Social relationships in physicians’ work moderate relationship between workload and wellbeing—9-year follow-up study

Social relationships in physicians’ work moderate relationship between workload and... Abstract Background Increasing wellbeing problems among physicians may lead to serious consequences in health care and means to prevent such development are called for. This study examined longitudinal associations between workload and changes in distress, sleep quality and workability in physicians and whether positive social relations at work would protect from such problems. Methods A baseline survey was conducted in 2006 for a random sample of 5000 physicians (n = 2841, response rate 57%). In 2015, the follow-up survey was sent to those 2 206 physicians who gave their consent (n = 1462, response rate 68.3%). The survey included scales for distress, sleeping problems, workability, workload, team climate, collegial support and questions for background information. Results Increased workload was associated with increased psychological distress, sleeping problems and decreased workability during the 9-year follow-up. Good team climate and collegial support were related to decreased distress and sleep quality and enhanced workability. Good collegial support buffered the associations of workload changes on distress and sleep quality changes. Team climate was more strongly associated with changes in sleep quality and workability among younger and middle aged physicians than older physicians. Also collegial support had a stronger association with sleep quality change among younger or middle aged physicians than older physicians. These associations were robust to adjustments for age, gender, specialization, leadership position, marital status and baseline wellbeing. Conclusions Health care organizations should take measures to decrease workload and to increase availability of social support for physicians in order to protect physicians from declining wellbeing. Introduction Poor work satisfaction and decreasing wellbeing among physicians is a recognized problem. Such development may lead to serious consequences in health care, such as lower quality of care or physician turnover. In UK, 30–50% of general practitioners reported high levels of psychological distress.1 Shanafelt et al.2 have reported that compared with general population physicians are more likely to have symptoms of burnout and be dissatisfied with work-life balance. Physician burnout and distress have been associated with decreased professional effort3 and lower productiveness4 as well as lower patient satisfaction.5 Poor wellbeing among physicians may even endanger patient safety due to medical errors.6 Physician’s work comprises various stressors such as clinical workload,7 adverse patient events8 and on-call work.9 In the Finnish Health Care Professionals Study, patient-related stress, poorly functioning patient record systems, and job demands in terms of time-pressure burdened physicians in hospitals and health centres.10 Time pressure, patient-related stress, and work interference with family were particularly associated with working in public primary care. Job strain has in general been associated with poor health consequences11,12 and poor psychological wellbeing,13 absenteeism14 and early retirement.15 A large body of evidence shows that high levels of social support and social connectedness are related to better health16–18 Social support may have a direct positive effect on health and wellbeing because it provides positive affect and stability and because it may help to avoid negative experiences. Social support may also buffer against negative effects of adverse events or circumstances. Social support may change the appraisal of the stressful situation or activate adaptive responses in the situation.19 In occupational health research social support has been incorporated as a coping resource into the extended job strain model.20 However, in addition to serving as adaptive resources, social relationships may also be a source of stress. Dense social networks or close relations may potentially cause social conflicts and therefore may have negative consequences for wellbeing.18 The importance of social relationships for health may also vary during the life-course. Hakulinen et al.18 have shown that association between negative aspects of social relationships, such as interpersonal conflicts and poor health strengthened by increasing age while association between strong social support and good mental health among women weakened by age. Social relationships have been found to be an important psychosocial resource in physicians work as well. Several studies have indicated that physicians who lack social support are at greater risk of impaired wellbeing or burnout.8,21–23 Another aspect of social relationships in medical work is working in teams. Team climate has been associated with better quality of care in health care setting24 and problems in team climate have been associated with intentions to leave among hospital staff25 and sickness absenteeism among physicians.26 This study focuses on the role of social relationships in the associations between physicians’ workload and wellbeing in a longitudinal design. More specifically, we examined whether perceived collegial support and team climate in the work unit are able to buffer the negative effects of increased workload on psychological distress, sleeping problems and workability among physicians in 9-year follow-up setting. In addition, we examine whether age buffers the importance of social support on wellbeing. Methods Study sample The present study is a part of the on-going Finnish Health Care Professionals Study, in which we drew a random sample of 5000 physicians in Finland (30% of the whole physician population) from the 2006 database of physicians maintained by the Finnish Medical Association. The register covers all licenced physicians in Finland. Phase 1 data were gathered with postal questionnaires in 2006. Non-respondents were sent a reminder and a copy of the questionnaire up to two times. Responses were received from 2841 physicians (response rate 57%). The sample is representative of the eligible population in terms of age, gender, and employment sector.27 Nine years later, in 2015 the follow-up questionnaire was sent to those who gave their consent in 2006 to participate to the follow-up survey (n = 2206, 77.6% of the respondents in the 2006 survey). Those who had died during the follow-up or who had an unknown address in 2015 were excluded leaving 2159 physicians to the follow-up sample in 2015. Of these 1462 physicians responded (response rate 68.3% among those consented to follow-up (29% of the original sample in 2006), 59% women, mean age 54.8, SD = 9.74). When compared with all physicians in Finland the 9-year follow-up respondents were more often aged 55 years or older (34% of all Finnish physicians and 41% in the follow-up sample) and specialized (62% of all Finnish physicians and 85% in the follow-up sample), but there were no differences in terms of gender. In addition the follow-up attrition was analysed by comparing the follow-up respondents to those who responded in the baseline but not in the 9-year follow-up in terms of baseline survey information. There were no differences according to gender and baseline work load between the follow-up respondents and non-responders. However, those who responded in the follow-up were in the baseline older, less often under specialization process and more often in superior position. Measurements Wellbeing indicators ‘Psychological distress’ was measured with the four items from the GHQ-12 (Goldberg 1972) representing anxiety/depression dimension of GHQ-12 from the Graetz28,29 three-factor structure model for GHQ-12. The response options ranged from 1 to 4. In this study, the scale was used as a continuous variable. Cronbach’s alpha coefficients for this sample were α = 0.83 in 2006 and 0.84 in 2015. ‘Sleeping problems’ were measured with four-item Jenkins scale30 (α = 0.77 in 2006 and 0.78 in 2015). Respondents were asked how often during the last four weeks they had troubles falling asleep, were waking up several times per night, had troubles staying asleep including waking up too early, and felt tired after usual amount of sleep. The response scale ranged from 1 (never) to 6 (every night). ‘Workability’ was assessed with a single item question included in the Workability Index31: ‘Assume that your workability at its best has a value of 10 and 0 would mean that you could not work at all. How many points would you give to your current workability (range 0-10)?’ Workload ‘Workload’ was measured with three items derived from the Harris Nurse Stress index, which have been developed based on previous research among nurses and health care staff and which have shown adequate psychometric properties in terms of internal consistency and construct validity.32 This scale measures stress due to time shortages at work and scheduling problems. An example item: ‘How often have you been distracted, worried or stressed about (during the past half-year period) not being able to do your work properly.’ The items were rated on a 5-point Likert-scale ranging from 1 (never) to 5 (very often) (α = 0.83 in 2006 and 0.87 in 2015). ‘Social relations’ variables were measured in 2015. ‘Team climate’ was measured with a Team Climate Inventory’s33 participative safety subscale that measures team participation, such as interaction frequency and information sharing (four items, α = 0.88; e.g ‘People feel understood and accepted by each other’). The items were rated on a 5-point Likert-scale, ranging from 1 (I totally disagree) to 5 (I totally agree). ‘Collegial support’ was measured with three items measuring measuring tangible social support at work in terms of consultation possibilities and co-operation in the team (α = 0.67). Example of the items includes: ‘How often have you been distracted, worried or stressed about (during the past half-year period) lack of consultation possibilities’ (reverse coded). The items were rated on a five-point Likert-scale ranging from 1 (never) to 5 (very often). Background variables The background variables used were gender, specialization status, leadership position, marital status, having children and baseline level of outcome variables. The information about specialization status and leadership position were based on self-report. The respondents were asked directly to indicate whether they were (i) not specialized, (ii) under specialization, (iii) specialized and whether they were in supervisory position (yes/no). Statistical analysis We constructed change scores for wellbeing indicators and workload (by subtracting the 2006 scores from the 2015 scores). Workload change score was constructed by subtracting the 2006 scores from the 2015 scores and further categorizing the change score into three groups (i) workload levels had decreased (negative change score (ii) stayed the same (change score = 0) and (iii) increased (positive change score). For outcome variables continuous change scores were used, higher scores indicating that distress, sleeping problems and workability had increased during this time period. The associations of workload change and social relations variables on distress, sleeping problems and workability change were examined using the analyses of covariance (in separate analyses). The analyses of main effects were conducted in two steps. First, the analyses were adjusted for gender and age. Second, specialization status, leadership position, marital status, having children and baseline level of outcome variable were additionally adjusted for. In analyses of work load change baseline work load at 2006 was additionally adjusted for. The interactions of social relations variables (team climate and collegial support; in separate analyses) with workload change and age were examined with the analyses of covariance adjusted for main effects, age, gender, specialization status, leadership position, marital status, having children and baseline level of outcome variable and in analyses of interaction between work load change and social support variables the baseline workload was additionally adjusted for. In the original analyses, age was used as continuous variable, but in order to illustrate the interactions in figures age was dichotomized (younger vs. older) using median as the cut-off point (≤55 years/>55 years). Similarly, social support variables were used as continuous variables in the analyses, but they were dichotomized to illustrate the interaction effects in the figures. Values 1 and 2 indicated poor team climate and values above 2 indicated good team climate. In the collegial support variable scores 1–4 were combined to indicate low collegial support and scores 4–5 indicated high collegial support. Results Main effects Table 1 shows the sample characteristics. Majority of respondents were women and had specialized. The mean age in the follow-up was 55 years. Table 2 presents the results of covariance analyses regarding changes of distress, sleeping problems and workability. Adjusting for age and gender, workload change was associated with changes in distress, sleeping problems and workability. Distress and sleeping problems had increased and workability decreased most among those whose workload levels had increased during the follow up. Higher age was associated with decreased distress and sleeping problems and decreased workability. Poor team climate and low collegial support were associated with increased distress and sleeping problems. Further adjustments for baseline distress or sleeping problems, specialization, leadership position and marital status did not attenuate the associations of team climate and collegial support with distress change and sleeping problems. Team climate and collegial support were not associated with changes in workability when adjusted only for age and gender; however, after adjusting also for baseline workability, specialization, leadership position and marital status better team climate and stronger collegial support were associated with increased workability. Table 1 Sample characteristics     n  %  Gender  Men  585  40  Women  856  59  Marital status  Single  81  6  Married/co-habiting  1191  83  Divorced  138  10  Widowed  34  2  Specialization  No  161  11  In training  53  4  Specialized  1232  85  Workload change 2006–2015  Increased  693  26  Same  197  14  Decreased  385  47      Mean  SD  Age 2015    54.8  9.74  Distress 2006a    1.89  0.60  Distress 2015a    1.77  0.65  Sleeping problems 2006b  2.32  1.01  Sleeping problems 2015b  2.44  1.02  Work ability 2006c    8.58  1.32  Work ability 2015c    8.20  1.54  Team climate 2015d  3.91  0.73  Collegial support 2015d  3.91  0.78      n  %  Gender  Men  585  40  Women  856  59  Marital status  Single  81  6  Married/co-habiting  1191  83  Divorced  138  10  Widowed  34  2  Specialization  No  161  11  In training  53  4  Specialized  1232  85  Workload change 2006–2015  Increased  693  26  Same  197  14  Decreased  385  47      Mean  SD  Age 2015    54.8  9.74  Distress 2006a    1.89  0.60  Distress 2015a    1.77  0.65  Sleeping problems 2006b  2.32  1.01  Sleeping problems 2015b  2.44  1.02  Work ability 2006c    8.58  1.32  Work ability 2015c    8.20  1.54  Team climate 2015d  3.91  0.73  Collegial support 2015d  3.91  0.78  a Scale 1–4, higher score indicates higher distress. b Scale 1–6, higher score indicates more sleeping problems. c Scale 0–10, higher score indicates better work ability. d Scale 1–5, higher score indicates better support or team climate. Table 1 Sample characteristics     n  %  Gender  Men  585  40  Women  856  59  Marital status  Single  81  6  Married/co-habiting  1191  83  Divorced  138  10  Widowed  34  2  Specialization  No  161  11  In training  53  4  Specialized  1232  85  Workload change 2006–2015  Increased  693  26  Same  197  14  Decreased  385  47      Mean  SD  Age 2015    54.8  9.74  Distress 2006a    1.89  0.60  Distress 2015a    1.77  0.65  Sleeping problems 2006b  2.32  1.01  Sleeping problems 2015b  2.44  1.02  Work ability 2006c    8.58  1.32  Work ability 2015c    8.20  1.54  Team climate 2015d  3.91  0.73  Collegial support 2015d  3.91  0.78      n  %  Gender  Men  585  40  Women  856  59  Marital status  Single  81  6  Married/co-habiting  1191  83  Divorced  138  10  Widowed  34  2  Specialization  No  161  11  In training  53  4  Specialized  1232  85  Workload change 2006–2015  Increased  693  26  Same  197  14  Decreased  385  47      Mean  SD  Age 2015    54.8  9.74  Distress 2006a    1.89  0.60  Distress 2015a    1.77  0.65  Sleeping problems 2006b  2.32  1.01  Sleeping problems 2015b  2.44  1.02  Work ability 2006c    8.58  1.32  Work ability 2015c    8.20  1.54  Team climate 2015d  3.91  0.73  Collegial support 2015d  3.91  0.78  a Scale 1–4, higher score indicates higher distress. b Scale 1–6, higher score indicates more sleeping problems. c Scale 0–10, higher score indicates better work ability. d Scale 1–5, higher score indicates better support or team climate. Table 2 Estimated marginal means and results of the analyses of covariance for the 9-year changes of distress (GHQ) and sleeping problems   Step 1a   Step 2b     Meanc (95% CI)  F  P  Meanc (95% CI)  F  P  Distress change                  Workload    37.13  <0.001    51.20  <0.001          Decreased  −0.23 (−0.28, −0.18)      −0.22 (−0.32, −0.12)              Stayed the same  0.02 (−0.07, 0.12)      0.01 (−0.11, 0.13)              Increased  0.14 (0.07, 0.21)      0.22 (0.11, 0.33)      Gender    0.22  0.639    9.91  0.002  Age    21.17  <0.001    10.80  <0.001  Team climate    31.85  <0.001    67.84  <0.001  Collegial support    24.50  <0.001    88.02  <0.001  Sleep quality change                  Workload    13.56  <0.001    19.27  <0.001          Decreased  0.06 (−0.01, 0.13)      −0.02 (−0.16, 0.13)              Stayed the same  0.29 (0.15, 0.43)      0.22 (0.04, 0.39)              Increased  0.36 (0.26, 0.46)      0.38 (0.21, 0.54)      Gender    1.87  0.172    1.22  0.269  Age    41.45  <0.001    10.11  0.001  Team climate    12.65  <0.001    26.82  <0.001  Collegial support    16.91  <0.001    42.83  <0.001  Work ability change      Time pressure    10.09  <0.001    18.86  <0.001          Decreased  −0.13 (−0.24, −0.02)      0.01 (−0.21, 0.24)              Stayed the same  −0.27 (−0.48, −0.06)      −0.11 (−0.36, 0.15)              Increased  −0.56 (−0.71, −0.41)      −0.57 (−0.81, −0.33)      Gender    0.14  0.713    0.72  0.395  Age    5.75  0.017    12.92  <0. 001  Team climate    0.51  0.476    9.74  <0.001  Collegial support    2.17  0.141    18.24  <0.001    Step 1a   Step 2b     Meanc (95% CI)  F  P  Meanc (95% CI)  F  P  Distress change                  Workload    37.13  <0.001    51.20  <0.001          Decreased  −0.23 (−0.28, −0.18)      −0.22 (−0.32, −0.12)              Stayed the same  0.02 (−0.07, 0.12)      0.01 (−0.11, 0.13)              Increased  0.14 (0.07, 0.21)      0.22 (0.11, 0.33)      Gender    0.22  0.639    9.91  0.002  Age    21.17  <0.001    10.80  <0.001  Team climate    31.85  <0.001    67.84  <0.001  Collegial support    24.50  <0.001    88.02  <0.001  Sleep quality change                  Workload    13.56  <0.001    19.27  <0.001          Decreased  0.06 (−0.01, 0.13)      −0.02 (−0.16, 0.13)              Stayed the same  0.29 (0.15, 0.43)      0.22 (0.04, 0.39)              Increased  0.36 (0.26, 0.46)      0.38 (0.21, 0.54)      Gender    1.87  0.172    1.22  0.269  Age    41.45  <0.001    10.11  0.001  Team climate    12.65  <0.001    26.82  <0.001  Collegial support    16.91  <0.001    42.83  <0.001  Work ability change      Time pressure    10.09  <0.001    18.86  <0.001          Decreased  −0.13 (−0.24, −0.02)      0.01 (−0.21, 0.24)              Stayed the same  −0.27 (−0.48, −0.06)      −0.11 (−0.36, 0.15)              Increased  −0.56 (−0.71, −0.41)      −0.57 (−0.81, −0.33)      Gender    0.14  0.713    0.72  0.395  Age    5.75  0.017    12.92  <0. 001  Team climate    0.51  0.476    9.74  <0.001  Collegial support    2.17  0.141    18.24  <0.001  a Adjusted for age and gender. b Adjusted for age, gender, having children, baseline level of outcome, specialization status, leadership position, and marital status. In the analyses between work load and outcomes the baseline work load (in 2006) was additionally adjusted for. c Estimated marginal means. Table 2 Estimated marginal means and results of the analyses of covariance for the 9-year changes of distress (GHQ) and sleeping problems   Step 1a   Step 2b     Meanc (95% CI)  F  P  Meanc (95% CI)  F  P  Distress change                  Workload    37.13  <0.001    51.20  <0.001          Decreased  −0.23 (−0.28, −0.18)      −0.22 (−0.32, −0.12)              Stayed the same  0.02 (−0.07, 0.12)      0.01 (−0.11, 0.13)              Increased  0.14 (0.07, 0.21)      0.22 (0.11, 0.33)      Gender    0.22  0.639    9.91  0.002  Age    21.17  <0.001    10.80  <0.001  Team climate    31.85  <0.001    67.84  <0.001  Collegial support    24.50  <0.001    88.02  <0.001  Sleep quality change                  Workload    13.56  <0.001    19.27  <0.001          Decreased  0.06 (−0.01, 0.13)      −0.02 (−0.16, 0.13)              Stayed the same  0.29 (0.15, 0.43)      0.22 (0.04, 0.39)              Increased  0.36 (0.26, 0.46)      0.38 (0.21, 0.54)      Gender    1.87  0.172    1.22  0.269  Age    41.45  <0.001    10.11  0.001  Team climate    12.65  <0.001    26.82  <0.001  Collegial support    16.91  <0.001    42.83  <0.001  Work ability change      Time pressure    10.09  <0.001    18.86  <0.001          Decreased  −0.13 (−0.24, −0.02)      0.01 (−0.21, 0.24)              Stayed the same  −0.27 (−0.48, −0.06)      −0.11 (−0.36, 0.15)              Increased  −0.56 (−0.71, −0.41)      −0.57 (−0.81, −0.33)      Gender    0.14  0.713    0.72  0.395  Age    5.75  0.017    12.92  <0. 001  Team climate    0.51  0.476    9.74  <0.001  Collegial support    2.17  0.141    18.24  <0.001    Step 1a   Step 2b     Meanc (95% CI)  F  P  Meanc (95% CI)  F  P  Distress change                  Workload    37.13  <0.001    51.20  <0.001          Decreased  −0.23 (−0.28, −0.18)      −0.22 (−0.32, −0.12)              Stayed the same  0.02 (−0.07, 0.12)      0.01 (−0.11, 0.13)              Increased  0.14 (0.07, 0.21)      0.22 (0.11, 0.33)      Gender    0.22  0.639    9.91  0.002  Age    21.17  <0.001    10.80  <0.001  Team climate    31.85  <0.001    67.84  <0.001  Collegial support    24.50  <0.001    88.02  <0.001  Sleep quality change                  Workload    13.56  <0.001    19.27  <0.001          Decreased  0.06 (−0.01, 0.13)      −0.02 (−0.16, 0.13)              Stayed the same  0.29 (0.15, 0.43)      0.22 (0.04, 0.39)              Increased  0.36 (0.26, 0.46)      0.38 (0.21, 0.54)      Gender    1.87  0.172    1.22  0.269  Age    41.45  <0.001    10.11  0.001  Team climate    12.65  <0.001    26.82  <0.001  Collegial support    16.91  <0.001    42.83  <0.001  Work ability change      Time pressure    10.09  <0.001    18.86  <0.001          Decreased  −0.13 (−0.24, −0.02)      0.01 (−0.21, 0.24)              Stayed the same  −0.27 (−0.48, −0.06)      −0.11 (−0.36, 0.15)              Increased  −0.56 (−0.71, −0.41)      −0.57 (−0.81, −0.33)      Gender    0.14  0.713    0.72  0.395  Age    5.75  0.017    12.92  <0. 001  Team climate    0.51  0.476    9.74  <0.001  Collegial support    2.17  0.141    18.24  <0.001  a Adjusted for age and gender. b Adjusted for age, gender, having children, baseline level of outcome, specialization status, leadership position, and marital status. In the analyses between work load and outcomes the baseline work load (in 2006) was additionally adjusted for. c Estimated marginal means. Interactions The interaction terms between age and team climate (F = 4.62, P = 0.032) and age and collegial support (F = 4.54, P = 0.033) were significant for sleeping problems change. Collegial support and team climate were more strongly associated with the change in sleeping problems among young and middle aged respondents than among older respondents (figure 1c and d). Interaction between age and team climate was significant also for workability change (F = 7.47, P = 0.006). Again, team climate was more strongly associated with workability change among young and middle respondents compared with older respondents (figure 1e). There were no significant interaction effect of age and social support variables (team climate and collegial support) for distress change (figure 1a and b) figures or between age and collegial support for workability change (figure 1f). Figure 1 View largeDownload slide Summary of the moderating effect of age on the associations of social relationships (team climate and collegial support) with well-being factors Figure 1 View largeDownload slide Summary of the moderating effect of age on the associations of social relationships (team climate and collegial support) with well-being factors The interaction of workload change with team climate (F = 3.37, P = 0.035) was significant for distress change. Team climate was associated with distress change more clearly among those who reported decreased or similar level of workload during the follow-up compared with those reporting increased workload (figure 2a). Interaction between workload change and collegial support was significant for distress change (F = 5.00, P = 0.007) and change in sleeping problems (F = 3.76, P = 0.024). High collegial support was more strongly associated with changes in distress and sleeping problems among those who reported similar or increased of workload compared with those reporting decreased workload (figure 2b and c). There was no significant interaction effect between work load change team climate for sleeping problems (figure 2c). Furthermore, work load change and social support variables (team climate and collegial support) did not have interaction effect for change in workability (figure 2e and 2f). Figure 2 View largeDownload slide Summary of the moderating effects of social relationships (team climate and collegial support) on the associations of workload change with well-being factors Figure 2 View largeDownload slide Summary of the moderating effects of social relationships (team climate and collegial support) on the associations of workload change with well-being factors Discussion The results of the current 9-year follow-up suggest that increased workload is associated with increased psychological distress, sleeping problems and decreased workability. Instead, good team climate and collegial support seem to be related to decline in distress and increase of sleep quality and workability. Good team climate may also buffer the negative effects of workload increase for distress and sleep quality. The role of team climate and collegial support for physicians seems to vary according to age, since team climate was more strongly associated with changes in sleep quality and workability among younger and middle aged physicians compared with older ones. Furthermore, also collegial support was more clearly associated with sleep quality changes among younger or middle aged physicians than older, more experienced physicians. Our finding that increased workload was associated with increased distress and sleeping problems and decreased workability in physicians are in accordance with body of research showing negative effects of workload for distress10 and burnout,22 job satisfaction23 and even poorer quality of care.34 Previous studies have also addressed the role of social relationships in physician work. Somville et al.8 reported from a cross sectional study that strong support from colleagues and supervisors was negatively related to psychological distress, fatigue and positively related to job satisfaction in a sample of 346 emergency physicians. In a German study conducted in hospital departments (n = 435) in a cross sectional setting high social support was related to less depressive symptoms and better ‘workability’ among medical residents.21 Wang et al.22 reported from a cross sectional study among 457 physicians from 21 hospitals in Shanghai that low work-related social support was related to higher levels of physicians’ burn out. Linzer et al.23) reported from a large cross-sectional sample (n = 2326) of primary care and subspecialty physicians that higher colleague support for work/home balance, higher support by spouse for career and lower social isolation due to gender or cultural differences were related to lower experienced stress in physicians. Supporting these previous studies, we showed in a longitudinal setting that collegial support was related to more positive changes in distress, sleep quality and workability. However, our results add to the previous results in showing that the associations of collegial support with distress and sleep quality were more pronounced at constant or increased levels of workload during the follow-up. Team-climate is another aspect of social relationships that has been studied in health care context. Team work has been shown to be increasingly important in various sectors of health care such as in chronic care35 and palliative care36 and rehabilitation.37 Some studies have associated good team climate with better quality of care24 although also null findings have been reported.38 In this study, we showed that good team climate was also related to positive changes in workability and sleep quality. Good team climate was also associated with positive distress change, particularly at lower levels of workload. The buffering effect hypothesis postulates social support to have beneficial impact particularly at demanding and stressful situations.19 The moderating effects we found showed that social support does not seem to bring an additional benefit for those reporting increased workload compared with those reporting constant level of workload during the follow-up. One possible explanation may be reversed causality between wellbeing and social support18: high levels of reported social support may indicate that other people respond by offering assistance for a person who’s wellbeing or self-efficacy has deteriorated. We tried to control the results for reversed causality by adjusting the results for baseline values of wellbeing. However, subsequent impaired wellbeing due to for example increased workload may still have had an impact on the availability of social support at work. Another explanation could be that high workload may have deteriorating effect on interpersonal relationships at work. Poor team climate could be an additional stressor in itself and poor team climate has in fact been associated with negative outcomes, such as greater risk of absenteeism26 and intention to leave25 in health care context. Increased workload may therefore attenuate the beneficial potential of social relationships on higher levels of workload. Our study indicated that social relationships are more important predictors of wellbeing among younger and middle aged physicians than older physicians. The strength of the association between social support and health has been shown to vary over the adult life course.18 The mean age in our sample of physicians was relatively high due to 9-year follow-up, hence even the youngest participants have been licenced to practice at least 9 years. Dyrbye et al.39 reported that particularly physicians in middle career worked more hours, took more overnight calls, had the lowest satisfaction with their specialty choice and their work-life balance, and had the highest rates of emotional exhaustion and burnout. Therefore, need for support from colleagues and work team may be particularly high in this group. We were able to examine workload, distress, sleep quality and workability changes in a 9-year longitudinal setting which enhances the validity of our results. Another strength of our study is a relatively large and comprehensive random sample of all physicians in Finland. Previous studies on the role of social support in physician work have been mostly based on small selected samples of physicians. However, the response rate could have been higher although comparable to many other studies on physicians, and it may be that the most stressed physicians have been less motivated to respond. This could lead to smaller variance in stress outcome measures used in the study and also to attenuated results between psychosocial predictors and outcomes since the effects are usually more pronounced at the extremes of the scales. A further limitation of the study is that we had to rely on self-reported measures, which may lead to problems associated with an inflation of the strengths of relationships due to the common method variance. We used validated instruments that have shown good reliability and have been commonly used in occupational stress research to minimize the problems with self-reports. We also controlled the results for age, gender, marital status and specialization status, but we cannot rule out the possibility of residual confounding. One obvious potential confounder is change of job or work place during the follow-up period that could account for changes in working conditions. Our follow-up period was relatively long, so even several changes of work place are possible during the study period. This is particularly important limitation since it does not allow us to control for potential reversed causality between study variables: poor health status may affect availability of social support and experience of work load. Finally, even though the analyses on the relationships between work load and outcome variables were based on longitudinal setting social support indicators were not measured in the baseline and therefore results concerning social support are based on cross sectional setting. Therefore common method variance may partly explain the associations between social support and outcome variables. In conclusion, the current findings highlight the importance of functioning interpersonal relationships at work and social support from colleagues for wellbeing of physicians. Therefore, it is important that health care organizations will pay more attention to interpersonal relations and psychosocial working environment in general and take measures to enhance availability of social support for physicians. One important method could be organizing clinical supervision for physicians. Torppa et al.40 reported that 36% of family physicians working in health centres expressed an unmet need for clinical supervision and the researchers suggest that clinical supervision should be integrated to physicians into continuing professional development of physicians. However, in addition to enhancing socials support it is important to tackle the important stressors such as work overload of the staff in health care context. Key points Increased workload in physicians was associated with increased distress and sleeping problems as well as decreased workability Strong social support and good team climate were associated with decreased distress and sleep quality and increased workability. Social support and team climate mitigated the effects of constant or increased workload on distress. Social support also mitigated effects of workload on sleeping problems. Health-care organizations should pay more attention to workload in physicians and try to promote availability of collegial support and good team climate. Funding This study was supported by the Academy of Finland (265977). Conflicts of interest: None declared. References 1 Appleton K, House A, Dowell A. A survey of job satisfaction, sources of stress and psychological symptoms among general practitioners in Leeds. Br J Gen Pract  1998; 48: 1059– 63. Google Scholar PubMed  2 Shanafelt TD, Boone S, Tan L, et al.   Burnout and satisfaction with work-life balance among US physicians relative to the general US population. Arch Intern Med  2012; 172: 1377– 85. http://dx.doi.org/10.1001/archinternmed.2012.3199 Google Scholar CrossRef Search ADS PubMed  3 Shanafelt TD, Mungo M, Schmitgen J, et al.   Longitudinal study evaluating the association between physician burnout and changes in professional work effort. Mayo Clinic Proc  2016; 91: 422– 31. http://dx.doi.org/10.1016/j.mayocp.2016.02.001 Google Scholar CrossRef Search ADS   4 Dewa CS, Loong D, Bonato S, et al.   How does burnout affect physician productivity? A systematic literature review. BMC Health Serv Res  2014; 14: 325. http://www.biomedcentral.com/1472-6963/14/325. Google Scholar CrossRef Search ADS PubMed  5 Anagnostopoulos F, Liolios E, Persefonis G, et al.   Physician burnout and patient satisfaction with consultation in primary health care settings: evidence of relationships from a one-with-many design. J Clin Psychol Med Settings  2012; 19: 401– 10. http://dx.doi.org/10.1007/s10880-011-9278-8 Google Scholar CrossRef Search ADS PubMed  6 Shanafelt TD, Balch CM, Bechamps G, et al.   Burnout and medical errors among american surgeons. Ann Surg  2010; 251: 995– 1000. http://dx.doi.org/10.1097/SLA.0b013e3181bfdab3 Google Scholar CrossRef Search ADS PubMed  7 Hobbs DFR, Bankhead C, Mukhtar T, et al.   Clinical workload in UK primary care: a retrospective analysis of 100 million consultations in England, 2007-14. Lancet  2016; 387: 2323– 30. http://dx.doi.org/10.1016/S0140-6736(16)00620-6 Google Scholar CrossRef Search ADS PubMed  8 Somville JF, Gucht V, Maes S. The impact of occupational hazards and traumatic events among Belgian emergency physicians. Int J Behav Med Nutr Phys Activity  2016; 24: 59. 9 Heponiemi T, Kouvonen A, Vanska J, et al.   Effects of active on-call hours on physicians' turnover intentions and well-being. Scand J Work Environ Health  2008; 34: 356– 63. http://dx.doi.org/10.5271/sjweh.1278 Google Scholar CrossRef Search ADS PubMed  10 Elovainio M, Salo P, Jokela M, Heponiemi T, Linna A, Virtanen M, et al.   Psychosocial factors and well-being among Finnish GPs and specialists: a 10-year follow-up. Occup Environ Med  2013; 70: 246– 51. 11 Ahola K, Hakanen J. Job strain, burnout, and depressive symptoms: a prospective study among dentists. J Affect Disord  2007; 104: 103– 10. http://dx.doi.org/10.1016/j.jad.2007.03.004 Google Scholar CrossRef Search ADS PubMed  12 Lallukka T, Chandola T, Hemingway H, et al.   Job strain and symptoms of angina pectoris among British and Finnish middle-aged employees. J Epid Comm Health  2009; 63: 980– 5. http://dx.doi.org/10.1136/jech.2008.085878 Google Scholar CrossRef Search ADS   13 Elovainio M, Heponiemi T, Kuusio H, Jokela M, Aalto A-M, Pekkarinen L, et al.   Job demands and job strain as risk factors for employee wellbeing in elderly care: an instrumental-variables analysis. Eur J Public Health  2015; 25: 103– 8. Google Scholar CrossRef Search ADS PubMed  14 Virtanen M, Vahtera J, Pentti J, et al.   Job strain and psychologic distress influence on sickness absence among Finnish employees. Am J Prev Med  2007; 33: 182– 7. http://dx.doi.org/10.1016/j.amepre.2007.05.003 Google Scholar CrossRef Search ADS PubMed  15 Vahtera J, Laine S, Virtanen M, et al.   Employee control over working times and risk of cause-specific disability pension: the Finnish Public Sector Study. Occ Envir Med  2010; 67: 479– 85. http://dx.doi.org/10.1136/oem.2008.045096 Google Scholar CrossRef Search ADS   16 House JS, Umberson D, Landis KR. Structures and processes of social support. Annu Rev Sociol  1988; 14: 293– 318. http://dx.doi.org/10.1146/annurev.so.14.080188.001453 Google Scholar CrossRef Search ADS   17 Barth J, Schneider S, von Känel R. Lack of social support in the etiology and the prognosis of coronary heart disease: a systematic review and meta-analysis. Psychosom Med  2010; 72: 229– 38. Google Scholar CrossRef Search ADS PubMed  18 Hakulinen C, Pulkki-Råback L, Jokela M, et al.   Structural and functional aspects of social support as predictors of mental and physical health trajectories: Whitehall II cohort study. J Epidemiol Commun Health  2016; 70: 710– 5. Google Scholar CrossRef Search ADS   19 Cohen S, Wills T. Stress, social support, and then buffering hypothesis. Psychol Bull  1985; 98: 310– 57. http://dx.doi.org/10.1037/0033-2909.98.2.310 Google Scholar CrossRef Search ADS PubMed  20 Van der Doef M, Maes S. The Job Demand-Control (-Support) Model and psychological well-being: a review of 20 years of empirical research. Work Stress  1999; 13: 87– 114. http://dx.doi.org/10.1080/026783799296084 Google Scholar CrossRef Search ADS   21 Bernburg M, Vitzthum K, Groneberg DA, Mache S. Physicians’ occupational stress, depressive symptoms and workability in relation to their working environment: a cross-sectional study of differences among medical residents with various specialties working in German hospitals. BMJ Open  2016; 6:e011369. 1. 22 Wang Z, Xie Z, Dai J, et al.   Physician burnout and its associated factors: a cross-sectional study in Shangh. J Occup Health  2014; 56: 73– 83. http://dx.doi.org/10.1539/joh.13-0108-OA Google Scholar CrossRef Search ADS PubMed  23 Linzer M, Gerrity M, Douglas JA, et al.   Physician stress: results from the physician worklife study. Stress and Health  2002; 18: 37– 42. http://dx.doi.org/10.1002/smi.917 Google Scholar CrossRef Search ADS   24 Bower P, Campbell S, Bojke C, Sibbald B. Team structure, team climate and the quality of care in primary care: an observational study. Qual Saf Health Care  2003; 12: 273– 9. http://dx.doi.org/10.1136/qhc.12.4.273 Google Scholar CrossRef Search ADS PubMed  25 Kivimaki M, Vanhala A, Pentti J, et al.   Team climate, intention to leave and turnover among hospital employees: prospective cohort study. BMC Health Serv Res  2007; 7: 170. http://dx.doi.org/10.1186/1472-6963-7-170 Google Scholar CrossRef Search ADS PubMed  26 Kivimäki M, Sutinen R, Elovainio M, et al.   Sickness absence in hospital physicians: 2 year follow up study on determinants. Occup Environ Med  2001; 58: 361– 6. Google Scholar CrossRef Search ADS PubMed  27 Elovainio M, Heponiemi T, Vänskä J, Sinervo T, Kujala S, Laakso E, et al.   Miten Suomalaisnen lääkäri voi 2000-luvulla. Suomen Lääkärilehti  2007; 62:2071–76. 28 Graetz B. Multidimensional properties of the General Health Questionnaire. Soc Psychiatry Psychiatr Epidemiol  1991; 26: 132– 8. http://dx.doi.org/10.1007/BF00782952 Google Scholar CrossRef Search ADS PubMed  29 Penninkilampi-Kerola V, Miettunen J, Ebeling H. A comparative assessment of the factor structures and psychometric properties of the GHQ-12 and the GHQ-20 based on data from a Finnish population-based sample. Scand J Psychol  2006; 47: 431– 40. Google Scholar CrossRef Search ADS PubMed  30 Jenkins CD, Stanton BA, Niemcryk SJ, Rose RM. A scale for the estimation of sleep problems in clinical research. J Clin Epidemiol  1988; 41: 313– 21. http://dx.doi.org/10.1016/0895-4356(88)90138-2 Google Scholar CrossRef Search ADS PubMed  31 Ilmarinen J, Tuomi K, Klockars M. Changes in the workability of active employees over an 11-year period. Scand J Work Environ Health  1997; 23(Suppl 1): 49– 57. Google Scholar PubMed  32 Elovainio M, Sinervo T. Psychosocial stressors at work, psychological stress and musculoskeletal symptoms in the care for the elderly. Work Stress  1997; 11: 351– 61. http://dx.doi.org/10.1080/02678379708252998 Google Scholar CrossRef Search ADS   33 Anderson NR, West MA. Measuring climate for work group innovation: development and validation of the team climate inventory. J Organ Behav  1998; 19: 235– 58. http://dx.doi.org/10.1002/(SICI)1099-1379(199805)19:3<235::AID-JOB837>3.0.CO;2-C Google Scholar CrossRef Search ADS   34 Klein J, Grosse Frie KG, Blum K, von dem Knesebeck O. Psychosocial stress at work and perceived quality of care among clinicians in surgery. BMC Health Serv Res  2011; 11: 109.http://www.biomedcentral.com/1472-6963/11/109. Google Scholar CrossRef Search ADS PubMed  35 Katon WJ, Lin EHB, Korff M, et al.   Collaborative Care for Patients with Depression and Chronic Illnesses. N Engl J Med  2010; 363: 2611– 20. http://dx.doi.org/10.1056/NEJMoa1003955 Google Scholar CrossRef Search ADS PubMed  36 Crawford GB, Price SD. Team working: palliative care as a model of interdisciplinary practice. Med J Aust  2003; 179: S32– 4. Google Scholar PubMed  37 Bettger JA, Stineman MG. Effectiveness of multidisciplinary rehabilitation services in postacute care: state-of-the-science. A review. Arch Phys Med Rehabil  2007; 88: 1526– 34. Google Scholar CrossRef Search ADS PubMed  38 Goh TT, Eccles M, Steen N. Factors predicting team climate, and its relationship with quality of care in general practice. BMC Health Serv Res  2009; 9: 138. http://dx.doi.org/10.1186/1472-6963-9-138 Google Scholar CrossRef Search ADS PubMed  39 Dyrbye LN, Varkey P, Boone SL, et al.   Physician satisfaction and burnout at different career stages. Mayo Clin Proc  2013; 88: 1358– 67. http://dx.doi.org/10.1016/j.mayocp.2013.07.016 Google Scholar CrossRef Search ADS PubMed  40 Torppa MA, Kuikka L, Nevalainen M, Pitkälä KH. Family physician experiences with and needs for clinical supervision: associations between work experiences, professional issues and social support at work. Patient Educ Couns  2016; 99: 1198– 202. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The European Journal of Public Health Oxford University Press

Social relationships in physicians’ work moderate relationship between workload and wellbeing—9-year follow-up study

Loading next page...
 
/lp/ou_press/social-relationships-in-physicians-work-moderate-relationship-between-wocltUhPck
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved.
ISSN
1101-1262
eISSN
1464-360X
D.O.I.
10.1093/eurpub/ckx232
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Increasing wellbeing problems among physicians may lead to serious consequences in health care and means to prevent such development are called for. This study examined longitudinal associations between workload and changes in distress, sleep quality and workability in physicians and whether positive social relations at work would protect from such problems. Methods A baseline survey was conducted in 2006 for a random sample of 5000 physicians (n = 2841, response rate 57%). In 2015, the follow-up survey was sent to those 2 206 physicians who gave their consent (n = 1462, response rate 68.3%). The survey included scales for distress, sleeping problems, workability, workload, team climate, collegial support and questions for background information. Results Increased workload was associated with increased psychological distress, sleeping problems and decreased workability during the 9-year follow-up. Good team climate and collegial support were related to decreased distress and sleep quality and enhanced workability. Good collegial support buffered the associations of workload changes on distress and sleep quality changes. Team climate was more strongly associated with changes in sleep quality and workability among younger and middle aged physicians than older physicians. Also collegial support had a stronger association with sleep quality change among younger or middle aged physicians than older physicians. These associations were robust to adjustments for age, gender, specialization, leadership position, marital status and baseline wellbeing. Conclusions Health care organizations should take measures to decrease workload and to increase availability of social support for physicians in order to protect physicians from declining wellbeing. Introduction Poor work satisfaction and decreasing wellbeing among physicians is a recognized problem. Such development may lead to serious consequences in health care, such as lower quality of care or physician turnover. In UK, 30–50% of general practitioners reported high levels of psychological distress.1 Shanafelt et al.2 have reported that compared with general population physicians are more likely to have symptoms of burnout and be dissatisfied with work-life balance. Physician burnout and distress have been associated with decreased professional effort3 and lower productiveness4 as well as lower patient satisfaction.5 Poor wellbeing among physicians may even endanger patient safety due to medical errors.6 Physician’s work comprises various stressors such as clinical workload,7 adverse patient events8 and on-call work.9 In the Finnish Health Care Professionals Study, patient-related stress, poorly functioning patient record systems, and job demands in terms of time-pressure burdened physicians in hospitals and health centres.10 Time pressure, patient-related stress, and work interference with family were particularly associated with working in public primary care. Job strain has in general been associated with poor health consequences11,12 and poor psychological wellbeing,13 absenteeism14 and early retirement.15 A large body of evidence shows that high levels of social support and social connectedness are related to better health16–18 Social support may have a direct positive effect on health and wellbeing because it provides positive affect and stability and because it may help to avoid negative experiences. Social support may also buffer against negative effects of adverse events or circumstances. Social support may change the appraisal of the stressful situation or activate adaptive responses in the situation.19 In occupational health research social support has been incorporated as a coping resource into the extended job strain model.20 However, in addition to serving as adaptive resources, social relationships may also be a source of stress. Dense social networks or close relations may potentially cause social conflicts and therefore may have negative consequences for wellbeing.18 The importance of social relationships for health may also vary during the life-course. Hakulinen et al.18 have shown that association between negative aspects of social relationships, such as interpersonal conflicts and poor health strengthened by increasing age while association between strong social support and good mental health among women weakened by age. Social relationships have been found to be an important psychosocial resource in physicians work as well. Several studies have indicated that physicians who lack social support are at greater risk of impaired wellbeing or burnout.8,21–23 Another aspect of social relationships in medical work is working in teams. Team climate has been associated with better quality of care in health care setting24 and problems in team climate have been associated with intentions to leave among hospital staff25 and sickness absenteeism among physicians.26 This study focuses on the role of social relationships in the associations between physicians’ workload and wellbeing in a longitudinal design. More specifically, we examined whether perceived collegial support and team climate in the work unit are able to buffer the negative effects of increased workload on psychological distress, sleeping problems and workability among physicians in 9-year follow-up setting. In addition, we examine whether age buffers the importance of social support on wellbeing. Methods Study sample The present study is a part of the on-going Finnish Health Care Professionals Study, in which we drew a random sample of 5000 physicians in Finland (30% of the whole physician population) from the 2006 database of physicians maintained by the Finnish Medical Association. The register covers all licenced physicians in Finland. Phase 1 data were gathered with postal questionnaires in 2006. Non-respondents were sent a reminder and a copy of the questionnaire up to two times. Responses were received from 2841 physicians (response rate 57%). The sample is representative of the eligible population in terms of age, gender, and employment sector.27 Nine years later, in 2015 the follow-up questionnaire was sent to those who gave their consent in 2006 to participate to the follow-up survey (n = 2206, 77.6% of the respondents in the 2006 survey). Those who had died during the follow-up or who had an unknown address in 2015 were excluded leaving 2159 physicians to the follow-up sample in 2015. Of these 1462 physicians responded (response rate 68.3% among those consented to follow-up (29% of the original sample in 2006), 59% women, mean age 54.8, SD = 9.74). When compared with all physicians in Finland the 9-year follow-up respondents were more often aged 55 years or older (34% of all Finnish physicians and 41% in the follow-up sample) and specialized (62% of all Finnish physicians and 85% in the follow-up sample), but there were no differences in terms of gender. In addition the follow-up attrition was analysed by comparing the follow-up respondents to those who responded in the baseline but not in the 9-year follow-up in terms of baseline survey information. There were no differences according to gender and baseline work load between the follow-up respondents and non-responders. However, those who responded in the follow-up were in the baseline older, less often under specialization process and more often in superior position. Measurements Wellbeing indicators ‘Psychological distress’ was measured with the four items from the GHQ-12 (Goldberg 1972) representing anxiety/depression dimension of GHQ-12 from the Graetz28,29 three-factor structure model for GHQ-12. The response options ranged from 1 to 4. In this study, the scale was used as a continuous variable. Cronbach’s alpha coefficients for this sample were α = 0.83 in 2006 and 0.84 in 2015. ‘Sleeping problems’ were measured with four-item Jenkins scale30 (α = 0.77 in 2006 and 0.78 in 2015). Respondents were asked how often during the last four weeks they had troubles falling asleep, were waking up several times per night, had troubles staying asleep including waking up too early, and felt tired after usual amount of sleep. The response scale ranged from 1 (never) to 6 (every night). ‘Workability’ was assessed with a single item question included in the Workability Index31: ‘Assume that your workability at its best has a value of 10 and 0 would mean that you could not work at all. How many points would you give to your current workability (range 0-10)?’ Workload ‘Workload’ was measured with three items derived from the Harris Nurse Stress index, which have been developed based on previous research among nurses and health care staff and which have shown adequate psychometric properties in terms of internal consistency and construct validity.32 This scale measures stress due to time shortages at work and scheduling problems. An example item: ‘How often have you been distracted, worried or stressed about (during the past half-year period) not being able to do your work properly.’ The items were rated on a 5-point Likert-scale ranging from 1 (never) to 5 (very often) (α = 0.83 in 2006 and 0.87 in 2015). ‘Social relations’ variables were measured in 2015. ‘Team climate’ was measured with a Team Climate Inventory’s33 participative safety subscale that measures team participation, such as interaction frequency and information sharing (four items, α = 0.88; e.g ‘People feel understood and accepted by each other’). The items were rated on a 5-point Likert-scale, ranging from 1 (I totally disagree) to 5 (I totally agree). ‘Collegial support’ was measured with three items measuring measuring tangible social support at work in terms of consultation possibilities and co-operation in the team (α = 0.67). Example of the items includes: ‘How often have you been distracted, worried or stressed about (during the past half-year period) lack of consultation possibilities’ (reverse coded). The items were rated on a five-point Likert-scale ranging from 1 (never) to 5 (very often). Background variables The background variables used were gender, specialization status, leadership position, marital status, having children and baseline level of outcome variables. The information about specialization status and leadership position were based on self-report. The respondents were asked directly to indicate whether they were (i) not specialized, (ii) under specialization, (iii) specialized and whether they were in supervisory position (yes/no). Statistical analysis We constructed change scores for wellbeing indicators and workload (by subtracting the 2006 scores from the 2015 scores). Workload change score was constructed by subtracting the 2006 scores from the 2015 scores and further categorizing the change score into three groups (i) workload levels had decreased (negative change score (ii) stayed the same (change score = 0) and (iii) increased (positive change score). For outcome variables continuous change scores were used, higher scores indicating that distress, sleeping problems and workability had increased during this time period. The associations of workload change and social relations variables on distress, sleeping problems and workability change were examined using the analyses of covariance (in separate analyses). The analyses of main effects were conducted in two steps. First, the analyses were adjusted for gender and age. Second, specialization status, leadership position, marital status, having children and baseline level of outcome variable were additionally adjusted for. In analyses of work load change baseline work load at 2006 was additionally adjusted for. The interactions of social relations variables (team climate and collegial support; in separate analyses) with workload change and age were examined with the analyses of covariance adjusted for main effects, age, gender, specialization status, leadership position, marital status, having children and baseline level of outcome variable and in analyses of interaction between work load change and social support variables the baseline workload was additionally adjusted for. In the original analyses, age was used as continuous variable, but in order to illustrate the interactions in figures age was dichotomized (younger vs. older) using median as the cut-off point (≤55 years/>55 years). Similarly, social support variables were used as continuous variables in the analyses, but they were dichotomized to illustrate the interaction effects in the figures. Values 1 and 2 indicated poor team climate and values above 2 indicated good team climate. In the collegial support variable scores 1–4 were combined to indicate low collegial support and scores 4–5 indicated high collegial support. Results Main effects Table 1 shows the sample characteristics. Majority of respondents were women and had specialized. The mean age in the follow-up was 55 years. Table 2 presents the results of covariance analyses regarding changes of distress, sleeping problems and workability. Adjusting for age and gender, workload change was associated with changes in distress, sleeping problems and workability. Distress and sleeping problems had increased and workability decreased most among those whose workload levels had increased during the follow up. Higher age was associated with decreased distress and sleeping problems and decreased workability. Poor team climate and low collegial support were associated with increased distress and sleeping problems. Further adjustments for baseline distress or sleeping problems, specialization, leadership position and marital status did not attenuate the associations of team climate and collegial support with distress change and sleeping problems. Team climate and collegial support were not associated with changes in workability when adjusted only for age and gender; however, after adjusting also for baseline workability, specialization, leadership position and marital status better team climate and stronger collegial support were associated with increased workability. Table 1 Sample characteristics     n  %  Gender  Men  585  40  Women  856  59  Marital status  Single  81  6  Married/co-habiting  1191  83  Divorced  138  10  Widowed  34  2  Specialization  No  161  11  In training  53  4  Specialized  1232  85  Workload change 2006–2015  Increased  693  26  Same  197  14  Decreased  385  47      Mean  SD  Age 2015    54.8  9.74  Distress 2006a    1.89  0.60  Distress 2015a    1.77  0.65  Sleeping problems 2006b  2.32  1.01  Sleeping problems 2015b  2.44  1.02  Work ability 2006c    8.58  1.32  Work ability 2015c    8.20  1.54  Team climate 2015d  3.91  0.73  Collegial support 2015d  3.91  0.78      n  %  Gender  Men  585  40  Women  856  59  Marital status  Single  81  6  Married/co-habiting  1191  83  Divorced  138  10  Widowed  34  2  Specialization  No  161  11  In training  53  4  Specialized  1232  85  Workload change 2006–2015  Increased  693  26  Same  197  14  Decreased  385  47      Mean  SD  Age 2015    54.8  9.74  Distress 2006a    1.89  0.60  Distress 2015a    1.77  0.65  Sleeping problems 2006b  2.32  1.01  Sleeping problems 2015b  2.44  1.02  Work ability 2006c    8.58  1.32  Work ability 2015c    8.20  1.54  Team climate 2015d  3.91  0.73  Collegial support 2015d  3.91  0.78  a Scale 1–4, higher score indicates higher distress. b Scale 1–6, higher score indicates more sleeping problems. c Scale 0–10, higher score indicates better work ability. d Scale 1–5, higher score indicates better support or team climate. Table 1 Sample characteristics     n  %  Gender  Men  585  40  Women  856  59  Marital status  Single  81  6  Married/co-habiting  1191  83  Divorced  138  10  Widowed  34  2  Specialization  No  161  11  In training  53  4  Specialized  1232  85  Workload change 2006–2015  Increased  693  26  Same  197  14  Decreased  385  47      Mean  SD  Age 2015    54.8  9.74  Distress 2006a    1.89  0.60  Distress 2015a    1.77  0.65  Sleeping problems 2006b  2.32  1.01  Sleeping problems 2015b  2.44  1.02  Work ability 2006c    8.58  1.32  Work ability 2015c    8.20  1.54  Team climate 2015d  3.91  0.73  Collegial support 2015d  3.91  0.78      n  %  Gender  Men  585  40  Women  856  59  Marital status  Single  81  6  Married/co-habiting  1191  83  Divorced  138  10  Widowed  34  2  Specialization  No  161  11  In training  53  4  Specialized  1232  85  Workload change 2006–2015  Increased  693  26  Same  197  14  Decreased  385  47      Mean  SD  Age 2015    54.8  9.74  Distress 2006a    1.89  0.60  Distress 2015a    1.77  0.65  Sleeping problems 2006b  2.32  1.01  Sleeping problems 2015b  2.44  1.02  Work ability 2006c    8.58  1.32  Work ability 2015c    8.20  1.54  Team climate 2015d  3.91  0.73  Collegial support 2015d  3.91  0.78  a Scale 1–4, higher score indicates higher distress. b Scale 1–6, higher score indicates more sleeping problems. c Scale 0–10, higher score indicates better work ability. d Scale 1–5, higher score indicates better support or team climate. Table 2 Estimated marginal means and results of the analyses of covariance for the 9-year changes of distress (GHQ) and sleeping problems   Step 1a   Step 2b     Meanc (95% CI)  F  P  Meanc (95% CI)  F  P  Distress change                  Workload    37.13  <0.001    51.20  <0.001          Decreased  −0.23 (−0.28, −0.18)      −0.22 (−0.32, −0.12)              Stayed the same  0.02 (−0.07, 0.12)      0.01 (−0.11, 0.13)              Increased  0.14 (0.07, 0.21)      0.22 (0.11, 0.33)      Gender    0.22  0.639    9.91  0.002  Age    21.17  <0.001    10.80  <0.001  Team climate    31.85  <0.001    67.84  <0.001  Collegial support    24.50  <0.001    88.02  <0.001  Sleep quality change                  Workload    13.56  <0.001    19.27  <0.001          Decreased  0.06 (−0.01, 0.13)      −0.02 (−0.16, 0.13)              Stayed the same  0.29 (0.15, 0.43)      0.22 (0.04, 0.39)              Increased  0.36 (0.26, 0.46)      0.38 (0.21, 0.54)      Gender    1.87  0.172    1.22  0.269  Age    41.45  <0.001    10.11  0.001  Team climate    12.65  <0.001    26.82  <0.001  Collegial support    16.91  <0.001    42.83  <0.001  Work ability change      Time pressure    10.09  <0.001    18.86  <0.001          Decreased  −0.13 (−0.24, −0.02)      0.01 (−0.21, 0.24)              Stayed the same  −0.27 (−0.48, −0.06)      −0.11 (−0.36, 0.15)              Increased  −0.56 (−0.71, −0.41)      −0.57 (−0.81, −0.33)      Gender    0.14  0.713    0.72  0.395  Age    5.75  0.017    12.92  <0. 001  Team climate    0.51  0.476    9.74  <0.001  Collegial support    2.17  0.141    18.24  <0.001    Step 1a   Step 2b     Meanc (95% CI)  F  P  Meanc (95% CI)  F  P  Distress change                  Workload    37.13  <0.001    51.20  <0.001          Decreased  −0.23 (−0.28, −0.18)      −0.22 (−0.32, −0.12)              Stayed the same  0.02 (−0.07, 0.12)      0.01 (−0.11, 0.13)              Increased  0.14 (0.07, 0.21)      0.22 (0.11, 0.33)      Gender    0.22  0.639    9.91  0.002  Age    21.17  <0.001    10.80  <0.001  Team climate    31.85  <0.001    67.84  <0.001  Collegial support    24.50  <0.001    88.02  <0.001  Sleep quality change                  Workload    13.56  <0.001    19.27  <0.001          Decreased  0.06 (−0.01, 0.13)      −0.02 (−0.16, 0.13)              Stayed the same  0.29 (0.15, 0.43)      0.22 (0.04, 0.39)              Increased  0.36 (0.26, 0.46)      0.38 (0.21, 0.54)      Gender    1.87  0.172    1.22  0.269  Age    41.45  <0.001    10.11  0.001  Team climate    12.65  <0.001    26.82  <0.001  Collegial support    16.91  <0.001    42.83  <0.001  Work ability change      Time pressure    10.09  <0.001    18.86  <0.001          Decreased  −0.13 (−0.24, −0.02)      0.01 (−0.21, 0.24)              Stayed the same  −0.27 (−0.48, −0.06)      −0.11 (−0.36, 0.15)              Increased  −0.56 (−0.71, −0.41)      −0.57 (−0.81, −0.33)      Gender    0.14  0.713    0.72  0.395  Age    5.75  0.017    12.92  <0. 001  Team climate    0.51  0.476    9.74  <0.001  Collegial support    2.17  0.141    18.24  <0.001  a Adjusted for age and gender. b Adjusted for age, gender, having children, baseline level of outcome, specialization status, leadership position, and marital status. In the analyses between work load and outcomes the baseline work load (in 2006) was additionally adjusted for. c Estimated marginal means. Table 2 Estimated marginal means and results of the analyses of covariance for the 9-year changes of distress (GHQ) and sleeping problems   Step 1a   Step 2b     Meanc (95% CI)  F  P  Meanc (95% CI)  F  P  Distress change                  Workload    37.13  <0.001    51.20  <0.001          Decreased  −0.23 (−0.28, −0.18)      −0.22 (−0.32, −0.12)              Stayed the same  0.02 (−0.07, 0.12)      0.01 (−0.11, 0.13)              Increased  0.14 (0.07, 0.21)      0.22 (0.11, 0.33)      Gender    0.22  0.639    9.91  0.002  Age    21.17  <0.001    10.80  <0.001  Team climate    31.85  <0.001    67.84  <0.001  Collegial support    24.50  <0.001    88.02  <0.001  Sleep quality change                  Workload    13.56  <0.001    19.27  <0.001          Decreased  0.06 (−0.01, 0.13)      −0.02 (−0.16, 0.13)              Stayed the same  0.29 (0.15, 0.43)      0.22 (0.04, 0.39)              Increased  0.36 (0.26, 0.46)      0.38 (0.21, 0.54)      Gender    1.87  0.172    1.22  0.269  Age    41.45  <0.001    10.11  0.001  Team climate    12.65  <0.001    26.82  <0.001  Collegial support    16.91  <0.001    42.83  <0.001  Work ability change      Time pressure    10.09  <0.001    18.86  <0.001          Decreased  −0.13 (−0.24, −0.02)      0.01 (−0.21, 0.24)              Stayed the same  −0.27 (−0.48, −0.06)      −0.11 (−0.36, 0.15)              Increased  −0.56 (−0.71, −0.41)      −0.57 (−0.81, −0.33)      Gender    0.14  0.713    0.72  0.395  Age    5.75  0.017    12.92  <0. 001  Team climate    0.51  0.476    9.74  <0.001  Collegial support    2.17  0.141    18.24  <0.001    Step 1a   Step 2b     Meanc (95% CI)  F  P  Meanc (95% CI)  F  P  Distress change                  Workload    37.13  <0.001    51.20  <0.001          Decreased  −0.23 (−0.28, −0.18)      −0.22 (−0.32, −0.12)              Stayed the same  0.02 (−0.07, 0.12)      0.01 (−0.11, 0.13)              Increased  0.14 (0.07, 0.21)      0.22 (0.11, 0.33)      Gender    0.22  0.639    9.91  0.002  Age    21.17  <0.001    10.80  <0.001  Team climate    31.85  <0.001    67.84  <0.001  Collegial support    24.50  <0.001    88.02  <0.001  Sleep quality change                  Workload    13.56  <0.001    19.27  <0.001          Decreased  0.06 (−0.01, 0.13)      −0.02 (−0.16, 0.13)              Stayed the same  0.29 (0.15, 0.43)      0.22 (0.04, 0.39)              Increased  0.36 (0.26, 0.46)      0.38 (0.21, 0.54)      Gender    1.87  0.172    1.22  0.269  Age    41.45  <0.001    10.11  0.001  Team climate    12.65  <0.001    26.82  <0.001  Collegial support    16.91  <0.001    42.83  <0.001  Work ability change      Time pressure    10.09  <0.001    18.86  <0.001          Decreased  −0.13 (−0.24, −0.02)      0.01 (−0.21, 0.24)              Stayed the same  −0.27 (−0.48, −0.06)      −0.11 (−0.36, 0.15)              Increased  −0.56 (−0.71, −0.41)      −0.57 (−0.81, −0.33)      Gender    0.14  0.713    0.72  0.395  Age    5.75  0.017    12.92  <0. 001  Team climate    0.51  0.476    9.74  <0.001  Collegial support    2.17  0.141    18.24  <0.001  a Adjusted for age and gender. b Adjusted for age, gender, having children, baseline level of outcome, specialization status, leadership position, and marital status. In the analyses between work load and outcomes the baseline work load (in 2006) was additionally adjusted for. c Estimated marginal means. Interactions The interaction terms between age and team climate (F = 4.62, P = 0.032) and age and collegial support (F = 4.54, P = 0.033) were significant for sleeping problems change. Collegial support and team climate were more strongly associated with the change in sleeping problems among young and middle aged respondents than among older respondents (figure 1c and d). Interaction between age and team climate was significant also for workability change (F = 7.47, P = 0.006). Again, team climate was more strongly associated with workability change among young and middle respondents compared with older respondents (figure 1e). There were no significant interaction effect of age and social support variables (team climate and collegial support) for distress change (figure 1a and b) figures or between age and collegial support for workability change (figure 1f). Figure 1 View largeDownload slide Summary of the moderating effect of age on the associations of social relationships (team climate and collegial support) with well-being factors Figure 1 View largeDownload slide Summary of the moderating effect of age on the associations of social relationships (team climate and collegial support) with well-being factors The interaction of workload change with team climate (F = 3.37, P = 0.035) was significant for distress change. Team climate was associated with distress change more clearly among those who reported decreased or similar level of workload during the follow-up compared with those reporting increased workload (figure 2a). Interaction between workload change and collegial support was significant for distress change (F = 5.00, P = 0.007) and change in sleeping problems (F = 3.76, P = 0.024). High collegial support was more strongly associated with changes in distress and sleeping problems among those who reported similar or increased of workload compared with those reporting decreased workload (figure 2b and c). There was no significant interaction effect between work load change team climate for sleeping problems (figure 2c). Furthermore, work load change and social support variables (team climate and collegial support) did not have interaction effect for change in workability (figure 2e and 2f). Figure 2 View largeDownload slide Summary of the moderating effects of social relationships (team climate and collegial support) on the associations of workload change with well-being factors Figure 2 View largeDownload slide Summary of the moderating effects of social relationships (team climate and collegial support) on the associations of workload change with well-being factors Discussion The results of the current 9-year follow-up suggest that increased workload is associated with increased psychological distress, sleeping problems and decreased workability. Instead, good team climate and collegial support seem to be related to decline in distress and increase of sleep quality and workability. Good team climate may also buffer the negative effects of workload increase for distress and sleep quality. The role of team climate and collegial support for physicians seems to vary according to age, since team climate was more strongly associated with changes in sleep quality and workability among younger and middle aged physicians compared with older ones. Furthermore, also collegial support was more clearly associated with sleep quality changes among younger or middle aged physicians than older, more experienced physicians. Our finding that increased workload was associated with increased distress and sleeping problems and decreased workability in physicians are in accordance with body of research showing negative effects of workload for distress10 and burnout,22 job satisfaction23 and even poorer quality of care.34 Previous studies have also addressed the role of social relationships in physician work. Somville et al.8 reported from a cross sectional study that strong support from colleagues and supervisors was negatively related to psychological distress, fatigue and positively related to job satisfaction in a sample of 346 emergency physicians. In a German study conducted in hospital departments (n = 435) in a cross sectional setting high social support was related to less depressive symptoms and better ‘workability’ among medical residents.21 Wang et al.22 reported from a cross sectional study among 457 physicians from 21 hospitals in Shanghai that low work-related social support was related to higher levels of physicians’ burn out. Linzer et al.23) reported from a large cross-sectional sample (n = 2326) of primary care and subspecialty physicians that higher colleague support for work/home balance, higher support by spouse for career and lower social isolation due to gender or cultural differences were related to lower experienced stress in physicians. Supporting these previous studies, we showed in a longitudinal setting that collegial support was related to more positive changes in distress, sleep quality and workability. However, our results add to the previous results in showing that the associations of collegial support with distress and sleep quality were more pronounced at constant or increased levels of workload during the follow-up. Team-climate is another aspect of social relationships that has been studied in health care context. Team work has been shown to be increasingly important in various sectors of health care such as in chronic care35 and palliative care36 and rehabilitation.37 Some studies have associated good team climate with better quality of care24 although also null findings have been reported.38 In this study, we showed that good team climate was also related to positive changes in workability and sleep quality. Good team climate was also associated with positive distress change, particularly at lower levels of workload. The buffering effect hypothesis postulates social support to have beneficial impact particularly at demanding and stressful situations.19 The moderating effects we found showed that social support does not seem to bring an additional benefit for those reporting increased workload compared with those reporting constant level of workload during the follow-up. One possible explanation may be reversed causality between wellbeing and social support18: high levels of reported social support may indicate that other people respond by offering assistance for a person who’s wellbeing or self-efficacy has deteriorated. We tried to control the results for reversed causality by adjusting the results for baseline values of wellbeing. However, subsequent impaired wellbeing due to for example increased workload may still have had an impact on the availability of social support at work. Another explanation could be that high workload may have deteriorating effect on interpersonal relationships at work. Poor team climate could be an additional stressor in itself and poor team climate has in fact been associated with negative outcomes, such as greater risk of absenteeism26 and intention to leave25 in health care context. Increased workload may therefore attenuate the beneficial potential of social relationships on higher levels of workload. Our study indicated that social relationships are more important predictors of wellbeing among younger and middle aged physicians than older physicians. The strength of the association between social support and health has been shown to vary over the adult life course.18 The mean age in our sample of physicians was relatively high due to 9-year follow-up, hence even the youngest participants have been licenced to practice at least 9 years. Dyrbye et al.39 reported that particularly physicians in middle career worked more hours, took more overnight calls, had the lowest satisfaction with their specialty choice and their work-life balance, and had the highest rates of emotional exhaustion and burnout. Therefore, need for support from colleagues and work team may be particularly high in this group. We were able to examine workload, distress, sleep quality and workability changes in a 9-year longitudinal setting which enhances the validity of our results. Another strength of our study is a relatively large and comprehensive random sample of all physicians in Finland. Previous studies on the role of social support in physician work have been mostly based on small selected samples of physicians. However, the response rate could have been higher although comparable to many other studies on physicians, and it may be that the most stressed physicians have been less motivated to respond. This could lead to smaller variance in stress outcome measures used in the study and also to attenuated results between psychosocial predictors and outcomes since the effects are usually more pronounced at the extremes of the scales. A further limitation of the study is that we had to rely on self-reported measures, which may lead to problems associated with an inflation of the strengths of relationships due to the common method variance. We used validated instruments that have shown good reliability and have been commonly used in occupational stress research to minimize the problems with self-reports. We also controlled the results for age, gender, marital status and specialization status, but we cannot rule out the possibility of residual confounding. One obvious potential confounder is change of job or work place during the follow-up period that could account for changes in working conditions. Our follow-up period was relatively long, so even several changes of work place are possible during the study period. This is particularly important limitation since it does not allow us to control for potential reversed causality between study variables: poor health status may affect availability of social support and experience of work load. Finally, even though the analyses on the relationships between work load and outcome variables were based on longitudinal setting social support indicators were not measured in the baseline and therefore results concerning social support are based on cross sectional setting. Therefore common method variance may partly explain the associations between social support and outcome variables. In conclusion, the current findings highlight the importance of functioning interpersonal relationships at work and social support from colleagues for wellbeing of physicians. Therefore, it is important that health care organizations will pay more attention to interpersonal relations and psychosocial working environment in general and take measures to enhance availability of social support for physicians. One important method could be organizing clinical supervision for physicians. Torppa et al.40 reported that 36% of family physicians working in health centres expressed an unmet need for clinical supervision and the researchers suggest that clinical supervision should be integrated to physicians into continuing professional development of physicians. However, in addition to enhancing socials support it is important to tackle the important stressors such as work overload of the staff in health care context. Key points Increased workload in physicians was associated with increased distress and sleeping problems as well as decreased workability Strong social support and good team climate were associated with decreased distress and sleep quality and increased workability. Social support and team climate mitigated the effects of constant or increased workload on distress. Social support also mitigated effects of workload on sleeping problems. Health-care organizations should pay more attention to workload in physicians and try to promote availability of collegial support and good team climate. Funding This study was supported by the Academy of Finland (265977). Conflicts of interest: None declared. References 1 Appleton K, House A, Dowell A. A survey of job satisfaction, sources of stress and psychological symptoms among general practitioners in Leeds. Br J Gen Pract  1998; 48: 1059– 63. Google Scholar PubMed  2 Shanafelt TD, Boone S, Tan L, et al.   Burnout and satisfaction with work-life balance among US physicians relative to the general US population. Arch Intern Med  2012; 172: 1377– 85. http://dx.doi.org/10.1001/archinternmed.2012.3199 Google Scholar CrossRef Search ADS PubMed  3 Shanafelt TD, Mungo M, Schmitgen J, et al.   Longitudinal study evaluating the association between physician burnout and changes in professional work effort. Mayo Clinic Proc  2016; 91: 422– 31. http://dx.doi.org/10.1016/j.mayocp.2016.02.001 Google Scholar CrossRef Search ADS   4 Dewa CS, Loong D, Bonato S, et al.   How does burnout affect physician productivity? A systematic literature review. BMC Health Serv Res  2014; 14: 325. http://www.biomedcentral.com/1472-6963/14/325. Google Scholar CrossRef Search ADS PubMed  5 Anagnostopoulos F, Liolios E, Persefonis G, et al.   Physician burnout and patient satisfaction with consultation in primary health care settings: evidence of relationships from a one-with-many design. J Clin Psychol Med Settings  2012; 19: 401– 10. http://dx.doi.org/10.1007/s10880-011-9278-8 Google Scholar CrossRef Search ADS PubMed  6 Shanafelt TD, Balch CM, Bechamps G, et al.   Burnout and medical errors among american surgeons. Ann Surg  2010; 251: 995– 1000. http://dx.doi.org/10.1097/SLA.0b013e3181bfdab3 Google Scholar CrossRef Search ADS PubMed  7 Hobbs DFR, Bankhead C, Mukhtar T, et al.   Clinical workload in UK primary care: a retrospective analysis of 100 million consultations in England, 2007-14. Lancet  2016; 387: 2323– 30. http://dx.doi.org/10.1016/S0140-6736(16)00620-6 Google Scholar CrossRef Search ADS PubMed  8 Somville JF, Gucht V, Maes S. The impact of occupational hazards and traumatic events among Belgian emergency physicians. Int J Behav Med Nutr Phys Activity  2016; 24: 59. 9 Heponiemi T, Kouvonen A, Vanska J, et al.   Effects of active on-call hours on physicians' turnover intentions and well-being. Scand J Work Environ Health  2008; 34: 356– 63. http://dx.doi.org/10.5271/sjweh.1278 Google Scholar CrossRef Search ADS PubMed  10 Elovainio M, Salo P, Jokela M, Heponiemi T, Linna A, Virtanen M, et al.   Psychosocial factors and well-being among Finnish GPs and specialists: a 10-year follow-up. Occup Environ Med  2013; 70: 246– 51. 11 Ahola K, Hakanen J. Job strain, burnout, and depressive symptoms: a prospective study among dentists. J Affect Disord  2007; 104: 103– 10. http://dx.doi.org/10.1016/j.jad.2007.03.004 Google Scholar CrossRef Search ADS PubMed  12 Lallukka T, Chandola T, Hemingway H, et al.   Job strain and symptoms of angina pectoris among British and Finnish middle-aged employees. J Epid Comm Health  2009; 63: 980– 5. http://dx.doi.org/10.1136/jech.2008.085878 Google Scholar CrossRef Search ADS   13 Elovainio M, Heponiemi T, Kuusio H, Jokela M, Aalto A-M, Pekkarinen L, et al.   Job demands and job strain as risk factors for employee wellbeing in elderly care: an instrumental-variables analysis. Eur J Public Health  2015; 25: 103– 8. Google Scholar CrossRef Search ADS PubMed  14 Virtanen M, Vahtera J, Pentti J, et al.   Job strain and psychologic distress influence on sickness absence among Finnish employees. Am J Prev Med  2007; 33: 182– 7. http://dx.doi.org/10.1016/j.amepre.2007.05.003 Google Scholar CrossRef Search ADS PubMed  15 Vahtera J, Laine S, Virtanen M, et al.   Employee control over working times and risk of cause-specific disability pension: the Finnish Public Sector Study. Occ Envir Med  2010; 67: 479– 85. http://dx.doi.org/10.1136/oem.2008.045096 Google Scholar CrossRef Search ADS   16 House JS, Umberson D, Landis KR. Structures and processes of social support. Annu Rev Sociol  1988; 14: 293– 318. http://dx.doi.org/10.1146/annurev.so.14.080188.001453 Google Scholar CrossRef Search ADS   17 Barth J, Schneider S, von Känel R. Lack of social support in the etiology and the prognosis of coronary heart disease: a systematic review and meta-analysis. Psychosom Med  2010; 72: 229– 38. Google Scholar CrossRef Search ADS PubMed  18 Hakulinen C, Pulkki-Råback L, Jokela M, et al.   Structural and functional aspects of social support as predictors of mental and physical health trajectories: Whitehall II cohort study. J Epidemiol Commun Health  2016; 70: 710– 5. Google Scholar CrossRef Search ADS   19 Cohen S, Wills T. Stress, social support, and then buffering hypothesis. Psychol Bull  1985; 98: 310– 57. http://dx.doi.org/10.1037/0033-2909.98.2.310 Google Scholar CrossRef Search ADS PubMed  20 Van der Doef M, Maes S. The Job Demand-Control (-Support) Model and psychological well-being: a review of 20 years of empirical research. Work Stress  1999; 13: 87– 114. http://dx.doi.org/10.1080/026783799296084 Google Scholar CrossRef Search ADS   21 Bernburg M, Vitzthum K, Groneberg DA, Mache S. Physicians’ occupational stress, depressive symptoms and workability in relation to their working environment: a cross-sectional study of differences among medical residents with various specialties working in German hospitals. BMJ Open  2016; 6:e011369. 1. 22 Wang Z, Xie Z, Dai J, et al.   Physician burnout and its associated factors: a cross-sectional study in Shangh. J Occup Health  2014; 56: 73– 83. http://dx.doi.org/10.1539/joh.13-0108-OA Google Scholar CrossRef Search ADS PubMed  23 Linzer M, Gerrity M, Douglas JA, et al.   Physician stress: results from the physician worklife study. Stress and Health  2002; 18: 37– 42. http://dx.doi.org/10.1002/smi.917 Google Scholar CrossRef Search ADS   24 Bower P, Campbell S, Bojke C, Sibbald B. Team structure, team climate and the quality of care in primary care: an observational study. Qual Saf Health Care  2003; 12: 273– 9. http://dx.doi.org/10.1136/qhc.12.4.273 Google Scholar CrossRef Search ADS PubMed  25 Kivimaki M, Vanhala A, Pentti J, et al.   Team climate, intention to leave and turnover among hospital employees: prospective cohort study. BMC Health Serv Res  2007; 7: 170. http://dx.doi.org/10.1186/1472-6963-7-170 Google Scholar CrossRef Search ADS PubMed  26 Kivimäki M, Sutinen R, Elovainio M, et al.   Sickness absence in hospital physicians: 2 year follow up study on determinants. Occup Environ Med  2001; 58: 361– 6. Google Scholar CrossRef Search ADS PubMed  27 Elovainio M, Heponiemi T, Vänskä J, Sinervo T, Kujala S, Laakso E, et al.   Miten Suomalaisnen lääkäri voi 2000-luvulla. Suomen Lääkärilehti  2007; 62:2071–76. 28 Graetz B. Multidimensional properties of the General Health Questionnaire. Soc Psychiatry Psychiatr Epidemiol  1991; 26: 132– 8. http://dx.doi.org/10.1007/BF00782952 Google Scholar CrossRef Search ADS PubMed  29 Penninkilampi-Kerola V, Miettunen J, Ebeling H. A comparative assessment of the factor structures and psychometric properties of the GHQ-12 and the GHQ-20 based on data from a Finnish population-based sample. Scand J Psychol  2006; 47: 431– 40. Google Scholar CrossRef Search ADS PubMed  30 Jenkins CD, Stanton BA, Niemcryk SJ, Rose RM. A scale for the estimation of sleep problems in clinical research. J Clin Epidemiol  1988; 41: 313– 21. http://dx.doi.org/10.1016/0895-4356(88)90138-2 Google Scholar CrossRef Search ADS PubMed  31 Ilmarinen J, Tuomi K, Klockars M. Changes in the workability of active employees over an 11-year period. Scand J Work Environ Health  1997; 23(Suppl 1): 49– 57. Google Scholar PubMed  32 Elovainio M, Sinervo T. Psychosocial stressors at work, psychological stress and musculoskeletal symptoms in the care for the elderly. Work Stress  1997; 11: 351– 61. http://dx.doi.org/10.1080/02678379708252998 Google Scholar CrossRef Search ADS   33 Anderson NR, West MA. Measuring climate for work group innovation: development and validation of the team climate inventory. J Organ Behav  1998; 19: 235– 58. http://dx.doi.org/10.1002/(SICI)1099-1379(199805)19:3<235::AID-JOB837>3.0.CO;2-C Google Scholar CrossRef Search ADS   34 Klein J, Grosse Frie KG, Blum K, von dem Knesebeck O. Psychosocial stress at work and perceived quality of care among clinicians in surgery. BMC Health Serv Res  2011; 11: 109.http://www.biomedcentral.com/1472-6963/11/109. Google Scholar CrossRef Search ADS PubMed  35 Katon WJ, Lin EHB, Korff M, et al.   Collaborative Care for Patients with Depression and Chronic Illnesses. N Engl J Med  2010; 363: 2611– 20. http://dx.doi.org/10.1056/NEJMoa1003955 Google Scholar CrossRef Search ADS PubMed  36 Crawford GB, Price SD. Team working: palliative care as a model of interdisciplinary practice. Med J Aust  2003; 179: S32– 4. Google Scholar PubMed  37 Bettger JA, Stineman MG. Effectiveness of multidisciplinary rehabilitation services in postacute care: state-of-the-science. A review. Arch Phys Med Rehabil  2007; 88: 1526– 34. Google Scholar CrossRef Search ADS PubMed  38 Goh TT, Eccles M, Steen N. Factors predicting team climate, and its relationship with quality of care in general practice. BMC Health Serv Res  2009; 9: 138. http://dx.doi.org/10.1186/1472-6963-9-138 Google Scholar CrossRef Search ADS PubMed  39 Dyrbye LN, Varkey P, Boone SL, et al.   Physician satisfaction and burnout at different career stages. Mayo Clin Proc  2013; 88: 1358– 67. http://dx.doi.org/10.1016/j.mayocp.2013.07.016 Google Scholar CrossRef Search ADS PubMed  40 Torppa MA, Kuikka L, Nevalainen M, Pitkälä KH. Family physician experiences with and needs for clinical supervision: associations between work experiences, professional issues and social support at work. Patient Educ Couns  2016; 99: 1198– 202. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved.

Journal

The European Journal of Public HealthOxford University Press

Published: Jan 19, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off