Scientists from Sechenov University investigated how daily remote monitoring using a voice AI assistant is associated with changes in indicators in people with type 2 diabetes and obesity. After a month, in the group that additionally interacted with the system, fasting glucose levels and indicators related to visceral obesity decreased, and quality of life according to the used scale increased, the university's press service told “Pervy Tekhnichesky”.

121 people participated in the study. 71 patients received standard treatment and additionally interacted with a voice AI assistant every day, while another 50 people received only standard medical care. The observation lasted one month.

The system itself called patients at a pre-agreed time. During the conversation, the person reported glucose levels and waist circumference, answered questions about symptoms of hypo- and hyperglycemia, and received reminders about medication intake. If indicators went beyond the set limits or alarming symptoms appeared, the information was transmitted to the doctor, who contacted the patient and, if necessary, adjusted the treatment.

After one month, in the remote monitoring group, the lipid accumulation index (LAP) decreased by 25.9%, and the visceral adiposity index decreased by 31.2%. Fasting glucose levels decreased by 15.1%. In the standard observation group, no statistically significant changes in these indicators were found during the same period.

Changes also affected quality of life

The average SF-36 score in the AI group increased by seven points compared to one point in the control group. The authors also report differences in adherence to medical recommendations. 42% of patients from the AI assistant group fully followed the prescriptions, compared to 18% among participants in standard observation. However, the observation period was short — one month. 

At the same time, the voice format allowed for dispensing with a separate medical application. It was enough for the patient to answer a phone call and report the necessary indicators.

This approach has already been used in another project of Sechenov University, where voice AI daily contacted patients with various chronic diseases, including diabetes. In the new study, scientists focused specifically on patients with type 2 diabetes and obesity.