Not applicable. Not applicable.
Conditions
Interventions
AI-enabled informational assistant for ICU rounds.
Sponsors
Erasmus MC, Universitair Medisch Centrum Rotterdam
Eligibility
Age
18 Years to 99 Years
Inclusion criteria
Inclusion criteria: ICU physician (nurse practicioner, resident, or staff intensivist) at the Erasmus MC.Signed informed-consent for study participation.
Exclusion criteria
Exclusion criteria: Physicians not expected to work on the ICU during the study period will not be approached.
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| The primary outcomes are acceptability, appropriateness and feasibility, assessed using the standardized instruments Acceptability of Intervention Measure (AIM), Intervention Appropriateness Measure (IAM) and Feasibility of Intervention Measure (FIM). The mean scores and standard deviations of the 5-point Likert scale (1 = strongly disagree, 2 = disagree, 3 = undecided, 4 = agree, 5 = strongly agree) closed-ended questions of the survey on the physicians’ perceptions on the acceptability, appropriateness and feasibility of the AI-based informational assistant will be calculated. A difference in mean scores over the four time points will be assessed with repeated measurement ANOVA.The interview data on acceptability, appropriateness and feasibility will undergo qualitative synthesis using thematic analysis, which involves stepwise data familiarisation, coding, and theme generation. Two researchers will independently code the interview transcripts to ensure the accuracy and consistency of the identified themes. Any discrepancies between the researchers’ codes will be resolved through discussion, or if necessary, with the involvement of a third researcher. Emerging themes will be iteratively refined, and illustrative examples will be drawn from the data to ensure they accurately reflect participants' perspectives and the overall dataset. | — |
Secondary
| Measure | Time frame |
|---|---|
| Secondary outcomes include adoption and fidelity, physicians’ self-assessed retrieval speed, factual accuracy and protocol adherence, clinician–AI interaction patterns, trust, workload, and performance and safety evaluation of the output of the AI-based informational assistant .Fidelity will be determined by the misusage per participant, reported as counts and proportions. Adoption will be determined by frequency of use (interactions per participant per week) and retention (continued use over time), expressed as counts and proportions. Adoption and fidelity will be aggregated at both participant and cohort level.The mean scores and standard deviations of the 5-point Likert scale closed-ended questions of the survey on the physicians’ perceptions on retrieval speed, factual accuracy and protocol adherence will be calculated.The overall NASA-TLX workload score for each participant was calculated as the average of the six subscale scores. The mean and standard deviations of the six NASA-TLX subscales and overall scale (mental, physical, temporal, performance, effort, and frustration) were calculated. The trust score for each participant will be determined by taking the means of the twelve 7-point Likert scale rated items (1 = Strongly Disagree, 2 = Disagree, 3 = Somewhat Disagree, 4 = Neutral, 5 = Somewhat Agree, 6 = Agree, 7 = Strongly Agree). The mean per item and overall trust score will be calculated.Qualitative interview and documentation output data will undergo thematic analysis to characterize interaction styles, use patterns, and user experience. Thematic analysis involves stepwise data familiarisation, coding, and theme generation. Two researchers will independently code the interview transcripts and output to ensure the accuracy and consistency of the identified themes. Any discrepancies between the researchers’ codes will be resolved through discussion, or if necessary, with the involvement of a third researcher. Emerging themes will be iterativel | — |
Countries
Netherlands
Contacts
Public ContactM.E. Genderen
Erasmus MC, Universitair Medisch Centrum Rotterdam
Outcome results
None listed