Evidence item

The packet directly supports the sample, randomized GPT-4o comparison, country-specific performance gains, and unassessed harms. DATASET_RESULT

Authenticity UNVERIFIED · extraction UNVERIFIED · independence group none (unreviewed) · created seq 17989

Source: Impact of LLM assistance on physician decision-making: a multi-country randomized controlled trial · locator {} · excerpt hash sha256:f236af82cbdf9d0ebfcc2105e25ca979a593b1fcaec2ecb9ed610a0647265425

Exact excerpt

Participants (N=249) were randomly assigned via simple randomisation to control or intervention (GPT-4o access). ... LLM assistance improved physician performance ... The largest difference ... in the Kenyan sample 18 percentage points (95% CI: 12.7 to 23.2, p < 0.001), followed by Indonesia 10.7 percentage points (95% CI: 5.7 to 15.7, p < 0.001) and the Netherlands 7.2 percentage points (95% CI: 3.7 to 10.7, p < 0.001). ... Harms were not assessed.

Linked claims

Claim	Direction	Relevance	Steps	Counted
ece240d6	SUPPORT	DIRECT	0	yes

Audits

None.