Use of Voice-Based Conversational Artificial Intelligence for Basal Insulin Prescription Management Among Patients With Type 2 Diabetes: A Randomized Clinical Trial
Authors' conclusion
In this randomized clinical trial of a voice-based conversational AI application that provided autonomous basal insulin management for adults with type 2 diabetes, participants in the AI group had significantly improved time to optimal insulin dose, insulin adherence, glycemic control, and diabetes-related emotional distress compared with those in the standard of care group. These findings suggest that voice-based digital health solutions can be useful for medication titration.
Study context
- Population
- Adults with T2D requiring initiation or adjustment of once-daily basal insulin; mean (SD) age 55.1 (12.7) years, 59.4% women
- Participants
- 32 adults
- Comparison
- Voice-based conversational artificial intelligence (AI) application vs standard of care for basal insulin management
- Voice-based conversational AIn = 16
- Standard of caren = 16reference
- Primary endpoint
- Time to optimal insulin dose (days to glycemic control), insulin adherence, and change in composite survey scores (diabetes-related emotional distress, attitudes toward health technology and medication adherence)
- Follow-up
- 8 weeks
Methodological details4 primary care clinics, academic medical center, 1:1 randomization
- 4 primary care clinics
- Academic medical center
- 1:1 randomization
Primary result
As reported in the abstract
Time to optimal insulin dose
Voice-based conversational AI
15
median · IQR 6-27 days · days
Standard of care
>56
median · IQR >29.5 to >56 days · days
P = 0.006
Secondary outcomes
| Outcome | Voice-based conversational AI | Standard of care | Difference (95% CI) | P value | |
|---|---|---|---|---|---|
| Insulin adherence (primary outcome)Mean | 82.9% (SD 20.6%) | 50.2% (SD 43.0%) | 32.7% (8.0%–57.4%) | P = 0.01 | |
| Glycemic control | 81.3% (95% CI 53.7%–95.0%) | 25.0% (95% CI 8.3%–52.6%) | 56.3% (21.4%–91.1%) | P = 0.005 | |
| −50050100Difference (linear scale) |
Points
| Outcome | Voice-based conversational AI | Standard of care | Difference (points) (95% CI) | P value | |
|---|---|---|---|---|---|
| Diabetes-related emotional distress, change in composite survey score (primary outcome) | −1.9 | 1.7 | −3.6 (−6.8 to −0.4 points) | P = 0.03 | |
| −10−505Difference, points (linear scale) |
| Outcome | Voice-based conversational AI | Standard of care | Difference (mg/dL) (95% CI) | P value | |
|---|---|---|---|---|---|
| Change in fasting blood glucoseMean, mg/dL | −45.9 (SD 45.9) | 23.0 (SD 54.7) | −68.9 (−107.1 to −30.7 mg/dL) | P = 0.001 | |
| 95% CI | −70.4 to −21.5 mg/dL | −8.6 to 54.6 mg/dL | |||
| −150−100−50050Difference, mg/dL (linear scale) |
Original publication
Use of Voice-Based Conversational Artificial Intelligence for Basal Insulin Prescription Management Among Patients With Type 2 Diabetes: A Randomized Clinical Trial. JAMA Netw Open. Published 1 December 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →