Back to Agenda
Assessing Comparative Effectiveness Research Feasibility and Interpretability A Priori
Session Chair(s)
Matthew D. Rotelli, PHD, MS
Vice President, Bioethics, Eli Lilly and Company, United States
Observational studies and comparative effectiveness research (CER) in settings outside of randomized clinical trials have an important role in informing health care decisions about alternative therapies, but are not without their challenges. CER design and analytic features can influence results. Researchers should understand the feasibility of adequately controlling for bias and confounding prior to launching a study. We recommend steps to be conducted in advance of CER to assess feasibility in light of inherent bias and confounding. Researchers must understand their data source and whether outcomes, exposures and confounding factors are captured sufficiently in the database to address the research question. Taking such steps will help ensure that such studies yield results that are interpretable and valid. Recommended a priori analyses include assessments of confounding by indication, robustness to unmeasured confounding, and construction of empirical null distributions based on negative controls.
This session has been developed by the DIA Comparative Effectiveness Scientific Working Group in association with the Evidence Based Medicine Community and the DIA Ethics and the Medicines Lifecycle Community.
Learning Objective : Explain the critical importance of sensitivity analyses on various design feature decisions in comparative effectiveness research; Describe a framework by which more assessment of design features could be better assessed prior to study conduct as part of study feasibility assessment.
Speaker(s)
Douglas E Faries, PHD
Research Fellow, Global Statistical Sciences, Eli Lilly and Company, United States
Feasibility of Comparative Effectiveness Research: Unmeasured Confounding and Operational Characteristics
Cynthia J. Girman, DrPH
President, CERobs Conmsulting, LLC, United States
Additional Steps for Feasibility and Pre-Indentifying Sensitivity Analyses Prior to Launching CER
Robert T. O'Neill, PHD
Senior Statistical Advisor, Office of Translational Sciences, CDER, FDA, United States
Discussant
Have an account?