Population-level Screening for Nonalcoholic Fatty Liver Disease (POPS-NAFLD)
PI(s): Jaideep Behari, University of Pittsburgh; Kathleen McTigue, University of Pittsburgh
Purpose: Building upon our previous studies of non-alcoholic steatohepatitis (NASH), this industry sponsored project aims to develop an evidence-based, comprehensive, risk-stratification strategy for non-alcoholic fatty liver disease (NAFLD)/NASH. Existing algorithms that could be used to screen clinical populations for possible NAFLD have limited utility. In this study, we aim to use predictive modeling to leverage existing EHR data to identify adults at high risk of NAFLD and its complications.
Objective 1. Develop a predictive model independent of variables used in FIB-4 calculation (age, ALT, AST, platelets) for identifying high-risk NAFLD patients at the population level using retrospective UPMC EHR data. Objective 2. Compare the diagnostic accuracy of population level risk stratification of NAFLD with FIB-4 versus an expanded pool of patients with missing FIB-4 variables using a data imputation strategy using retrospective UPMC EHR data.
Study Design: PaTH Data-only Observational Study
PCORnet Partners:
PaTH Partners:
Sponsor: Industry partner
Coordinating Center: University of Pittsburgh