Table 6 Ablation study showing the impact of agent roles on AD prediction at -10-year prediction

From: CARE-AD: a multi-agent large language model framework for Alzheimer’s disease prediction using longitudinal clinical notes

Agents

AD cases (P/R/F)

Controls (P/R/F)

Accuracy

CARE-AD

0.20 (0.18, 0.21)/0.38 (0.35, 0.41)/0.26 (0.24, 0.28)

0.77 (0.76, 0.78)/0.57 (0.55, 0.59)/0.65 (0.64, 0.67)

0.53

(0.51, 0.54)

Zero-shot (No agents)

0.09 (0.08, 0.09)/0.25 (0.23, 0.28)/0.13 (0.11, 0.14)

0.56 (0.54, 0.57)/0.26 (0.25, 0.27)/0.35 (0.34, 0.37)

0.26

(0.25, 0.27)

Exclude neurologist

0.17 (0.16, 0.18)/0.33 (0.30, 0.36)/0.23 (0.21, 0.24)

0.75 (0.74, 0.76)/0.55 (0.53, 0.57)/0.64 (0.62, 0.65)

0.50

(0.49, 0.52)

Exclude clinical psychologist

0.19 (0.17, 0.20)/0.36 (0.33, 0.39)/0.25 (0.23, 0.27)

0.76 (0.75, 0.77)/0.57 (0.55, 0.59)/0.65 (0.64, 0.67)

0.52

(0.51, 0.54)

Exclude psychiatrist

0.17 (0.15, 0.18)/0.34 (0.31, 0.37)/0.23 (0.21, 0.24)

0.75 (0.73, 0.75)/0.53 (0.51, 0.55)/0.62 (0.61, 0.63)

0.49

(0.47, 0.50)

Exclude primary care physician

0.18 (0.17, 0.20)/0.37 (0.34, 0.40)/0.25 (0.23, 0.27)

0.76 (0.75, 0.77)/0.55 (0.53, 0.57)/0.64 (0.63, 0.65)

0.51

(0.50, 0.53)

Exclude geriatrician

0.18 (0.16, 0.19)/0.34 (0.31, 0.37)/0.23 (0.21, 0.25)

0.75 (0.74, 0.76)/0.56 (0.54, 0.58)/0.64 (0.63, 0.66)

0.51

(0.50, 0.53)

Back to article page