Table 4 Chapter-level coding performance in micro-F1 score of CliniCoCo and its variant without adopting collaborative annotation strategy
From: Optimising the paradigms of human AI collaborative clinical coding
Setting | Dataset | Chapter-level diagnosis | ||||||
|---|---|---|---|---|---|---|---|---|
I | E | J | K | Z | N | D | ||
CliniCoCo | HPH-50 | 0.8473 | 0.8689 | 0.7943 | 0.8244 | 0.8573 | 0.8254 | 0.8241 |
APH-50 | 0.8238 | 0.8582 | 0.8252 | 0.7992 | 0.7211 | 0.7864 | 0.8672 | |
Collaborative annotation w/o | HPH-50 | 0.8475 | 0.8693 | 0.8012 | 0.8210 | 0.8302 | 0.8067 | 0.8087 |
APH-50 | 0.8218 | 0.8611 | 0.8179 | 0.7956 | 0.6883 | 0.7745 | 0.8423 | |
Difference value | HPH-50 | −0.0002 | −0.0004 | −0.0069 | +0.0034 | +0.0271 | +0.0187 | +0.0154 |
APH-50 | +0.0021 | −0.0029 | +0.0073 | +0.0036 | +0.0328 | +0.0119 | +0.0249 | |
O | C | R | G | B | F | |||
CliniCoCo | HPH-50 | 0.7188 | 0.8117 | 0.7824 | N/A | N/A | N/A | |
APH-50 | N/A | 0.7455 | 0.7486 | 0.8287 | 0.8213 | 0.8286 | ||
Collaborative annotation w/o | HPH-50 | 0.6925 | 0.7946 | 0.7506 | N/A | N/A | N/A | |
APH-50 | N/A | 0.7222 | 0.7264 | 0.8043 | 0.8002 | 0.8019 | ||
Difference value | HPH-50 | +0.0263 | +0.0171 | +0.0318 | N/A | N/A | N/A | |
APH-50 | N/A | +0.0233 | +0.0222 | +0.0244 | +0.0211 | +0.0267 | ||