Prediction of Readmission Following Sepsis Using Social Determinants of Health.

OBJECTIVES: To determine the predictive value of social determinants of health (SDoH) variables on 30-day readmission following a sepsis hospitalization as compared with traditional clinical variables. DESIGN: Multicenter retrospective cohort study using patient-level data, including demographic, clinical, and survey data. SETTINGS: Thirty-five hospitals across the United States from 2017 to 2021. PATIENTS: Two hundred seventy-one thousand four hundred twenty-eight individuals in the AllofUs initiative, of which 8909 had an index sepsis hospitalization. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Unplanned 30-day readmission to the hospital. Multinomial logistic regression models were constructed to account for survival in determination of variables associate with 30-day readmission and are presented as adjusted odds rations (aORs). Of the 8909 sepsis patients in our cohort, 21% had an unplanned hospital readmission within 30 days. Median age (interquartile range) was 54 years (41-65 yr), 4762 (53.4%) were female, and there were self-reported 1612 (18.09%) Black, 2271 (25.49%) Hispanic, and 4642 (52.1%) White individuals. In multinomial logistic regression models accounting for survival, we identified that change to nonphysician provider type due to economic reasons (aOR, 2.55 [2.35-2.74]), delay of receiving medical care due to lack of transportation (aOR, 1.68 [1.62-1.74]), and inability to afford flow-up care (aOR, 1.59 [1.52-1.66]) were strongly and independently associated with a 30-day readmission when adjusting for survival. Patients who lived in a ZIP code with a high percentage of patients in poverty and without health insurance were also more likely to be readmitted within 30 days (aOR, 1.26 [1.22-1.29] and aOR, 1.28 [1.26-1.29], respectively). Finally, we found that having a primary care provider and health insurance were associated with low odds of an unplanned 30-day readmission. CONCLUSIONS: In this multicenter retrospective cohort, several SDoH variables were strongly associated with unplanned 30-day readmission. Models predicting readmission following sepsis hospitalization may benefit from the addition of SDoH factors to traditional clinical variables.

Ransomware Cyberattack Associated With Cardiac Arrest Incidence and Outcomes at Untargeted, Adjacent Hospitals.

OBJECTIVES: Healthcare ransomware cyberattacks have been associated with major regional hospital disruptions, but data reporting patient-oriented outcomes in critical conditions such as cardiac arrest (CA) are limited. This study examined the CA incidence and outcomes of untargeted hospitals adjacent to a ransomware-infected healthcare delivery organization (HDO). DESIGN SETTING AND PATIENTS: This cohort study compared the CA incidence and outcomes of two untargeted academic hospitals adjacent to an HDO under a ransomware cyberattack during the pre-attack (April 3-30, 2021), attack (May 1-28, 2021), and post-attack (May 29, 2021-June 25, 2021) phases. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Emergency department and hospital mean daily census, number of CAs, mean daily CA incidence per 1,000 admissions, return of spontaneous circulation, survival to discharge, and survival with favorable neurologic outcome were measured. The study evaluated 78 total CAs: 44 out-of-hospital CAs (OHCAs) and 34 in-hospital CAs. The number of total CAs increased from the pre-attack to attack phase (21 vs. 38; p = 0.03), followed by a decrease in the post-attack phase (38 vs. 19; p = 0.01). The number of total CAs exceeded the cyberattack month forecast (May 2021: 41 observed vs. 27 forecasted cases; 95% CI, 17.0-37.4). OHCA cases also exceeded the forecast (May 2021: 24 observed vs. 12 forecasted cases; 95% CI, 6.0-18.8). Survival with favorable neurologic outcome rates for all CAs decreased, driven by increases in OHCA mortality: survival with favorable neurologic rates for OHCAs decreased from the pre-attack phase to attack phase (40.0% vs. 4.5%; p = 0.02) followed by an increase in the post-attack phase (4.5% vs. 41.2%; p = 0.01). CONCLUSIONS: Untargeted hospitals adjacent to ransomware-infected HDOs may see worse outcomes for patients suffering from OHCA. These findings highlight the critical need for cybersecurity disaster planning and resiliency.

Emergency Department Take-Home Naloxone Improves Access Compared with Pharmacy-Dispensed Naloxone.

BACKGROUND: Opioid overdose is a major cause of mortality in the United States. In spite of efforts to increase naloxone availability, distribution to high-risk populations remains a challenge. OBJECTIVE: To assess the effects of multiple different naloxone distribution methods on patient obtainment of naloxone in the emergency department (ED) setting. METHODS: Naloxone was provided to patients in three 12-month phases between February 2020 and February 2023. In Phase 1, physicians could offer patients electronic prescriptions, which were filled in a nearby in-hospital discharge pharmacy. In Phase 2, physicians directly provided patients with take-home naloxone at discharge. In Phase 3, distribution was expanded to allow ED staff to hand patients take-home naloxone at time of discharge. The total number of prescriptions, rate of prescription filling, and amount of take-home naloxone kits provided to patients were then statistically analyzed using 95% confidence intervals (CI) and chi-squared testing. RESULTS: In Phase 1, 348 naloxone prescriptions were written, with 133 (95% CI 112.5-153.5) filled. In Phase 2, 327 (95% CI 245.5-408.5) take-home naloxone kits were given to patients by physicians. In Phase 3, 677 (95% CI 509.5-844.5) take-home naloxone kits were provided to patients by ED staff. There were statistically significant increases in naloxone distribution from Phase 1 to Phase 2, and Phase 2 to Phase 3. CONCLUSIONS: Take-home naloxone increases access when compared with naloxone prescriptions in the ED setting. A multidisciplinary approach combined with the removal of regulatory and administrative barriers allowed for further increased distribution of no-cost naloxone to patients.

Temporal Trends in Methamphetamine Use in Patients Admitted to the Hospital: A Retrospective Cohort Study.

OBJECTIVES: Although methamphetamine use is common, the scope of methamphetamine use and outcomes for patients admitted to the hospital is unclear. This study aims to identify the prevalence of methamphetamine use from January 2012 to January 2022, coingestions, hospital course, and readmission rate of admitted patients. METHODS: This was a retrospective cohort study conducted on patients admitted to our center with the following inclusions: age older than 18 years, positive/"pending confirm" value for methamphetamine on urine drug screen, and/or an International Classification of Diseases, Tenth Revision, code related to stimulant use disorder as an active issue. Urine drug screen data are reported as methamphetamine +/- and polysubstance (PS) +/-. Patient demographics, admission diagnosis, and hospital course were extracted. Statistical tests used included t tests and Mann-Whitney U tests. RESULTS: A total of 19,159 encounters were included, representing 12,057 unique patients. The median (interquartile range) age was 43 (33-54) years. Of all encounters, 35.3% were methamphetamine + and PS -, and 46.3% were methamphetamine + and PS +. Hospitalizations increased from 883 in 2012 to 2532 in 2021. The median (IQR) hospital stay was 48 (48-120) hours. Of all encounters, 16.8% included an intensive care unit (ICU) admission, and the median ICU stay was 42 (21-87) hours. A total of 2988 patients (24.7%) were readmitted within the study period, and 4988 (71.5%) returned within 1 year of the previous encounter. In context of all emergency department admissions from 2013 to 2022, 13.1% had a urine drug screen + for methamphetamine. CONCLUSIONS: Hospitalizations with recent methamphetamine use doubled at our institution from 2012 to 2022. In addition, 1 in 4 is readmitted (typically within 1 year), and a minority requires ICU care.

Impact of a deep learning sepsis prediction model on quality of care and survival.

Sepsis remains a major cause of mortality and morbidity worldwide. Algorithms that assist with the early recognition of sepsis may improve outcomes, but relatively few studies have examined their impact on real-world patient outcomes. Our objective was to assess the impact of a deep-learning model (COMPOSER) for the early prediction of sepsis on patient outcomes. We completed a before-and-after quasi-experimental study at two distinct Emergency Departments (EDs) within the UC San Diego Health System. We included 6217 adult septic patients from 1/1/2021 through 4/30/2023. The exposure tested was a nurse-facing Best Practice Advisory (BPA) triggered by COMPOSER. In-hospital mortality, sepsis bundle compliance, 72-h change in sequential organ failure assessment (SOFA) score following sepsis onset, ICU-free days, and the number of ICU encounters were evaluated in the pre-intervention period (705 days) and the post-intervention period (145 days). The causal impact analysis was performed using a Bayesian structural time-series approach with confounder adjustments to assess the significance of the exposure at the 95% confidence level. The deployment of COMPOSER was significantly associated with a 1.9% absolute reduction (17% relative decrease) in in-hospital sepsis mortality (95% CI, 0.3%-3.5%), a 5.0% absolute increase (10% relative increase) in sepsis bundle compliance (95% CI, 2.4%-8.0%), and a 4% (95% CI, 1.1%-7.1%) reduction in 72-h SOFA change after sepsis onset in causal inference analysis. This study suggests that the deployment of COMPOSER for early prediction of sepsis was associated with a significant reduction in mortality and a significant increase in sepsis bundle compliance.

Predicting Hospital Readmission among Patients with Sepsis Using Clinical and Wearable Data.

Sepsis is a life-threatening condition that occurs due to a dysregulated host response to infection. Recent data demonstrate that patients with sepsis have a significantly higher readmission risk than other common conditions, such as heart failure, pneumonia and myocardial infarction and associated economic burden. Prior studies have demonstrated an association between a patient's physical activity levels and readmission risk. In this study, we show that distribution of activity level prior and post-discharge among patients with sepsis are predictive of unplanned rehospitalization in 90 days (P-value<1e-3). Our preliminary results indicate that integrating Fitbit data with clinical measurements may improve model performance on predicting 90 days readmission.Clinical relevance Sepsis, Activity level, Hospital readmission, Wearable data.

Making the Improbable Possible: Generalizing Models Designed for a Syndrome-Based, Heterogeneous Patient Landscape.

Syndromic conditions, such as sepsis, are commonly encountered in the intensive care unit. Although these conditions are easy for clinicians to grasp, these conditions may limit the performance of machine-learning algorithms. Individual hospital practice patterns may limit external generalizability. Data missingness is another barrier to optimal algorithm performance and various strategies exist to mitigate this. Recent advances in data science, such as transfer learning, conformal prediction, and continual learning, may improve generalizability of machine-learning algorithms in critically ill patients. Randomized trials with these approaches are indicated to demonstrate improvements in patient-centered outcomes at this point.

A Review of Bicarbonate Use in Common Clinical Scenarios.

BACKGROUND: The use of sodium bicarbonate to treat metabolic acidosis is intuitive, yet data suggest that not all patients benefit from this therapy. OBJECTIVE: In this narrative review, we describe the physiology behind commonly encountered nontoxicologic causes of metabolic acidosis, highlight potential harm from the indiscriminate administration of sodium bicarbonate in certain scenarios, and provide evidence-based recommendations to assist emergency physicians in the rational use of sodium bicarbonate. DISCUSSION: Sodium bicarbonate can be administered as a hypertonic push, as a resuscitation fluid, or as an infusion. Lactic acidosis and cardiac arrest are two common scenarios where there is limited benefit to routine use of sodium bicarbonate, although certain circumstances, such as patients with concomitant acute kidney injury and lactic acidosis may benefit from sodium bicarbonate. Patients with cardiac arrest secondary to sodium channel blockade or hyperkalemia also benefit from sodium bicarbonate therapy. Recent data suggest that the use of sodium bicarbonate in diabetic ketoacidosis does not confer improved patient outcomes and may cause harm in pediatric patients. Available evidence suggests that alkalinization of urine in rhabdomyolysis does not improve patient-centered outcomes. Finally, patients with a nongap acidosis benefit from sodium bicarbonate supplementation. CONCLUSIONS: Empiric use of sodium bicarbonate in patients with nontoxicologic causes of metabolic acidosis is not warranted and likely does not improve patient-centered outcomes, except in select scenarios. Emergency physicians should reserve use of this medication to conditions with clear benefit to patients.

Disparities Exist in the Application of Low Tidal-volume Ventilation in the Emergency Department.

INTRODUCTION: Low tidal-volume ventilation (LTVV), defined as a maximum tidal volume of 8 milliliters per kilogram (mL/kg) of ideal body weight, is a key component of lung protective ventilation. Although emergency department (ED) initiation of LTVV has been associated with improved outcomes, disparities in LTVV application exist. In this study our aim was to evaluate whether rates of LTVV are associated with demographic and physical characteristics in the ED. METHODS: We conducted a retrospective observational cohort study using a dataset of patients who underwent mechanical ventilation at three EDs in two health systems from January 2016-June 2019. Demographic, mechanical ventilation, and outcome data including mortality and hospital-free days were abstracted by automatic query. A LTVV approach was defined as a tidal volume ≤8 mL/kg ideal body weight. We performed descriptive statistics and univariate analysis as indicated, and created a multivariate logistic regression model. RESULTS: Of 1,029 patients included in the study, 79.5% received LTVV. Tidal volumes of 400-500 mL were used in 81.9% of patients. Approximately 18% of patients had tidal volumes changed in the ED. Female gender (adjusted odds ratio [aOR] 4.17, P< 0.001), obesity (aOR 2.27, P< 0.001), and first-quartile height (aOR 12.2, P < 0.001) were associated with receiving non-LTVV in multivariate regression analysis. Hispanic ethnicity and female gender were associated with first quartile height (68.5%, 43.7%, P < 0.001 for all). Hispanic ethnicity was associated with receiving non-LTVV in univariate analysis (40.8% vs 23.0%, P < 0.001). This relationship did not persist in sensitivity analysis controlling for height, weight, gender, and body mass index. Patients who received LTVV in the ED had 2.1 more hospital-free days compared to those who did not (P = 0.040). No difference in mortality was observed. CONCLUSION: Emergency physicians use a narrow range of initial tidal volumes that may not meet lung-protective ventilation goals, with few corrections. Female gender, obesity, and first-quartile height are independently associated with receiving non-LTVV in the ED. Using LTVV in the ED was associated with 2.1 fewer hospital-free days. If confirmed in future studies, these findings have important implications for achieving quality improvement and health equality.

Representation Learning and Spectral Clustering for the Development and External Validation of Dynamic Sepsis Phenotypes: Observational Cohort Study.

BACKGROUND: Recent attempts at clinical phenotyping for sepsis have shown promise in identifying groups of patients with distinct treatment responses. Nonetheless, the replicability and actionability of these phenotypes remain an issue because the patient trajectory is a function of both the patient's physiological state and the interventions they receive. OBJECTIVE: We aimed to develop a novel approach for deriving clinical phenotypes using unsupervised learning and transition modeling. METHODS: Forty commonly used clinical variables from the electronic health record were used as inputs to a feed-forward neural network trained to predict the onset of sepsis. Using spectral clustering on the representations from this network, we derived and validated consistent phenotypes across a diverse cohort of patients with sepsis. We modeled phenotype dynamics as a Markov decision process with transitions as a function of the patient's current state and the interventions they received. RESULTS: Four consistent and distinct phenotypes were derived from over 11,500 adult patients who were admitted from the University of California, San Diego emergency department (ED) with sepsis between January 1, 2016, and January 31, 2020. Over 2000 adult patients admitted from the University of California, Irvine ED with sepsis between November 4, 2017, and August 4, 2022, were involved in the external validation. We demonstrate that sepsis phenotypes are not static and evolve in response to physiological factors and based on interventions. We show that roughly 45% of patients change phenotype membership within the first 6 hours of ED arrival. We observed consistent trends in patient dynamics as a function of interventions including early administration of antibiotics. CONCLUSIONS: We derived and describe 4 sepsis phenotypes present within 6 hours of triage in the ED. We observe that the administration of a 30 mL/kg fluid bolus may be associated with worse outcomes in certain phenotypes, whereas prompt antimicrobial therapy is associated with improved outcomes.

Predicting Hospital Readmission among Patients with Sepsis using Clinical and Wearable Data.

Sepsis is a life-threatening condition that occurs due to a dysregulated host response to infection. Recent data demonstrate that patients with sepsis have a significantly higher readmission risk than other common conditions, such as heart failure, pneumonia and myocardial infarction and associated economic burden. Prior studies have demonstrated an association between a patient's physical activity levels and readmission risk. In this study, we show that distribution of activity level prior and post-discharge among patients with sepsis are predictive of unplanned rehospitalization in 90 days (P-value<1e-3). Our preliminary results indicate that integrating Fitbit data with clinical measurements may improve model performance on predicting 90 days readmission.

Prediction of Unplanned Hospital Readmission using Clinical and Longitudinal Wearable Sensor Features.

Predictive models have been suggested as potential tools for identifying highest risk patients for hospital readmissions, in order to improve care coordination and ultimately long-term patient outcomes. However, the accuracy of current predictive models for readmission prediction is still moderate and further data enrichment is needed to identify at risk patients. This paper describes models to predict 90-day readmission, focusing on testing the predictive performance of wearable sensor features generated using multiscale entropy techniques and clinical features. Our study explores ways to incorporate pre-discharge and post-discharge wearable sensor features to make robust patient predictions. Data were used from participants enrolled in the AllofUs Research program. We extracted the inpatient cohort of patients and integrated clinical data from the electronic health records (EHR) and Fitbit sensor measurements. Entropy features were calculated from the longitudinal wearable sensor data, such as heart rate and mobility-related measurements, in order to characterize time series variability and complexity. Our best performing model acheived an AUC of 83%, and at 80% sensitivity acheived 75% specificity and 57% positive predictive value. Our results indicate that it would be possible to improve the ability to predict unplanned hospital readmissions by considering pre-discharge and post-discharge wearable features.

Passing the SNF Test: A Secondary Analysis of a Sepsis Transition Intervention Trial Among Patients Discharged to Post-Acute Care.

OBJECTIVES: Sepsis survivors discharged to post-acute care facilities experience high rates of mortality and hospital readmission. This study compared the effects of a Sepsis Transition and Recovery (STAR) program vs usual care (UC) on 30-day mortality and hospital readmission among sepsis survivors discharged to post-acute care. DESIGN: Secondary analysis of a multisite pragmatic randomized clinical trial. SETTING AND PARTICIPANTS: Sepsis survivors discharged to post-acute care. METHODS: We conducted a secondary analysis of patients from the IMPACTS (Improving Morbidity During Post-Acute Care Transitions for Sepsis) randomized clinical trial who were discharged to post-acute care. IMPACTS evaluated the effectiveness of STAR, a nurse-navigator-led program to deliver best practice post-sepsis care. Subjects were randomized to receive either STAR or UC. The primary outcome was 30-day readmission and mortality. We also evaluated hospital-free days alive as a secondary outcome. RESULTS: Of 691 patients enrolled in IMPACTS, 175 (25%) were discharged to post-acute care [143 (82%) to skilled nursing facilities, 12 (7%) to long-term acute care hospitals, and 20 (11%) to inpatient rehabilitation]. Of these, 87 received UC and 88 received the STAR intervention. The composite 30-day all-cause mortality and readmission endpoint occurred in 26 (29.9%) patients in the UC group vs 18 (20.5%) in the STAR group [risk difference -9.4% (95% CI -22.2 to 3.4); adjusted odds ratio 0.58 (95% CI 0.28 to 1.17)]. Separately, 30-day all-cause mortality was 8.1% in the UC group compared with 5.7% in the STAR group [risk difference -2.4% (95% CI -9.9 to 5.1)] and 30-day all-cause readmission was 26.4% in the UC group compared with 17.1% in the STAR program [risk difference -9.4% (95% CI -21.5 to 2.8)]. CONCLUSIONS AND IMPLICATIONS: There are few proven interventions to reduce readmission among patients discharged to post-acute care facilities. These results suggest the STAR program may reduce 30-day mortality and readmission rates among sepsis survivors discharged to post-acute care facilities.

Optimizing the Implementation of Clinical Predictive Models to Minimize National Costs: Sepsis Case Study.

BACKGROUND: Sepsis costs and incidence vary dramatically across diagnostic categories, warranting a customized approach for implementing predictive models. OBJECTIVE: The aim of this study was to optimize the parameters of a sepsis prediction model within distinct patient groups to minimize the excess cost of sepsis care and analyze the potential effect of factors contributing to end-user response to sepsis alerts on overall model utility. METHODS: We calculated the excess costs of sepsis to the Centers for Medicare and Medicaid Services (CMS) by comparing patients with and without a secondary sepsis diagnosis but with the same primary diagnosis and baseline comorbidities. We optimized the parameters of a sepsis prediction algorithm across different diagnostic categories to minimize these excess costs. At the optima, we evaluated diagnostic odds ratios and analyzed the impact of compliance factors such as noncompliance, treatment efficacy, and tolerance for false alarms on the net benefit of triggering sepsis alerts. RESULTS: Compliance factors significantly contributed to the net benefit of triggering a sepsis alert. However, a customized deployment policy can achieve a significantly higher diagnostic odds ratio and reduced costs of sepsis care. Implementing our optimization routine with powerful predictive models could result in US $4.6 billion in excess cost savings for CMS. CONCLUSIONS: We designed a framework for customizing sepsis alert protocols within different diagnostic categories to minimize excess costs and analyzed model performance as a function of false alarm tolerance and compliance with model recommendations. We provide a framework that CMS policymakers could use to recommend minimum adherence rates to the early recognition and appropriate care of sepsis that is sensitive to hospital department-level incidence rates and national excess costs. Customizing the implementation of clinical predictive models by accounting for various behavioral and economic factors may improve the practical benefit of predictive models.

The Knowledge Gap: Mentorship in Emergency Medicine Residency.

STUDY OBJECTIVE: Studies of mentorship in emergency medicine show that mentored residents are twice as likely to describe their career preparation as excellent as compared to nonmentored peers. There has been significant interest in the mentor-mentee relationship in medicine; however, there is minimal guidance and published literature specific to emergency medicine residents. METHODS: In this narrative review, we described the emergency medicine mentor-mentee relationship, discussed alternatives to the traditional dyadic model, and highlighted current barriers to effective mentorship. We conducted a structured literature review to identify relevant published articles regarding the mentoring of emergency medicine residents. Additional studies from general mentoring literature were included based on relevancy. RESULTS: We identified 39 studies in emergency medicine literature based on our search criteria. Additional studies from general medicine literature were included based on relevancy to this review. Based on the limited available literature, we recommend maximizing the resident mentoring relationship by developing formal mentoring programs, supporting the advancement of women and underrepresented minority mentors, and moving toward team mentoring, including peer, near-peer, and collaborative mentorship. The development of a mentoring network is a logical strategy for residents to work with a diverse group of individuals to maximize benefits in multiple areas. CONCLUSION: Alternative approaches to the traditional and hierarchal dyadic mentoring style (eg, team mentoring) are effective methods that residencies may promote to increase effective mentoring. Future efforts in mentoring emergency medicine residents emphasize these strategies, which are increasingly beneficial given the constraints and use of technology highlighted by the COVID-19 pandemic.

Assessing Barriers to Implementation of Machine Learning and Artificial Intelligence-Based Tools in Critical Care: Web-Based Survey Study.

BACKGROUND: Although there is considerable interest in machine learning (ML) and artificial intelligence (AI) in critical care, the implementation of effective algorithms into practice has been limited. OBJECTIVE: We sought to understand physician perspectives of a novel intubation prediction tool. Further, we sought to understand health care provider and nonprovider perspectives on the use of ML in health care. We aim to use the data gathered to elucidate implementation barriers and determinants of this intubation prediction tool, as well as ML/AI-based algorithms in critical care and health care in general. METHODS: We developed 2 anonymous surveys in Qualtrics, 1 single-center survey distributed to 99 critical care physicians via email, and 1 social media survey distributed via Facebook and Twitter with branching logic to tailor questions for providers and nonproviders. The surveys included a mixture of categorical, Likert scale, and free-text items. Likert scale means with SD were reported from 1 to 5. We used student t tests to examine the differences between groups. In addition, Likert scale responses were converted into 3 categories, and percentage values were reported in order to demonstrate the distribution of responses. Qualitative free-text responses were reviewed by a member of the study team to determine validity, and content analysis was performed to determine common themes in responses. RESULTS: Out of 99 critical care physicians, 47 (48%) completed the single-center survey. Perceived knowledge of ML was low with a mean Likert score of 2.4 out of 5 (SD 0.96), with 7.5% of respondents rating their knowledge as a 4 or 5. The willingness to use the ML-based algorithm was 3.32 out of 5 (SD 0.95), with 75% of respondents answering 3 out of 5. The social media survey had 770 total responses with 605 (79%) providers and 165 (21%) nonproviders. We found no difference in providers' perceived knowledge based on level of experience in either survey. We found that nonproviders had significantly less perceived knowledge of ML (mean 3.04 out of 5, SD 1.53 vs mean 3.43, SD 0.941; P<.001) and comfort with ML (mean 3.28 out of 5, SD 1.02 vs mean 3.53, SD 0.935; P=.004) than providers. Free-text responses revealed multiple shared concerns, including accuracy/reliability, data bias, patient safety, and privacy/security risks. CONCLUSIONS: These data suggest that providers and nonproviders have positive perceptions of ML-based tools, and that a tool to predict the need for intubation would be of interest to critical care providers. There were many shared concerns about ML/AI in health care elucidated by the surveys. These results provide a baseline evaluation of implementation barriers and determinants of ML/AI-based tools that will be important in their optimal implementation and adoption in the critical care setting and health care in general.

Ketamine is not associated with more post-intubation hypotension than etomidate in patients undergoing endotracheal intubation.

INTRODUCTION: Emergency department (ED) patients undergoing emergent tracheal intubation often have multiple physiologic derangements putting them at risk for post-intubation hypotension. Prior work has shown that post-intubation hypotension is independently associated with increased morbidity and mortality. The choice of induction agent may be associated with post-intubation hypotension. Etomidate and ketamine are two of the most commonly used agents in the ED, however, there is controversy regarding whether either agent is superior in the setting of hemodynamic instability. The goal of this study is to determine whether there is a difference in the rate of post-intubation hypotension who received either ketamine or etomidate for induction. Additionally, we provide a subgroup analysis of patients at pre-existing risk of cardiovascular collapse (identified by pre-intubation shock index (SI) > 0.9) to determine if differences in rates of post-intubation hypotension exist as a function of sedative choice administered during tracheal intubation in these high-risk patients. We hypothesize that there is no difference in the incidence of post-intubation hypotension in patients who receive ketamine versus etomidate. METHODS: A retrospective cohort study was conducted on a database of 469 patients having undergone emergent intubation with either etomidate or ketamine induction at a large academic health system. Patients were identified by automatic query of the electronic health records from 1/1/2016-6/30/2019. Exclusion criteria were patients <18-years-old, tracheal intubation performed outside of the ED, incomplete peri-intubation vital signs, or cardiac arrest prior to intubation. Patients at high risk for hemodynamic collapse in the post-intubation period were identified by a pre-intubation SI > 0.9. The primary outcome was the incidence of post-intubation hypotension (systolic blood pressure < 90 mmHg or mean arterial pressure < 65 mmHg). Secondary outcomes included post-intubation vasopressor use and mortality. These analyses were performed on the full cohort and an exploratory analysis in patients with SI > 0.9. We also report adjusted odds ratios (aOR) from a multivariable logistic regression model of the entire cohort controlling for plausible confounding variables to determine independent factors associated with post-intubation hypotension. RESULTS: A total of 358 patients were included (etomidate: 272; ketamine: 86). The mean pre-intubation SI was higher in the group that received ketamine than etomidate, (0.97 vs. 0.83, difference: -0.14 (95%, CI -0.2 to -0.1). The incidence of post-intubation hypotension was greater in the ketamine group prior to SI stratification (difference: -10%, 95% CI -20.9% to -0.1%). Emergency physicians were more likely to use ketamine in patients with SI > 0.9. In our multivariate logistic regression analysis, choice of induction agent was not associated with post-intubation hypotension (aOR 1.45, 95% CI 0.79 to 2.65). We found that pre-intubation shock index was the strongest predictor of post-intubation hypotension. CONCLUSION: In our cohort of patients undergoing emergent tracheal intubation, ketamine was used more often for patients with an elevated shock index. We did not identify an association between the incidence of post-intubation hypotension and induction agent between ketamine and etomidate. Patients with an elevated shock index were at higher risk of cardiovascular collapse regardless of the choice of ketamine or etomidate.