A methodology to generate longitudinally updated ACLF prognostication scores from electronic health record data

Abstract

AbstractBackground and AimsQueries of electronic health record (EHR) data repositories allow for automated data collection. These techniques have not been utilized in hepatology due to previous inability to capture hepatic encephalopathy (HE) grades, which are inputs for acute-on-chronic liver failure (ACLF) models. Here, we describe a methodology to utilizing EHR data to calculate rolling ACLF scores.MethodsWe examined 239 patient-admissions with end-stage liver disease 7/2014-6/2019. We mapped EHR flowsheet data to determine HE grades and calculated two longitudinally updated ACLF scores. We validated HE grades and ACLF diagnoses via chart review; and calculated sensitivity, specificity, and Cohen’s kappa.ResultsOf 239 patient-admissions analyzed, 37% women, 46% non-Hispanic White, median age 60 years, median MELD-Na at admission. Of the 239, 7% were diagnosed with NACSELD-ACLF at admission, 27% during the hospitalization, and 9% at discharge. Forty percent diagnosed with CLIF-C-ACLF at admission, 51% during the hospitalization, and 34% at discharge.From chart review of 51 admissions, we found sensitivities and specificities for any HE (grades 1-4) were 92-97% and 76-95%, respectively; for severe HE (grades 3-4) were 100% and 78-98%, respectively. Cohen’s kappa between flowsheet and chart review HE grades ranged 0.55-0.72. Sensitivities and specificities for NACSELD-ACLF diagnoses were 75-100% and 96-100%, respectively; for CLIF-C-ACLF diagnoses were 91-100% and 96-100%, respectively. We generated approximately 28 unique ACLF scores per patient per admission-day.ConclusionIn this study, we developed an informatics-based methodology for to calculate longitudinally updated ACLF scores. This opens new analytic potentials, such big data methods to develop electronic phenotypes for ACLF patients.