Suicidal thoughts are believed to be part of a set of suicidal behaviours that are most serious when they are associated with suicidal plans and actions. Suicidal thoughts thus represent one stage of a continuum that begins with low mood, progresses towards passive death wishes, suicidal thoughts and finally suicidal actions [1–3]. Among individuals who harbour suicidal plans, the probability of actually undertaking a suicide attempt was, according to Scocco (2008), nearly 50 percent [3].

The healthy worker effect refers to the consistent tendency of the actively employed to have a lower mortality rate than the population at large due to selection effects [4]. However, this does not apply to physicians when it comes to death by suicide. Compared with other occupations [5, 6], other university graduates [7] and people in general [8, 9], suicide and suicidal thoughts are significantly more common among physicians. The mortality rate by all causes of death except suicide is however lower for physicians than for the general public [10], which is in line with the healthy worker effect.

To prevent suicides among physicians, it is essential to identify potential risk for, and protective factors against, suicidal thoughts. The present study focuses on the association between work-and health-related factors and suicide ideation among resident and specialist university physicians. The difference between these work positions involves specialists having greater responsibility at work than residents, whereas the work of residents is inherently of a more dependent nature. This hierarchical difference between the two groups makes separate analysis of work-related correlates of suicide ideation necessary.

Among physicians who commit suicide, cases associated with the recent death of a friend or family member or recent crises are less common than for the general public [11]. Instead, it is more likely among physicians for a problem at work to have contributed to the suicide, as was shown in a study concerning the details of suicide from 17 different states in the USA during the years 2003-2008 [11]. Time pressure, work environment, communication, career and goal achievement have previously been found to be work-stressors predictive of high levels of dissatisfaction at work for general practitioners [12].

At university hospitals, physicians’ patient-work is entwined with an academic career. Because of this, the work climate at university hospitals is often extremely competitive and demanding of a high degree of commitment [13]. The presence of a competitive work climate may in turn lead to an oppressive kind of interaction amongst colleagues, or even harassment [14, 15]. In a study on Italian and Swedish female physicians, harassment at work was found to be one of the variables that had the largest association with suicidal thoughts [16]. The same study also showed that having meetings to discuss demanding experiences at work was associated with a lower risk of suicidal thoughts among both Italian and Swedish female physicians [16]. Harassment was also found to be related to suicidal thoughts among Italian and Swedish male university physicians, but control over one’s own work seemed to decrease the risk of harbouring suicidal thoughts [17]. In line with this, work autonomy has been seen to have positive effects on the wellbeing of physicians in Sweden, Norway and Iceland [18]. Control over one’s own work was furthermore emphasized in a study of Finnish anaesthesiologists at a university hospital. Workers’ ability to make decisions concerning when and how to apply their own skills, along with organisational justice, were the most important determinants of work-related wellbeing [19].

Residency is the period following internship during which a physician undergoes further clinical training, usually in one medical speciality under the supervision of a specialist. A number of studies have indicated that residents’ psychological wellbeing may also be at risk. The level of depressive symptoms moreover seems to increase with time in residency according to a study using the Patient Health Questionnaire [20]. Scores of depressive symptoms increased from 2.38 at baseline (before entering internship) to 6.26 after 12 months of internship Another study found the lifetime prevalence of suicidal thoughts among residents to be 43.1%, with perceived job stress being one of the factors related to the occurrence of suicidal thoughts during the first postgraduate year [21]. Additionally, perceived medical school stress has been shown to be a predictor of postgraduate mental health problems [22]. However, findings regarding this have been somewhat conflicting, as a more recent study showed that lower perceived medical school stress was a predictor of a transition from suicidal thoughts to suicidal planning after graduating [23].

Moreover, medical residents who changed their position (that is, residents who became specialists) experienced an increase in job satisfaction [24]. Decreased job insecurity was believed to be an explanation for this, which suggests that career stage can have an influence on wellbeing at work.

Younger age is in itself a risk factor for suicidal ideation [25, 26], and residents are as a group relatively young, so age may be an explanatory factor. In keeping with this, psychological distress has been found to be significantly more common among younger doctors and residents than among their older colleagues [27]. Furthermore, not having children or having just one, compared with having more than one child, is a demographic factor that has been found to be related to work distress among younger doctors [28]. However, this has not been associated with suicidal thoughts among specialists [16, 17].

Another possible correlate to suicidal thoughts among residents is sickness presenteeism, that is, working even when sick, which has been found to be common among residents [29]. Sickness presenteeism is also a well-known problem among specialists [30], as it has been seen to be a risk factor for health problems among physicians [31], as well as among the general public [32]. Particularly, the risk for mental health problems is larger among those who work despite being sick [32].

Furthermore, emotional distress is reported to affect physicians’ patient work through an increased irritability, inability to concentrate and reduced work capacity [27]. A way of coping with feelings of exhaustion may be by distancing oneself from patients [33]. Work disengagement can be defined as an uncaring, distanced and cynical attitude toward the work itself or the people with whom one works. It has been found to promote demanding and threatening patient behaviour, which can lead to a vicious cycle as the patient-physician relationship deteriorates [33].

According to previous research, part of the explanation for the greater levels of mental illness, suicidal thoughts and suicides among physicians compared with those in the general public could lie in factors related to their work situation [11]. In addition, the work conditions, hierarchical positions and socio-demographics of residents and specialists are inherently different. Accordingly, their work-related health differs, as indicated by previous research [24, 27]. At university hospitals, the different work situations of residents and specialists may be of even greater importance owing to the higher level of engagement that is required at these hospitals [13]. Furthermore, a more competitive work climate may most adversely affect the residents, who have a weaker position at such workplaces. As residents represent the younger group of doctors upon whom we as a public are dependent for care in the future, focusing on this group is especially important. However, as they are still at the starting point of their careers, they may also be more susceptible to interventions. It is essential that the work situation of physicians is acceptable so that psychological distress does not hinder them from caring for their patients, as well as for themselves. Unfortunately, there is a lack of studies that compare residents and specialists with regard to work-related correlates of suicide ideation. For proper examination of the work factors that could protect against, or predict, suicidal thoughts among physicians in different hierarchical levels within their profession, there is a need to study these two groups of physicians separately.

The aim of the present study is thus to investigate the possible correlates of suicide ideation among university hospital resident and specialist physicians separately. For this purpose, logistic regression analysis of variables related to suicide ideation will be performed. It is hypothesized that harassment, sickness presenteeism, work disengagement, exhaustion and the need to demonstrate competence are associated with an increased level of suicide ideation. Control of one’s own work, an empowering leadership and the organisational culture, as well as having meetings to discuss demanding experiences at work, are hypothesized to be associated with a lower level of suicidal ideation. We want to investigate if there is variation in these relationships between residents and specialists.

The present study used cross-sectional data from physicians working at a university hospital in Sweden. The data collection in 2005 was organised through a web-based survey by the Department of Research and Development at St Olav’s University Hospital, Trondheim, Norway. The study was presented to the physicians as a research program concerning work-related health, organisational culture, career paths and working conditions in university hospitals. Associate professor, psychologist and psychotherapist Ann Fridner, and professor in psychiatry KG Götestam administered the questionnaires. All participants used personal log-in information to enter their responses to the web-based survey anonymously. Participation was voluntary and the act of logging in served as informed consent. The questionnaire was written in English, which can be considered as a nearly congruent language among university hospital physicians in Sweden. Familiarity with the English language is a prerequisite for completion of medical education, since nearly all of the texts in the medical education program are written in English. The questionnaire took 15-25 minutes to complete and contained 116 questions. Four reminders were sent via e-mail, and a paper version was also sent to all of the physicians. Confidentiality was guaranteed and a website was provided that displayed further information. The Regional Ethics Board in Stockholm examined and approved the project (reference number 04-913/2).

The outcome variable suicidal ideation was assessed by numerical counts and percentages. χ ² tests were used to identify differences between residents and specialists. There were no significant differences in the combined outcome variable: recent suicidal thoughts and/or recent thoughts of suicidal actions (Table 1).

Each of the demographic variables was also analysed in terms of its separate connection to suicide ideation, with results showing that there were no significant associations between age, civil status and suicide ideation. However, having children was associated with a lower level of suicide ideation among residents (OR = 0.63, 95% CI: 0.41-0.97, p = 0.34).

The aim of the current study was to investigate separately possible correlates of suicide ideation among resident and specialist university physicians. Comparisons between the participating resident physicians and specialist physicians revealed significant differences with respect to demographic characteristics and work-related factors. There was however no difference between the two groups in terms of the prevalence of recent suicide ideation.

According to previous research, autonomy at work is an important factor for wellbeing and work-related health [18, 19], including suicidal thoughts [17]. However, the findings of the present study do not support the hypothesis that control over one’s own work is associated with a lower level of suicide ideation, as this factor was not related to suicidal thoughts for either of the two groups of physicians. Even though residents perceived significantly less control over their own work than specialists, there was no connection between this variable and suicidal thought. Exhaustion was also not related to suicidal thoughts in the present study. This lack of correlation was unexpected as exhaustion is considered to be one of the core elements of burnout [33], which in turn has been identified as being connected to suicidal thoughts [38].

A positive organisational culture and the need to demonstrate competence were, in the present study, regarded as concepts related to organisational justice, which has previously been shown to be one of the most important determinates of work-related wellbeing [19]. Previous research has furthermore shown that the work climate at university hospitals is highly competitive [13]. However, the hypothesis that a positive organisational culture would be associated with a lower level of suicide ideation among the participating physicians was not supported in the present study. The hypothesis that the need to demonstrate competence would be related to suicide ideation also failed to gain support. As hypothesized in the present study, having been subjected to harassment was associated with a higher level of suicide ideation among specialists. However, this did not apply to residents, for whom there was no association between harassment and suicide ideation. According to previous research, harassment is a major risk factor for suicidal thoughts among physicians [16, 17], including medical students [39], as well as for the general public [40]. The association found in the current study between harassment and suicide ideation for specialists alone might suggest that the negative impact of harassment increases with time and/or career advancement. Perhaps this is because competition regarding positions and research funding also increases as one ascends through the hierarchy, as shown by Fridner (2004) [13]. To be involved in research for more than 50% of one’s working hours, which is limited to only a small number of physicians high up in the hierarchy, has also been found to increase the level of harassment experienced [41].

An empowering leadership was associated with a lower level of suicide ideation among residents, but not among specialists. That an empowering leadership is not associated with suicide ideation for specialists could be explained by the differing positions that the two groups of physicians hold. Specialists are more autonomous than residents, as they are more skilled, and are hence not as dependent upon their bosses. The results obtained in this study might thus be understood to support the assertion that different kinds of leadership are needed for different kinds of worker, as suggested in the Hersey and Blanchard model (1969) [42], in which the leadership style matches the maturity and skill of the employees. For example, if the maturity level of the employees is high, the model suggests a delegating style of leadership with minimal guidance from the leader, whereas the leadership of less mature employees should be of a more supportive kind [42]. Assuming that this leadership approach was applied at the current hospital, this could explain why specialists do not seem to benefit from encouraging leadership, while residents do. The result of the present study is encouraging as it indicates that certain important individuals at a workplace can have a major positive impact on residents.

Having frequent meetings to discuss demanding experiences at work was associated with a lower level of suicide ideation among specialists before and after adjusting for demographic variables. But among residents these meetings were not significantly associated with a lower level of suicide ideation when we adjusted for demographics. Previous research shows that the prevalence of frequent meetings to discuss demanding situations at work is associated with a lower level of suicidal thoughts among female physicians [16]. That residents do not seem to benefit from meetings in this respect may entail that they perceive barriers in engaging in discussions about difficulties that they are having at work, something that could possibly be due to job insecurity. To a significantly higher degree than specialists, residents find it necessary to demonstrate competence at work. This might explain why they do not seem to be relieved by talking about demanding experiences. That is, it is possible that residents fear that engaging in a discussion about their difficulties at work could make them seem incompetent. Personal vulnerability and a feeling that the need for supervision and support could indicate incompetence have previously been reported as factors hindering the implementation of successful group-based supervision among psychiatric nurses [43]. In consonance with this are findings showing that group-based clinical supervision is experienced as difficult because of uncomfortable exposure to one’s peers and superiors [44]. Perhaps individual supportive talks would be a more beneficial alternative for residents, as the ones in most need of guidance and support might also be the ones who find it most difficult to discuss problems openly. In other words, residents may feel freer to talk about difficulties with one or a few trusted individuals, instead of an entire work-team.

Sickness presenteeism appeared to induce an elevated risk of suicide ideation in both groups. This result is in consonance with previous research [31, 32], demonstrating a connection between sickness presenteeism and poorer health. Working while sick is believed to be part of a larger behavioural pattern where university hospital physicians neglect their own health. Factors associated with a competitive climate and myths about the healthy doctor are assumed to contribute to this [31]. In addition, the current study further demonstrates the highly detrimental correlates of work disengagement, known as one of the core elements of burnout [33], as it seems to be closely related to recent suicide ideation.

The demographic factors of age and civil status have previously internationally been shown to be connected to suicidal thoughts as an older age and being married are protective against suicidal thoughts among the general public [25, 26]. Contrary to this, in the present study, no connection was found between age and civil status and suicide ideation. Only number of children had a significant association with recent suicide ideation, and only among residents, for whom the estimated risk of having had recent suicide ideation seemed to be decreased by having children. This finding is in line with previous research showing a connection between fewer children and work distress among younger physicians [28]. It is possible that children have a positive effect on the wellbeing of residents because they are a distraction from the supposedly negative focus on competitive work and intense medical training.

Suicidal ideation was not associated with work control, organisational culture, the need to demonstrate competence and exhaustion. However, harassment at work, having meetings to discuss demanding experiences, empowering leadership, sickness presenteeism and work disengagement were all related to suicidal ideation. The current study further demonstrates that the work-related factors that are most important for the occurrence of recent suicide ideation among physicians largely seem to depend on position. Harassment at work, having meetings to discuss demanding experiences at work and an empowering leadership were related to suicidal ideation in different ways for residents and specialists.

Physicians’ thoughts of suicide and/or thoughts of suicidal actions are alarming as this group may be more aware of effective methods for actually performing these actions than the general population. Establishing which factors are related to suicidal thoughts is central to the process of preventing suicides among physicians. The current study demonstrates that these factors differ depending on one’s employment position within the medical hierarchy, suggesting that, when dealing with work-related distress among physicians, residents and specialists require separate interventions. However, this study should be extended by obtaining longitudinal data on the work-related predictors of suicide ideation among resident and specialist university hospital physicians. Future similar studies are recommended to stratify the subjects by medical speciality.