Interactive Virtual Assistant for Health Promotion Among Older Adults With Type 2 Diabetes: The IVAM-ED Randomized Clinical Trial.

Abstract

Older patients with diabetes often experience high levels of mental distress, which may complicate diabetes management and negatively impact health outcomes. To evaluate whether an interactive virtual assistant device built with a behavioral intervention model can improve mental health and diabetes-related outcomes in older adults with diabetes. This open randomized clinical trial was conducted at an academic medical center in Brazil between June 22, 2023, and February 8, 2024. Participants aged 65 years or older with type 2 diabetes were included. Participants were randomized 1:1 to receive the interactive virtual assistant device (Smart Speaker Echo Dot, 3rd generation [Amazon]) for home use over a 12-week period or to continue with usual care. The device was programmed with a behavioral intervention model aimed at enhancing mental health and diabetes management. The primary outcome was the between-group mean difference (MD) in mental distress, assessed using the Self-Reporting Questionnaire, in which scores range from 0 to 20, with higher scores indicating greater mental distress. Secondary outcomes included quality of life (assessed using the 36-Item Short Form Health Survey), perceived stress (assessed using the Perceived Stress Scale), adherence to diabetes self-care behaviors (assessed using the Self-Care Inventory Revised questionnaire), and glycemic control (hemoglobin A1c). All outcomes were assessed at the 12-week follow-up. Among 112 participants (mean [SD] age, 72.5 [5.7] years; 71 females [63.4%]), the mean (SD) hemoglobin A1c level was 7.9% (1.5%), and outcome measures were available for 103 participants (52 in the intervention group and 51 in the usual care group). At 12 weeks, the fully adjusted mean (SE) Self-Reporting Questionnaire score was 6.29 (0.44) for participants in the intervention group and 7.75 (0.42) for those in the usual care group, resulting in a fully adjusted MD of -1.46 (95% CI, -2.73 to -0.19; P = .02), suggesting a reduction in mental distress among the intervention group. Subgroup analyses showed consistent improvement favoring the intervention across all subgroups. The intervention was also associated with improvements in quality of life (MD, 9.46 [95% CI, 3.65 to 15.26]; P = .001), adherence to diabetes self-care behaviors (MD, 3.40 [95% CI, 1.61 to 5.19]; P < .001), and glycemic control (MD, -0.48% [95% CI, -0.85 to -0.11]; P = .01). No difference was observed regarding perceived stress (MD, -3.00 [95% CI, -6.20 to 0.20]; P = .07), and no participants withdrew due to adverse events. In this randomized clinical trial of older adults with diabetes, participants in the smart speaker group showed significant improvements in mental distress, quality of life, diabetes self-care, and glycemic control. These findings suggest that this easily implemented self-management intervention could enhance health outcomes in this population. ClinicalTrials.gov Identifier: NCT05329376.