Table 5 Odds-ratios (or) for each endpoint, estimated separately for experts and non experts. We report in brackets the \(95\%\) confidence intervals.
From: Experimental evidence of effective human–AI collaboration in medical decision-making
Endpoint | or | Expert | Non-expert |
|---|---|---|---|
1. Influence of the ai | \(\omega _I\) | 2.88 [2.48, 3.34] | 3.20 [2.80, 3.65] |
2. Diagnostic accuracy | \(\omega _A\) | 1.15 [1.01, 1.30] | 1.61 [1.44, 1.79] |
3. Effectiveness | \(\omega _E\) | 3.22 [2.64, 3.93] | 3.65 [3.11, 4.28] |
4. Safety | \(\omega _S\) | 0.45 [0.37, 0.54] | 0.63 [0.54, 0.75] |
