| Lupus100 | ChatGPT-4 | p-value | |
|---|---|---|---|
| Word count [mean (SD)] | 241 (135) | 372 (52) | 0.001 |
| Quality score [mean (SD) [CI]] | 4.31 (0.72) [4.23 – 4.39] | 4.55 (0.65) [4.48 – 4.62] | 0.001 |
| responses scored ‘poor’ or ‘very poor’ [n (%)] | 5 (0.8%) | 4 (0.7%) | |
| Empathy score [mean (SD) [CI]] | 4.07 (0.84) [3.97 – 4.17] | 4.14 (0.82) [4.05 – 4.23] | 0.27 |
| responses scored ‘not empathetic’ [n (%)] | 1 (0.2%) | 0 (0%) | |
| Responses preferred [n (%)] | 129 (43%) | 171 (57%) | 0.02 |