Depression is common in general practice and lack of concordance is an important problem in its treatment. Homeopathy has few side effects and is generally associated with good compliance. We report a study investigating the feasibility of a trial to compare the effectiveness of homeopathy with a conventional antidepressant and placebo.
Objectives: To assess the feasibility of a general practice-based clinical trial comparing the effectiveness of individualised homeopathic treatment vs Fluoxetine (Prozac) vs placebo in the treatment of major depressive episodes of moderate severity.
Design: Randomised, double-dummy, double-blind parallel group clinical trial.
Setting: Lower Clapton Group Practice, East London.
Method: Patients were recruited through their general practitioners as they presented during a 9 month period. Recruitment target was 30 patients. Eligibility was confirmed by a consultant psychiatrist using standard criteria (DSM-IV) and instruments Hamilton Depression Scale (HAMD). Suicidal and psychotic patients were excluded, additional precautions against suicide were incorporated. There was a 1 week run-in period and patients showing spontaneous improvement were excluded. Homeopathic treatment was prescribed by a GP qualified in homeopathy, from a 'limited list' of 30 homeopathic medicines, with the help of decision support software. Patients were randomised to receive verum Fluoxetine and placebo homeopathy, or verum homeopathy and placebo Fluoxetine, or placebo homeopathy and placebo Fluoxetine. Treatment duration was 12 weeks. The outcomes were: adverse drug reactions, clinical global impression (CGI); HAMD; mini international psychiatric Interview; Pittsburgh sleep quality index; Side-effects checklist; Short Form 12; treatment credibility questionnaire; work and social disability scale. The primary outcome measures were HAMD and CGI.
Results: A recruitment calculation indicated that over 230 suitable patients would be expected to attend the practice during the recruitment phase. Thirty one patients were referred for possible inclusion in the trial by their GPs. Twenty three met the entry criteria, 11 were randomised and 6 completed the study. Of the completers, one received homeopathy, 2 placebo and 3 Fluoxetine.
Conclusions: A trial of this design in general practice is not feasible, because of recruitment difficulties, many of them linked to patient preference. Different approaches are required to recruit adequate patient numbers to trials of this sort.