Table 4 Clinician validation of LLM-generated specialties: accuracy, acceptability, and error rates

From: Evaluating large language model workflows in clinical decision support for triage and referral and diagnosis

Clinician

Accurate [%]

Acceptable [%]

Accurate & Acceptable [%]

Error Rate [%]

Clinician 1

93.77

6.23

100

0

Clinician 2

82.05

8.79

90.84

9.15

Clinician 3

81.91

17.06

98.98

1.02

Clinician 4

68.94

30.34

98.98

1.02

Average

81.5

15.53

97.03

2.63

  1. Average Accuracy [%], Acceptability [%], Combined Accuracy and Acceptability [%], and Error Rate [%] for the ground truth specialties generated by the LLM, as evaluated by clinicians. The results are shown for each clinician and the overall average across all clinicians.
Back to article page