Keywords: bias, confounders, generalizability, propensity scores, Bayesian statistics
In this era of “Big Data” there are many sources of possible data that could be leveraged in the regulatory environment. The effort to convert such data into evidence and then utilize such evidence in an efficient manner can be a challenge but the payoff can be enormous in terms of savings of time and resources. A real concern in any application is the assessment of the quality of the real world data. In addition, there are inherent biases (especially selection biases) of observational data that need to be addressed as well as the confounders to be adjusted for due to different population characteristics. Also at issue is the generalizability of the studies. The potential uses of such real-world evidence in the premarket regulatory environment for expansion of an indication or data for a control group can rely on such statistical methodologies such as propensity scores and Bayesian statistics. Real-world evidence can also be useful in the post-market in mandated surveillance studies and in condition-of-approval studies. The use of real-world evidence in the post-market is being actualized through a medical device National Evaluation System for health Technology (NEST) coordinated by the Medical Device Innovation Consortium (MDIC), a public-private partnership.