Conducting research in individual patients: lessons learnt from two series of N-of-1 trials

BACKGROUND: Double-blind randomised N-of-1 trials (N-of-1 trials) may help with decisions concerning treatment when there is doubt regarding the effectiveness and suitability of medication for individual patients. The patient is his or her own control, and receives the experimental and the control treatment during several periods of time in random order. Reports of N-of-1 trials are still relatively scarce, and the research methodology is not as firmly established as that of RCTs. Recently, we have conducted two series of N-of-1 trials in general practice. Before, during, and after data-collection, difficulties regarding outcome assessment, analysis of the results, the withdrawal of patients, and the follow-up had to be dealt with. These difficulties are described and our solutions are discussed. DISCUSSION: To prevent or anticipate difficulties in N-of-1 trials, we argue that that it is important to individualise the outcome measures, and to carefully consider the objective, type of randomisation and the analysis. It is recommended to use the same dosages and dosage forms that the patient used before the trial, to start the trial with a run-in period, to formulate both general and individualised decision rules regarding the efficacy of treatment, to adjust treatment policies immediately after the trial, and to provide adequate instructions and support if treatment is adjusted. SUMMARY: Because of the specific characteristics of N-of-1 trials it is difficult to formulate general 'how to do it' guidelines for designing N-of-1 trials. However, when the design of each N-of-1 trial is tailored to the specific characteristics of each individual patient and the underlying medical problem, most difficulties in N-of-1 trials can be prevented or overcome. In this way, N-of-1 trials may be of help when deciding on drug treatment for individual patients.