Table 2 The scoring for ChatGPT’ s diagnosis output

1

All diagnosis is incorrect

0-20% diagnoses are included

More than 5 text errors

2

Chief complaint related diagnosis is incorrect;

Partial chief complaint unrelated diagnoses are correct

20-40% diagnoses are included

3 ~ 4 text errors

3

Chief complaint related diagnosis is incorrect;

All chief complaint unrelated diagnoses are correct

40-60% diagnoses are included

2 text errors

4

Chief complaint related diagnosis is correct;

Partial chief complaint unrelated diagnoses are correct

60-80% diagnoses are included

1 text error

5

All diagnoses are correct

80-100% diagnoses are included

No text error