The phenomenon of improved diabetes self-management following participation in a clinical trial, with subsequent improvement of glycaemic control, has been acknowledged in literature but has received little attention. Also, the potential implications of such a 'study effect' for clinical research are poorly explored. We review the literature and describe the effects on glycaemic and psychological outcomes in long-term poorly controlled type 1 diabetes patients participating in a qualification phase of a Good Clinical Practice (GCP) trial. Improved glycaemic control following participation in a clinical trial is best understood as the result of improved patients' instrumental coping behaviours, including increased self-monitoring of blood glucose (SMBG). Such improvement in self-care with ensuing improved glycaemic control has important consequences for trial design. Firstly, benefits seen in uncontrolled trials should be interpreted with extreme caution. Secondly, unspecific study effects and the effect of a given intervention may not simply be additive. Therefore, it is wise to include a run-in or qualification phase of adequate length before randomization in a clinical trial. A stable baseline HbA1c can thus be reached, upon which the specific effect of an intervention can be properly judged. Also, in a multi-centre trial, a qualification phase of sufficient length will help diminish differences in terms of intensity of care provided in participating centres.
Copyright 2003 John Wiley & Sons, Ltd.