2017/05/26 by Jason Black, Black, Jason, Amanda Terry +3
Health Professions · #Artificial Intelligence in Healthcare
paper · pdf · doi:10.48550/arxiv.1705.09563
Background-Prognostic predictive models are used in the delivery of primary\ncare to estimate a patients risk of future disease development. Electronic\nmedical record, EMR, data can be used for the construction of these models.\nObjectives- To provide a framework for those seeking to develop prognostic\npredictive models using EMR data, and to illustrate these steps using\nosteoarthritis risk estimation as an example. FRAMR-EMR-The FRAmework for\nModelling Risk from EMR data, FRAMR-EMR, was created, which outlines\nstep-by-step guidance for the construction of a prognostic predictive model\nusing EMR data. Throughout these steps, several potential pitfalls specific to\nusing EMR data for predictive purposes are described and methods for addressing\nthem are suggested. Case Study-We used the DELPHI, DELiver Primary Healthcare\nInformation, database to develop our prognostic predictive model for estimation\nof osteoarthritis risk. We constructed a retrospective cohort of 28447 eligible\nprimary care patients. Patients were included if they had an encounter with\ntheir primary care practitioner between 1 January 2008 and 31 December 2009.\nPatients were excluded if they had a diagnosis of osteoarthritis prior to\nbaseline. Construction of a prognostic predictive model following FRAMR-EMR\nyielded a predictive model capable of estimating 5-year risk of osteoarthritis\ndiagnosis. Logistic regression was used to predict osteoarthritis based on age,\nsex, BMI, previous leg injury, and osteoporosis. Internal validation of the\nmodels performance demonstrated good discrimination and moderate calibration.\nConclusions-This study provides guidance to those interested in developing\nprognostic predictive models based on EMR data. The production of high quality\nprognostic predictive models allows for practitioner communication of\naccurately estimated risks of developing future disease among primary care\npatients.\n