Abstract
Objectives: Standard randomized controlled trials of interventions for chronic conditions that involve behavioural change, or that are highly desired by participants, are difficult to undertake because of problems with recruitment and contamination. Alternatives include cluster-randomized trials or pre-randomization designs such as the Zelen design. The aim here was to develop a pre-randomization design that would overcome ethical and methodological problems associated with the conventional Zelen design, and permit the rigorous evaluation of a complex package of care, involving physical therapy and behavioural changes, for patients with painful patello-femoral osteoarthritis of the knee joint.
Methods: Eligible patients were first consented to a one-year observational study of their arthritis. They were subsequently randomized into intervention and control arms. Those in the intervention arm were then asked if they were willing to participate in a further study involving regular sessions with a physiotherapist. Those in the control arm were not told about this, but were followed up as agreed.
Results: Eighty-seven patients consented to the observational study, 43 of whom were subsequently randomized to the intervention arm. All 43 consented to the intervention, although five of these did not receive the full package of care. Assessments were carried out at five months and one year on 82 patients, and concealment was satisfactorily maintained in the majority.
Conclusions: We conclude that this study design could potentially offer an acceptable compromise between the need for scientific rigour and the ethical imperative of fully informed consent in trials that involve behavioural change or interventions that patients might want to obtain.
Get full access to this article
View all access options for this article.