Use of a medication control officer to reduce bias in a clinical trial: lessons learned from the scleroderma lung study.

Background Scleroderma Lung Study (SLS) was designed to evaluate the efficacy and safety of oral cyclophosphamide (CYC) versus placebo taken for 1 year for scleroderma-associated interstitial lung disease. An independent medication control officer (MCO), usually a physician, at each center was assigned to monitor laboratory and clinical toxicity of study medication and regulate its dosing based on these results. By having an MCO who watched and managed toxicity, the study investigators were free to care for study patients and to assess study outcomes without the potential bias of knowing toxicity data (toxicity from cyclophosphamide is distinctive - cytopenias and hematuria in particular). Purpose To assess the usefulness of an MCO, whose chief role was to maintain safety while retaining the blinding in the clinical trial. Methods Patients had safety laboratory testing every 2-4 weeks and results were sent directly to the MCO within 2 days of the test. Other clinical adverse events (AEs) were reported by the patient to a nurse coordinator who reported them to the MCO who then managed the AEs to preserve the blinding of investigators caring for the patients. The MCO was provided pre-determined algorithms for dose adjustments of test medication based on the presence and severity of laboratory abnormalities. Results Safety monitoring by the MCO was effective in the early detection of drug toxicity with provision of appropriate medical intervention on a timely basis. At the same time, investigator blinding appeared to be maintained. Limitations The testing of MCO effectiveness in maintaining blinding and consistency was not defined as an a priori hypothesis and thus complete data relating to the efficacy of the MCO were not collected in a prospective fashion. Conclusion An MCO and pre-specified monitoring and dosing guidelines, coupled with uniform pre-specified responses to AEs, may be used effectively to preserve investigator blinding and provide consistency in response to AEs in a clinical trial setting, even when AEs of the test medication are distinctive. Clinical Trials 2010; 7: 85-89. http://ctj.sagepub.com.