*.mp. http://bmjopen.bmj.com/ 33 47. training.mp. 34 35 48. coaching.mp. 36 37 49. facilitat*.mp. 38 39 50. 27 or 28 or 29 or 30 or 31 or 32 or 33 or 34 or 35 or 36 or 37 or 38 or 39 or 40 or 41 or 42 or 43 or 44 or 45 or 46 or 47 or 48 or 49 40 on September 26, 2021 by guest. Protected copyright. 41 51. 10 and 26 and 50 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 3 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 41 of 44 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

26 ES_BG = Mean change_BG / SD* pooled_BG 27 28 29 30 Legend. ES = Effect size, WG = Within group, BG = Between group, Exp = experimental group, Ctrl = 31 control group, N = sample size, pre = pre-intervention measure, post = post-intervention measure, SQRT = 32

square root, ^2 = raised to the power of two. http://bmjopen.bmj.com/ 33 34 35 36 37 38 39 40 41 on September 26, 2021 by guest. Protected copyright. 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 4 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 42 of 44 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

1 2 3 Supplement 3. Effect sizes for outcome measures and sub-measures 4 5 Between group 6 Within group ES Hedges' g absolute mean 7 absolute mean Exp (Delta change SD* ES Hedges' g Name, change (post - SD* Exp mean/ SD* (experiment – Baseline (Delta Mean / 8 n pre/post Measure pre experiment) pooled pool) control group) pooled SD* pool) 9 10 With control 11 group 12 AMUTIO* FFMQ 0.37 0.48 0.78 0.54 0.36 1.51 13 n=21/21 Observe 0.65 0.62 1.05 0.84 0.75 1.12 14 Describe 0.25 0.67 0.37 0.31 0.63 0.49 15 Awareness 0.32 0.77 0.42 0.34 0.79 0.43 16 Non-judgement 0.36 0.65 0.55 0.39 0.58 0.67 Non-reactiveFor peer 0.29review 0.57 only0.51 0.26 0.54 0.48 17 SRSI-3 Relaxation 0.54 0.57 0.95 0.46 0.52 0.89 18 SRSI-3 Pos energy 0.71 0.74 0.97 0.71 0.63 1.13 19 SRSI-3 Mindfulness 0.71 0.81 0.88 0.76 0.91 0.84 20 SRSI-3 Transcend 0.6 0.93 0.65 0.84 0.88 0.96 21 VERWEIJ** FFMQ 6.87 14.49 0.47 6.9 13.82 0.50 22 n=43/43 exp, 23 n=20/20 control Observe 2.04 4.49 0.45 1.56 4.28 0.36 24 Describe 1.51 4.69 0.32 1.64 4.52 0.36 25 Awareness 0.98 4.59 0.21 1.57 4.47 0.35 Non-judgement 1.2 5.25 0.23 1.37 5.04 0.27 26 Non-reactive 1.13 3.71 0.30 0.76 3.56 0.21 27 JSPE total 1.95 9.90 0.20 0.2 9.61 0.02 28 29 ASSUERO FFMQ 12.40 19.50 0.64 11.3 19.65 0.57 30 n=43/43 exp 31 n=25/25 control Observe 4.30 7.07 0.61 4.7 7.33 0.64 32 Describe 0.70 5.50 0.13 0.6 5.52 0.11

46 GARDINER WRM-7 total 2.79 6.46 0.43 2 6.63 0.30 47 n=86/77 exp, 48 n=24/19 control QoWL-6 total 2.92 6.50 0.45 1.76 6.48 0.27 49 50 SCHROEDER MAAS 0.62 0.99 0.63 0.76 0.86 0.88 51 N=16/13 exp, 52 n=17/13 control BRS 2.53 4.93 0.51 2.95 4.81 0.61 SCBC 1.53 4.33 0.35 3.46 4.75 0.73 53 54 MARGALIT*** 55 n=44/44 Self-esteem 4.4 13.72 0.32 56 57 58 59 5 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 43 of 44 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

1 2 3 4

5 6 7 Without control 8 group 9 10 MONTERO- MARIN 11 n=30/30 PANAS 1 session/wk 1.25 6.10 0.20 12 13 PANAS 2 session/wk 2.97 5.69 0.52 14 MAAS 1 session/wk 2.67 11.49 0.23 MAAS 2 session/wk 4.59 12.28 0.37 15 CD-RISC total 1 16 session/wkFor peer 1.23review 7.31 only0.17 17 CD-RISC total 2 18 session/wk 2.48 6.79 0.37

55 56 57 58 59 6 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 44 of 44 BMJ Open: first published as 10.1136/bmjopen-2020-046599 on 18 August 2021. Downloaded from

Page 45 of 44 BMJ Open