Effectiveness of selective serotonin reuptake inhibitors

[The authors respond:]

We selected leaving the study early for any reason as the primary outcome for our systematic review1 because patients frequently stop taking or change their antidepressant medication. In Italy, for example, a recent survey showed that of more than 2800 adults observed for 6 months after receiving their first antidepressant prescription, 60% received only occasional prescriptions after their first one.2 We therefore reasoned that treatment adherence might represent a clinically useful outcome measure in meta-analyses of randomized controlled trials, as we thought that under experimental conditions this outcome might integrate patients' and clinicians' judgments of efficacy, safety and tolerability into a global measure of effectiveness and acceptability. Similar reasoning was recently used in a clinical trial of antipsychotic drugs.3

We acknowledge that this outcome measure may only offer a “down-to-earth” evaluation of a drug's effectiveness and acceptability, but this limitation can be seen as a strength in a field of research where efficacy is typically quantified as a score on a rating scale: in clinical practice, physicians seldom define patient improvement with rating scales.

For patients with moderate to severe major depression, one of the first goals is to keep them on treatment. Therefore, the main clinical question of our systematic review was whether paroxetine is better than placebo at keeping patients on treatment. Staying on treatment can also be seen as a hard measure with little measurement error. In addition, we investigated the effectiveness of paroxetine in those who actually took it and we also used standard measures of depression.

The main clinical message of our analysis is that the effect of antidepressants in patients with moderate to severe depression is modest. Physicians should consider combining pharmacologic and nonpharmacologic treatments such as psychological and psychosocial interventions backed by scientific evidence.4^,5 Similarly, patients should not receive the message that modifications of thought, mood and behaviour can be achieved by pharmacologic means only.