As mentioned above, one difficulty in the study of depression therapy is that unsatisfactory treatment may reflect a combination of poor patient adherence and medical advice that is inconsistent with expert guidelines. To be clinically effective, interventions should ideally deal with both aspects of quality improvement. In 1999, Peveler et al. were able to show that two brief sessions of counselling provided by a primary care nurse could greatly reduce rates of discontinuation of treatment at 12 weeks (from 61% to 37%), but clinical benefit was only seen in a post hoc analysis of the subgroup of patients receiving adequate doses of medication (5). A small feasibility study also suggested that similar benefits could be obtained by telephone counselling (22). Information alone, provided by leaflet (5) or by repeated mailings (23), did not appear to be effective in improving rates of adherence (see also Table 3).

Most other studies have tested complex, multi-faceted, interventions designed to improve the overall quality of care. For example, Katon et al. (24 - 27) evaluated the impact of increased involvement of secondary care specialist staff and closer surveillance of patients receiving treatment in primary care. They reported improved adherence, boosting the proportion of patients receiving an adequate dose of their medication at 90 days to 75%, but although this group initially had better clinical outcomes, these benefits were no longer evident at 19-month follow-up. Subsequent work has shown that a relapse prevention programme can also improve longer-term outcome (28).

Table 3 Factors affecting adherence to treatment for depression and interventions for improving it, listed by the five dimensions and the interventions used to improve adherence

(+) Factors having a positive effect on adherence; (-) factors having a negative effect on adherence.