A second opinion tested across three health systems.
A 249-physician study in Kenya, Indonesia, and the Netherlands tested GPT-4o assistance on clinical vignettes and reported higher guideline-based scores in each setting.
- Maturity
- Demonstrated, stage 3 of 4
- Support
- 2 sources · paper, registry
- Evidence detail
- How we know ↓
Clinicians must apply changing guidance under time and information pressure, with uneven access to specialist support.
General-purpose clinical assistants may perform differently across languages, guidelines, and health systems, and many evaluations do not include practicing clinicians across countries.
GPT-4o supplied information and reasoning support while physicians retained the task and answer.
What was shown, and what wasn’t
Shown
Among 249 physicians, guideline scores increased by 18 percentage points in Kenya, 10.7 in Indonesia, and 7.2 in the Netherlands. These were vignette scores, not patient outcomes.
Not shown · limits
- The tasks were clinical vignettes, not live care.
- Guideline-score gains do not establish safety, diagnostic accuracy, or patient benefit.
- One model version and study interface may not generalize to other tools or changing models.
Bright editorial interpretation
The study asks whether assistance transfers across settings while making the geographic differences visible.
Still open
Which specialties and case types account for errors or gains?
How does assistance affect time, overreliance, and disagreement in real practice?
Do language and local guideline differences change safety?
How we know2 sources · checked 2026-09-19 · no corrections
Original sources
- Impact of LLM assistance on physician decision-making: a multi-country randomized controlled trial ↗ · paper
- The Big Unknown: A Journey Into Generative AI's Transformative Effect on Meical Professions ↗ · registry
Institutions: Study sites in Kenya, Indonesia, and the Netherlands
- Maturity
- Demonstrated
- Claim confidence
- high
- Event date
- 2026-09-09
- Source published
- 2026-09-09
- Captured
- 2026-09-19
- Last source review
- 2026-09-19
- Editorial method
- AI-assisted source review
- Place / relevance
- Kenya, Indonesia, and the Netherlands · global-study
Bright compared this account with the linked original and supporting sources and kept reported, budgeted, projected, and observed claims distinct. Bright did not independently audit the underlying records.
Maturity describes the tested or operational setting. Confidence describes support for the particular claim; one does not determine the other.
Revision & correction history
2026-09-19 · Initial three-country vignette-study draft.
No corrections recorded.
Keep exploring
Explore the shared question in another setting. These connections do not imply replication.