Letters

Problem solving treatment for depression

Does the paper really prove that problem solving treatment is helpful?

Study should have included placebo group

Cost effectiveness is not clear

Author's reply

Does the paper really prove that problem solving treatment is helpful?

EDITORPlease tell me if I am missing something, but I am not convinced that the paper by Mynors-Wallis et al shows that problem solving is an effective treatment for depression.¹ As I read it, patients were allocated to one of four groups. Of the two groups treated with problem solving alone, up to a quarter (10 of 39 treated by the doctor and six of 41 treated by the nurse) withdrew from the trial because the treatment was not working. Those left in the trialthat is, those for whom the treatment was workingwere compared with those given antidepressant treatment, and it was found that the treatment was working. I note that none of the antidepressant group withdrew because the treatment was not working.

It may well be that problem solving is helpful for depressed patients. It may well be that an hour's initial treatment followed by up to six sessions of half an hour is helpful for depressed patients. But I am not convinced that this paper proves it.

Charles West, general practitioner. 
Medical Centre, Church Stretton, Shropshire SY6 6BL gorswen{at}dircon.co.uk

Study should have included placebo group

EDITORMynors-Wallis et al conclude that problem solving is an effective treatment for depression.¹ They base this statement on the effectiveness of selective serotonin reuptake inhibitors and on the lack of difference that they found between their study groups. We base our critique of their argument on the lack of a placebo group.

The authors compared four groups of depressed patients receiving problem solving treatment given by a general practitioner, problem solving treatment given by a nurse, a selective serotonin reuptake inhibitor, or a combined treatment (antidepressant plus problem solving treatment). Although the authors referred to historical probabilities of recovery in placebo groups in studies that used selective serotonin reuptake inhibitors, they did not use this information in their analysis. They used analysis of variance for continuous variables and found no differences between the groups at baseline, 6, 12, and 52 weeks.

A much more relevant answer to the study's main question entails analysis of the categorical variable recovered subjects versus not recovered subjects at the end of the 12 week trial. The ² test showed no differences between the groups.

Since in this test of independence we can use only marginals to calculate expected values, the test does not permit us to obtain expected values under different modelsfor example, historical placebo rates. We believe that the results of the analysis used in this paper do not address the question of the effectiveness of these interventions. If the authors' line of reasoning was accepted then there would remain the unsupported statement that problem solving is effective.

We explored recovered versus not recovered subjects for each group, using ² with a 47% baseline recovery rate quoted in the paper. We found P=0.59 for problem solving treatment given by a general practitioner, P=0.39 for such treatment given by a nurse, P=0.02 for antidepressant treatment, and P=0.12 for antidepressant plus problem solving treatment. This suggests that problem solving alone or in combination is not effective in the treatment of depression.

We believe that concluding that problem solving is effective on the basis of no differences between the groups is unwarranted and that this study fails to show that problem solving is effective in treating depression in primary care.

D Marchevsky, consultant psychiatrist. 
david.march{at}ukgateway.net

O Adebajo, senior house officer in psychiatry. 
Department of Psychiatry, Campbell Centre, Milton Keynes MK6 5NG

Cost effectiveness is not clear

EDITORIt is disappointing that Mynors-Wallis et al did not include cost data in their paper.¹ If two treatments are equally effective the general principle is that the treatment of choice should be the least expensive, so the following three inferences may be drawn if we accept the paper's conclusions.

Firstly, combining problem solving treatment with antidepressant treatment is not justified as there is no added value.

Secondly, since appropriately trained practice nurses are as effective as general practitioners in problem solving treatment they should be the professionals of choice to deliver this treatment if it is deemed appropriate.

Thirdly, it is unclear whether problem solving treatment is more or less expensive than antidepressant treatment. Given that it typically amounted to 3.5 hours' professional time per patient plus an unspecified amount of training, supervision, and administrative input, however, it is by no means cheap. If it is a more expensive choice although of equal efficacy, antidepressants should remain the treatment of first choice. Problem solving treatment should then be reserved for those patients who cannot or will not take antidepressants and who express a preference for a psychotherapeutic intervention.

As other authors have suggested in their electronic responses to this paper, the question remains as to whether it is feasible to introduce training in problem solving treatment into a typical general practice.

Gerry Waldron, consultant in public health medicine. 
Northern Health and Social Services Board, Ballymena, Northern Ireland BT42 1QB Gerry.waldron{at}nhssb.n-i.nhs.uk

Author's reply

EDITORWest asked to be told if he is missing something. The short answer is yes. The results are analysed on the basis of intention to treat with the last available result carried forward. Thus results from patients who drop out are included in the analysis.

Marchevsky and Adebajo criticise the lack of a placebo group. I agree that it would be helpful to have had a placebo group. In the design of the study, however, it was thought to be unethical to have a placebo group when both drug treatment and problem solving treatment have been shown to be significantly better then placebo. Categorical outcomes are given in the paper (that is, recovered, not recovered). There were no significant differences between the four groups at 12 or 52 weeks.

I would not disagree with the inferences that Waldron draws from the study. In an earlier study of the treatment of emotional disorders in primary care, problem solving was more expensive than drug treatment in terms of treatment costs but resulted in significant savings from indirect costs.¹ Colleagues and I have submitted a bid to evaluate fully the cost effectiveness and cost utility of problem solving and drug treatment in a more naturalistic study of depressive disorders in primary care.

L Mynors-Wallis, consultant psychiatrist. 
Oxford University Department of Psychiatry, Warneford Hospital, Oxford OX3 7BJ lmmw{at}soton.ac.uk