Cumulative Effect in Observational Study: Challenge and Impact of Statistical Models

Abstract

There have been more and more applications of secondary database for investigating causal relationship between drugs and adverse effects. One group of publications studied the cumulative effect of drug on their corresponding adverse effects. Upon researching the statistical methods and recent published methodologies on this particular topic, to our surprises, we found that most of publications reported the biased results based on cumulative effect. In this talk, I will describe the statistical methodological development on this topic and point out the power of correct statistical modeling. Furthermore, there have been a growing criticism against the applications of secondary database but I believe that this example signify the role of data scientists/statisticians and what we can contribute to the medical/clinical fields.