Table 1.
Grading of responses generated by GPT-4 to nutrition questions related to IBD stratified by question category.
| Question Category and Grading Scale | Number of responses (%) |
|---|---|
| Nutrition and diet needs for surgery (N = 16) | |
| 1. Comprehensive | 10 (62.5) |
| 2. Correct but inadequate | 4 (25.0) |
| 3. Mixed with correct and incorrect/outdated data | 2 (12.5) |
| 4. Completely incorrect | 0 |
| Tube feeding and parenteral nutrition (N = 13) | |
| 1. Comprehensive | 12 (92.3) |
| 2. Correct but inadequate | 0 |
| 3. Mixed with correct and incorrect/outdated data | 1 (7.7) |
| 4. Completely incorrect | 0 |
| General diet questions (N = 17) | |
| 1. Comprehensive | 11 (64.7) |
| 2. Correct but inadequate | 5 (29.4) |
| 3. Mixed with correct and incorrect/outdated data | 1 (5.8) |
| 4. Completely incorrect | 0 |
| Diet for reducing symptoms/inflammation (N = 20) | |
| 1. Comprehensive | 10 (50.0) |
| 2. Correct but inadequate | 2 (10.0) |
| 3. Mixed with correct and incorrect/outdated data | 8 (40.0) |
| 4. Completely incorrect | 0 |
| Micronutrients/supplementation needs (N = 22) | |
| 1. Comprehensive | 18 (81.8) |
| 2. Correct but inadequate | 1 (4.5) |
| 3. Mixed with correct and incorrect/outdated data | 3 (13.6) |
| 4. Completely incorrect | 0 |