Managing Insulin With a Voice AI (MIVA)

Using an AI-based Voice Assistant to Manage Insulin in Diabetes: a Randomized-Control Trial

This study randomizes participants to have their basal insulin titrated either through standard of care or by receiving prompts through interactions with an AI-enabled smart speaker device. The primary objective of this study is to investigate the feasibility of an AI-enabled smart speaker device and whether such a device facilitates insulin titration management, increases insulin adherence and decreases time to optimal insulin dose. The secondary objective of the study is to explore whether the device improves glycemic control as defined by improvements in fasting blood sugar.

Study Overview

• Status: Completed
• Conditions: Medication Adherence
• Intervention / Treatment: Device: Voice Assistant Device; Device: Voice Assistant Device

• Study Type: Interventional
• Enrollment (Actual): 39
• Phase: Not Applicable

Contacts and Locations

Study Locations

• United States
  • California
    • Stanford, California, United States, 94305
      • Stanford University

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 14 years and older (Adult, Older Adult)
• Accepts Healthy Volunteers: Yes

Description

Inclusion Criteria:

• Patients with Type 2 Diabetes
• Patients clinically indicated to be taking daily long-acting insulin
• Patients currently taking long-acting insulin but necessitating active dose adjustments

Exclusion Criteria:

• Patients who do not speak English
• Patients who do not own a smart phone
• Patients who do not have stable wireless internet connection at home

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Treatment
• Allocation: Randomized
• Interventional Model: Parallel Assignment
• Masking: None (Open Label)

Number of Arms

2

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
Experimental: Voice Assistant Device; Subjects have their insulin titrated by a voice assistant device for 8 weeks. The Voice Assistant Device software algorithm is used to provide daily insulin dose and titration instructions.	Device: Voice Assistant Device; Voice AI-enabled smart speaker device delivers daily custom insulin titration instructions.; Device: Voice Assistant Device; Voice AI-enabled smart speaker device delivers daily generic insulin reminders.
Active Comparator: Control; Subjects have their insulin titrated via standard of care for 8 weeks. The Voice Assistant Device is limited to providing a daily reminder to take medication.	Device: Voice Assistant Device; Voice AI-enabled smart speaker device delivers daily custom insulin titration instructions.; Device: Voice Assistant Device; Voice AI-enabled smart speaker device delivers daily generic insulin reminders.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Time to Optimal Insulin Dose	Number of days between study start date and goal fasting sugar	8 weeks
Insulin Medication Adherence	Percentage of adherence to taking insulin based on logs over 8 weeks	8 weeks
Attitudes Toward Diabetes	Change in Composite Score of Attitudes Toward Diabetes (PAID-5). A 5-item psychometrically validated self-report questionnaire on a 5 point Likert scale (0= not problem; 4= a serious problem). The scores of each item are summed to create an overall score (Minimum value = 0; Maximum value = 20). Higher scores denote higher levels of diabetes related distress.	Baseline and 8 weeks
Attitudes Toward Health Technology	Change in Composite Score of Attitudes Toward Health Technology. A 2-item self-report questionnaire on a 5 point Likert scale, scored 0 to 4. The scores are summed to create an overall score (Minimum value = 0, Maximum value = 8). Higher scores indicated more favorable attitudes toward health technology.	Baseline and 8 weeks
Attitudes Toward Medication Adherence	Change in Composite Score of Attitudes Toward Medication Adherence. A 5-item self-report questionnaire on a 5-point Likert scale (0 to 4). The scores are summed to create an overall score (Minimum value = 0, Maximum value = 20). Higher scores indicate more favorable attitudes toward medication adherence.	Baseline and 8 weeks

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Glycemic Improvement	Change in 3-day average fasting sugars between first 3 days and last 3 days of study	Day 3 and 8 weeks (assessed at the end of each 3 day period)
Percentage of Participants Who Achieved Glycemic Control	Proportion (percentage) of participants with fasting sugar within goal at end of trial.	8 weeks

Collaborators and Investigators

• Sponsor: Stanford University
• Investigators: Principal Investigator: Kevin Schulman, MD, MBA, Stanford University; Principal Investigator: Ashwin Nayak, MD, Stanford University; Principal Investigator: Sharif Vakili, MD, MS, MBA, Stanford University

Study record dates

Study Major Dates

• Study Start (Actual): March 24, 2021
• Primary Completion (Actual): December 1, 2022
• Study Completion (Actual): December 1, 2022

Study Registration Dates

• First Submitted: October 5, 2021
• First Submitted That Met QC Criteria: October 5, 2021
• First Posted (Actual): October 18, 2021

Study Record Updates

• Last Update Posted (Actual): January 2, 2024
• Last Update Submitted That Met QC Criteria: December 13, 2023
• Last Verified: December 1, 2023

More Information

Terms related to this study

• Keywords: Artificial Intelligence; AI; Medication Management; Voice Assistant; Insulin Management; Smart Speaker; Autonomous Prescriptions
• Other Study ID Numbers: 58276

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: Yes
• product manufactured in and exported from the U.S.: No