Association between Secure Messaging and Primary Care Face-to-Face Visits and Phone Visits

To study the association between secure message (SM) utilization and the subsequent use of face-to-face and telephone visits. MyHealtheVet is a web-based application allowing Veterans and their Veterans Health Administration (VHA) care team to communicate securely online. SMs may be sent 24 hours, 7 days/week. Two analyses were performed: (1) a retrospective cohort study comparing changes in utilization after initial SM use and (2) a matched cohort difference in difference analysis comparing utilization changes among SM users to non-SM users. Veterans registered in MyHealtheVet, who sent an index SM in calendar year 2016, were identified. A subset of these patients was matched to patients not using SM in calendar year 2016 on the same provider panel. Administrative data were abstracted via the VA Informatics Computing Infrastructure one year prior and after index SM for users and matched non-users. Utilization outcomes included annual primary care face-to-face and telephone visits. Firstly, a paired t test compared the within SM user difference in primary care face-to-face and telephone visits between the pre- and post-periods. Secondly, we matched SM users directly to non-users on the same panel based on age, gender, service-connected percent, and urban-rural designation. Controls were selected within 0.20 standard deviations of predicted propensity of SM use based upon Nosos comorbidity risk score and drive time to the nearest VA clinic. 154,053 Veterans who initiated secure messaging during calendar year 2016 with 25,683 of these matched to controls with no secure messaging use in 2016 (N = 49,266). Compared to the VHA population, SM users were younger (54.4 years vs 62.8) with a higher proportion of females (15.2% vs 8.0%) and urban residents (71.4% vs 63.7%). Among SM users, the annual primary care face-to-face visit rate decreased by 13% from 1.6 to 1.4 visits per SM user per year (P < .001). Similarly, annual telephone visit rates increased by 14%, from 2.7 to 3.1 visits per SM user per year (P < .001). Matched SM users and control samples did not statistically differ by service connection percentage, marital status, rurality, or drive time. However, when comparing cases to controls statistical differences are noted in age (62.3 vs 65.3), female gender (4.2% vs 2.5%), and Nosos comorbidity risk score (0.7 vs 0.6), respectively. Among Veterans in the matched control study, the annual primary care face-to-face visit rate decreased by 16%, or 0.23 visits per SM user per year (P < .001) and 9%, or 0.1 visits (P < .001) for controls. Likewise, the annual telephone visit rate increased by 11%, or 0.27 visits per SM user per year (P < .001) and 4%, or 0.08 visits (P < .001) for controls. For both visit types, the between-group difference was 7% (P < .001). The initiation of secure messaging was associated with a decrease in face-to-face primary care visits and an increase in telephone visits. The association between SM use and reduced face-to-face visits may improve the availability of clinical appointment slots while increasing the impact and time commitment required of non-traditional forms of enhanced access. Department of Veterans Affairs.