Process guide for inferential studies using healthcare data from routine clinical practice to evaluate causal effects of drugs (PRINCIPLED): considerations from the FDA Sentinel Innovation Center
Rishi Desai, Shirley Wang, Sushama Kattinakere Sreedhara, Luke E. Zabotka, Farzin Khosrow‐Khavar, Jennifer C. Nelson, Xu Shi, Sengwee Toh, Richard Wyss, Elisabetta Patorno, Sarah K. Dutcher, Jie Li, Hana Lee, Robert Ball, Gerald J. Dal Pan, Jodi B Segal, Samy Suissa, Kenneth J. Rothman, Sander Greenland, Miguel A. Hernán, Patrick J. Heagerty, Sebastian Schneeweiß
Abstract
This report proposes a stepwise process covering the range of considerations to systematically consider key choices for study design and data analysis for non-interventional studies with the central objective of fostering generation of reliable and reproducible evidence. These steps include (1) formulating a well defined causal question via specification of the target trial protocol; (2) describing the emulation of each component of the target trial protocol and identifying fit-for-purpose data; (3) assessing expected precision and conducting diagnostic evaluations; (4) developing a plan for robustness assessments including deterministic sensitivity analyses, quantitative bias analyses, and net bias evaluation; and (5) inferential analyses.