Clinical benefit of Artificial Intelligence (AI)-assisted lung ultrasound in a resource-limited intensive care unit
Authors' conclusion
AI-assisted LUS can help non-expert clinicians in an LMIC ICU improve their performance in interpreting LUS features more accurately, more quickly and more confidently.
Study context
- Population
- Clinicians interpreting point-of-care lung ultrasound (LUS) in a low-resource LMIC intensive care unit (ICU)
- Exposure
- Non-expert clinicians interpreting LUS with vs without a bespoke AI tool
- Outcome
- Accuracy of LUS interpretation, prospective real-time testing in ICU
Methodological detailsthree-phase
- Three-phase
Main finding
As reported in the abstract
P < 0.001
Observational data: associations, not effects.
Further findings
Time-to-interpret clips
P < 0.001
| Outcome | With AI tool | Without AI tool | P value |
|---|---|---|---|
| LUS interpretation accuracy by clinical user group (first phase)Experts vs Beginners, Intermediate, Advanced · experts significantly better | P < 0.001 | ||
| LUS interpretation accuracy, prospective real-time testing in the ICU (non-expert clinicians) | 93.4% (95% CI 89.0–97.8%) | 68.1% (95% CI 57.9–78.2%) | P < 0.001 |
| Clinicians' confidence level | median 4 out of 4 | median 3 out of 4 |
- LUS interpretation accuracy by clinical user group (first phase): Beginners average 68.7% (95% CI 66.8–70.7%)
- LUS interpretation accuracy by clinical user group (first phase): Intermediate average 72.2% (95% CI 70.0–75.6%)
- LUS interpretation accuracy by clinical user group (first phase): Advanced average 73.4% (95% CI 62.2–87.8%)
- LUS interpretation accuracy by clinical user group (first phase): Experts average 95.0% (95% CI 88.2–100.0%)
Original publication
Clinical benefit of AI-assisted lung ultrasound in a resource-limited intensive care unit. Crit Care. Published 1 July 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →