Predicting the impact of an electronic health record on practice patterns using computational modeling and simulation.

摘要

ospitals are under intense pressure by private and public agencies to improve clinical performance outcomes (JCAHO Web Site, 2005). The wide variation across hospitals in clinical outcomes is, in part, fueling this demand (Dartmouth Atlas of Health Care, 2006). In an attempt to reduce variation professional associations have funded the development of evidence based clinical practice guidelines (CPGs). Recently, electronic health records (EHR) that incorporate computerized decision support systems (DSS), point-of-contact clinical documentation (POCD), Internet access to guidelines and provider order entry (CPOE) features have embedded electronic order sets and documentation templates into their applications.;The overall purpose of this research study is to discover and apply new knowledge regarding methods to predict the impact of an EHR on clinical practice guidelines in complex systems such as hospitals. Specifically, the aims of this study are: (1) to build, simulate and validate the accuracy of a computational model representing the current practice patterns in a sample of patients diagnosed with heart failure and treated in a community hospital; and (2) using computational modeling and simulation, develop a method to predict the effects of clinical practice guidelines on practice patterns after implementation of an EHR.;A computational model representing the practice patterns of providers (lab, radiology, medications, patient days, ECGs, Echos, and respiratory care ordered) treating heart failure patients in a community hospital was constructed using Arena a commercially available software tool. The training model was validated (no significant difference at the p.05 level) against best practices clinical guidelines and a best practices physician group (65 patients) using a hold-out group (88 patients) and a test group (245 patients). The model results showed a significant decline (p.05) in net tests and treatments of 26.22%, total annual process delay hours of 12.56% and total annual costs by