Predictive Modeling of Health ConditionsIdentification and Validation of a Sickle Cell Disease Cohort Within Electronic Health Records

ObjectiveTo develop and validate a computable phenotype algorithm for identifying patient populations with sickle cell disease.MethodsIn this retrospective study we used electronic health record data from the Children's Hospital of Wisconsin to develop a computable phenotype algorithm for sickle cell disease. The algorithm was on the basis of the International Classification of Diseases, Ninth Revision codes, number of visits, and hospital admissions for sickle cell disease. Using Informatics for Integrating Biology and the Bedside queries, the algorithm was refined in an iterative process. The final algorithm was verified using manual medical records review and by comparison with a gold standard set of confirmed sickle cell cases. The algorithm was then validated at Froedtert Hospital, a neighboring health system for adults.ResultsFrom the Children's Hospital of Wisconsin, our computable phenotype algorithm identified patients with confirmed sickle cell disease with a positive predictive value of 99.4% and a sensitivity of 99.4%. Additionally, using data from Froedtert, the computable phenotype algorithm identified patients with confirmed sickle cell disease with a positive predictive value of 95.8% and a sensitivity of 98.3%.ConclusionsThe computable phenotype algorithm developed in this study had a high sensitivity and positive predictive value when identifying patients with sickle cell disease in the electronic health records of the Children's Hospital of Wisconsin and Froedtert, a neighboring health system for adults. Our algorithm allows us to harness data provided by the electronic health record to rapidly and accurately identify patient with sickle cell disease and is a rich resource for future clinical trials.