Psychosocial health risk factors and resources of medical students and physicians: a cross-sectional study
© Voltmer et al. 2008
Received: 21 February 2008
Accepted: 02 October 2008
Published: 02 October 2008
Skip to main content
© Voltmer et al. 2008
Received: 21 February 2008
Accepted: 02 October 2008
Published: 02 October 2008
Epidemiological data indicate elevated psychosocial health risks for physicians, e. g., burnout, depression, marital disturbances, alcohol and substance abuse, and suicide. The purpose of this study was to identify psychosocial health resources and risk factors in profession-related behaviour and experience patterns of medical students and physicians that may serve as a basis for appropriate health promoting interventions.
The questionnaire -Related Behaviour and Experience "Work administered in cross-sectional surveys to students in the first (n = 475) and in the fifth year of studies (n = 355) in required courses at three German universities and to physicians in early professional life in the vicinity of these universities (n = 381).
Scores reflecting a healthy behaviour pattern were less likely in physicians (16.7%) compared to 5th year (26.0%) and 1st year students (35.1%) while scores representing unambitious and resigned patterns were more common among physicians (43.4% vs. 24.4% vs. 41.0% and 27.3% vs. 17.2% vs. 23.3 respectively). Female and male responders differed in the domains professional commitment, resistance to stress and emotional well-being. Female physicians on average scored higher in the dimensions resignation tendencies, satisfaction with life and experience of social support, and lower in career ambition.
The results show distinct psychosocial stress patterns among medical students and physicians. Health promotion and prevention of psychosocial symptoms and impairments should be integrated as a required part of the medical curriculum and be considered an important issue during the further training of physicians.
Health, stress and impairment of physicians have attracted increased attention in health care systems during the past years. Burnout rates in physicians in European and Anglo-American countries are estimated to be between 20 and 45% [1, 2]. In a survey among U.S. internal medicine residents, the prevalence was as high as 76% . Physicians as compared to the general population or other academic professions are also at elevated risk for other psychosocial health problems such as alcohol and substance abuse, marital disturbances or suicide [4–6]. There is increasing evidence that the development of psychosocial symptoms and impairment result from a complex interplay of several factors. Early longitudinal studies showed that personality traits in medical students evaluated before starting their course of study were important predictors for later impairment and burnout [7–9]. A highly demanding educational process and professional life frequently fostered by a competitive environment and a hierarchic culture of medicine are regarded additional relevant factors [4, 10, 11]. However, people respond differently to such challenges and strains. In a recent longitudinal study, Buddeberg-Fischer et al.  reported that while almost three quarters of evaluated Swiss physicians during residency showed no or decreasing stress levels, about 25% reported increasing or persistently high levels of stress as the extrinsic and 17% overcommitment as the intrinsic part of the effort-reward imbalance model. Individual perceptions of strain and coping styles are therefore important parameters for predicting whether health can be maintained or impairment will occur [13–16]. Research in health psychology revealed various attitudes, traits and behaviour patterns which either aggravate stress or have protecting effects as coping resources, e. g., perfectionism, compulsiveness, social support, optimism, hardiness, self-efficacy, and sense of coherence [17–19].
Tyssen et al. point out, that types of personality are more reliable in identifying those at risk than are dimensions of personality alone . It has also been reported that male and female students and physicians differ in their experience and response to stress and psychosocial symptoms [21–23].
Germany has no longstanding tradition in addressing physicians' health and impairments. Resch and Hagge , two German psychologists, note that care for physicians' health in Germany is still at the level of the USA in the '60's. The aim of this study was therefore to investigate work-related experiences and behaviour patterns in medical students and physicians in Germany at different points in their medical education and professional life. Based on previous work by Schaarschmidt & Fischer, we aimed to identify risk factors or health resources for subsequent psychosocial impairments. We used the personality typology described by these authors derived from the cluster analysis of eleven health relevant dimensions from the domains of professional ambition, resistance to stress and emotional well-being, which comprises four health relevant patterns (the healthy pattern G, the unambitious pattern S, the overcommitted risk pattern A und the resigned risk pattern B) [25–27].
Our hypotheses were:
A majority in first year students presents a healthy pattern.
Physicians are less likely to present healthy patterns and more likely to present risk patterns and less favourable scores in the health relevant dimensions compared to first and fifth year students.
Differences between male and female students and physicians were expected, in particular in the domains of professional ambition, resistance to stress and social support.
We conducted three cross-sectional surveys among medical students and physicians. In order to achieve a high response rate in students of more than one university, we recruited subjects at a teaching unit in required courses at three universities in three different federal states in Germany. Students in their first and fifth year of study were presented with the study goals and a questionnaire was administered. Participation was voluntary. The students were given sufficient time for completion and the questionnaires were collected immediately afterwards. Return after completion was deemed informed consent. There was no dependency relationship between students and researchers (e. g., evaluation, supervision). Physicians' data were gathered by means of a mail-in survey in the vicinity of the three universities. Addresses were selected with the support of the local Medical Associations or the university. Selection criteria were the number of postgraduate years in professional work-life. Because the study was associated with minimal risks and complied with the data protection rules, it was approved by the ethics commission of the University of Freiburg in a minimal risk review and exempted from a full formal evaluation.
Demographic characteristics of students and physicians
Students 1st year
n = 471
Students 5th year
n = 345
Physicians 3rd–8th year
n = 344
with partner (%)
In addition to a set of demographic questions, the questionnaire "Work-Related Behaviour and Experience Patterns" (AVEM)[25, 26, 28] was administered. The instrument has been developed to gather self-reported data about personal experiences with work-related stress and typical behavioural responses to cope with stress. In detail, the questionnaire covers the following three major domains: a) professional commitment, b) resistance to stress, and c) emotional well-being (in the context of work) which are assessed with 11 separate scales (Additional file 1). Each scale consists of 6 items with response options presented as a 5-point Likert-scale ranging from "I strongly disagree" to "I strongly agree". In order to determine an individual's pattern of behavioural response to occupational stress, it is necessary to go beyond the analysis of the single scales. Information about the configuration of the characteristics has a more predictive value than distinct parameters for each dimension . For example, high professional commitment in itself does not constitute a health risk. However, if a tendency to high professional commitment is coupled with reduced capacity to cope, the combination of these two traits may put the individual at a higher risk for developing health problems. In order to identify characteristic configurations of behavioural traits that help to estimate a person's health status and risk, respectively, Schaarschmidt & Fischer  analysed the data of the initial AVEM sample group (N = 1,598), a group comprising representatives of diverse professions. A cluster analysis revealed a four-cluster solution. The same cluster solution was replicated with sufficient concurrence in 10 random samples drawn from the entire sample of 1,598 test subjects (average κ > 0.80) . Scale reliability was assessed in samples of different professions. The median Cronbach's α as a measure of internal consistency was 0.81 (minimum: 0.79, maximum: 0.86). The validity of the measure was also supported by moderate to good correlations with scales measuring related constructs (e. g., FPI, MBI). The health relevance of the patterns was supported by expected correlations with a broad set of criteria such as depression, well-being, sickness, absence, blood pressure, heart rate, type-A behaviour or burnout.
This pattern represents a healthy attitude towards work. The individuals are ambitious at work, but also able to maintain emotional distance from work. They score high in the dimensions that represent resistance to stress and in all dimensions related to positive emotions.
A rather unambitious attitude towards work is characteristic for this pattern with the lowest scores in the dimensions describing commitment to work and highest scores in capacity for detachment. Nevertheless, low scores with a tendency to resignation and medium to high scores for inner balance, satisfaction with life, and the experience of social support suggest a generally positive experience with life. The challenge of this pattern is less in health than in promoting motivation.
In contrast to patterns G and S, the following two are repeatedly shown to be linked to illness. They therefore play a key role in preventing impairment and promoting health.
This pattern is characterised by excessive commitment to work and difficulties with emotional distancing from work. Limited coping abilities in stressful situations and negative emotions also characterise this exhausting pattern. Individuals with risk pattern A show many similarities to the concept of type-A behaviour described for coronary artery disease and myocardial infarction . High positive correlations have been observed for risk type A and type A behaviour.
Individuals with this pattern show low scores for the dimensions related to professional commitment. They attain high scores for the tendency to resignation and correspondingly low scores for emotional distancing and active coping. Their emotional status is characterised by low scores for balance and mental stability, satisfaction with work, and satisfaction with life, and shows limited experience of social support. This pattern represents the core symptoms of burnout syndrome and correlates with burnout scores measured with standard instruments (MBI) .
All analyses were carried out with SPSS for Windows release 10.0 (SPSS Inc., Chicago, IL, USA). Descriptive statistics are presented in terms of counts, percentages, means and standard deviations. For hypothesis (1), the relative fractions of the different behavioural patterns were analysed. For hypothesis (2) and (3), the differences between study groups in the AVEM-dimensions were analysed in a two-factorial MANCOVA using general linear model (GLM) estimations with AVEM-dimensions as dependent variables, study groups and gender as independent variables and marital status as covariate. Samples were not adjusted for age due to the high natural correlation between age and study groups. Differences in the categorical behaviour and experience patterns were tested with χ2 test. The Bonferroni method was used to adjust for multiple tests.
Health-relevant dimensions in men and women
(n = 464)
(n = 176)
(n = 288)
(n = 339)
(n = 113)
(n = 226)
(n = 341)
(n = 165)
(n = 176)
subjective significance of work
tendency to exert
striving for perfection
offensive coping with problems
balance and mental stability
satisfaction with work
satisfaction with life
experience of social support
The main purpose of our study was to identify psychosocial health risks and resources in the study- or work-related behaviour and experience in medical students and physicians. The AVEM questionnaire identifies not only important domains and dimensions of work-related individual behaviour, but also allows the identification of patterns that indicate serious health risks.
The majority of first-year students displayed the healthy and ambitious pattern G. However, it is important to note that different from our first hypothesis, this was not the predominant pattern: approximately one fourth of the students presented the overambitious risk pattern A. While this may have been expected taking into account the challenging education and professional life, one has to recall that this risk pattern represents a tendency to overexert and difficulties to distance oneself from work. A comparable fraction presented with the unambitious S type, and 17% of students with a resigned and burnout-related risk pattern B are determined as early as at the beginning of their studies. This finding is supported by other studies that describe a similar degree of psychological vulnerability at this early stage of study and emphasise the importance of personality traits for the development of stress and psychosocial symptoms [20, 30, 31].
In concordance with our second hypothesis, physicians presented lower scores of the healthy behaviour and experience pattern than both student groups, and correspondingly higher scores for the burnout-related risk pattern B. This confirms results from other researchers who describe an accumulation and worsening of effects in medical education and professional life over time [4, 32, 33]. In Swiss medical residents both men and women reported a significantly worse physical and psychological well-being as well as life satisfaction after their first year of residency compared to the time directly before their graduation from medical school . Schaarschmidt and Fischer  describe a correlation between risk pattern B and standard measurements of burnout (MBI , BHD ). Other researchers therefore applied the AVEM to assess burnout in relation to the psychopathological and psychosomatic symptom load (SCL90R). Risk pattern B showed the highest scores of psychiatric symptoms (in particular depression) in the SCL90R . It seems therefore appropriate to compare the scores for resigned risk pattern B of the physicians surveyed in our study with burnout rates reported elsewhere. The prevalence of risk pattern B among physicians in our study is quite similar to estimates of burnout in Germany in general . Lower scores have been reported for general practitioners in Switzerland , higher scores in a survey among Canadian physicians . The observed fraction of the healthy pattern G among physicians is among the lowest reported for other professions such as teachers, business founders or nurses [36, 38].
The most impressive result of this study is certainly the high fraction of unambitious pattern S observed in fifth-year students and physicians. In contrast to the popular stereotype of physicians being highly motivated high achievers in a challenging profession, this type is characterised by low self-perceived significance of work and professional ambition and high emotional distancing from work. Taking into account the lower values for the dimensions of emotional well-being, in particular in male physicians, this seems to reflect frustration and overwork. The high fraction of resigned risk pattern B among physicians (27%) which represents the core symptoms of the burnout syndrome supports this interpretation. While the relatively low response rate in physicians may have influenced this result, comparisons between responders and mail-in group do not indicate a strong selection bias. The low response rate of physicians might itself be an additional clue of demotivation and poorer coping. Other researchers confirm increasing demotivation even in younger physicians [39, 40]. The poor coping styles observed in the physician sample therefore suggest that both primary medical education as well as continuing medical training in physicians should target the emotional well-being and health concerns relating to the profession. This is emphasised by the fact that psychiatric morbidity in students and physicians is common but students and physicians are reluctant to seek help [9, 41, 42].
As hypothesised, females scored lower than males on career ambition in all study groups. Especially in fifth year students and physicians this may reflect differences in career aspirations and coping with professional experiences [43, 44] as well as efforts to combine family and professional life [45, 46]. It is striking that women scored lower than men particularly in the domain of resistance to stress that covers the dimensions resignation tendencies, offensive coping with problems as well as balance and mental stability. This is confirmed by other researchers who report higher scores of depression and worse mental well-being in women than in men in medical student or physician samples [47, 48]. Because the majority of medical students in Germany are currently female, this issue needs to be addressed by interventions that support effective coping.
Compared to both student groups, physicians presented lower scores in the dimension experience of social support. Withdrawal from social relationships in response to educational or professional stress in medical students and physicians has been reported by others [49–51]. We observed no significant gender-related differences in first-year students relating to social support but among fifth-year students and physicians, females scored significantly higher than their male counterparts. Since social support has been shown to be one of the most important supportive factors for dealing with stress and impairment in medical students and physicians [52, 53] and is even more important to women , this issue should be addressed in medical education and postgraduate training.
This study has also a number of limitations: firstly, we surveyed three distinct populations at different stages in their career, and thus the data are not longitudinal. While there is no indication for this, differences in patterns may be explained by the differences between populations rather than an outcome of career advancement. Secondly, the response rate among physicians is comparable to the one obtained in other studies, but remains low. We therefore cannot exclude the possibility that our results are affected by response bias. However, as there were no differences in the gender distribution and only small differences in age between responding and addressed physicians, this bias – if any – should be small.
In summary, our data show a consistently lower prevalence of the healthy pattern G in fifth-year students and physicians compared to first-year students, and a higher prevalence of the resigned risk pattern B. Differences between the student samples were observed in the domains of professional ambition and resistance to stress. Women show less favourable scores in the dimensions of professional ambition and resistance to stress, but score higher on social support.
A large number of studies, including the one presented here, demonstrate the potential health risks resulting from the strains and burden associated with the medical education and profession. It seems paradoxical, that the training and professional activity aimed at helping and healing others show little awareness towards potentially harmful effects, and little willingness to engage in the prevention of a declining health in candidates of the medical profession themselves. This is essential not only for the benefit of medical students and physicians, but also for the provision of high quality care to patients.
We are grateful for the generous support of our study by the Universities of Lübeck (Prof. Westermann) and Witten/Herdecke (Prof. Schnell) and the Medical Associations of Schleswig-Holstein and Southern Baden.
The valuable comments of two referees are appreciated.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.