Table 4. Association of AI Use by Frequency With Burnout by Demographic Status and Workload and AI Acceptance.
| Characteristics | Burnouta | Emotional exhaustion | Depersonalization | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Regularly AI use, OR (95% CI)b | Consistently AI use, OR (95% CI)b | P value for trendc | Regularly AI use, OR (95% CI)b | Consistently AI use, OR (95% CI)b | P value for trendc | Regularly AI use, OR (95% CI)b | Consistently AI use, OR (95% CI)b | P value for trendc | |
| Sex | |||||||||
| Male | 1.10 (0.97-1.24) | 1.68 (1.24-2.28)d | .002 | 1.08 (0.95-1.23) | 1.63 (1.21-2.18)d | .005 | 1.06 (0.93-1.20) | 1.84 (1.40-2.41)d | .001 |
| Female | 1.37 (1.15-1.63)d | 1.35 (0.96-1.91) | .001 | 1.38 (1.15-1.64)d | 1.37 (0.97-1.93) | <.001 | 0.96 (0.79-1.18) | 1.24 (0.83-1.86) | .66 |
| Age group | |||||||||
| <40 y | 1.23 (1.05-1.43)d | 1.53 (1.10-2.12)d | <.001 | 1.24 (1.06-1.45)d | 1.54 (1.11-2.14)d | <.001 | 1.03 (0.87-1.21) | 1.48 (1.07-2.05)d | .08 |
| ≥40 y | 1.16 (1.01-1.32)d | 1.48 (1.09-2.01)d | .003 | 1.15 (1.00-1.31)d | 1.43 (1.05-1.94)d | .006 | 1.08 (0.94-1.25) | 1.72 (1.27-2.34)d | .005 |
| Education levele | |||||||||
| Low | 1.19 (1.07-1.33)d | 1.57 (1.22-2.02)d | <.001 | 1.18 (1.06-1.32)d | 1.55 (1.21-1.98)d | <.001 | 1.09 (0.97-1.23) | 1.68 (1.32-2.15)d | <.001 |
| High | 1.26 (0.97-1.64) | 1.10 (0.68-1.77) | 0.24 | 1.33 (1.02-1.73) | 1.23 (0.75-2.10) | .08 | 0.90 (0.69-1.20) | 1.19 (0.73-1.95) | .91 |
| Hospital type | |||||||||
| Secondary | 1.14 (0.96-1.34) | 1.42 (0.89-2.28) | .04 | 1.06 (0.89-1.25) | 1.36 (0.85-2.20) | .24 | 1.06 (0.89-1.27) | 1.37 (0.85-2.20) | .22 |
| Tertiary | 1.23 (1.08-1.40)d | 1.48 (1.15-1.92)d | <.001 | 1.29 (1.13-1.47)d | 1.52 (1.18-1.96)d | <.001 | 1.06 (0.92-1.22) | 1.60 (1.24-2.05)d | .003 |
| Workload score | |||||||||
| Low score | 1.09 (0.94-1.27) | 1.73 (1.20-2.49)d | .01 | 1.07 (0.91-1.25) | 1.53 (1.06-2.21)d | .06 | 1.06 (0.90-1.25) | 1.89 (1.29-2.77)d | .02 |
| Medium score | 1.32 (1.11-1.58)d | 1.24 (0.87-1.74) | .006 | 1.30 (1.10-1.54)d | 1.23 (0.87-1.74) | .007 | 1.14 (0.95-1.37) | 1.35 (0.93-1.95) | .06 |
| High score | 1.27 (1.01-1.61) | 1.60 (1.00-2.56)d | .01 | 1.35 (1.07-1.69)d | 1.74 (1.10-2.75)d | .002 | 0.94 (0.76-1.17) | 1.71 (1.14-2.58)d | .16 |
| AI acceptance | |||||||||
| Low | 1.55 (1.24-1.92)d | 2.50 (1.33-4.71)d | <.001 | 1.50 (1.21-1.86)d | 2.68 (1.43-5.02)d | <.001 | 1.43 (1.13-1.79)d | 1.45 (0.81-2.58) | .004 |
| High | 1.10 (0.98-1.24) | 1.30 (1.02-1.65) | .01 | 1.12 (0.99-1.25) | 1.26 (0.99-1.60) | .02 | 0.97 (0.86-1.10) | 1.54 (1.22-1.96)d | .04 |
Abbreviations: AI, artificial intelligence; OR, odds ratio.
a
Burnout was defined as having at least 1 symptom of the emotional exhaustion (≥27) or depersonalization (≥10).
b
The regressions with inverse probability weighting were fitted by taking the frequency of AI use as categorical variable, adjusting for covariates and random effect of provinces, with the exception of stratified variables, which were mutually adjusted.
c
P for trend was calculated by the same model using frequency of AI use as continuous variable.
d
P < .05 with multiple testing correction by Hochberg method.
e
Education level was categorized as bachelor’s degree and below and graduate degree and above.