In search of FH patients

Application 2017265 Ronen Durst, Hadassah Hebrew University Medical C enter Joshua W. Knowles, Stanford University Ran Balicer, Clalit Health Services SHORT ABSTRACT IN LAY TERMS An effort to identify Familial Hypercholesterolemia patients in the databases of Clalit Health Services using machine learning algorithms.

FH (Familial Hypercholesterolemia) is a common b ut vastly underdiagnosed, inherited form of high cholesterol that is potentially devastating if undiagnosed, but once diagnosed can be treated and prevent i ts complications. Our project aims to identify previously undiagnosed individuals with FH within electronic health records (EHRs) at Clalit Health Services (Clalit) using a novel computer algorithm s o that these patients and their families c an receive potentially l ife-saving p reventive c are. This a lgorithm will be adapted from o ne created at Stanford University using machine learning techniques that has already been validated as having u tility. Our collaborative team including data scientists, lipid experts and g eneticists from Stanford and Clalit will collaborate to scan over 6 million health records together.

Individuals with possible FH within Clalit will be c ontacted for screening and therapy and we will assess the effect on the h ealth of t hose individuals. Besides the important implication of this project for FH, t his project is also significant in that i t is a n ew technological approach for case identification that could be applied to many other d iseases.