Editorials

Managing depression in primary care

The type of treatment matters less than ensuring it is done properly and followed up

Primary care p 772

Several recent studies have evaluated alternative approaches to managing depression in primary care. The range of disease and the treatments examined have varied widely, no doubt contributing to the variation in results. Nevertheless, randomised trials leave little doubt that antidepressant drugs are efficacious in major depression, ^{1 2} and recent evidence suggests efficacy in dysthymia and subsyndromal depression as well.³ But what role does counselling play in the primary care management of patients with various forms of depression? Recent trials in primary care have produced conflicting results and conclusions.

The paper in this issue by Chilvers et al (p 772)⁴ and an earlier report from the same study⁵ address three important questions about treating major depression in primary care. Is there a difference in the effectiveness of drugs versus counselling? Is the non-standardised counselling provided by most mental health providers effective? Does matching treatment with patient preferences increase effectiveness? In Chilvers et al's study only the first question is addressed using a randomised design. Unfortunately, small sample sizes and difficulties in follow up urge caution in interpreting the results. Regarding the second and third questions, we must settle for non-experimental comparisons within this sample and with previous reports.

Chilvers et al conclude that generic counselling appears to be as effective as antidepressant drugs for major depression, though patients given drugs may recover more quickly. There may be differences in longer term effects as well. Tables 3 and 4 in the paper show that patients randomised to drugs were 16% more likely to have a "good" global outcome, 10% more likely to ever remit, and 30% less likely to be depressed by research diagnostic criteria. These differences in 12 month outcomes, none of which reached statistical significance, raise a conundrum. Are the differences between drugs and counselling in the randomised group large enough to have implications for practice?

Randomised controlled trials on both sides of the Atlantic now provide evidence that different approaches to counsellingcognitive-behavioural,⁶ interpersonal,¹ and problem solving² have equivalent efficacy to drugs in treating major depression. But in these studies the "talking therapy" is applied by protocol using specially trained counsellors who are often monitored for adherence to the protocol. Chilvers et al's study placed few constraints on either the drug treatment or the type of counselling other than that the counselling should be provided by an experienced mental health professional in six sessions. In effect therefore they compared non-standardised antidepressant use with non-standardised counselling by experienced mental health professionals in general practice. Because statistical tests showed no significant differences in effectiveness the authors conclude that generic counselling is effective. Recent comparisons of more rigorously applied non-directive and cognitive-behavioural counselling with usual general practitioner care among a broader range of depressed patients found both specific therapies to be better than usual care at four months but not at 12.⁷ This may suggest advantages for more specific, standardised counselling over more generic approaches. Only direct comparisons of generic counselling with more standardised, specific approaches will resolve this question.

As to the implications for practice, the results in the patient preference group may be relevant. Over two thirds of the patients refused randomisation because they preferred a particular form of treatment, and nearly two thirds of them preferred counselling. Both the high proportion of people with a preference and the high proportion of them preferring counselling are consistent with other recent findings. ^{7 8} Within the patient preference group there were no differences in outcomes between the groups treated with counselling or drugs. Thus, regardless of one's interpretation of the randomised results, patient selected counselling or drugs appear to be equally effective if the counselling is provided by an experienced therapist.

It remains possible that patients without preferences will have better long term outcomes with drugs under real world circumstances where follow up may be sporadic. The major differences between usual care and protocol driven care for depression are the assurance of adequate intensity of treatment, whether counselling or drugs, and the consistency of follow up. ^{9 10} The low rates of assessment at 12 months in this study illustrate the difficulties with follow up in everyday practice. When care is organised to assure intensity and continuity of treatment, then the totality of evidence strongly indicates no difference between specific counselling or drugs. Giving patients with major depression their choice of treatment and then assuring adequate intensity of treatment and follow up represent high quality care.

Edward H Wagner, director, McColl Institute for Healthcare Innovation. 
Gregory E Simon, investigator. 

Center for Health Studies, Group Health Cooperative, 1730 Minor Avenue, Suite 1300, Seattle, WA 98101, USA

Acknowledgments

GES has received research funds from Pfizer and Eli Lilly, EHW from Parke-Davies.