OBJECTIVES: Comparative effectiveness research using real-world patient data collected outside of clinical trials is expected to increase dramatically in the coming decade as electronic medical records (EMRs) are more widely adopted in routine patient care settings. Retrospective analyses of such data to answer clinical questions will be easier than ever to perform, but with greater ease comes greater threats to validity. Our objective here is to describe a methodology we developed to move beyond the traditional chart review towards a way to use clinical practice data for reliable comparative effectiveness assessments. METHODS: We recently planned and performed an analysis of retrospective data from routine patient care settings from several liver clinics (with varying use of clinical databases) to compare the relative effectiveness of two commonly used treatment regimens in adults with chronic HCV infection. Participating clinicians in Europe and the US were initially identified via authorship of published abstracts of single center case series of HCV patients. A project protocol was prospectively developed to define site and patient eligibility criteria, and to provide sites with a template for transmission of specific data elements for analysis. Data privacy, security, and IRB issues were addressed, and a statistical analysis plan included sample size calculations and hierarchical linear modeling methods to adjust both site and patient level variables concurrently. Agreement to disseminate results regardless of findings was obtained up-front from all investigators and the corporate sponsor. RESULTS: N/A CONCLUSION: A rigorous and standardized methodology is urgently needed to advance the science of clinical effectiveness research using real-world patient records, especially those obtained from EMR systems. The approach described herein could serve as a useful case in point.