Compliance therapy: a randomised controlled trial in schizophrenia
O'Donnell C, Donohoe G, Sharkey L, Owens N, Migone M, Harries R, et al.;
Commented by , 27 Oct 2003
Aim of the study
Non-adherence to antipsychotic medication is an important issue in the treatment of schizophrenia. It has been estimated that two thirds of people with schizophrenia readmitted to hospital are at least partially non-compliant with their prescribed medication.
Therefore, Kemp and colleagues had developed a cognitive behaviour intervention with techniques adapted from motivational interviewing and other cognitive therapies including psychoeducation which they called “compliance therapy”. In the only randomised controlled trial about compliance therapy, Kemp and co-workers had shown significant advantages in terms of improved compliance and longer “survival” in the community after compliance therapy. The authors of the current study tried to replicate these results in a subsequent randomised trial among patients with schizophrenia.
Methods
This was a randomised controlled, open trial. Of 94 consecutive admissions of patients with schizophrenia from an urban catchment area psychiatric service, 56 could be included in the study. They randomly received either compliance therapy and non-specific counselling, each consisting of 5 sessions lasting 30-60 minutes.
For the compliance therapy group the same manual produced by Kemp for their previous positive study was used. The primary outcome parameter was compliance with drug treatment at one year. In addition attitudes to treatment was assessed with the “Drug Attitude Inventory”, symptomatology with the “Positive and Negative Symptoms Scale”, insight with the “Schedule for Assessment of Insight” and quality of life with the “Heinrich Scale”. The following service use parameters were examined: length of "survival" in the community, bed days, and rehospitalisation rates at two years.
Results
The study could not replicate the previous findings by Kemp’s group. Compliance therapy was not significantly more effective than non-specific counseling in improving compliance at one year (43% (12/28) v 54% (15/28), difference -11% (95% confidence interval -37% to 15%).
Nor was there any significant advantage in terms of the secondary outcome measures - symptomatology, attitudes to treatment, insight, global assessment of functioning, and quality of life. Attitudes to treatment at baseline predicted adherence one year later and may be a clinically useful tool.
Discussion
At first glance the results appear to be sobering. The authors did not find any statistically significant superiority of compliance therapy and therefore concluded that compliance therapy may not be of benefit to patients with schizophrenia.
They describe differences compared to the previous study by Kemp’s group that may explain the differences: notably, evaluators at baseline and at endpoint were blind to the intervention the patients received and the study only included patients with schizophrenia whereas Kemp examined a mixed group.
However, one important limitation of the study may be the small sample size which was insufficient to detect statistical significance. Indeed, an absolute risk reduction of non-compliance of 11 percent points by compliance therapy may well be a clinically important effect. Larger trials are needed to verify whether this may be a true clinical effect. These trials could use this 11 % difference for their case number calculations.