Development of a Hospital Outcome Measure Intended for Use With Electronic Health Records. Issue 9 (September 2015)

• Record Type: Journal Article
• Authors: McNamara, Robert L.; Wang, Yongfei; Partovian, Chohreh; Montague, Julia; Mody, Purav; Eddy, Elizabeth; Krumholz, Harlan M.; Bernheim, Susannah M.
• Is Part Of: Medical care. Volume 53:Issue 9(2015)
• Publication Date: 2015-09
• Subjects: Economics, Medical -- Periodicals; Insurance, Health -- Periodicals; Santé, Services de -- Administration -- Périodiques; Soins médicaux -- Périodiques; Medical economics -- Periodicals; Health insurance -- Periodicals; Medical economics -- United States -- Periodicals; Health insurance -- United States -- Periodicals; Comprehensive Health Care -- Periodicals; Personal Health Services -- Periodicals; Gezondheidszorg; Économie de la santé -- Périodiques; Santé, Services de -- Périodiques; Health insurance; Medical economics; United States; Periodicals; 362.10973
• DOI: 10.1097/MLR.0000000000000402
• Languages: English
• ISSNs: 0025-7079

Abstract

<abstract> <title> <x xml:space="preserve">Abstract</x> </title> <sec> <title>Background:</title> <p>Electronic health records (EHRs) offer the opportunity to transform quality improvement by using clinical data for comparing hospital performance without the burden of chart abstraction. However, current performance measures using EHRs are lacking.</p> </sec> <sec> <title>Methods:</title> <p>With support from the Centers for Medicare & Medicaid Services (CMS), we developed an outcome measure of hospital risk-standardized 30-day mortality rates for patients with acute myocardial infarction for use with EHR data. As no appropriate source of EHR data are currently available, we merged clinical registry data from the Action Registry—Get With The Guidelines with claims data from CMS to develop the risk model (2009 data for development, 2010 data for validation). We selected candidate variables that could be feasibly extracted from current EHRs and do not require changes to standard clinical practice or data collection. We used logistic regression with stepwise selection and bootstrapping simulation for model development.</p> </sec> <sec> <title>Results:</title> <p>The final risk model included 5 variables available on presentation: age, heart rate, systolic blood pressure, troponin ratio, and creatinine level. The area under the receiver operating characteristic curve was 0.78. Hospital risk-standardized mortality rates ranged from 9.6% to 13.1%, with a median of 10.7%. The odds of mortality for a high-mortality hospital (+1 SD) were 1.37 times those for a low-mortality hospital (−1 SD).</p> </sec> <sec> <title>Conclusions:</title> <p>This measure represents the first outcome measure endorsed by the National Quality Forum for public reporting of hospital quality based on clinical data in the EHR. By being compatible with current clinical practice and existing EHR systems, this measure is a model for future quality improvement measures.</p> </sec> </abstract>