Generative AI-Assisted Clinical Decision Support for Medical Intensive Care Unit Physicians
19 juli 2026 uppdaterad av: Seoul National University Hospital
Evaluation of the Feasibility and Effectiveness of Generative AI-Assisted Multidisciplinary Decision Support in Medical Intensive Care: A Pilot Randomized Controlled Trial
This pilot study evaluated the feasibility and usefulness of generative artificial intelligence (AI) as a clinical decision-support tool for physicians working in a medical intensive care unit.
Participating physicians were assigned by work period to either use a generative AI system in addition to usual clinical information resources or to use usual resources without generative AI.
The assigned condition was then switched so that participants experienced both approaches.
During the AI-assisted periods, physicians used de-identified clinical information and considered the AI-generated responses as reference information.
All final clinical decisions remained the responsibility of the treating physicians.
The study assessed acceptability, usability, satisfaction, perceived decision support, workload, confidence, and learning experience through repeated questionnaires.
Studieöversikt
Status
Avslutad
Betingelser
Intervention / Behandling
Detaljerad beskrivning
This was a single-center, open-label, pilot cluster-randomized crossover study involving physicians working in a medical intensive care unit. Each participating physician was observed during a scheduled one-month rotation in the medical intensive care unit. At the beginning of each monthly rotation, participating physicians were divided into two clusters. The clusters were randomized to begin with either the ChatGPT-assisted condition or the control condition. After approximately two weeks, each cluster crossed over to the alternate condition for the remainder of the one-month rotation. This design allowed participating physicians to experience both study conditions within the same rotation.
During the AI-assisted condition, physicians were encouraged to use ChatGPT (OpenAI) as a reference tool to support clinical information review and decision-making. Only non-identifiable clinical information was permitted to be entered into ChatGPT. Patient names, medical record numbers, contact information, and other information that could directly identify an individual patient were not entered. Physicians summarized clinically relevant information in their own words and considered the responses generated by ChatGPT when planning patient management. The Situation-Background-Assessment-Recommendation framework was recommended as an optional structure for organizing clinical information, but its use was not mandatory. Physicians were otherwise free to formulate their queries and interact with ChatGPT according to their clinical needs. A suggested prompt encouraged ChatGPT to present multiple management options, together with their rationale, potential benefits and risks, relevant supporting evidence, and areas of uncertainty. ChatGPT did not make or implement clinical decisions.
During the control condition, physicians used usual information resources, including discussions with other clinicians, multidisciplinary rounds, consultations, textbooks, clinical practice guidelines, PubMed, and other established clinical reference services, without using ChatGPT or other generative AI tools for study-related clinical decision support.
All diagnostic and treatment decisions were made independently by the treating physicians. Repeated questionnaires assessed satisfaction, decision-making experience, confidence, perceived efficiency, workload, educational value, and other aspects of clinical decision support. At study completion, participants also evaluated usability, satisfaction, perceived learning, reliance on ChatGPT, intention for future use, and the extent to which ChatGPT-generated suggestions were reflected in their clinical plans.
Studietyp
Interventionell
Inskrivning (Faktisk)
15
Fas
- Inte tillämpbar
Kontakter och platser
Det här avsnittet innehåller kontaktuppgifter för dem som genomför studien och information om var denna studie genomförs.
Studieorter
-
-
Seoul
-
Seoul, Seoul, Sydkorea, 03080
- Seoul National University Hospital
-
-
Deltagandekriterier
Forskare letar efter personer som passar en viss beskrivning, så kallade behörighetskriterier. Några exempel på dessa kriterier är en persons allmänna hälsotillstånd eller tidigare behandlingar.
Urvalskriterier
Åldrar som är berättigade till studier
- Vuxen
- Äldre vuxen
Tar emot friska volontärer
Ja
Beskrivning
Inclusion Criteria:
- Age 19 years or older.
- Physicians, including residents, fellows, and attending physicians, working in the medical intensive care unit at Seoul National University Hospital.
- Scheduled to work as a primary treating physician for at least 5 days during a planned observation period.
- Able and willing to provide written informed consent.
Exclusion Criteria:
- Did not provide written informed consent.
- Withdrew consent from study participation.
Studieplan
Det här avsnittet ger detaljer om studieplanen, inklusive hur studien är utformad och vad studien mäter.
Hur är studien utformad?
Designdetaljer
- Primärt syfte: Hälsovårdsforskning
- Tilldelning: Randomiserad
- Interventionsmodell: Crossover tilldelning
- Maskning: Ingen (Open Label)
Vapen och interventioner
Deltagargrupp / Arm |
Intervention / Behandling |
|---|---|
|
Experimentell: ChatGPT-Assisted Condition First, Then Control Condition
Physician clusters used ChatGPT-assisted clinical decision support during the first approximately two weeks of their one-month medical intensive care unit rotation.
They then crossed over to the control condition and used usual clinical information resources without generative AI for the remainder of the rotation.
|
During the assigned period, physicians were encouraged to use ChatGPT (OpenAI) as a generative AI-based reference tool to support clinical information review and decision-making.
During the control period, physicians used usual clinical information resources, including discussions with other clinicians, multidisciplinary rounds, specialty consultations, textbooks, clinical practice guidelines, PubMed, and established clinical reference services.
No generative AI tool was used for clinical decision support during this period.
|
|
Experimentell: Control Condition First, Then ChatGPT-Assisted Condition
Physician clusters used usual clinical information resources without generative AI during the first approximately two weeks of their one-month medical intensive care unit rotation.
They then crossed over to the ChatGPT-assisted clinical decision-support condition for the remainder of the rotation.
|
During the assigned period, physicians were encouraged to use ChatGPT (OpenAI) as a generative AI-based reference tool to support clinical information review and decision-making.
During the control period, physicians used usual clinical information resources, including discussions with other clinicians, multidisciplinary rounds, specialty consultations, textbooks, clinical practice guidelines, PubMed, and established clinical reference services.
No generative AI tool was used for clinical decision support during this period.
|
Vad mäter studien?
Primära resultatmått
Resultatmått |
Åtgärdsbeskrivning |
Tidsram |
|---|---|---|
|
Daily Physician Satisfaction Score
Tidsram: At the end of each working day during the one-month medical intensive care unit rotation
|
The score of 10 questionnaire items assessing physicians' satisfaction with their daily clinical work, modified from Shore and Franks (1986) and Suchman et al. (1993).
Each item was rated on a 5-point Likert scale from -2 (strongly disagree) to +2 (strongly agree).
Negatively worded items were reverse-scored.
The mean score ranges from -2 to +2, with higher scores indicating greater satisfaction.
|
At the end of each working day during the one-month medical intensive care unit rotation
|
|
Daily Clinical Decision-Making Score
Tidsram: At the end of each working day during the one-month medical intensive care unit rotation
|
The score of 6 questionnaire items assessing satisfaction with the clinical decision-making process, perceived decision difficulty, clarity of the preferred treatment, availability of relevant information, and identification of factors affecting the decision.
Items were modified from Gedney (1994) and Dolan (1999) and rated on a 5-point Likert scale from -2 (strongly disagree) to +2 (strongly agree).
Negatively worded items were reverse-scored.
The mean score ranges from -2 to +2, with higher scores indicating a more favorable decision-making experience.
|
At the end of each working day during the one-month medical intensive care unit rotation
|
|
Perceived Quality Score for ChatGPT
Tidsram: At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
The score of 8 questionnaire items assessing the perceived information quality, system quality, and service quality of ChatGPT, modified from Pillong et al. (2025).
Each item was rated on a 5-point Likert scale from -2 (strongly disagree) to +2 (strongly agree).
The negatively worded response-time item was reverse-scored.
The mean score ranges from -2 to +2, with higher scores indicating better perceived quality.
|
At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
|
Generative AI Usability Score
Tidsram: At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
The score of 3 questionnaire items assessing ease of use, ease of learning, and clarity of interaction with Generative AI (ChatGPT), modified from Pillong et al. (2025).
Each item was rated on a 5-point Likert scale from -2 (strongly disagree) to +2 (strongly agree).
The mean score ranges from -2 to +2, with higher scores indicating greater usability.
|
At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
|
Satisfaction Score for Generative AI Use
Tidsram: At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
The score of 6 questionnaire items assessing the perceived usefulness, productivity, effectiveness, overall satisfaction, appropriateness, and intention to reuse Generative AI (ChatGPT), modified from Pillong et al. (2025).
Each item was rated on a 5-point Likert scale from -2 (strongly disagree) to +2 (strongly agree).
The mean score ranges from -2 to +2, with higher scores indicating greater satisfaction.
|
At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
Andra resultatmått
Resultatmått |
Åtgärdsbeskrivning |
Tidsram |
|---|---|---|
|
Confidence
Tidsram: At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
A single questionnaire item assessing whether use of Generative AI (ChatGPT) increased the physician's confidence in clinical decision-making.
The item was rated on a 5-point Likert scale from -2 (strongly disagree) to +2 (strongly agree), with higher scores indicating a greater perceived increase in confidence.
|
At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
|
Perceived Acquisition of New Knowledge or Clinical Insight
Tidsram: At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
A single questionnaire item assessing whether the physician acquired new knowledge or clinical insight while using Generative AI (ChatGPT).
The item was rated on a 5-point Likert scale from -2 to +2, with higher scores indicating greater perceived learning.
|
At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
|
Application of Generative AI-Derived Knowledge to Other Clinical Situations
Tidsram: At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
A single questionnaire item assessing whether knowledge obtained through Generative AI (ChatGPT) was applied to the care of other patients.
The item was rated on a 5-point Likert scale from -2 to +2, with higher scores indicating greater transfer of learning.
|
At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
|
Intention to Continue Using Generative AI (ChatGPT) in Future Clinical Practice
Tidsram: At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
A single questionnaire item assessing the physician's intention to continue using Generative AI (ChatGPT) in future clinical practice.
The item was rated on a 5-point Likert scale from -2 to +2, with higher scores indicating stronger intention for continued use.
|
At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
|
Perceived Reliance on Generative AI (ChatGPT)
Tidsram: At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
A single questionnaire item assessing whether the physician perceived increased reliance on Generative AI (ChatGPT) during clinical care.
The item was rated on a 5-point Likert scale from -2 to +2, with higher scores indicating greater perceived reliance.
|
At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
|
Perceived Reduction in Clinical Workload
Tidsram: At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
A single questionnaire item assessing whether use of Generative AI (ChatGPT) reduced the physician's perceived clinical workload.
The item was rated on a 5-point Likert scale from -2 to +2, with higher scores indicating greater perceived workload reduction.
|
At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
|
Clinical Specialty Area in Which Generative AI (ChatGPT) Was Most Helpful
Tidsram: At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
The clinical specialty area in which the physician perceived the greatest practical benefit from Generative AI (ChatGPT), selected from predefined categories or reported as free text, including cardiology, infectious diseases, pulmonology, and nephrology.
|
At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
|
Percentage of Generative AI (ChatGPT) Suggestions Reflected in Clinical Plans
Tidsram: At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
The physician-reported percentage of Generative AI (ChatGPT)-generated suggestions that were reflected in actual clinical management plans.
Scores range from 0% to 100%, with higher percentages indicating greater incorporation of ChatGPT suggestions.
|
At the end of the one-month medical intensive care unit rotation, after completion of both crossover periods
|
Samarbetspartners och utredare
Det är här du hittar personer och organisationer som är involverade i denna studie.
Utredare
- Huvudutredare: Minju Han, M.D., Seoul National University Hospital
Studieavstämningsdatum
Dessa datum spårar framstegen för inlämningar av studieposter och sammanfattande resultat till ClinicalTrials.gov. Studieposter och rapporterade resultat granskas av National Library of Medicine (NLM) för att säkerställa att de uppfyller specifika kvalitetskontrollstandarder innan de publiceras på den offentliga webbplatsen.
Studera stora datum
Studiestart (Faktisk)
1 december 2025
Primärt slutförande (Faktisk)
31 maj 2026
Avslutad studie (Faktisk)
31 maj 2026
Studieregistreringsdatum
Först inskickad
12 juli 2026
Först inskickad som uppfyllde QC-kriterierna
12 juli 2026
Första postat (Faktisk)
16 juli 2026
Uppdateringar av studier
Senaste uppdatering publicerad (Faktisk)
21 juli 2026
Senaste inskickade uppdateringen som uppfyllde QC-kriterierna
19 juli 2026
Senast verifierad
1 juli 2026
Mer information
Termer relaterade till denna studie
Nyckelord
Andra studie-ID-nummer
- 2510-145-1689
Plan för individuella deltagardata (IPD)
Planerar du att dela individuella deltagardata (IPD)?
NEJ
IPD-planbeskrivning
Individual participant data will not be shared because of the small sample size, the limited number of physicians working in the study setting, and the potential risk of re-identification even after removal of direct identifiers.
External sharing of individual-level data was not included in the participant consent or institutional review board-approved data management plan.
Läkemedels- och apparatinformation, studiedokument
Studerar en amerikansk FDA-reglerad läkemedelsprodukt
Nej
Studerar en amerikansk FDA-reglerad produktprodukt
Nej
Denna information hämtades direkt från webbplatsen clinicaltrials.gov utan några ändringar. Om du har några önskemål om att ändra, ta bort eller uppdatera dina studieuppgifter, vänligen kontakta register@clinicaltrials.gov. Så snart en ändring har implementerats på clinicaltrials.gov, kommer denna att uppdateras automatiskt även på vår webbplats .