Table 2.
Responses to statements. P values in italics represent a significant difference (P<.05) between the Leiden University Medical Center and Amsterdam University Medical Center respondents. Results are reported on a 5-point Likert scale, ranging from 1 (strongly disagree) to 5 (strongly agree). Scores >3 indicate median agreement with the statement and results <3 median disagreement.
| Question | Total, median (IQR) | Leiden University Medical Center, median (IQR) | Amsterdam University Medical Center, median (IQR) | P valuesa | |||||
| Domain 1: Physicians’ current decision-making behavior with respect to discharging ICUb patients | |||||||||
|
|
Q1: “The decision to discharge a patient to a lower care ward is complex” | 3 (2-4) | 3 (2-3) | 4 (2-4) | .04 | ||||
|
|
Q2: “A patient’s ICU readmission risk is an important factor in my decision to discharge” | 4 (4-4) | 4 (4-4) | 4 (4-5) | .09 | ||||
|
|
Q3: “I take bed availability into account for my decision to discharge a patient” | 4 (3-4) | 4 (3-4) | 4 (3.5-4) | .43 | ||||
| Domain 2: Physicians’ perspectives on the use of artificial intelligence (AIc)–based clinical decision support tools in general | |||||||||
|
|
Q4: “I am familiar with the concept of AI” | 4 (4-4.25) | 4 (4-4) | 4 (4-5) | .004 | ||||
|
|
Q5: “I believe AI could support me in my work as physician” | 4 (4-4) | 4 (4-4) | 4 (4-4.5) | .006 | ||||
|
|
Q6: “I believe that AI will take over my job in the future” | 2 (2-3) | 2 (2-3) | 2 (2-2.5) | .22 | ||||
|
|
Q7: “I believe AI understands my work sufficiently in order to support me” | 3 (3-4) | 3 (3-4) | 3 (3-4) | .39 | ||||
|
|
Q8: “I believe in the added value of AI based decision support at the ICU” | 4 (4-4) | 4 (4-4) | 4 (4-4) | .41 | ||||
| Domain 3: Physicians’ willingness to incorporate the discharge decision support tool in daily clinical practice | |||||||||
|
|
Q9: “An AI based decision support for ICU readmission could be of positive value in the decision to discharge a patient” | 4 (4-4) | 4 (4-4) | 4 (4-4) | .02 | ||||
|
|
Q10: “It is important for me to have insight in the contributing factors to the predicted chance of readmission” | 4 (4-4.25) | 4 (4-5) | 4 (4-4) | .03 | ||||
|
|
Q18: “I assume that no readmission risk prediction score could influence my behavior” | 2 (2-2) | 2 (2-2) | 2 (2-2) | .11 | ||||
|
|
Q19: “I’m willing to consult the prediction of the decision support tool before making my decision to discharge a patient” | 4 (4-4) | 4 (4-4) | 4 (4-4) | .47 | ||||
|
|
Q20: “Taking into account the current workload at my department, I have time to take in the prediction score provided by the decision support tool and to take this into account for my decision to discharge a patient” | 4 (4-4) | 4 (3-4) | 4 (4-4) | .11 | ||||
aP values were calculated with the Mann-Whitney U test.
bICU: intensive care unit.
cAI: artificial intelligence.