Home | DTI | 2006–07 funded proposals | Bonnie L. Westra, Karen Dorman Marek, David Muhovich, Der-Fa Lu, Hwanjo Yu
Bonnie L. Westra, Karen Dorman Marek, David Muhovich, Der-Fa Lu, Hwanjo Yu
Using Electronic Health Record Data to Predict Medical Emergencies for Homecare Patients
The Institute of Medicine (IOM) estimates that medical errors result in 44,000 — 98,000 deaths for hospitalized patients annually, and another 777,000 are injured as a result in adverse drug events. Moreover, $300 billion are spent annually on treatments that may not improve care, are redundant, or inappropriate. Although care based on the most recent scientific evidence improves outcomes, it takes approximately 17 years to translate scientific findings into practice. The IOM called for the implementation of Evidence-Based Practice (EBP) guidelines to improve patient safety and outcomes. The overall goal of this multiphase study is to develop and test prediction models for EBP that extracts data from home care agencies’ electronic health records to determine the best interventions to decrease hospital readmission and emergent care. KDD is an emerging informatics research method for data mining and knowledge development and will be used in this phase one study, focusing on the best model for predicting these outcomes. The overall goal of this multiphase study is to develop and test prediction models for evidence-based practice to guide clinicians in improving outcomes for patients in homecare. This proposal is Phase One, which includes selection, preprocessing, and transforming the data to conduct data mining for development of models to predict two patient outcomes: reducing hospital readmissions and reducing emergency care utilization.
Digital Technology Center
499 Walter Library, 117 Pleasant Street SE, Minneapolis, MN 55455
P: (612) 624-9510