Self-Report Medication Adherence and Clinical Outcomes after Implementing a Telemedicine Intervention to Individuals with Severe Mental Illness

Background/Purpose: Individuals with severe mental illnesses (SMI) have elevated risk for adverse health and social outcomes. Pharmacological treatments have limited effectiveness because individuals with SMI have adherence rates ranging from 50-70%, whereas 80% is considered adherent. The Medherent device was designed to promote adherence among this population by alerting participants to their next scheduled dose and sounding an alarm if the dose is missed, until the medication is dispensed. This study uses self-report data from structured interviews to demonstrate trends in perceived adherence and clinical outcomes from pre to post-intervention.

Method: All clients (N=185) were interviewed at multiple time points: enrollment, baseline, and 1, 6, and 12-month follow-ups. Medherent devices were installed immediately following baseline interviews, making enrollment and baseline interviews pre-treatment, with all subsequent interviews post-treatment. Interviews used a structured questionnaire consisting of 8 validated self-report measures of psychiatric symptoms, medication adherence, and therapeutic relationships; demographic data were collected at enrollment. Interviews were conducted in-person (pre-pandemic) and were later adapted to phone or virtual teleconferencing in Fall 2020. In this study, we used all available survey data to evaluate the intervention impact on adherence and clinical outcomes. We used descriptive data and Generalized Estimating Equations (GEE) to examine trends and measure within-subject treatment effects. Additionally, we used random intercept models to identify agency-level influence on adherence.

Results: Descriptive data indicate adherence was high at enrollment (79-85%), decreased at one-month follow-up (66% according to one measure), and increased again at 6-month follow-up to pre-treatment levels. However, GEE models indicate no significant within-subject effect for self-reported adherence, psychiatric symptoms, or therapeutic relationships from pre to post-treatment. The percentage of participants who reported “taking medications is difficult to remember everyday” gradually declined over time, from 24% at intake to 11% at 6-month follow-up, and this change was significant from pre to post-treatment (F_(1,179)=7.5, p < .01). Lastly, the random intercept model demonstrates that 12% of the variance in adherence can be attributed to differences between agencies.

Conclusions/Implications: Findings indicate that the Medherent device supports adherence for SMI individuals who receive services from community mental health agencies. Adherence rates were high at intake (85%), indicating an atypically adherent sample of SMI individuals. The brief drop in adherence may be attributed to adjustments to the Medherent device, and the return to pre-treatment adherence levels demonstrates that participants were able to resume high adherence after adjusting. The non-significant within-subject effects further support that adherence remained high from pre to post-treatment. As such, psychiatric symptoms and therapeutic relationships remained relatively stable over time as well. Interestingly, participants reported a significant improvement in their ability to remember to take their daily medications which contradicts our adherence findings and leads to more questions about measuring adherence behaviors. The influence of agency on adherence suggests there may be agency-level factors impacting participant adherence to medications and/or use of the Medherent device, warranting further investigation. Overall, these findings demonstrate that a telemedicine device can support perceived adherence, psychiatric symptoms, and therapeutic relationships for SMI individuals.