VA device a valuable add-on to diabetes self-care regimens

A smart speaker appears to have gone beyond its initial capabilities, improving mental health and diabetes-related outcomes in older adults with type 2 diabetes (T2D) in the Brazilian IVAM-ED* trial.

“This [study] found that an interactive virtual assistant (VA) device with a voice interface, built with a behavioural intervention model, improved mental distress among older individuals with T2D compared with usual care (UC),” the investigators said. “Exploratory analyses also suggested improvements in quality of life (QoL), adherence to diabetes-related self-care behaviours, and glycaemic control.”

Mental health, QoL

At the end of the intervention period, the baseline-adjusted mean SRQ-20** score was lower in the VA than the UC group (6.39 vs 7.66; mean difference [MD], –1.28; p=0.04). The difference remained significant after adjusting for sociodemographic variables (mean SRQ-20 score, 6.29 vs 7.75; adjusted MD [aMD], –1.46; p=0.02). [JAMA Netw Open 2026;9:e2553508]

QoL was also better with VA than with UC, as reflected by the higher SF-36*** score after the intervention period (57.84 vs 48.39; aMD, 9.46; p=0.001). Stress levels as measured by the PSS^#, albeit numerically lower with VA than UC, failed to reach statistical significance (19.29 vs 22.29; aMD, –3.00; p=0.07).

“The improvement in mental distress … supports the hypothesis that smart speakers can provide emotional support for older people with diabetes,” the researchers noted.

Moreover, about a third of participants in a qualitative study of VA users reported feeling humanized toward the device. [Acta Diabetol 2025;62:523-529] “This perception of the device as quasi-human may have contributed to the observed improvements in mental distress and QoL in this trial,” they added.

Diabetes-related outcomes

At 12 weeks, the VA group had a higher SCI-R^## score than the UC group (39.80 vs 36.40; aMD, 3.40; p<0.001), suggesting greater adherence to diabetes-related self-care behaviour. The lower adjusted mean Hb_A1c level in the VA group indicates better glycaemic control than in the UC group (7.56 percent vs 8.04 percent; aMD, –0.48 percent; p=0.01).

“The continuous support and personalized medication reminders may have contributed to better adherence to diabetes management practices, ultimately improving diabetes-related outcomes,” they said. The improved mental health may have also positively influenced diabetes self-care behaviours, as reduced mental distress can boost motivation and capacity for disease management. [Diabetes Care 2016;39:2126-2140]

Sensitivity analysis, safety

Greater MDs were also observed among women (−1.98), younger participants (−1.93), those with a negative screening for cognitive decline (−3.27), and those with a history of anxiety (−3.72) on sensitivity analysis.

“The greater benefit in women is an important finding, given that anxiety and mood disorders have consistently been shown to be more prevalent among women,” the researchers said.  Older adults may not have been able to optimize the VA’s benefit, potentially due to challenges in adapting to and engaging with it, they added.

Apart from one participant complaining about the frequency of device interactions, no adverse events or episodes of severe hypoglycaemia were reported in either group during follow-up.

Further studies

A total of 112 participants (mean age 72.5 years, 63.4 percent women) were randomized 1:1 to the VA device for home use over a 12-week period or to continue with UC.

The findings underscore the device’s potential clinical relevance and the need for studies with longer follow-ups and larger proportions of older participants. It would also be imperative to evaluate its capacity to extend its benefits to those with cognitive limitations or limited educational backgrounds to assess its impact across a broader population.