Table 1.

Studies included in systematic review

First author, year	Population/country	Sample size/response rate	Study design	Definition of resilience	Aim of study	Results associated with resilience
Quantitative studies
Cooke, 20137	GP registrars (junior doctors), hospital and GP rotations. In four regions of Australia	148 invited: participants reported: 128 (90%)	Cross-sectional	Dynamic, evolving process of positive attitudes and effective strategies, as responses to life stressors	To measure resilience, burnout, compassion satisfaction, personal meaning in patient care, and intolerance of uncertainty	Compassion satisfaction/personal meaning in patient care; negative association with burnout/secondary traumatic stress/inhibitory anxiety/general intolerance to uncertainty/concern about bad outcomes and reluctance to disclose uncertainty to patients No relationship between burnout or resilience and reluctance to disclose uncertainty to physicians Lower resilience scores associated with higher risk of burnout
Eley, 201310	Family practitioners and (10%) international medical graduates across all states in Australia	785 invited: participants 479 (61%)	Cross-sectional	Resilience can be considered as a process of adaptation to adversity and stress. Resilient individuals tend to recover from setbacks or trauma and portray a common set of characteristics that help them cope with challenges in life	To investigate the relationship between the pattern of personality traits and measures of resilience in a sample of family practitioners	Multiple regression found significant association between resilience and high self-directedness, high persistence, and low levels of harm avoidance (full model not shown)
Gerber, 201412	Healthcare workers and social insurance officers, in Sweden	4434 invited: response rate 2705 (61%), final sample after outliers removed	Cross-sectional	Those with high stress and good mental health are defined as resilient	To examine whether employees with differing occupational stress and mental health profiles differ in their self-reported levels of physical activity	Physical activity associated with healthy and resilient profiles among the public service sector individuals, independent of social and demographic background
Glasberg, 200716	Healthcare personnel including primary healthcare centres and one hospital, in rural northern Sweden	625 invited: response rate 423 (68%)	Cross-sectional	Not given by authors Examples from Wagnild and Young (1993) and Wagnild (2009)21,22	To determine the factors associated with burnout in healthcare personnel	Univariate analysis: higher levels of EE and DP both associated with low resilience. Multivariate analysis: EE associated with resilience (not DP)
Keeton, 200715	Doctors: general obstetrician–gynaecologists, subspecialty obstetrician–gynaecologists, general internal medicine, general paediatrics, general surgery, and family medicine, in US	2000 invited: completed surveys 935 (48%)	Cross-sectional	Not explicitly reported	To explore factors associated with physician career satisfaction, work–life balance, and burnout focusing on differences across age and specialty	Career satisfaction: females more satisfied, personal accomplishment and emotional resilience significantly associated with career satisfaction. Associations with work–life balance: some control of schedule and hours, total weekly hours, being older, fewer children. Associations with emotional resilience: control over schedule and hours, being older, fewer children. Associations with personal accomplishment: being older, some control over schedule and hours, total weekly hours, gross annual household income, and being an obstetrician–gynaecologist
Rossouw, 201311	Medical doctors in community healthcare clinics and district hospitals, in South Africa	147 invited: response rate 135 (92%), 3 excluded, total participants 132 (90%)	Cross-sectional	Resilience defined in discussion as a measure of a person’s stress coping ability and as a target for interventional measures when combating burnout	To investigate burnout and depression in medical doctors in context of work-related conditions and the role of resilience as a modifiable factor. To define magnitude of the problem to motivate the development of an action within reach of doctors working in these settings	CD-RISC score correlated negatively with emotional exhaustion, depersonalisation, and BDI score. CD-RISC score correlated positively with personal accomplishment and delivered quality of care. Participants using medication had lower CD-RISC score. No correlations with sociodemographic data
Taku, 201313	Doctors: mixture of internal medicine; family medicine, radiology, and other. Mixture of residents, attending, and fellows, in US	839 invited: total participants 290 (34.5%)	Cross-sectional	Resilience is a personality attribute that moderates the negative effects of stress and promotes the ability to ‘bounce back’ after adversity.20 Resilience moderates the negative effects of stress and promotes the ability to ‘bounce back’ following adversity (Wagnild & Young 1993)21	To examine the relationships between perceived growth as a physician and burnout after controlling for the effects of perceived family support, dispositional resilience, age, and marital status	PTG significantly associated with burnout in all three burnout domains. EE: higher PTG associated with lower levels of EE-resilience and family support not significant in final model. DP: PTG/family support associated with DP, interaction between PTG and family support, that is, high DP had low levels of PTG family support. Personal accomplishment: resilience and PTG associated, interaction between resilience and PTG, effect of PTG stronger for physicians with lower resilience
Unrath, 201217	GPs in outpatient sector (primary care) in Rhineland-Palatinate, Germany	2092 invited: response rate 808 (39%)	Cross-sectional	Resilience relates to the capacity to remain healthy in the face of strain and is therefore a sort of hardiness	To identify possible risk factors for AUDs among GPs working in the outpatient sector	Resilience negatively associated with AUD. Results suggest certain resilience level is protective against AUD; does not alter much with increased resilience
Qualitative studies
Bowden, 201414	Frontline mental health professionals (locally known as ‘linkworkers’) working mainly in general practice settings, in UK	10 invited: 9 (90%) involved in individual interviews and focus groups	Qualitative	Not defined but referred to in guidance papers as ‘linked with adaptability to change’ 23,24	To investigate frontline mental health professionals’ perceptions of work stress and the rewards and demands associated with their work	Factors related to caring role: stressors control and responsibility for care, demands/managing boundaries/ownership. Creativity related to developing service. Range of coping seen in individual and team strategies. Linkworkers: aspects of work positive but in excess could become sources of stress
Jensen, 200818	Family physicians (GPs) in Hamilton, Ontario, Canada	20 invited to attend focus groups: participants 17 (85%)	Qualitative	No working definition — in guidance papers ‘A dynamic evolving process of positive attitudes and effective strategies’	To explore the dimensions of family physician resilience	Four main aspects of physician resilience were: 1) attitudes and perspectives, includes valuing physician role, maintaining interest, developing self-awareness, accepting personal limitations; 2) balance and prioritisation, includes setting limits, taking effective approaches to continuing professional development, honouring the self; 3) practice management style, includes sound business management, good staff, using effective practice arrangements; 4) supportive relations, includes positive personal relationships, effective professional relationships, and good communication
Stevenson, 201119	Primary healthcare physicians working in Aboriginal health, prisons, drug and alcohol medicine, or youth and refugee health, in Australia	15 (100%)	Qualitative	No explicit definition of resilience given	To explore job satisfaction and resilience among primary care doctors who have worked for sustained periods in medically underserved populations in Australia	Emergent theory: reflectiveness, respectful engagement with job, and clear sense of boundaries and limits important in mitigating impact of work volume and intensity that organisations often struggled to contain. Also linked to and argued as predictors of resilience: job satisfaction, respect for patients, sense of control, and intellectual interest. Referred from existing literature, construct of ‘prosociality’ links to ‘vicarious resilience’
Zwack, 20139	Psychiatrists, surgeons, GPs, and other physicians, across Germany	Recruiting followed pyramid approach; 200 interviews conducted	Qualitative	Not defined	To identify health-promoting strategies employed by experienced physicians in order to define prototypical resilience processes and key aspects of resilience-fostering preventive actions	30 subcodes/three dimensions: 1) job-related gratifications derived from treatment interactions; 2) practices: leisure-time activities, self-demarcation, limitation of working hours, continuous professional development; 3) attitudes: acceptance of professional and personal boundaries, focus on positive aspects of work, personal reflexivity Conclusion: ‘In relation to Conservation of Resources Theory, physician resilience emerged as the ability to invest personal resources in a way that initiates positive resource spirals in spite of stressful working conditions.’
Intervention
Fortney, 20138	Primary care clinicians Wisconsin-Madison, US	Number invited unknown: 30 recruited	Intervention	The ability to recover from adversity	To investigate whether an abbreviated mindfulness intervention could increase job satisfaction, quality of life, and compassion among primary care clinicians	Reductions in job burnout, depression, anxiety, and stress associated with participation in mindfulness training courses adapted for primary care physicians. Resilience score not significantly changed at any follow-up (1 day, 8 weeks, or 9 months)

AUD = alcohol use disorder. BDI = Beck Depression Inventory. CD-RISC = Connor Davidson Resilience Scale. DP = depersonalisation. EE = emotional exhaustion. PTG = post-traumatic growth.