Gajic, Ognjen

Critically ill patients presenting with acute on chronic liver failure (ACLF) represent a particularly vulnerable population due to various considerations surrounding the syndrome definition, lack of robust prospective evaluation of outcomes, and allocation of resources such as organs for transplantation. Ninety-day mortality related to ACLF is high and patients who do leave the hospital are frequently readmitted. Artificial intelligence (AI), which encompasses various classical and modern machine learning techniques, natural language processing, and other methods of predictive, prognostic, probabilistic, and simulation modeling, has emerged as an effective tool in various areas of healthcare. These methods are now being leveraged to potentially minimize physician and provider cognitive load and impact both short-term and long-term patient outcomes. However, the enthusiasm is tempered by ethical considerations and a current lack of proven benefits. In addition to prognostic applications, AI models can likely help improve the understanding of various mechanisms of morbidity and mortality in ACLF. Their overall impact on patient-centered outcomes and countless other aspects of patient care remains unclear. In this review, we discuss various AI approaches being utilized in healthcare and discuss the recent and expected future impact of AI on patients with ACLF through prognostic modeling and AI-based approaches.

ABSTRACT Objective: To identify and describe characteristics of patients with sepsis who could be treated with minimally invasive sepsis (MIS) approach without intensive care unit (ICU) admission and to develop a prediction model to select candidates for MIS approach. Methods: A secondary analysis of the electronic database of patients with sepsis at Mayo Clinic, Rochester, MN. Candidates for the MIS approach were adults with septic shock and less than 48 hours of ICU stay, who did not require advanced respiratory support and were alive at hospital discharge. Comparison group consisted of septic shock patients with an ICU stay of more than 48 hours without advanced respiratory support at the time of ICU admission. Results: Of 1795 medical ICU admissions, 106 patients (6%) met MIS approach criteria. Predictive variables (age >65 years, oxygen flow >4 L/min, temperature <37°C, creatinine >1.6 mg/dL, lactate >3 mmol/L, white blood cells >15 × 109/L, heart rate >100 beats/min, and respiration rate >25 breaths/min) selected through logistic regression were translated into an 8-point score. Model discrimination yielded the area under the receiver operating characteristic curve of 79% and was well fitted (Hosmer-Lemeshow P = 0.94) and calibrated. The MIS score cutoff of 3 resulted in a model odds ratio of 0.15 (95% confidence interval, 0.08–0.28) and a negative predictive value of 91% (95% confidence interval, 88.69–92.92). Conclusions: This study identifies a subset of low-risk septic shock patients who can potentially be managed outside the ICU. Once validated in an independent, prospective sample our prediction model can be used to identify candidates for MIS approach.

Abstract Background CERTAIN (Checklist for Early Recognition and Treatment of Acute Illness and iNjury) education program was developed to accelerate the global dissemination of a standardized, systemic, structured approach to critical care delivery. The coronavirus disease (COVID-19) pandemic prompted the evolution of this program from a live in-person course to a blended synchronous and asynchronous learning experience, including virtual simulation. Objectives We describe our experience and insights gained through this digital program transformation and highlight areas in need of further research to advance the delivery of high-quality online education offerings to global interprofessional audiences. Methods The CERTAIN education program was delivered to a broad international audience first in person (2016–2019) and then virtually during the COVID-19 global pandemic (2020–present). During this transition, we adopted a flipped classroom model to deliver the core content asynchronously using an online learning management system, supplemented by a novel synchronous online experience to provide learners with the opportunity to apply these concepts using a series of simulated clinical cases. Results A total of 400 participants attended 11 CERTAIN courses. We transitioned our 10-hour live course to a 3-hour virtual workshop. The duration of simulation activities (admission, rounding, and shared decision-making) remained constant. Didactic lectures were eliminated from the synchronous online course and presented as recorded videos in precourse materials. We collected 306 postcourse surveys (response rate, 76.5%). The majority of the overall course ratings were excellent (147 [49.5%]) and very good (97 [32.7%]), and learner responses were similar to live and online courses. Simulation activities were consistently the most popular elements of our program. Access to digital learning platforms and language barriers during simulation activities proved to be the greatest challenges during our transition. Delivering mobile-friendly online content and close coordination between dedicated bilingual faculty and local champions helped overcome these challenges. Conclusion Critical care education and case-based simulation workshops can be delivered to international interprofessional audiences with similar, high degrees of learner satisfaction to in-person offerings.

ImportanceThe conflict in Ukraine has forced civilian hospitals with limited trauma and battlefield medicine experience to care for casualties of war, placing significant strain on the health care system. Using the Checklist for Early Recognition and Treatment of Acute Illness and Injury (CERTAIN) program, a multimodal trauma critical care knowledge-exchange platform was created for clinicians practicing in these institutions.ObjectivesTo describe the development and implementation of the CERTAIN for Ukraine program and to evaluate the reach of this intervention, together with participant engagement and satisfaction.Design, Setting, and ParticipantsThis quality improvement study included clinicians caring for critically ill patients during the ongoing Ukrainian conflict who were part of a community developed using a messaging app. The program was implemented by a group of international trauma and critical care experts in collaboration with critical care leaders from the Shupyk National Healthcare University in Kyiv, Ukraine. This study evaluates data collected from the CERTAIN for Ukraine program from its launch on April 9, 2022, to August 31, 2022.InterventionsThe initiative comprised a longitudinal series of interactive tele-education sessions, a webpage containing the CERTAIN approach and current trauma critical care guidelines translated into Ukrainian and Russian, and a private messaging chat for asynchronous discussion.Main Outcomes and MeasuresParticipant engagement and satisfaction were tracked using multimedia analytics and a post-session survey.ResultsSince program launch, 838 participants have joined the messaging group, and 6 tele-education sessions have been delivered, with 1835 total views. The CERTAIN website has had 3527 visits, mainly from Ukraine (1378 [39%]) and the United States (1060 [30%]). Of the 74 completed postsession surveys, 65 respondents (88%) rated the course content excellent or very good, and 73 (99%) recommended it to others.Conclusions and RelevanceThe findings of this quality improvement study indicate that, using widely available and low-cost platforms, knowledge was shared rapidly and efficiently to a large community of clinicians practicing in a wartime environment with broad-based engagement and a high level of learner satisfaction.

Background: There is a paucity of multicenter data describing the impact of coronavirus disease 2019 (COVID-19) on hospitalized pediatric oncology patients. Using a large, multicenter, Society of Critical Care Medicine (SCCM) Discovery Viral Infection and Respiratory Illness University Study (VIRUS) database, we aimed at assessing outcomes of COVID-19 infection in this population. Method: This is a matched-cohort study involving children below 18 years of age hospitalized with COVID-19 between March 2020 and January 2021. Using the VIRUS; COVID-19 Registry database, children with oncologic diseases were compared with propensity score matched (age groups, sex, race, and ethnicity) cohort of children without oncologic diseases for the prevalence of Multisystem Inflammatory Syndrome in Children (MIS-C), intensive care unit (ICU) admission, interventions, hospital, and ICU length of stay. Results: The number of children in the case and control groups was 45 and 180, respectively. ICU admission rate was similar in both groups ([47.7 vs 51.7%], P=0.63). The proportion of children requiring noninvasive and invasive mechanical ventilation, and its duration were similar between groups, same as hospital mortality. Interestingly, MIS-C was significantly lower in the oncology group compared with the control (2.4 vs 24.6%; P=0.0002). Conclusions: In this study using a multicenter VIRUS database, ICU admission rate, interventions, and outcomes of COVID-19 were similar in children with the oncologic disease compared with control patients. The incidence of MIS-C is lower in oncologic patients.

AbstractStudy ObjectiveIn critically ill patients, adequacy of early antibiotic exposure has been incompletely evaluated. This study characterized factors associated with inadequate cefepime exposure in the first 24 h of critical illness.DesignProspective cohort study.SettingAcademic Medical Center.PatientsCritically ill adults treated with cefepime. Patients with acute kidney injury or treated with kidney replacement therapy or extracorporeal membrane oxygenation were excluded.InterventionNone.MeasurementsA nonlinear mixed‐effects pharmacokinetic (PK) model was developed to estimate cefepime concentrations for each patient over time. The percentage of time the free drug concentration exceeded 8 mg/L during the first 24 h of therapy was calculated (%ƒT>8; appropriate for the susceptible breakpoint for Pseudomonas aeruginosa). Factors predictive of low %ƒT>8 were explored with multivariable regression.Main ResultsIn the 100 included patients, a one‐compartment PK model was developed with first‐order elimination with covariates for weight and estimated glomerular filtration rate based on creatinine and cystatin C (eGFRSCr‐CysC). The median (interquartile range) %ƒT>8 for cefepime in the first 24 h of therapy based on this model was 85% (66%, 100%). Less than 100% ƒT>8 during first 24 h of therapy occurred in 70 (70%) individuals. Lower Sequential Organ Failure Assessment score (p = 0.032) and higher eGFRSCr‐CysC (p < 0.001) predicted a lower %ƒT>8. Central nervous system infection source was protective (i.e., associated with a higher %ƒT>8; p = 0.008).ConclusionsDuring early critical illness, cefepime concentrations were inadequate in a significant proportion of patients. Antimicrobial optimization is needed to improve the precision of pharmacotherapy in the critically ill patients.