Study designs of evaluations included in the review
Qualitative articles such as editorials and review papers were excluded. Randomised controlled trials (RCTs), non-randomised studies, meta-analyses and descriptive epidemiological studies were included. The duration of observation ranged from 2 to 104 weeks (median: 12 weeks).
Specific interventions included in the review
Studies of interventions affecting adherence to antidepressant treatments were eligible. The included studies examined the following types of treatment: psychological treatment, patient education, education of the patient's family, training of the physician, training of the nurses, changes in patient management and medication clinics. Almost half of the studies were conducted in the USA. The studies were conducted in out-patient psychiatric services (the majority of the studies), psychiatric hospitals or primary care.
Participants included in the review
Studies of patients with unipolar depression, either as the exclusive diagnostic category or as a substantial part of the diagnosis, were eligible. The included studies were of patients with major depression, major and minor depression, mixed diagnoses with depression (including schizophrenia and bipolar disorders) and unspecified depressive disorders.
Outcomes assessed in the review
Studies that assessed the adherence rates as the primary outcome were eligible. In the individual studies, adherence was measured according to appointments kept, pills taken, plasma levels, protocol deviations and ad hoc composite indexes.
How were decisions on the relevance of primary studies made?
Two reviewers independently assessed articles that were judged to be potentially relevant, according to the eligibility criteria.