Mild depression in general practice: is the automatism of antidepressant prescribing an evidence‐based approach?

Most patients with depressive episodes receive care in the general practice, where depression is intended in terms of fluctuating mood disturbances occurring in response to life situations and chronic difficulties, physical illness, insecure relationships and deprivation (1). Most such patients in this setting just reach the minimum diagnostic criteria for major depression and are described as having mild depression (2). Rambelomanana et al. (3), who published, in this issue of Acta Psychiatrica Scandinavica, a survey describing general practitioners’ (GPs) opinions on antidepressant drugs and their prescriptions to depressed patients, showed that the degree of coherence between prescriber opinions (prescription intentions) and real practice for actual patients was substantial. However, little coherence can be inferred between actual prescriber practices and the scientific evidence that, theoretically, should guide physicians in reaching decisions about optimal care. In this survey all general practice patients with depressive symptoms were treated with antidepressants. Surprisingly, however, this practice is not backed by strong evidence. Only a minority of randomized controlled trials investigating the efficacy of antidepressants in depressed patients was carried out in the general practice and these trials typically enrolled patients with major depression of at least moderate severity. A recently performed systematic review of trials investigating whether antidepressants are effective in general practice patients with major depression identified only 10 studies comparing tricyclic antidepressants (TCA) with placebo, three comparing selective-serotonin reuptake inhibitors (SSRIs) with placebo and two comparing both drug treatments with placebo (4). Studies were short-term, only 535 patients received treatment with TCA and 552 with SSRIs (in the efficacy analysis), and outcome measures typically included rating scales with items concerning sleep and anxiety, which could have influenced the comparisons, and with other psychometric problems (5). In addition to these limitations, in most such studies scores on rating scales were dichotomized and results were presented in terms of proportion of responders. Although from a practical viewpoint this seems very reasonable because it allows physicians to make reasoning in terms of proportion of patients (and not in terms of means and standard deviations), it has been noted that this approach systematically magnifies the effect of new medicines against placebo (6). Additionally, in terms of continuous outcome measures, only a modest effect for TCAs [for depression scores the standardized mean difference versus placebo was −0.42 (95% confidence interval −0.55 to −0.3)] was observed, while no data could be summarized for SSRIs on a continuous outcome. In general, a standardized mean difference of at least 0.5 is required for a claim of medium effect size. Even assuming that most patients included in this analysis were suffering from mild depression, a balanced interpretation of these results suggests that the efficacy of antidepressants in the treatment of general practice patients with mild depression is sparse and inconclusive. A second interesting aspect emerged from the Rambelomanana et al. (3) survey is that SSRIs are considered more effective than TCAs. Again, this is not supported by evidence. MacGillivray et al. (7), who carried out a systematic review of trials comparing SSRIs with TCAs in the general practice, showed that, although only 11 studies of variable quality were identified, no differences could be detected, in terms of efficacy and in the short-term, between the most commonly prescribed classes of antidepressants. If anything, the contrary seems true, that is old antidepressants have the edge over SSRIs in terms of efficacy, as shown by systematic reviews of trials conducted in patients with major depression recruited in specialized settings (8, 9). Additionally, methodological considerations suggest that comparative trials tended to exaggerate the effect of SSRIs over TCAs (10) and that the confounding role of sponsorship might have biases many comparisons (11–13). These considerations lead to a crucial point. The scientific evidence that should guide GPs in reaching decisions is not a powerful determinant of prescriber opinions and practices. In the field of antidepressants, in recent years, opinions and practices have been more effectively influenced by a strong marketing campaign that succeeded in decreasing the threshold of antidepressant prescribing and in inducing the choice of new antidepressants over old ones. Robust evidence indicates that frequent contacts with drug representatives are significantly associated with a greater willingness to prescribe new drugs and to agree to patients’ requests to prescribe a drug that is not clinically indicated (14–16). Although some qualitative studies exploring how GPs decide to prescribe antidepressants suggested that the preferred strategy, theoretically, is to ‘wait and see’ (17), the dramatic increase in antidepressant sales and consumption (18, 19) indicates that prescriptions have progressively become automatic answers to patients’ depressive symptoms. This automatism, unfortunately, has negative consequences. First, it induces a passive attitude among GPs, who tend to consider antidepressants as their only therapeutic strategy, therefore not taking into consideration the possibility of developing more specific and individualized treatment plans, including, for example, watchful waiting or psychological/psychosocial interventions backed by scientific evidence (20). Second, a passive attitude is inevitably induced in patients, who receive a message suggesting that modifications of thought, mood and conduct can be achieved by pharmacological means only. Action is required to tackle these trends. At the individual level GPs should consider that the efficacy/safety profile of antidepressants has been demonstrated in adults with moderate to severe depression only, and that extrapolating these data to other patient populations, such as individuals with mild depressive symptoms, or adolescents with psychological problems, might be inappropriate. Psychotherapies should be considered as treatment of milder depression or as an adjunct to antidepressant drugs in more severe illness (20). At the epidemiological level, we need more data collected under real word circumstances to establish the effect of these prescribing practices on the course and outcome of depression in the general practice, in order to ascertain the degree of coherence between beneficial and adverse outcomes measured in experimental conditions and outcomes measured in real-word settings.