The Association of Patient Portal Engagement with Critical Diabetes Care Indicators and Healthcare Utilization Among Adults with Type 2 Diabetes

Abstract

Type 2 diabetes mellitus (T2DM) remains a significant health problem that is increasing in prevalence. Despite recent advances in drug development and well-publicized guidelines for glycemic, blood pressure (BP), and cholesterol management, achieving the major goals of T2DM remains suboptimal. Patient portals can facilitate disease self-management by allowing patients to access and interact with their health records and communicate with their health care providers. Patient portals provide patients with T2DM the information and tools needed to achieve goal hemoglobin A1c (HbA1c), low-density lipoprotein cholesterol (LDL-c), and BP measurements. Methods This project aimed to comprehensively evaluate the long-term effects of patient portal use on critical T2DM care indicators (HbA1c, LDL-c, and BP assessed using mean arterial pressure [MAP]) and healthcare utilization metrics over four years of follow-up. A retrospective observational cohort study of patients with T2DM treated at UPMC was conducted. Propensity score matching was used to create comparable high and low portal engagement groups. Linear mixed-effect modeling was used to examine the longitudinal association between portal engagement and critical T2DM care indicators. Further, a comprehensive portal feature analysis and its association with improved HbA1c over time were completed using linear mixed-effect modeling. Negative binomial regression modeling was employed to evaluate the association of patient portal use with the utilization of the healthcare system. Results The study included 96,391 patients and identified 10,573 (11%) high-engagement users. This group, compared to 21,146 matched low-engagement users, was predominantly female (53 vs 51%), White (90 vs 82%), and exhibited a more favorable socioeconomic profile. Engaged portal users presented with lower baseline HbA1c (7.4 vs 7.7%) and a greater number of prescriptions (27 vs 25). The longitudinal analysis revealed a modest yet statistically significant negative association between portal engagement and HbA1c affecting both the baseline and the rate of change over time (P<0.01). The analysis of LDL-c showed an association between higher portal engagement and improved LDL-c trajectories (P<0.01). No significant association was found between MAP and portal use. The portal features of clinical data review and patient-provider communication had the strongest associations with better glycemic control. Portal use was also associated with significantly higher outpatient encounters but lower emergency visits. Discussion Patient portal engagement can lead to modest but meaningful improvements in HbA1c and LDL-c for patients with T2DM. These improvements assist patients in meeting treatment goals and potentially reduce the risk of complications. The proactive care pattern among engaged users can lead to consistent management of T2DM, limit complications, and ultimately lead to potential cost savings for both patients and the healthcare system. Future research should explore the development and testing of new portal functionalities and include patients and nurses in the process of design and implementation of new portal features.