Results: Purpose: We conducted this observational study to assess whether clinics that use electronic medical records (EMRs) provide better quality diabetes care than clinics that do not use EMRs.
Methods: Project QUEST included 19 medical groups with 84 participating clinics, and 2,117 randomly sampled adults with diabetes. We compared levels of glycated hemoglobin (A1c), systolic blood pressure (SBP) and low-density lipoprotein (LDL) in the 455 diabetes patients receiving care at EMR clinics versus the 1111 diabetes patients receiving care at non-EMR clinics and with sufficient data for analysis. Multivariate analysis utilized general linear mixed models via MLWin software to compare three consecutive years of outcomes, accounting for the nested data and controlling for patient age, sex, education, duration of diabetes, heart disease status, physician specialty, body mass index, and Charlson comorbidity score.
Results: In bivariate analysis, EMR use by clinics was not associated with better A1c (7.2% at EMR clinics vs. 7.3% at non-EMR clinics, p=.40), LDL (98.2 mg/dl vs. 102.4 mg/dl, p=.12), or systolic BP (133.1 mm Hg vs. 131.0 mm Hg, p=.06) in diabetes patients receiving care at those clinics. Multivariate models adjusting for patient and provider characteristics (see Methods above) and clustering of multiple outcomes within patients and patients within clinics confirmed no significant differences in levels of A1c (p=.56), SBP (p=.15), or LDL (p=.95) when comparing patients receiving care at EMR versus non-EMR clinics.
Conclusions: Patients at clinics that used EMRs had no better diabetes care than those at clinics without EMRs. More research is urgently needed to identify specific EMR applications and associated office workflows that may lead to better diabetes care.