Table 3. Bivariate comparisons of characteristics by willingness to adopt large language models among health care professionals. Associations were tested using t tests for continuous variables and Chi-square tests for categorical variables.
| Variables | Willing | Unwilling or uncertain | P valuea |
|---|---|---|---|
| Sex, n (%) | .927 | ||
| Male | 69/116 (59.5) | 72/124 (58.1) | |
| Female | 47/116 (40.5) | 52/124 (41.9) | |
| Profession, n (%) | .712 | ||
| Physician | 75/116 (64.7) | 84/124 (67.7) | |
| Nurse | 41/116 (35.3) | 40/124 (32.3) | |
| Department, n (%) | .751 | ||
| Internal medicine | 27/116 (23.3) | 27/124 (21.8) | |
| Surgery | 22/116 (19) | 19/124 (15.3) | |
| ICUb | 22/116 (19) | 25/124 (20.2) | |
| Emergency | 25/116 (21.6) | 35/124 (28.2) | |
| Other | 20/116 (17.2) | 18/124 (14.5) | |
| Hospital affiliation, n (%) | .137 | ||
| Nonaffiliated | 55/116 (47.4) | 46/124 (37.1) | |
| Affiliated | 61/116 (52.6) | 78/124 (62.9) | |
| Hospital type, n (%) | .938 | ||
| Oncology | 22/116 (19) | 22/124 (17.7) | |
| General | 94/116 (81) | 102/124 (82.3) | |
| Province’s economic level, n (%) | <.001 | ||
| Low | 11/116 (9.5) | 35/124 (28.2) | |
| Medium | 34/116 (29.3) | 59/124 (47.6) | |
| High | 71/116 (61.2) | 30/124 (24.2) | |
| Frequency of digital tool use, n (%) | <.001 | ||
| Rarely | 5/116 (4.3) | 33/124 (26.6) | |
| Occasionally | 36/116 (31) | 49/124 (39.5) | |
| Often | 75/116 (64.7) | 42/124 (33.9) | |
| Previous awareness of LLMsc, n (%) | <.001 | ||
| Yes | 93/116 (80.2) | 73/124 (58.9) | |
| No | 23/116 (19.8) | 51/124 (41.1) | |
| Previous use of LLMs, n (%) | <.001 | ||
| Yes | 56/116 (48.3) | 32/124 (25.8) | |
| No | 60/116 (51.7) | 92/124 (74.2) | |
| Received AId or informatics training, n (%) | .030 | ||
| Yes | 53/116 (45.7) | 39/124 (31.5) | |
| No | 63/116 (54.3) | 85/124 (68.5) | |
| Age (years), mean (SD) | 39.5 (6.4) | 40.3 (6.6) | .382 |
| Years of clinical experience, mean (SD) | 10.7 (5.1) | 9.2 (4.4) | .017 |
| Trust in LLMs (1-5), mean (SD) | 3.20 (1.22) | 2.78 (1.26) | .011 |
| Perceived usefulness (1-5), mean (SD) | 3.70 (1.11) | 2.52 (1.02) | <.001 |
| Privacy concerns (1-5), mean (SD) | 2.82 (1.20) | 3.62 (1.33) | <.001 |
| Clarity on legal responsibility (1-5), mean (SD) | 3.41 (1.29) | 2.84 (1.29) | <.001 |
a
P<.05 was considered statistically significant.
b
ICU: intensive care unit.
c
LLM: large language model.
d
AI: artificial intelligence.