Researchers developed an algorithm using encounter diagnoses, clinical history, pharmacy data and laboratory results to identify diabetes cases. The algorithm was then applied to a population of 160,872 unique patients at a public hospital using a comprehensive EHR system during the study period (January 2009 to April 2011). Physician chart review was used to confirm identified cases of diabetes.
Results showed the algorithm had a positive predictive value of 90 percent and a negative predictive value of 96 percent for the identification of diabetes patients. The algorithm was able to identify potential diabetes patients sooner, well within the three-month window for routine glucose testing by physicians.
More Articles on EHRs:
Go-Live is Just the Beginning: 5 Steps to Successful Clinical Optimization
NYC Pilot Aggregates EHR Data to Examine Public Health
The Case for Integrated Clinical Decision Support Software in EMRs
