PURPOSE:Large language models (LLMs) are promising artificial intelligence (AI) tools to support clinical decision-making. The ability of LLMs to evaluate medication regimens, identify drug-drug interactions (DDIs), and provide clinical recommendations has undergone limited evaluation. The purpose of this study was to compare the performance of 3 LLMs in recognizing DDIs, determining clinical relevance, and generating management recommendations. METHODS:A total of 15 patient cases with medication regimens were created; each contained a commonly encountered DDI. Two separate study phases were developed: (1) DDI identification and determination of clinical relevance; and (2) DDI identification and generation of a clinical recommendation. The primary outcome was the ability of the LLMs (GPT-4, Gemini 1.5, and Claude 3) to identify the DDI within each medication regimen. Secondary outcomes included the ability of the LLMs to identify the clinical relevance of each DDI and generate a recommendation of high quality relative to ground truth. RESULTS:Claude 3 identified all DDIs, followed by GPT-4 (14/15, 93.3%) and Gemini 1.5 (12/15, 80.0%). All LLMs were significantly more likely than clinical experts to categorize the DDI as clinically relevant (P < 0.01). DDI management recommendations provided by GPT-4 were rated as optimal in 8 of 13 (61.5%) of the cases (P = 0.05 for comparison to ground truth). Two recommendations from GPT-4 and one recommendation from Gemini 1.5 were deemed to result in potential patient harm. CONCLUSION:While LLMs demonstrate promising potential to identify DDIs, application to clinical cases requires ongoing development. Findings from this study may assist in future development and refinement of LLMs for clinical decision-making related to DDIs.

We report results from a survey of members of the Society of Critical Care Medicine to assess ICU clinicians' perceptions of artificial airway safety practices and unplanned extubation (UE) prevention. The survey was distributed between January and February 2024 and received 518 responses (68.5% response rate), with 87.5% from adult ICUs and 12.5% from Pediatric ICUs. Only 48% of adult ICU respondents tracked UE, compared with 73% tracking pressure injuries. Most respondents did not consider UE a "never event," with over half viewing it as unavoidable. In adult ICUs, delirium was ranked as the highest UE risk factor, and commercial securement devices were the primary endotracheal tube securement method (75.2%). Significant variations were observed in artificial airway management practices and responsibility assignments across ICU settings. The results highlight substantial disparities in airway safety management beliefs and practices, underscoring the need for standardized, evidence-based guidelines.

OBJECTIVE:Endotracheal intubation is a life-saving intervention for adults with acute respiratory failure (ARF) but may result in unplanned extubation (UE). The success of UE prevention efforts has varied. We describe the development, implementation and impact of an airway safety quality improvement programme (ASQIP) embedded in an existing ABCDEF (A-F) bundle on UE occurrence. DESIGN:Before-and-after evaluation of an ASQIP. SETTING:42-bed mixed intensive care unit (ICU) at a university-affiliated teaching hospital. PATIENTS:Consecutive adult patients, endotracheally intubated for ARF. INTERVENTION:The ASQIP, developed from a literature review, the results of a national clinician survey, local clinician focus group input and root cause analyses of prior UE events, included interprofessional rounding scripts and was embedded into an existing A-F bundle. Multiple implementation strategies were employed, including didactic education to all ICU nurses (registered nurse, RN) and respiratory care therapists (RTs), the daily posting of signs of the ASQIP on the doors of rooms with a patient deemed to be at high risk for UE, and daily reminders from managers to bedside RTs and RNs. MEASUREMENTS:ASQIP implementation was effective and was associated with a significantly lower incidence of UE per 100 MV days (before 0.43 vs after 0.29; p=0.04). CONCLUSIONS:A multidisciplinary quality improvement initiative that incorporates airway safety within the A-F bundle may help reduce UE rates in critically ill adults. Future research is needed to validate standardised communication and assess the long-term sustainability of such interventions.

BACKGROUND:Mortality prediction in ICU adults is only marginally improved when medication regimen complexity (MRC) data is incorporated into traditional regression models. Machine learning (ML) may improve this prediction. OBJECTIVE:To compare the performance of different ML approaches incorporating MRC data to both traditional and advanced regression approaches, with and without MRC data, to predict hospital mortality in ICU adults. DERIVATION COHORT:Nine hundred ninety-one ICU adults at the University of North Carolina (UNC) Health System. VALIDATION COHORT:A temporally distinct cohort of 4,878 ICU adults at UNC and an external cohort of 12,290 ICU adults at the Oregon Health and Science University. PREDICTION MODEL:Supervised, classification-based ML models (e.g., Random Forest, Support Vector Machine [SVM], and XGBoost) were developed. Twenty-seven variables at ICU baseline (age, sex, service, diagnosis) and 24 hours (illness severity, supportive care use, fluid balance, laboratory values, MRC-ICU, vasopressor use) associated with mortality, and 14 missingness indicator variables, were included in each ML model. Traditional and advanced (equipped with linear predictors, predictors in nature cubic splines, predictors in smoothing cubic splines, and local linear predictors) regression models were optimized using stepwise selection by Bayesian Information Criterion. Area under the receiver operating characteristic (AUROC) was compared among models. RESULTS:Random Forest, SVM, and XGBoost achieved AUROCs of 0.83, 0.85, and 0.82, respectively, on the test set. Traditional regression models based on Sequential Organ Failure Assessment, Acute Physiology and Chronic Health Evaluation (APACHE) II, MRC-ICU + Sequential Organ Failure Assessment + APACHE II with and without an interaction term, and a full model including all 27 available variables demonstrated AUROCs of 0.81, 0.72, 0.82, 0.83, and 0.86, respectively. Advanced regression models yielded AUROCs of 0.85, 0.86, 0.85, and 0.84, respectively. The MRC-ICU exhibited a moderate level of feature importance in both XGBoost and Random Forest models. Models demonstrated lower performance in the validation cohorts. CONCLUSIONS:Use of ML, compared with traditional and advanced regression methods, did not improve hospital mortality prediction despite medication data inclusion. The MRC-ICU demonstrates moderate feature importance in select ML models.

OBJECTIVES:Medication management in the ICU is causally linked to both treatment success and adverse drug events. The purpose of this evaluation was to explore the effect of comprehensive medication management (CMM) on mortality in critically ill patients. DESIGN:Retrospective, observational, propensity-matched cohort study. SETTING:Adult ICUs at the Oregon Health Sciences University. PATIENTS:Consecutive adults admitted to an ICU greater than or equal to 24 hours between June 1, 2020, and June 7, 2023, with available pharmacist intervention data. INTERVENTIONS:None. MEASUREMENTS AND MAIN RESULTS:CMM was measured by documented critical care pharmacist (CCP) medication interventions. Propensity score matching was performed to generate a balanced 1:1 matched cohort, and logistic regression was applied for estimating propensity scores. The primary outcome was the odds of hospital mortality. Hospital and ICU length of stay were also assessed. In a cohort of 10,441 ICU patients, the unadjusted mortality rate was 11% with a mean Acute Physiology and Chronic Health Evaluation II score of 9.54 ± 4.18 and Medication Regimen Complexity-ICU (MRC-ICU) score of 5.78 ± 4.09. Compared with CCP interventions less than 3, more CCP interventions was associated with a significantly reduced risk of mortality (estimate, -0.04; 95% CI, -0.06 to -0.03; p < 0.01) and shorter length of ICU stay (estimate, -2.77; 95% CI, -2.98 to -2.56; p < 0.01). CONCLUSIONS:The quantity of CCP-delivered CMM in the ICU is directly associated with reduced hospital mortality independent of patient characteristics and MRC.

BACKGROUND:Fluid overload (FO) in the intensive care unit (ICU) is common, serious, and may be preventable. Intravenous medications (including administered volume) are a primary cause for FO but are challenging to evaluate as a FO predictor given the high frequency and time-dependency of their use and other factors affecting FO. We sought to employ unsupervised machine learning methods to uncover medication administration patterns correlating with FO. METHODS:This retrospective cohort study included 927 adults admitted to an ICU for ≥72 h. FO was defined as a positive fluid balance ≥7% of admission body weight. After reviewing medication administration record data in 3-h periods, medication exposure was categorized into clusters using principal component analysis (PCA) and Restricted Boltzmann Machine (RBM). Medication regimens of patients with and without FO were compared within clusters to assess their temporal association with FO. RESULTS:FO occurred in 127 (13.7%) of 927 included patients. Patients received a median (interquartile range) of 31(13-65) discrete intravenous medication administrations over the 72-h period. Across all 47,803 intravenous medication administrations, 10 unique medication clusters, containing 121 to 130 medications per cluster, were identified. The mean number of Cluster 7 medications administered was significantly greater in the FO cohort compared with patients without FO (25.6 vs.10.9, p < 0.0001). A total of 51 (40.2%) of 127 unique Cluster 7 medications were administered in more than five different 3-h periods during the 72-h study window. The most common Cluster 7 medications included continuous infusions, antibiotics, and sedatives/analgesics. Addition of Cluster 7 medications to an FO prediction model including the Acute Physiologic and Chronic Health Evaluation (APACHE) II score and receipt of diuretics improved model predictiveness from an Area Under the Receiver Operation Characteristic (AUROC) curve of 0.719 to 0.741 (p = 0.027). CONCLUSIONS:Using machine learning approaches, a unique medication cluster was strongly associated with FO. Incorporation of this cluster improved the ability to predict FO compared to traditional prediction models. Integration of this approach into real-time clinical applications may improve early detection of FO to facilitate timely intervention.

Background In critically ill patients, complex relationships exist among patient disease factors, medication management, and mortality. Considering the potential for nonlinear relationships and the high dimensionality of medication data, machine learning and advanced regression methods may offer advantages over traditional regression techniques. The purpose of this study was to evaluate the role of different modeling approaches incorporating medication data for mortality prediction.Methods This was a single-center, observational cohort study of critically ill adults. A random sample of 991 adults admitted ≥ 24 hours to the intensive care unit (ICU) from 10/2015 to 10/2020 were included. Models to predict hospital mortality at discharge were created. Models were externally validated against a temporally separate dataset of 4,878 patients. Potential mortality predictor variables (n=27, together with 14 indicators for missingness) were collected at baseline (age, sex, service, diagnosis) and 24 hours (illness severity, supportive care use, fluid balance, laboratory values, MRC-ICU score, and vasopressor use) and included in all models. The optimal traditional (equipped with linear predictors) logistic regression model and optimal advanced (equipped with nature splines, smoothing splines, and local linearity) logistic regression models were created using stepwise selection by Bayesian information criterion (BIC). Supervised, classification-based ML models [e.g., Random Forest, Support Vector Machine (SVM), and XGBoost] were developed. Area under the receiver operating characteristic (AUROC), positive predictive value (PPV), and negative predictive value (NPV) were compared among different mortality prediction models.Results A model including MRC-ICU in addition to SOFA and APACHE II demonstrated an AUROC of 0.83 for hospital mortality prediction, compared to AUROCs of 0.72 and 0.81 for APACHE II and SOFA alone. Machine learning models based on Random Forest, SVM, and XGBoost demonstrated AUROCs of 0.83, 0.85, and 0.82, respectively. Accuracy of traditional regression models was similar to that of machine learning models. MRC-ICU demonstrated a moderate level of feature importance in both XGBoost and Random Forest. Across all ten models, performance was lower on the validation set.Conclusions While medication data were not included as a significant predictor in regression models, addition of MRC-ICU to severity of illness scores (APACHE II and SOFA) improved AUROC for mortality prediction. Machine learning methods did not improve model performance relative to traditional regression methods.### Competing Interest StatementThe authors have declared no competing interest.### Funding StatementFunding through Agency of Healthcare Research and Quality for Drs. Devlin, Murphy, Sikora, Smith, and Kamaleswaran was provided through R21HS028485 and R01HS029009.### Author DeclarationsI confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.YesThe details of the IRB/oversight body that provided approval or exemption for the research described are given below:University of Georgia Institutional Review BoardI confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.YesI understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).YesI have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.YesAll data produced in the present study are available upon reasonable request to the authors

Antipsychotics are an attractive option for adjunctive sedation in critically ill adults given their sedating effects, short duration of action, lack of effect on respiratory drive, and multiple routes of administration but high-quality evidence from RCTs is not yet available to guide the use of these agents as adjunct sedatives in the intensive care unit. The purpose of this article is to explore the role of antipsychotics as adjunctive sedatives in mechanically ventilated adults receiving continuous sedation.

Objective: Common data models provide a standard means of describing data for artificial intelligence (AI) applications, but this process has never been undertaken for medications used in the intensive care unit (ICU). We sought to develop a common data model (CDM) for ICU medications to standardize the medication features needed to support future ICU AI efforts. Materials and Methods: A 9-member, multi-professional team of ICU clinicians and AI experts conducted a 5-round modified Delphi process employing conference calls, web-based communication, and electronic surveys to define the most important medication features for AI efforts. Candidate ICU medication features were generated through group discussion and then independently scored by each team member based on relevance to ICU clinical decision-making and feasibility for collection and coding. A key consideration was to ensure the final ontology both distinguished unique medications and met Findable, Accessible, Interoperable, and Reusable (FAIR) guiding principles. Results: Using a list of 889 ICU medications, the team initially generated 106 different medication features, and 71 were ranked as being core features for the CDM. Through this process, 106 medication features were assigned to 2 key feature domains: drug product-related (n=43) and clinical practice-related (n=63). Each feature included a standardized definition and suggested response values housed in the electronic data library. This CDM for ICU medications is available online. Conclusion: The CDM for ICU medications represents an important first step for the research community focused on exploring how AI can improve patient outcomes and will require ongoing engagement and refinement. Lay Summary Medication data pose a unique challenge for interpretation by artificial intelligence (AI) because of its alphanumerical combinations (eg, ibuprofen 200 mg every 4 hours) and the technical detail associated with drug prescriptions (eg, ibuprofen 200 mg and acetaminophen 325 mg are both starting doses and round tablet sizes, so it would be incorrect for the machine to view 325 mg as "more" than 200 mg). Because AI has great potential to improve the safety and efficacy of medication use, a common data model for ICU medications (ICURx) is proposed to overcome these challenges and support AI efforts in medication analysis.

Delirium, an acute disorder of attention and cognition, has been well-recognized as a common, serious, and potentially preventable contributor to poor outcomes in older adults, including substantial morbidity, mortality, and loss of independence, along with tremendous healthcare costs.1, 2 Despite its importance, delirium remains relatively neglected as an area of scientific investigation. To address this gap, the NIA issued RFA-AG-16-009 in 2015, intended to create a transdisciplinary, collaborative, research network to advance the field—one of the earliest NIA research networks of this kind. The grant was intended to create a robust infrastructure to support and nurture collaboration in delirium research, to provide career development for junior investigators, and to catalyze delirium research. In response, the Network for Investigation of Delirium: Unifying Scientists (NIDUS) (R24AG054259; PI: SK Inouye) was proposed and funded in 2016, with the overarching goal of developing the collaborative network and infrastructure to advance scientific research and training on the causes, mechanisms, outcomes, diagnosis, prevention, and treatment of delirium in older adults.3 Through a dedicated collaborative effort of 28 investigators (12 NIDUS leaders and 16 Scientific Advisory Board members), working across 4 cores (Figure 1), NIDUS succeeded in unifying the field, catalyzing new projects, mentoring new generations of delirium researchers, and widely disseminating information about delirium across disciplines and to the general public. The Research Resources Core created the NIDUS Research Hub, which provides a detailed index to over 1800 studies in delirium—both active and completed. The Measurement and Harmonization Core created Delirium Information Cards on more than 45 validated delirium instruments and developed tools that directly harmonize the most commonly used measures for delirium identification and severity.4, 5 The Pilot Core awarded more than 20 small grants to catalyze exploratory projects or preliminary work for future grants (~3 to 4/year). The Career Development and Dissemination Core oversaw the highly successful annual Delirium Boot Camp, a 2.5-day intensive training course in delirium research. To date, 126 mentees from 9 countries have attended 11 Boot Camps. In addition, NIDUS synergized the field with a robust website, webinars, blogs, newsletters, White Papers,6, 7 mentorship, Delirium Bibliography, annual symposia, social media, and extensive collaborations with the American Delirium Society, American Geriatrics Society (AGS), AGS CoCare HELP, and international delirium societies in Europe, Australia/Asia, and Brazil. See Table 1 for details on NIDUS accomplishments (2016–2023). Currently includes 1510 human and 255 animal studies. Searchable by investigator, study status, data types and availability of resources NIH-sanctioned Resource Sharing Site. In 2020, NIDUS was renewed as NIDUS II under an R33 mechanism (R33AG071744, MPI: SK Inouye, RN Jones), based on its success in the first cycle. The structure of NIDUS remained essentially the same; however, the renewal focused on utilizing the existing infrastructure to advance science in dedicated ways, through harmonization projects, collaborative working groups, and White Papers. The NIDUS Delirium Boot Camp, now entering its 12th year, was renewed under a separate R13 mechanism (R13AG072860, MPI: J Devlin, J Busby-Whitehead) and has provided an invaluable career development resource that has inspired and trained new generations of delirium researchers. NIDUS alumni (former Boot Camp and Pilot awardees) remain integrally involved and engaged with NIDUS, serving as Boot Camp faculty, pilot grant referees, symposium moderators, coauthors of White Papers, and so forth. NIDUS has faced numerous challenges. Pilot grants have been relatively small ($40–50,000) and limited in number (typically funding <5% of submissions). Supplements were awarded in the first cycle to increase the number and size of pilot awards; however, these were not available during the renewal. Long-term sustainability of the network is a major issue without dedicated NIH funding. Parts of the infrastructure might be maintained through different mechanisms, such as professional societies and/or philanthropy, but supporting the synergistic whole may be challenging. Heightened international funding restrictions by NIH, along with a prolonged vetting process, hinder international pilot awards and collaborative projects. Data use and sharing agreements, so essential for harmonization or collaborative work, can pose nearly insurmountable obstacles and delays for junior investigators. Unfortunately, all of these issues work against the overarching goal of developing and nurturing collaborative research. Despite the challenges, NIDUS remains a strong, vibrant, and innovative research network. The greatest strength of NIDUS is the enormous dedication and commitment of the NIDUS leaders and mentees. Moreover, the robust infrastructure created and continually improved over the past 8 years provides valuable research resources and events. The NIA set the stage for innovation and collaboration through the development of research networks in its visionary call in 2015, and NIDUS has risen to fulfill the vision and to meet the call for advancing the field of delirium research globally. This paper is supported in part by Grant Nos. R33AG071744 and R13AG072860 from the National Institute on Aging.