Fig. 1 | Scientific Reports

Fig. 1

From: Evaluating the appropriateness and safety of generative AI in delivering lifestyle guidance for atrial fibrillation patients

Fig. 1

The distribution of Bedside Manner (empathy to the user) and Helpfulness across the four groups. Mean empathy scores (1 = highest, 5 = lowest) and distribution of ranking scores (1st–4th) among the four models. Lower ranking values indicate higher accuracy. Bars represent least-squares means and 95% confidence intervals obtained from the Generalized Linear Mixed Model (GLMM) analysis. For binary analysis, empathy scores of 1–2 were coded as “1” and 3–5 as “0”, while ranking scores of 1st–2nd were coded as “1” and 3rd–4th as “0”. p values were adjusted for multiple comparisons using Dunnett’s test with the electrophysiologist as the control.

Back to article page