Depression undertreatment: lost cohorts, lost opportunities?

Depression is in the news! In this country the Defeat Depression Campaign (Royal College of Psychiatrists, 1993) launched by the Royal Colleges of Psychiatrists and of General Practitioners is both consensus driven and media directed; to this may be added the emergence of consensus statements on depression treatment elsewhere in the world. Suicide reduction as a national health outcome target (Department of Health, 1992), in part through the more effective treatment of mental illness and particularly depression, is further evidence of a growing interest in this topic outside psychiatry. Apart from the enormous disability and economic cost of depression, its well established treatability makes it a priority topic for attention in any public health strategy programme that aims to deliver efficiency (financial) as well as efficacy (clinical, social and economic functioning). The welter of evidence from randomized control trials (RCTs) and other relevant data on efficacy has recently been digested by multidisciplinary groups given the task of achieving consensus statements on management. Clearer guidelines have emerged on what clinicians should do (Paykel & Priest, 1992; American Psychiatric Association, 1993; British Association for Psychopharmacology, 1993) in order to make depressed patients better (recovery, remission), to remain better (continuance and recurrence prevention), to avoid a future relapse (maintenance or new episode prevention), and on how to manage a failure to respond (Depression Guideline Panel, 1993). If the desired targets are to be achieved, the question that public health physicians must ask and which epidemiologists should be able to answer is 'Does existing treatment also work as well as one would expect in routine clinical practice in the "real" world?' The following fictitious conversation may serve to illustrate the question. A psychiatrist and an epidemiologist, formerly undergraduate students together, chancing upon one another on the train and both travelling to their respective national advisory committees soon engage in conversation about psychiatric claims for obviously realizable public health gains. There is a debilitating condition (major depression) affecting approximately 5% of the adult population; a choice of treatment options, relatively inexpensive (indeed free in most industrialized countries), increasingly safe, and if necessary with very simple instructions (take one when going to bed); and in RCTs treatment typically increases the recovery rate after 4-6 weeks from 1/3 to 2/3 of cases. The epidemiologist than asks: 'How much has the prevalence fallen since treatments became widely available?'. The psychiatrist, pausing for a moment, points out there are problems: poor case detection (an important but separate issue), high relapse rates and difficulties with achieving consistent prevalence estimation. But even allowing for these, it would be hard to say that prevalence has been noticeably reduced. The epidemiologist acknowledges that the incidence could be rising for other reasons and any secular trends may be too gradual for an effect to be detectable, so what evidence is there from prospective cohort surveys and data base records of routine practice that those treated have a better outcome than the untreated, as you would expect? The topic of conversation drifts elsewhere, the psychiatrist vowing to chase up the literature on that.... It is clear that retrospectively identified cohorts may be unreliable because patients who ' dropped out' following a decision to treat depression may fail to be identified and included. Reports from treatment services will have to be judged on how well they describe outcome in relation to the total pool of referrals originally received; ad hoc follow-up studies of subjects originally enrolled in