General Practice

Predicting stress in general practitioners: 10 year follow up postal survey

BMJ 1997; 315 doi: (Published 05 July 1997) Cite this as: BMJ 1997;315:34
  1. Jenny Firth-Cozensa, principal research fellow
  1. a Department of Psychology, University of Leeds, Leeds LS2 9JT
  • Accepted 21 November 1996


High levels of stress in general practitioners have been described in numerous cross sectional studies,1 but few have used longitudinal data to explore possible precursors that might allow early prevention or intervention. Such precursors may relate to previous signs of psychological distress or may involve individual differences, such as personality. The importance of a self critical or perfectionist disposition in predicting stress has been noted in preregistration doctors over a two year period,2 and Blatt and Zuroff have investigated its influence on depression.3 I followed up a group of general practitioners from their fourth undergraduate year to investigate the importance of early symptoms of stress and self criticism in predicting stress levels 10 years later.

Subjects, methods and results

During the autumn and winter of 1993-4, I sent a questionnaire on stress levels and work related factors to the 302 subjects who had been investigated for self criticism and stress as fourth year medical students in 1983-4. I received responses from 224 individuals (74%), of whom 131 were general practitioners. These general practitioners form the sample for this study. The design and measures used are described in detail elsewhere.2

In all, 43 (33%) general practitioners scored above threshold for stress symptoms. This proportion is considerably higher than in the general working population but lower than the 48% reported by Caplan,1 which might reflect either his older and more geographically concentrated sample or differences in assessment. Stress levels were not significantly correlated over the 10 years (r=0.15, P<0.1), and current stress levels were not correlated with hours worked in the past week (r=0.16, P<0.1). Self criticism as students, however, was highly correlated with current stress levels (r=0.34, P<0.0001), accounting for 12% of the variance.


It is clear from this that, although early stress levels are much less important, high self criticism is a strong predictor of stress symptoms over a long period during which environmental factors such as marriage, children, and several different jobs and homes are likely to have occurred—life changes that might have been expected to overrule the influence of disposition. Nevertheless, some work related stressors, such as tiredness, are likely to contribute independently to stress, whereas others might interact with self criticism to exacerbate its effect. Blatt and Zuroff, for example, have written about depression that is coloured by high self criticism, describing those affected as engaging in harsh self evaluation, striving for achievement, and having a strong fear of criticism.3 Such characteristics would undoubtedly be particularly difficult within a competitive, humiliating, or status-conscious work culture, which medicine sometimes is.4

This study shows that it should be possible to reduce stress symptoms in future general practitioners by recognising those students who may be vulnerable—for example, those in whom their tutors see signs of self blame in clinical discussions. High self criticism is a way of thinking, a cognitive style in which self blame occurs whenever things go wrong; it can therefore be changed by teaching how to allocate responsibility less destructively. This is not about blaming others, as particularly low self criticism is related longitudinally to having poor relationships with patients and colleagues4; rather it entails learning to judge events, both good and bad, more reasonably.5 This finding has important implications in terms of counselling interventions, both for undergraduates and for professionals, and a prevention strategy may also be used as part of the general curriculum, perhaps within the context of mistakes and accidents or as part of stress management.


Funding: NHS research and development mental health initiative.

Conflict of interest: None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract