Abstract 16536: Implementation Of Flag, Identify, Network, Deliver: Find FH Using The Electronic Medical Record To Identify Familial Hypercholesterolemia Within A Single Healthcare System

IntroductionFamilial hypercholesterolemia (FH) is a common underdiagnosed and undertreated condition that leads to premature cardiovascular disease. The FH Foundation developed a machine learning algorithm (MLA) to identify at-risk individuals for targeted screening for FH. FIND FH was implemented to evaluate the likelihood of patients with a high score (the MLA output) having an FH clinical diagnosis or FH-causing mutation in the preventive cardiology clinic of the University of Pennsylvania Healthcare System (UPHS).MethodsThe FIND FH algorithm generated a list of patients in the UPHS electronic medical record (EHR) with a FIND FH score above 0.20. The patients were organized into three sets. For ‘set 1’ of FIND FH patients, the patient’s primary provider was contacted to seek permission in order to contact the patient. Willing patients underwent a clinical assessment for FH. For ‘set 2’ of FIND-FH patients, text-mining was used to determine how many patients previously seen in the clinic had “familial hypercholesterolemia” in the EHR. For ‘set 3’ of FIND FH patients, a subset of patients from sets 1 and 2, genetic testing for FH was performed by Quest Diagnostics.ResultsIn set 1, a total of 8614/1.6 million individuals were identified by FIND FH. In set 1, 67 of these individuals were clinically evaluated. Median age (IQR) was 55 (44-65), 28 (42%) were female, and 49 (73%) were white. Of these, 26 (38.2%) had possible, probable, or definite FH by clinical diagnosis. In set 2, a total of 874/3674 individuals were existing preventive cardiology patients identified by FIND-FH with a score above 0.2. Of these, 36% (310 of 874) had FH in the EHR. In set 3, the number of FIND-FH patients with genetic testing for FH was 103 of which 22 (21%) had a pathogenic FH mutation while 5 (4.8%) and 2 (1.9%) others had a variant of unknown significance and hypocholesterolemic variant, respectively.ConclusionsIn this clinical application of FIND FH, a high percentage of patients seen in the preventive cardiology clinic were recognized to have a clinical diagnosis of FH and/or a positive FH-causing mutation. Further research to assess the utility of the MLA to improve the diagnosis rate, treatment patterns, and cardiovascular outcomes of FH is needed.