Do general practitioners' attitudes towards depression predict their clinical behaviour?
Background. GPs' attitudes towards depression vary, as do their abilities to detect and manage it effectively. Associations between attitudes and clinical behaviour have not yet been demonstrated directly. We tested two hypotheses: (1) that questionnaire measures of GPs' confidence in identifying depression predict their ability to identify depression in their patients; and (2) that GPs who prefer antidepressants prescribe more than those who prefer psychotherapy.
Methods. Forty GPs in Liverpool and Manchester completed the Depression Attitude Questionnaire (DAQ) and were asked for prescribing (PACT) information. Attender surveys using the General Health Questionnaire (GHQ-12), in combination with GP ratings of patients' psychological status, generated indices for GPs' case identification, bias and accuracy. We tested associations between these indices and the four DAQ components, in particular GPs' confidence in diagnosis, across a total of 1436 patients. We also compared the DAQ component on attitudes to treatment with relevant PACT data.
Results. The DAQ assessment of GPs' ability to identify cases of depression bore no relationship to their observed ability, as measured by accuracy, bias, or identification indices. However, there were significant associations between observed diagnostic ability and: preference for psychotherapy; ease in managing depression; and, belief in successful treatment. PACT data were available for 26 (65%) GPs. There was an association between preference for antidepressants and prescription of SSRIs (rs 0·3981, P < 0·044), but not for overall antidepressant or tricyclic prescribing, or for dose of dothiepin.
Conclusions. The DAQ measure of ease of identification is not valid when compared to actual clinical practice. The ability of GPs to identify depression may not be an independent variable, but may rather reflect other beliefs, attitudes and skills. This has considerable implications for educational interventions in primary care.
c1 Address for correspondence: Professor Christopher Dowrick, Department of Primary Care, University of Liverpool, Liverpool L69 3GB.