Changes in the lifetime prevalence of suicidal feelings and thoughts among Norwegian doctors from 2000 to 2010: a longitudinal study based on national samples

Since 1994 the Institute for Studies of the Medical Profession in Norway has regularly surveyed a representative cohort of about 1,270 active Norwegian doctors with postal questionnaires. The cohort was supplemented with approximately 400 randomly selected doctors in 2000, and with another 250 in 2008. Over the whole period, approximately 400 doctors have left the cohort due to death, retirement or voluntary withdrawal. The present study is based on data from 2000 and 2010.

The project is in compliance with the declaration of Helsinki on Ethical Principles for Medical Research Involving Human Subjects adopted by the World Medical Association (http://​www.​wma.​net/​en/​30publications/​10policies/​b3/​index.​html). According to the Regional Committee for Medical Research Ethics, the study based on “Norwegian Physician Survey - A bi-annual prospective questionnaire survey to a representative sample of Norwegian physicians” is exempt from review in Norway, cf. §§ 4 of The Act. The project can be implemented without the approval by the Regional Committee for Medical Research Ethics (IRB 0000 1870). Additionally, approval for data protection of the bi-annual prospective survey among Norwegian doctors was obtained from the Norwegian Social Science Data Service (Reference 19521).

The response rate was 86% (1,385/1,616) in 2000 and 67% (1,014/1,520) in 2010. The number of respondents who answered all three questions on suicidal feelings was 1,253 in 2000 and 962 in 2010. 715 doctors answered both in 2000 and 2010.

Table 1 shows gender, age and job situation for all respondents in 2000 and 2010, as well for the 715 doctors who answered both in 2000 and 2010, and how they compare with the total Norwegian doctor workforce in 2000.

Table 2 shows the responses to the three first Paykel items in 2000 and 2010. There is a trend towards a lower prevalence of suicidal feelings. When we calculate the means based on assigning the value 0 for “never”, 1 for “hardly ever”, 2 for “sometimes” and 3 for “often”, paired t-tests among those who responded both in 2000 and 2010 show significant reductions for all items.

In 2000, 9.5% (117/1,236) of the doctors reported suicidal planning (item 4 of Paykel’s instrument), and 1.3% (16/1,247) had attempted suicide (item 5 of Paykel’s instrument) on one or more occasions in lifetime. These questions were not included in the 2010 survey.

Figure 1 illustrates how the responses on suicidal thoughts changed from 2000 to 2010 among the 715 doctors who answered at both points in time. The prevalence of the response alternatives “often”, “sometimes” or “hardly ever” decreased, while the alternative “never” increased.

Of the 185 respondents who gave different answers, more changed to lower than to higher seriousness: from “often” to “sometimes”, from “sometimes” to “hardly ever” and “never”, and from “hardly ever” to “never”.

In 2010, 7.7% of the doctors reported that they had sometimes or often had suicidal thoughts (Table 2). Table 3 shows a multivariate logistic model with having ever had serious suicidal thoughts as response variable, age, gender, specialty group, job satisfaction, subjective well-being, self-rated health, psychosocial stress, and having answered or not in 2000 as possible simultaneous predictors. Subjective well-being, self-rated health and psychosocial work stress were all independent significant predictors of ever having serious suicidal thoughts. Gender, age, job satisfaction, speciality or having answered in 2000 showed no effect in the model.

In the German survey from 2006, the response rate for the hospital doctors was 58.2% (1,917/3,295). Of these 1,903 answered all three questions on suicidal feelings and thoughts. In the Norwegian 2010 survey, data from 528 hospital doctors was available. The gender distributions were similar in Germany and Norway, with 61.7 (52.4 to 60.8) % and 56.6 (58.9 to 64.5) % males respectively. The German doctors were significantly younger with a mean age of 39.8 (39.4 to 40.2) years vs. 47.8 (46.9 to 48.5) years. The distribution of medical specialities – with exception of internal medicine – differed between the two countries. In Norway there were fewer doctors in surgery (13.9% vs. 30.1%), gynaecology (6.3% vs. 7.1%) and anaesthesiology (7.8% vs. 14.9%) and more in psychiatry (17.8% vs. 2.7%) and other specialties (14.2% vs. 6.2%).

Norwegian hospital doctors compared with their German counterparts reported quite similar prevalences of ever felt life was not worth living (46.0, 41.8 to 50.3% vs. 39.9, 37.7 to 42.1%), ever wished to be dead (23.7, 20.1 to 27.3% vs. 24, 22.1 to 25.9%), and ever had thoughts of taking one’s life (25.9, 22.2 to 29.6% vs. 22.0, 20.1 to 23.9%).

To further explore the differences between Norwegian and German doctors, we performed multivariate logistic regressions with dichotomized versions of ever felt life was not worth living, ever wished to be dead, or ever having thoughts of taking life as response variables (“never” or “hardly ever” = 0, “sometimes” or “often” = 1). Possible predictors were country, age, gender and speciality (surgery, internal medicine, anaesthesiology, gynaecology, psychiatry and others).

Country and age were not significant predictors in any of the models. Working in psychiatry was significant in all three models: ever felt life was not worth living (OR 2.1, 95% CI 1.3 to 3.3), ever wished to be dead (2.0, 1.3 to 3.3) and ever had thoughts of taking life (2.3, 1.4 to 3.7). Being female was a significant predictor in the models of ever felt life was not worth living (1.4, 1.2 to 1.7) and ever wished to be dead (1.6, 1.3 to 2.0).

The lifetime prevalence of ever had feelings of life not worth living, ever wished own death and ever had thoughts of taking own life among Norwegian doctors decreased from 2000 to 2010. In 2010, significant predictors of serious suicidal thoughts were lower subjective well-being, poor or average self-rated health and high psychosocial work stress, but not age, gender, speciality or job satisfaction.

The main strength of our study lies first and foremost in the representative dataset, making the results generalizable to the entire population of doctors in Norway. The longitudinal design is also a strength. Similarities in survey methods and comparable items from Paykel’s instrument, [6] demographics and speciality should also be pointed out. The response rates were fairly good, 86% in 2000 and 67% in 2010. This is higher than in a number of other doctor studies, [16, 18, 32] but do not rule out non-respondents bias.

There is of course the possibility of non-responding doctors having a lower prevalence of suicidal feelings. In one study from the US the non-respondents did not have suicidal ideation during the prior twelve months [18]. In our 2010 sample, we found no significant differences between doctors who answered in 2000 and those who did not in reporting serious suicidal thoughts, suggesting no non-respondent bias (Table 3).

That lifetime prevalence estimates of suicidal feelings are not consistent over time funds support in the literature [33]–[35]. We have shown that some doctors reported suicidal feelings ever in 2000, but never in 2010. For instance, 43.3% (91/210) of the doctors with suicidal thoughts in 2000 reported no suicidal thoughts ever in 2010 (Figure 1). To further explore possible recall bias in our study, we performed analyses on the 1.3% (16/1,253) doctors reporting one or more suicidal attempts ever in the survey in 2000. We found that four of these had since voluntarily withdrawn from the panel. Six of the remaining twelve doctors answered the question on suicidal thoughts in 2010, and two of these reported no suicidal thoughts ever. The other four reported suicidal thoughts “hardly ever” or “sometimes”. Thus, to the extent that suicidal feelings correlate with actual suicide attempts, the prevalence of suicidal feelings in our study may be underestimated due to recall bias, or other reasons for not reporting. This is important since the lifetime prevalence of suicidal thoughts is a widely used measure in psychiatric epidemiology [33].

Further limitations include the lack of items on subjective well-being, self-rated health and psychosocial work stress in the 2000 survey, and the lack of the fourth and fifth items of the Paykel’s instrument [6] in 2010, which might further elucidate our findings. Other specific elements in doctors’ personality, mental disorders, physical illnesses, and other personal risk factors might also be useful, [3, 16, 18, 36] but such data were not available for present study.

Because the prevalence of suicidal thoughts varies between cultures, [13] and the number of foreign doctors in Norway is increasing, [36] it is also important to include this perspective in further research on suicidal ideation and thoughts of doctors.

Differences in methodology limit direct comparisons with other studies. However, it is possible to point out some general tendencies in suicidal thoughts.

Compared with another study among Norwegian doctors in 1993, [20] doctors in our 2000 and 2010 samples reported lower prevalence of suicidal thoughts ever: 36.3 (31.4 to 41.2) % in 1993, 30.8 (28.2 to 33.4) % in 2000 and 25.4 (22.7 to 28.2) % in 2010 (Table 2). Since the 95% confidence intervals in 1993 and 2010 do not overlap, the decrease over this 17 year period is statistically significant. Compared with data on suicidal thoughts among Norwegian police officers in 2000 [21] (22.6%, 21.1 to 24.2) and operational ambulance personnel in 2005, [22] (22.8%, 20.4 to 25.2), the doctors seem to converge with these other groups.

US surgeons in 2008 had significantly fewer cases of ever having thoughts of taking one’s own life, with 14.9 (14.1 to 15.7) % [16]. In contrast, suicidal ideation or suicidal thoughts reported by German hospital doctors in 2006 were similar to our 2010 findings.

It is challenging to compare the risk factors for serious suicidal thoughts (Table 3), because the definitions of risk group and the methods of analysis vary considerably between studies. However, our results seem to be in line with earlier studies showing that doctors with job-related stress, [3, 17, 19] low quality of life [18] and subjective mental or physical health complaints [3, 16, 18, 20, 37] were more at risk for reporting suicidal thoughts, while job satisfaction or age [20] had no effect. While female doctors in a previous study were more likely to have experienced serious suicidal feelings, [20] we found no such differences. There are also other studies reporting gender similarities [3, 16, 23]. In a study with data from 1993, anaesthesiologists had higher risk for serious suicidal thoughts [20]. In our multivariate models with ever having suicidal ideation or suicidal thought as response variables, psychiatry was a significant predictor, while in the model with ever having serious suicidal thoughts, there was no difference between medical specialties, suggesting that seriously thinking on suicide is less dependent on specialty.

The decreasing trend in suicidal feelings among Norwegian doctors from 2000 to 2010 may reflect a number of factors. The doctors who answered both in 2000 and 2010 have of course grown ten years older, from mean age 42.6 to 52.6 years. Literature suggests that mental well-being is U-shaped over the life course among Europeans, with a minimum in the mid-40s [38]. However, the observed reduction in the lifetime prevalence of suicidal feelings is approximately the same in the cohort that grew 10 years as in the two cross-sectional surveys in 2000 and 2010, where both groups are age representative (Table 2).

Another possible explanation for the changes in reporting suicidal feelings might be recall bias [33]–[35]. Doctors may consciously or unconsciously forget such details from the past (Figure 1).

There is a relationship between mental health and work conditions such as workload, stress and control over work [3, 17, 19, 27, 39]. Thus, a further reason for a decreasing trend in suicidal feelings among Norwegian doctors might lie in their work conditions. Some health care reforms have been implemented during the last decade in Norway [40]. The reforms were, at least in certain groups, often met with fear of declining professional autonomy [40, 41]. However, studies with data from the last decade show that Norwegian doctors have enjoyed a stable and high level of life satisfaction, [42, 43] high and increasing level of job satisfaction [40, 43] and stable weekly working hours [44, 45]. The fraction who perceive their workload as unacceptable is relatively small and has not increased significantly among junior doctors (19.2% in 2000, 18.5% in 2008) or consultants (26.8% in 2000, 32.2% in 2008), [44] and is actually reduced significantly among general practitioners (38.1% in 2000, 25.5% in 2008) [45]. It is possible that these trends could have a bearing on the marginal reduction in reporting suicidal feelings among Norwegian doctors.

High levels of stress and mental health disorders have been associated directly or indirectly with unfavourable lifestyle like low physical inactivity, [46] smoking [47] and heavy drinking [48]. It has also been documented that heavy drinking increases suicide risk [49]. We have recently shown that the drinking patterns of Norwegian doctors have changed from 2000 to 2010 towards more moderate alcohol consumption, fewer episodes of heavy drinking and less alcohol-related problems [50]. We have also shown that smoking decreased from 2000 to 2010, [44] and that physical activity increased in the period 1993–2010 [51]. A study based on mortality data from 1960 to 2000 documents that doctors, compared with other population groups, had lower mortality from all causes, including lifestyle-related diseases, only with the exception of suicide [1]. However, another Norwegian study suggests that the decrease in suicide rate among female doctors from 1960 to 1989 follows the general population trend [52]. Other studies from Norway, England and Wales suggest that the differences between doctors and population in suicide rates are receding [2, 53]. The further development of suicide remains to be documented. In our study, ever having suicidal thoughts was approximately as common among Norwegian doctors in 2010 as among other human Norwegian service professionals like police officers and ambulance personnel.

In 2010, 7.7% of doctors had sometimes or often had thoughts of taking their own life (Table 2). Unfortunately, we do not have data on whether serious suicidal thoughts were mainly attributed to personal, social, familial or professional factors. We have shown that both work-related factors, like high psychosocial work stress and poor self-reported health and low life satisfaction were associated with serious suicidal thoughts (Table 3), suggesting an unfortunate combination of several factors. However, work-related factors might be less important that what was found in previous studies on doctors, [20] police officers [21] and ambulance personal [22] in Norway. There is also evidence that German doctors, compared with their colleagues in Norway, had lower job satisfaction, longer working hours and higher work-related stress [30, 54, 55]. In the sample of German and Norwegian doctors, country was not a significant predictor in any of the multivariate regression models with ever having suicidal ideation or ever having suicidal thoughts as response variables, suggesting a low importance of cultural factors, at least between these two countries.

In terms of health care policy, regular preventive screening for mental health problems might be a possible strategy. Use of period prevalence through longitudinal follow-up rather than lifetime prevalence might increase the validity of the screening [33]. A recent study among university hospital doctors in Sweden and Italy shows that the majority of doctors with signs of psychological distress (78.3%) or with recent suicidal thoughts (106 of 155) did never seek professional help [56]. Since self-treatment for mental problems among doctors is common, [57] those with suicidal thoughts should be encouraged to seek psychiatric help. In Norway, initiatives financed by the Sickness Compensation Fund for Doctors (SOP) provide special programs for doctors with mental health problems [57]. Since psychosocial work stress is positively associated with serious suicidal thoughts for doctors, as shown in this study, their work organisations should constantly aim for a reduction in such stressors.