Heart failure & cardiomyopathiesRandomised controlled trial
Artificial Intelligence–Enabled Detection and Management of Physiological Instability in Heart Failure
- prospective, randomized, open-label, blinded-endpoint study
- 5 Veterans Affairs medical centers
- feasibility study
Authors' conclusion
AI-enabled continuous physiological monitoring was feasible and reduced AI-detected physiological instability. Although this feasibility study did not demonstrate a reduction in exploratory clinical events, these findings highlight that accurate physiological risk detection alone may be insufficient to modify clinical risk without mechanistically aligned and trusted therapeutic responses. (Detection of Physiological Anomaly Using Artificial Intelligence and Prevention of Heart Failure Hospitalization [LINK-HF2]; NCT04502563).
Study context
- Population
- Adults hospitalized for acute heart failure (HF), enrolled at discharge
- Participants
- 171 analyzed participants
- Comparison
- Artificial intelligence (AI)-guided monitoring with clinician notification vs standard care
- interventionn = 80
- controln = 91reference
Primary result
As reported in the abstract
Hazard ratio
0.90
95% CI 0.43–1.84
P = 0.76
Events by treatment group
Further findings
| Outcome | HR (95% CI) | P value | |
|---|---|---|---|
| Composite of all-cause unplanned rehospitalizations, emergency department visits, and all-cause mortality (secondary exploratory) · did not differ between groups | 1.03 (0.63–1.70) | P = 0.89 | |
| AI-detected physiological instability · lower rate with intervention | 0.76 (0.60–0.97) | P = 0.025 | |
| 0.512Hazard ratio (log scale) |
Monitoring adherence
- Median monitoring adherence (% of follow-up days)
- 72.4%
Clinician response time to communicated notifications
- Median time to response
- 1 day (range: 1-5 days)
Acceptability and barriers (qualitative analysis)
- High patient acceptance and clinician engagement
- Workflow integration and response uncertainty identified as the primary barriers to implementation
Original publication
Summarised from the published abstract; every number is checked against it