Evaluation of a Diabetes Screening Clinical Decision Support Tool

Abstract

Introduction

The authors evaluated whether an electronic health record clinical decision support system improves diabetes screening across a health system.

Methods

Study population included adults without diabetes attending a visit at 27 primary care clinics. Outcomes included the monthly screening laboratory order rate and completion rate among eligible patient visits. The authors performed logistic regression using a generalized estimating equations model and interrupted time series analysis to evaluate the change in the outcome from baseline to implementation and postimplementation periods.

Results

From the baseline to postimplementation period, screening laboratory order rates increased from 53% to 66%, and completion rates increased from 46% to 54%, respectively. The odds of laboratory order and completion increased significantly from the baseline to postimplementation period (test order: OR=3.7; 95% CI=3.4, 4.1, p<0.001; test completion: OR=2.1; 95% CI=2.0, 2.3, p<0.001). In the interrupted time series analysis, laboratory order and completion rates increased significantly from the baseline period (p<0.001 for both).

Conclusions

The authors developed and implemented a clinical decision support system alert that automatically identifies eligible patients and facilitates single-click ordering of a diabetes screening test. An easily implementable and scalable clinical decision support system alert can improve diabetes screening.