Development of a Hospital Outcome Measure Intended for Use With Electronic Health Records: 30-Day Risk-standardized...

Area of Science:

• Health Services Research
• Clinical Informatics
• Cardiovascular Medicine

Background:

• Electronic health records (EHRs) have the potential to enhance quality improvement by enabling hospital performance comparisons using clinical data.
• Current performance measures utilizing EHR data are insufficient for comprehensive quality assessment.

Purpose of the Study:

• To develop and validate a novel outcome measure for hospital risk-standardized 30-day mortality rates in acute myocardial infarction (AMI) patients, specifically designed for EHR data.
• To create a measure compatible with existing EHR systems and standard clinical practices, facilitating widespread adoption.

Main Methods:

• Merged clinical registry data (Action Registry-Get With The Guidelines) with Centers for Medicare & Medicaid Services (CMS) claims data for model development (2009) and validation (2010).
• Selected feasible variables from current EHRs, avoiding changes to standard clinical practice.
• Employed logistic regression with stepwise selection and bootstrapping for robust model development.

Main Results:

• The final risk model incorporated five key variables available at presentation: age, heart rate, systolic blood pressure, troponin ratio, and creatinine level.
• The model achieved an area under the receiver operating characteristic curve of 0.78, indicating good discriminative ability.
• Risk-standardized mortality rates varied across hospitals, with a median of 10.7% and a 1.37-fold higher odds of mortality for high-risk hospitals compared to low-risk hospitals.

Conclusions:

• This represents the first National Quality Forum-endorsed outcome measure for public reporting of hospital quality using EHR clinical data.
• The measure's compatibility with current clinical workflows and EHR systems positions it as a foundational model for future quality improvement initiatives.