Intensive care medicine: Overview

The “ICU Toolbox” collection provides practical advice on a range of special focus topics. It aims to be a beacon of practical insights, offering concise and insightful hands-on guidance for healthcare professionals navigating the intricate landscapes of intensive care, and providing them with tangible tools and strategies to enhance their daily practice.

Abstract
Background: Postoperative ICU admission is crucial in surgical patient management, impacting morbidity and mortality outcomes. Traditional prediction methods, such as the ASA physical status classification, are subjective and prone to variability. This study evaluates the effectiveness of ChatGPT-4 in predicting postoperative ICU admission needs using comprehensive preoperative patient data.

Methods: In this prospective multicentric study, data from 406 patients aged 18 and older were analyzed. Patients requiring emergency surgery and those with insufficient information were excluded. Preoperative data, including demographics, medical history, laboratory results, and imaging findings, were collected and anonymized. ChatGPT-4 was configured to predict ICU admission needs based on this data. The model's predictions were compared with actual ICU admissions using Chi-Square, confusion matrix, and one-sample t-tests.

Results: ChatGPT-4 model predicted 128 patients for ward care and 278 for ICU admission. Among the predicted ICU cases, 160 were correctly identified as ICU, while 118 were observed to need ward care. The overall accuracy of the model was 0.645, with a specificity of 0.464, sensitivity of 0.860, and an F1 score of 0.690. A chi-square test revealed a significant result (P=0.000).

Conclusions: The findings demonstrate that ChatGPT-4 can effectively predict postoperative ICU needs, providing a valuable tool for clinical decision-making. While the model showed strong agreement with actual ICU admissions, refinement is needed to improve the accuracy of ICU stay duration predictions. Integrating AI in preoperative assessments could enhance objectivity and efficiency, contributing to optimized patient care and resource allocation. Further validation across diverse patient populations is recommended.

Abstract in English, French
Purpose: The Brief Illness Perception Questionnaire (Brief-IPQ) is used to quantify a patient's understanding of their condition. In critical care settings, substitute decision makers (SDMs) are tasked with making decisions on behalf of patients, but it can be difficult to evaluate how they perceive the patient's illness. We developed the Illness Perception Questionnaire for Substitute Decision Makers (IPQ-SDM) by adapting the Brief-IPQ and testing it with SDMs of patients in critical care.

Methods: In a mixed methods study, we modified the Brief-IPQ for the context of SDMs of critically ill patients using qualitative feedback. We then approached SDMs of critically ill patients three days after intensive care unit (ICU) admission to complete the IPQ-SDM in a prospective observational study. We calculated summary statistics for the individual questions and the total score. We evaluated prespecified associations between individual questions as well as the total score with clinical data that were abstracted from the patient chart.

Results: We received 44/50 (88%) complete responses with a mean (standard deviation [SD]) total score of 43 (11). Forty-six percent reported that they did not have any in-person visits due to the restrictions during the COVID-19 pandemic. The correlation of total score with the nine equivalents of nursing manpower score (NEMS) was 0.42 (P = 0.005) but it was not correlated (P = 0.07) with the multiple organ dysfunction score (MODS). The total mean (SD) score was higher (i.e., worse) for patients who died in hospital (48.2 [9.7], n = 13) than for those who were discharged alive (40.5 [10.2], n = 31, P = 0.03). Concern regarding the illness was correlated with MODS (Spearman correlation coefficient [ρ] = 0.39, P = 0.01) and NEMS (ρ = 0.29, P = 0.04).

Conclusion: We observed a modest level of correlation of the IPQ-SDM score with clinical measures in the adult ICU setting. Further research is needed before such a tool is broadly implemented.

Abstract
Background: Critically ill patients often experience substantial stress during their ICU treatment. The ICU Feel Better App is a novel mobile application that patients can use to evaluate ICU-related stressors during their stay. We aimed to investigate if using the app, without feedback to the ICU staff, would be associated with changes in perceived acute stress.

Methods: This prospective cohort study used a before-and-after design. Data were collected at the beginning of ICU treatment (T1) and at discharge from ICU (T2). The comparison group (N.=20) received treatment as usual, i.e. standard critical care including assessment and treatment of pain and delirium. The App group (N.=20) used the ICU Feel Better App, which included 24 common ICU-related stressors, between T1 and T2. Secondary outcomes included app use extent and patients' ratings of acceptance and usability.

Results: No significant differences were observed in clinical characteristics or stress levels between treatment groups at T1, T2, and from T1 to T2. Each app session averaged a median of 1.82 minutes (IQR: 1.54-2.66), with users averaging 8.5 sessions over 3.5 days, totaling 16.16 minutes of use (IQR: 9.72-27.51). Patients found the app highly usable and acceptable, with assistance needed in only one session for three users and no premature terminations.

Conclusions: The lack of significant stress reduction suggests that the effectiveness of the app could be improved by incorporating feedback mechanisms to communicate patients' stressor evaluations to staff, prompting stress-reducing measures. Data on app use, acceptability, and usability indicate feasibility of the ICU Feel Better App.