Abstract:Background: Few studies have examined the association between patient-initiated electronic messaging (e-messaging) and clinical outcomes in fee-for-service settings. Objective: To estimate the association between patient-initiated e-messages and quality of care among patients with diabetes and hypertension. Design: Longitudinal observational study from 2009 to 2013. In March 2011, the medical group eliminated a $60/year patient user fee for e-messaging and established a provider payment of $3-5 per patient-initiated e-message. Quality of care for patients initiating e-messages was compared before and after March 2011, relative to nonmessaging patients. Propensity score weighting accounted for differences between e-messaging and nonmessaging patients in generalized estimating equations. Setting: Large multispecialty practice in California compensating providers' fee-for-service. Subjects: Patients with diabetes (N=4232) or hypertension (N=15,463) who had activated their online portal but not e-messaged before e-messaging became free. Measures: Quality of care included HEDIS-based process measures for hemoglobin (Hb) A1c, blood pressure, low-density lipoprotein (LDL), nephropathy, and retinopathy tests, and outcome measures for HbA1c, blood pressure, and LDL. E-messaging was measured as counts of patient-initiated e-message threads sent to providers. Patients were categorized into quartiles by e-messaging frequency. Results: The probability of annually completing indicated tests increased by 1%-7% for e-messaging patients, depending on the outcome and e-messaging frequency. E-messaging was associated with small improvements in HbA1c and LDL for some patients with diabetes. Conclusion: Patient-initiated e-messaging may increase the likelihood of completing recommended tests, but may not be sufficient to improve clinical outcomes for most patients with diabetes or hypertension without additional interventions.

Patient-initiated Electronic Messages and Quality of Care for Patients With Diabetes and Hypertension in a Large Fee-for-Service Medical Group: Results From a Natural Experiment