Physicians' occupational stress, depressive symptoms and work ability in relation to their working environment: a cross-sectional study of differences among medical residents with various specialties working in German hospitals.

OBJECTIVES: This study aimed to analyse and compare differences in occupational stress, depressive symptoms, work ability and working environment among residents working in various medical specialties.
METHODS: 435 German hospital residents in medical training working in 6 different medical specialties participated in a cross-sectional survey study. Physicians were asked about their working conditions and aspects of mental health and work ability. The Copenhagen Psychosocial Questionnaire, the Work Ability Index, the ICD-10 Symptom Rating and the Perceived Stress Questionnaire were used to measure working conditions, mental health and work ability.
RESULTS: Results show that up to 17% of the physicians reported high levels of occupational distress and 9% reported high levels of depressive symptoms. 11% of the hospital physicians scored low in work ability. Significant differences between medical specialties were demonstrated for occupational distress, depressive symptoms, work ability, job demands and job resources. Surgeons showed consistently the highest levels of perceived distress but also the highest levels of work ability and lowest scores for depression. Depressive symptoms were rated with the highest levels by anaesthesiologists. Significant associations between physicians' working conditions, occupational distress and mental health-related aspects are illustrated.
CONCLUSIONS: Study results demonstrated significant differences in specific job stressors, demands and resources. Relevant relations between work factors and physicians' health and work ability are discussed. These findings should be reinvestigated in further studies, especially with a longitudinal study design. This work suggests that to ensure physicians' health, hospital management should plan and implement suitable mental health promotion strategies. In addition, operational efficiency through resource planning optimisation and work process improvements should be focused by hospital management. Published by the BMJ Publishing Group Limited. For permission to use (where not already granted under a licence) please go to http://www.bmj.com/company/products-services/rights-and-licensing/

This is one of the first studies to analyse and compare differences in mental health outcomes, perceived distress and working conditions among medical specialties.

The main predictors for work ability and depressive symptoms within the work domain of physicians are illustrated.

The relatively large sample of physicians, together with few missing data, strengthens the study.

The cross-sectional design limits the value in supporting causal effects.

Data were assessed by self-report measures that limit the objectivity of the study results.

Background

German physicians’ working environment

The working situation for German hospital physicians is changing constantly: currently, all German doctors are confronted with cost-effective medicine, increasing legal regulations and clinical job demands, combined with a decrease in time for individual-oriented patient care.1 The pressure to balance budgets with less financial resources add extra burden on the hospital personnel.2 As a result, physicians experience growing demands and pressure to perform at work.3 4

Although there has been growing interest in horizontal comparisons of hospital doctors' working environment in different medical specialties and their job outcomes, only few studies focused on these assessments in German hospitals.

Tanner et al5 demonstrated significant differences between medical specialties for time pressure, uncertainty and social stressors with patients. Mache et al6–10 published several observational studies including work task analysis in different medical settings showing different workloads and working demands.11 In addition, von dem Knesebeck et al12 analysed psychosocial stress in the workplace among hospital doctors working in surgical fields in Germany showing that this group suffers from more severe stress at work than other occupational groups.

Associations between physicians' mental health, work ability and working environment

Physicians' mental health condition is a result of a complex and simultaneous interaction of individual constraints, individual resources and social and environmental factors.13 The relationship between mental health and physicians' working conditions has been previously demonstrated in several theoretical models such as in the job demands-resources model (JD-R model) and the demand–control–support model.14–16 These models build the rationale for this study.

According to the JD-R model, effects of certain job demands (ie, organisational job factors that require sustained physical and/or psychological efforts) are strongly associated with mental health outcomes. For example, when high levels of job demands and low levels of control and reward are combined, employees are at an increased risk of anxiety and depression.17 In addition, high levels of job demands are linked to the experience of fatigue.18

A key proposition of the JD-R model is that interactions between job demands and resources are important by the way that certain job resources can buffer negative effects of psychological distress.15 Research supports this assumption, and a number of cross-sectional studies have shown negative associations of job resources with depression and occupational distress. These job resources include job control,19 20 social support at work,21 22 reward, feedback and supervisory coaching.23

The demand–control–support model24 presumes that working situations have negative psychological or physical consequences in particular when high demands coincide with limited decision latitude and low social support at the workplace.24 Several studies showed that increased job demands are associated with depressive symptoms and distress.25–28

Tanner et al5 focused on analyses of hospital working environment and demonstrated that working conditions in hospital settings have an influence on physicians' health and their job performance. The authors stated that to keep healthcare availability in Germany at the desired level, it is important to prioritise and organise responsibilities, resources and job demands to keep hospital physicians healthy in their job and to prevent diseases or even absenteeism. Particularly in healthcare jobs, work ability is a precondition to cope with demanding tasks in different work conditions (eg, high patient loads, working under pressure). Work ability is defined as ‘the sum of factors enabling an employed person in a certain situation to manage his/her working demands successfully’.29

Decreased work ability is associated with diminished productivity at work,30 increased risk of long-term sickness absence31 and early retirement.32 Particularly in healthcare jobs, work ability is a precondition to cope with demanding tasks in different work conditions (eg, high patient loads, working under pressure).

Physicians' mental health has been investigated in several studies. A current review demonstrated the prevalence of depression or depressive symptoms being 29% among physicians.33 A study by Ruitenburg et al34 showed that a quarter of the physicians in their study were found to have high depression (29%) and anxiety (24%) symptoms, and one out of every six (15%) physicians had a high prevalence of post traumatic stress disorder (PTSD) symptoms.

There is some indication to suggest differences in mental health and work ability between clinicians working in different medical specialties. Only few studies already exist focusing on physicians' general health and differences between medical specialties. For example, a study by Hughes et al35 demonstrated higher rates of substance use by emergency physicians and psychiatrists compared to other medical specialties. Buddeberg-Fischer et al36 showed that surgeons have been shown to have a higher risk of cardiovascular diseases compared to other medical specialists.

Despite these research studies, we found no study comparing mental health issues and work ability among German medical specialties.

Study aim

To our knowledge, no study has evaluated differences in working conditions and mental health issues as well as work ability among hospital physicians working in different medical specialties. Therefore, this study aimed to examine physicians' working conditions as well as levels of physicians' occupational distress, depressive symptoms and work ability. Findings are compared for different medical specialties. A better understanding in this research field will become central for designing strategies and health promotion programmes to support their mental health and maintain physicians' productivity.

Research questions and hypotheses

In summary of the outlined frame and the cited empirical results, we work on the following questions and hypotheses:

Question I: How do German hospital physicians evaluate their working conditions across medical specialties?

Hypothesis I: Working conditions differ significantly among hospital physicians working in different medical specialties.

Question IIa: How do German hospital physicians evaluate their levels of occupational distress, depressive symptoms and work ability?

Question IIb: Do these levels of occupational distress, depressive symptoms and work ability differ across medical specialties?

Hypothesis II: Levels of physicians' occupational distress, depressive symptoms and work ability differ significantly across medical specialties.

Question III: Are there relations between working conditions and occupational distress, depressive symptoms and work ability?

Hypothesis IIIa: Perceptions of workplace demands (eg, quantitative demands) will be positively related to occupational distress and depressive symptoms and negatively related to physicians' work ability.

Hypothesis IIIb: Perceptions of workplace resources (eg, social support) will be negatively related to occupational distress and depressive symptoms and positively related to physicians' work ability.

Methods

Study design and setting

The study was designed as a cross-sectional questionnaire evaluation. Data collection took place between 2011 and 2014 in 12 hospital departments in the northern and eastern parts of Germany. Based on information from the German Federal Office of Statistics in 2012, the chosen hospitals are representative of other German hospitals.37 The hospitals included in the study were similar in the following variables: size, number of patients/beds as well as employed medical staff. As an inclusion criterion, physicians had to have a minimum score of 1-year working experience. Focus of this study was to study on residents in medical training. Attending physicians and chief physicians were not included in the study.

We asked full-time working physicians to fill out the questionnaire within 3 weeks and to return them to boxes placed in doctors' rooms. Reminders were sent by email after 3 weeks to increase the response rate. Seven hundred and three physicians were invited to participate in the study. In total, 435 questionnaires were evaluated in the analyses—split into six different medical specialties (Internal Medicine, Neurology, Surgery, Paediatrics, Anaesthesiology and Gynaecology & Obstetrics) (the response rate was 61.8%).

Dependent variables and instruments

Dependent variables were (1) depressive symptoms and (2) work ability.

Depressive symptoms are measured with the ICD-10 Symptom Rating.38 Items can be answered on a 5-point rating scale (range: 0–4). Higher scores are indicative of greater symptom dominance. The scale for depressive symptoms showed a satisfactory internal consistency (Cronbach's α=0.73).

Physicians' work ability was measured using a single item of the Work Ability Index.34 Participating physicians were asked to rate their current work ability against their work ability in their own best period on an 11-point scale from 0 to 10. Ten stands for their best work ability in their own best period of life. The cut-off value for work ability was six: lower scores are considered as deficient work ability. This cut-off point was defined by Ruitenburg et al.34 39

Independent variables included physicians' working conditions and occupational distress

The German version of the Copenhagen Psychosocial Questionnaire (COPSOQ) was used to evaluate job-related and psychosocial factors at work.40 41 The questionnaire includes 12 subscales, which measured job demands (emotional and quantitative demands), job resources (quality of leadership, opportunities for development and social support) and job-related outcome factors (cognitive stress symptoms and burnout). Items are scored on a 5-point Likert scale, where the first category represents the maximal value (ie, ‘always’) and the last the minimal value (ie, ‘never’). Item values were transformed into point values ranging from 0 (minimum) to 100 points (maximum).

The COPSOQ is a well-validated and internationally recognised survey instrument. Previous investigations proved quality criteria of the COPSOQ (reliability, validity).41

Distress was measured with the Perceived Stress Questionnaire including seven scales with 30 items (scales: Harassment, Overload, Irritability, Lack of Joy, Fatigue, Worries and Tension).42 These items can be answered on a 4-point rating scale (1=almost never, 2=sometimes, 3=often and 4=usually) and refer to the period of the last 4 weeks. According to Levenstein et al,42 the resulting total score was linearly transformed between 0 and 1. The Cronbach's α value of the Perceived Stress Questionnaire is high (α=0.89).

Sociodemographic factors

Sociodemographic variables were included in the questionnaire: age, gender, years of working experience, marital status and presence of children.

Marital status was differentiated between physicians who are married or living in a partnership and others. In addition, the presence of children in the household was measured.

Statistical data analysis

Data analysis included descriptive analyses, parametric analyses and regression analyses. Data were adjusted for age, gender, years of working experience, marital status and presence of children.

One-way analyses of variance (ANOVAs) were conducted to compare medical specialties. Post hoc tests were used to identify significant group differences. In the end, all p values were corrected for multiple testing (Bonferroni corrections). p Values of <0.05 were considered significant, and all p values given were two tailed.43 In addition, Cohen's d effect size was analysed and described.

We checked for violations of the assumptions (eg, normality, equality of variances, data homoscedasticity and removal of the outliers).

Two regression analyses were conducted to analyse associations between working conditions and the dependent outcome variables: (1) depressive symptoms and (2) work ability.

In the first step, control variables were included, to determine whether, independently of these variables, there are associations between the working conditions, occupational distress, depressive symptoms and work ability. In the next step, physicians' working conditions were included.

Data were calculated using the SPSS (V.21.0) software package for social sciences.

Results

Descriptive statistics of the study sample

Physicians' sociodemographic characteristics are given in table 1. Fifty-one per cent of the respondents were female physicians (49% men). The ages ranged from 27 to 57 years, with the vast majority aged 26–35 years. Of the respondents, 63% were married or cohabiting and 46% had children. On average, physicians had been working 4 years (SD=2 years). Significantly more female physicians had children (p<0.05) and were married (p<0.05) compared to their male colleagues. The number of physicians per medical specialty included in this study is shown in table 2. No significant group differences have been found with regard to the sociodemographic variables.

Table 1

Demographic characteristics of the participants (n=435)

	N	Percent
Gender
Female	222	51
Male	213	49
Age
<25	0	0
26–30	109	25
31–35	131	30
36–40	83	19
41–45	52	12
46–50	43	10
>50	17	4
Marital status
Single	161	37
Married/partnership	274	63
Children
Child/children	200	46
Without children	235	54

Table 2

Mean values and analyses of variance for the working conditions by medical specialty

	Surgery (n=78)	Internal medicine (n=83)	Paediatrics (n=61)	Neurology (n=65)	Gynaecology (n=78)	Anaesthesiology (n=70)	F
Quantitative demands	77.21	74.62	65.28	69.39	70.18	68.47	4.31**
Emotional demands	61.15	65.39	69.38	66.72	68.62	55.82	3.13**
Influence at work	51.24	56.33	51.53	57.38	55.46	50.02	1.98
Possibilities for development	66.63	65.42	63.79	65.18	64.37	63.25	1.85
Degree of freedom at work	48.25	48.82	50.33	49.16	50.29	51.72	1.74
Sense of community	58.18	57.50	64.61	61.73	58.56	59.65	1.91
Feedback	34.35	43.17	45.63	37.85	44.62	39.29	3.58**
Quality of leadership	46.72	50.72	52.44	57.39	49.13	48.51	4.01**
Social support	59.48	62.85	65.73	57.92	59.98	55.36	3.91**
Social relationships	61.27	63.51	67.39	58.43	60.25	57.12	3.85**

N=435.

**p<0.01.

Comparison of physicians’ working conditions, occupational distress, depressive symptoms and work ability across medical specialties

Several differences were found between the examined job demands and job resources (see table 2). As shown, quantitative job demands were rated highest by physicians in Surgery and Internal Medicine compared to the other disciplines (p<0.01). Emotional demands were rated highest by paediatricians, gynaecologists and neurologists (p<0.01).

Significant differences in job resources have also been found among medical specialties: social support and social relationships have been scored highest in Paediatrics. Getting feedback and quality of leadership were scored with the lowest scores in Surgery Medicine compared to their colleagues (see table 2).

An overview of the prevalence rates of the examined outcome variables (psychosocial distress, depressive symptoms and work ability) are presented in table 3. Mean score for psychosocial distress was 0.48 (SD=0.18). Cut-off score for moderate level of distress was estimated to be >0.40 to≤0.60. The prevalence of perceived psychosocial distress at a moderate level was 39.5%. The prevalence of perceived psychosocial distress at high levels was 17.1%.

Table 3

Perceived occupational distress, depressive symptoms and work ability among hospital doctors

		Occupationaldistress M (SD)	Depressivesymptoms M (SD)	Work abilityM (SD)
Medical specialty	n (%)	(range: 0–1)	(range: 0–1)	(range: 1–10)
Internal medicine	83 (19%)	0.42 (0.16)	0.43 (0.23)	7.8 (1.2)
Surgery	78 (18%)	0.56 (0.19)	0.29 (0.19)	8.9 (1.5)
Paediatrics	61 (14%)	0.49 (0.17)	0.31 (0.18)	8.2 (1.3)
Neurology	65 (15%)	0.52 (0.15)	0.45 (0.20)	7.5 (1.6)
Gynaecology	78 (18%)	0.41 (0.18)	0.35 (0.18)	7.9 (1.4)
Anaesthesiology	70 (16%)	0.45 (0.17)	0.49 (0.20)	7.4 (1.3)
Sum score	435 (100%)	0.48 (0.18)	0.43 (0.19)	7.9 (1.4)

M, mean; SD, standard deviation.

The mean score for depressive symptoms was 0.43 (SD=0.19). Nine per cent of physicians indicated high levels of depressive symptoms. Physicians had a mean work ability score of 7.9 (SD=1.4) although 11% of physicians rated their work ability as insufficient (n=48).

Significant differences among medical specialties were found for occupational distress symptoms (F=11.65, p<0.001, η2=0.035) and depressive symptoms (F=8.57, p<0.001, η2=0.026). Bonferroni post hoc tests showed that physicians in Surgery Medicine reported the highest levels of occupational distress compared with internal medicine (p=0.003, d=0.30), obstetrics and gynaecology (p=0.001, d=0.43), paediatrics (p=0.001, d=0.54), neurology (p=0.001, d=0.42) and anaesthesiology (p=0.001, d=0.56). According to Bonferroni post hoc tests, surgical physicians reported the lowest scores of depressive symptoms. Physicians of other medical specialties scored higher on depressive symptoms: internal medicine (p=0.003, d=0.39), obstetrics and gynaecology (p=0.003, d=0.35), paediatrics (p<0.001, d=0.63), neurology (p<0.001, d=0.48) or anaesthesiology (p<0.001, d=0.64).

An ANOVA revealed an effect of medical specialty on work ability (F=11.65, p<0.001, η2=0.032). Post hoc Bonferroni tests showed that surgeons reported higher work ability scores than physicians in internal medicine (p=0.001, d=0.39), paediatrics (p<0.001, d=0.61) (p<0.001, d=0.65) and obstetrics and gynaecology (p<0.001, d=0.41). Surgeons scored significantly higher than anaesthesiologists (p=0.001, d=0.42) and neurologists (p<0.001, d=0.38).

Relations between working conditions, depressive symptoms and work ability

Bivariate analyses revealed several significant correlations between working conditions and depressive symptoms and work ability (see table 4). We performed regression analyses in which depressive symptoms and work ability were set as the dependent variables. For both analyses, the variables age, work experience, cohabitation with a partner, professional position and medical specialty were statistically controlled. The final model of the regression analyses is illustrated in table 4.

Table 4

Regression analyses for depressive symptoms and work ability

	DV: depressive symptoms			DV: work ability
Independent variables	r	β	ΔR2	r	β	ΔR2
Step 1: Control variables			0.05			0.06
Age	0.15	0.08		−0.19**	−0.12*
Gender	0.10**	0.07		0.05	0.03
Marital status	0.11*	0.09		0.04	0.02
Presence of children	0.09	0.06		0.06	0.04
Medical specialty	0.04	0.02		0.07	0.05
Years of experience	0.11	0.07		−0.14*	−0.10*
Step 2: Working conditions			0.21			0.16
Quantitative demands	0.26**	0.20**		−0.28**	−0.22**
Emotional demands	0.25**	0.20**		−0.20**	−0.17**
Influence at work	−0.21**	−0.18**		0.15*	0.07
Possibilities for development	−0.17**	−0.14**		0.14*	0.06
Degree of freedom at work	−0.15**	−0.09		0.10*	0.08
Sense of community	−0.18*	−0.15**		0.09	0.05
Feedback	−0.11*	−0.08		0.12*	0.08
Quality of leadership	−0.13*	−0.11*		0.09	0.06
Social support	−0.18**	−0.15**		0.15**	0.12*
Social relationships	−0.12*	−0.10*		0.11**	0.06
Step 3			0.09			0.06
Occupational distress	0.31**	0.25**		−0.19**	−0.15**
Total R2			0.35			0.28

β, standardised β-coefficients from the final step of the model; ΔR2, change in explanation rate in each step; DV, dependent variable; R2, explanation rate.

*p<0.05; **p<0.01.

After adjustment for the sociodemographic variables, the job demands and resources were added in step 2, before occupational distress was entered into step 3 of the model.

The included sociodemographic and the work-related factors explained 26% of the variance for depressive symptoms and 22% of the variance for work ability.

As illustrated in table 4, occupational distress was positively associated with depressive symptoms but negatively with work ability. Occupational distress accounted for an additional 9.3% of the variance for depressive symptoms and an additional 6.1% of the variance for work ability (p<0.05).

Discussion

The aim of the current study was to investigate on possible differences in the working environment of resident hospital physicians as well as levels of physicians' occupational distress, depressive symptoms and work ability. This study demonstrated significant differences in working conditions, occupational distress, depressive symptoms and work ability among resident physicians. In addition, relations between physicians' working conditions and their perceived distress, depression and work ability have been analysed and illustrated.

In sum, all residents showed high rates for quantitative working demands. Comparable findings have been reported in former study reports focusing on residents in various medical specialties.12 44 45 Facing differences in medical specialties, our results show that physicians working in Surgery and Internal Medicine rated quantitative job demands more highly than their colleagues in other medical specialties. Ruitenburg et al46 found a comparable result: in this study, job demands of residents and specialists in Surgery were quantified for an average workday and compared with other hospital physicians, showing that surgeons perform significantly more job tasks than their colleagues working in other disciplines.

The highest levels for emotional demands have been rated by paediatricians. This result can be explained by the fact that paediatricians have to face children's illness that is very strainful and emotionally drained. Weigl et al47 also demonstrated that emotional work stressors are prevalent among residents working in Paediatric Medicine. Especially paediatricians have to develop adequate emotional strategies to deal with daily work demands and young patients and their families.

In addition, emotional demands have also been evaluated with high scores by physicians working in Neurology, Internal medicine and Gynaecology. Working in these specialties involves more communication, emotional support and direct patient care (ie, diagnosis talks) than, for example, in Anaesthesiology. A study performed by Tanner et al5 showed comparable findings for emotional stressors in these disciplines.

All physicians reported comparable low levels of receiving supportive feedback and quality of leadership (no significant differences). Former studies also reported low reward, work support and lack of feedback in the medical profession.48 As studies demonstrated, most employees suffer from a lack of feedback.49 However, getting feedback is essential for effective change and to develop and improve performance.50 So, as our study results illustrate, giving feedback and support should be improved by supervisors and colleagues in the clinical setting.

Our results also demonstrated moderate values for depressive symptoms, perceived occupational distress and work ability among the included physicians. The findings for depressive symptoms and distress are in line with former study results.44 51–53 Ruitenburg et al34 performed a study showing a comparable mean score for self-reported work ability of 8.1 but higher levels of mental health problems. They concluded that these findings should be a major concern and it is very important keeping physicians healthy at the job; so more attention should be given to the mental health of hospital physicians by the hospital management. As seen in previous studies, reduced self-reported work ability and mental health problems have been strongly associated with an increased risk of long-term sickness absence,54 workplace failure and reduced job performance.55–58

The results also showed significant differences in occupational distress, depressive symptoms, work ability and mental health between medical specialties.

We found that physicians working in Surgery Medicine rated occupational distress more highly than do physicians in other specialties. No comparable study has been found for German residents focusing on mental health differences between medical disciplines. However, these results support the findings of studies performed by Park et al59 showing comparable differences in burnout and exhaustion between residents with various specialties.

Surgeons reported the highest levels of quantitative working demands and occupational distress but reported the lowest levels of depressive symptoms and the highest levels of work ability. No comparable study has been found that focused on the identical topic. However, several studies demonstrated similar levels of job demands and distress among surgeons.44

Differences in the working environment that contribute to differential consequences in terms of perceived distress and mental health perceptions are argued in research findings as possible reasons.60

As demonstrated in this study, how physicians' evaluate their working conditions proved to have an impact on how they rate their work ability and depressive symptoms. In particular, our results showed that high levels of job demands are associated with high levels of depressive symptoms. This result is consistent with a former study showing relations between high job demands, low work control and heavy on call-duty with poor mental health of hospital physicians.61

Anagnostopoulos et al62 also demonstrated associations between job demands (eg, workload, emotional demands), job resources (eg, supervisor support, job autonomy) and residents' mental health outcomes. The authors also conclude that residency is a stressful period in a physician's development, characterised by long work hours, time pressure and excessive work load, which can exert negative effects on residents' mental health.62

In addition, several studies performed in other occupational settings confirmed relations between working conditions and mental health problems.63 64

In addition, relations between perceived influence at work, collegial support and depressive symptoms as well as work ability have been analysed. According to previous research, influence at work and autonomy at work are important factors for productivity, general job performance and well-being at work.65 66 As research demonstrated, residents perceive significantly less influence and control over their own work than specialists. Residents often report that they actually have little autonomous power in the clinical setting.67 Lack of power, few influence and control over work are associated with feelings of frustration, anger and helplessness. As possible long-term consequences, depressive symptoms and further mental issues can arise.68 In addition, residents often complain about insufficient support from supervisors and experienced colleagues. As research stated, perceived support is another important factor on job performance, work ability and mental health in the long term.69 These study results indicate that improving job autonomy and establishing a support system may be useful to increase job performance and job satisfaction. Appropriate suggestions on how to improve these working resources are listed in the Practical implication section.

The study result further supports the suggestion that perceived distress is a negative psychological factor for depressive symptoms. Consistent with this finding, occupational stress has been identified in several studies as a risk factor for depressive symptoms as it may influence general mental health and work ability.70–72

Limitations

Some limitations of this study should be addressed. First, the study's cross-sectional design limits the value in supporting causal effects and generalisability of the findings. Since data were assessed by self-report measures and no observers' data are available to cross-validate the data of participants, we can only claim limited objectivity of our results. The results might be influenced by common method bias. In addition, we did not include and ask for stressors perceived outside work that may have an influence on the study results. Furthermore, due to logistic and pragmatic reasons, our study sample includes only physicians working in clinic departments of the northern and eastern parts of Germany, which may also limit the generalisability of the study findings. Since we included only residents, age and clinical experience should be acknowledged as limiting factors for the external validity of the study results. In addition, the participating physicians worked in different hospitals may cause potential bias. A selection bias cannot be excluded.

Practical implications

Practical implications can be drawn to increase hospital physicians' mental health and work ability and subsequently the quality of treatment and patient satisfaction.

At the individual level, more health prevention programmes should be implemented in German hospitals. Hospitals should implement individual and/or group trainings and programmes including supervisory support and counselling for physicians. Such programmes may include elements to improve physicians' self-care and to develop coping strategies.

Moreover, a specific mental health support programme could be used to monitor physicians' mental health conditions and to provide interventions to avoid productivity loss and to increase quality of care.13 In addition, participative interventions on psychosocial work factors to prevent mental health problems in hospital settings should be implemented.

We also suggest organisational and work-design interventions that address the high working demands and the low levels of reward, feedback and quality of leadership.

At the organisational level, it would be useful to improve organisational procedures such as staff planning and/or by delegation of documentation, administrative work and duties to other professions. In addition, trainings for attending physicians and chief physicians may improve perceived quality of leadership and may create a culture of effective feedback.

In general, hospital organisation should create a climate for constant learning and exchange of knowledge. There are a number of individual and/or setting-based strategies and interventions to support residents in this way: for example, mentoring programmes and individual training. Three forms of support measures could be used: (1) emotional (verbal and nonverbal communication of support, (2) informational (provision of information used to guide or advice) and (3) instrumental (provision of material goods, physical assistance).73 Moreover, the improvement of job autonomy during residency could be established via implementing structured work schedules during medical training for residents in order to achieve anticipated performance and training goals. In addition, job task rotation, job enlargement and job enrichment could improve perceived job autonomy and individual development. All three forms give employees the opportunity to perform various tasks and to develop skills in more than one area of expertise.74 75

Conclusion

In this study, we presented differences in working conditions, occupational distress, depressive symptoms and work ability among physicians working in different medical specialties. In addition, the meaning of working conditions for aspects of work ability and mental health was described. Improving working conditions and mental health promotion, particularly reducing stressful working demands and organisation of work schedules, may have beneficial effects on physicians' mental health, job performance and quality of care.

Longitudinal studies are necessary to confirm our results. In addition, more empirical studies including other work, social and individual factors are needed to reach a better understanding of work-related mental health problems.