Table 3.
Associations of generative AI (GenAI) usage, perceived usefulness (PU), and perceived risk (PR) with physician burnout and high fulfillment: a fixed-effects logistic regression analysis.
|
|
Burnout | High fulfillment | ||||||
|
|
OR (95% CI) | P value | OR (95% CI) | P value | ||||
| Modela | ||||||||
|
|
The frequencies of using GenAI | |||||||
|
|
|
Mostly | Reference | Reference | Reference | Reference | ||
|
|
|
Occasionally | 1.46 (0.94-2.28) | .09 | 0.57 (0.36-0.91) | .02 | ||
|
|
|
Rarely | 1.31 (0.80-2.14) | .28 | 0.62 (0.37-1.05) | .08 | ||
|
|
|
Never | 1.24 (0.74-2.08) | .41 | 0.64 (0.37-1.11) | .11 | ||
|
|
PU | 0.79 (0.65-0.97) | .02 | 1.85 (1.48-2.33) | <.001 | |||
|
|
PR | 2.09 (1.74-2.51) | <.001 | 0.77 (0.64-0.92) | .005 | |||
| Modelb | ||||||||
|
|
The frequencies of using GenAI | |||||||
|
|
|
Mostly | Reference | Reference | Reference | Reference | ||
|
|
|
Occasionally | 1.19 (0.73-1.94) | .48 | 0.63 (0.37-1.06) | .08 | ||
|
|
|
Rarely | 1.02 (0.59-1.78) | .94 | 0.74 (0.41-1.35) | .33 | ||
|
|
|
Never | 1.10 (0.60-2.00) | .76 | 0.74 (0.39-1.37) | .34 | ||
|
|
PU | 0.84 (0.66-1.08) | .18 | 1.56 (1.17-2.08) | .003 | |||
|
|
PR | 1.80 (1.46-2.21) | <.001 | 0.86 (0.69-1.08) | .21 | |||
aCrude models.
bModels adjust for age, gender, marital status, ethnicity, salary satisfaction, professional title, years in practice, night shift, weekly patient volume, specialty, teaching assessment pressure, research assessment pressure, and province.