Astegolimab or Efmarodocokin Alfa in Patients With Severe COVID-19 Pneumonia: A Randomized, Phase 2 Trial.

OBJECTIVES: Severe cases of COVID-19 pneumonia can lead to acute respiratory distress syndrome (ARDS). Release of interleukin (IL)-33, an epithelial-derived alarmin, and IL-33/ST2 pathway activation are linked with ARDS development in other viral infections. IL-22, a cytokine that modulates innate immunity through multiple regenerative and protective mechanisms in lung epithelial cells, is reduced in patients with ARDS. This study aimed to evaluate safety and efficacy of astegolimab, a human immunoglobulin G2 monoclonal antibody that selectively inhibits the IL-33 receptor, ST2, or efmarodocokin alfa, a human IL-22 fusion protein that activates IL-22 signaling, for treatment of severe COVID-19 pneumonia. DESIGN: Phase 2, double-blind, placebo-controlled study (COVID-astegolimab-IL). SETTING: Hospitals. PATIENTS: Hospitalized adults with severe COVID-19 pneumonia. INTERVENTIONS: Patients were randomized to receive IV astegolimab, efmarodocokin alfa, or placebo, plus standard of care. The primary endpoint was time to recovery, defined as time to a score of 1 or 2 on a 7-category ordinal scale by day 28. MEASUREMENTS AND MAIN RESULTS: The study randomized 396 patients. Median time to recovery was 11 days (hazard ratio [HR], 1.01 d; p = 0.93) and 10 days (HR, 1.15 d; p = 0.38) for astegolimab and efmarodocokin alfa, respectively, versus 10 days for placebo. Key secondary endpoints (improved recovery, mortality, or prevention of worsening) showed no treatment benefits. No new safety signals were observed and adverse events were similar across treatment arms. Biomarkers demonstrated that both drugs were pharmacologically active. CONCLUSIONS: Treatment with astegolimab or efmarodocokin alfa did not improve time to recovery in patients with severe COVID-19 pneumonia.

The Reviewer Academy of the Society of Critical Care Medicine: Key Principles and Strategic Plan.

The Society of Critical Care Medicine (SCCM) Reviewer Academy seeks to train and establish a community of trusted, reliable, and skilled peer reviewers with diverse backgrounds and interests to promote high-quality reviews for each of the SCCM journals. Goals of the Academy include building accessible resources to highlight qualities of excellent manuscript reviews; educating and mentoring a diverse group of healthcare professionals; and establishing and upholding standards for insightful and informative reviews. This manuscript will map the mission of the Reviewer Academy with a succinct summary of the importance of peer review, process of reviewing a manuscript, and the expected ethical standards of reviewers. We will equip readers to target concise, thoughtful feedback as peer reviewers, advance their understanding of the editorial process and inspire readers to integrate medical journalism into diverse professional careers.

ClotCatcher: a novel natural language model to accurately adjudicate venous thromboembolism from radiology reports.

INTRODUCTION: Accurate identification of venous thromboembolism (VTE) is critical to develop replicable epidemiological studies and rigorous predictions models. Traditionally, VTE studies have relied on international classification of diseases (ICD) codes which are inaccurate - leading to misclassification bias. Here, we developed ClotCatcher, a novel deep learning model that uses natural language processing to detect VTE from radiology reports. METHODS: Radiology reports to detect VTE were obtained from patients admitted to Emory University Hospital (EUH) and Grady Memorial Hospital (GMH). Data augmentation was performed using the Google PEGASUS paraphraser. This data was then used to fine-tune ClotCatcher, a novel deep learning model. ClotCatcher was validated on both the EUH dataset alone and GMH dataset alone. RESULTS: The dataset contained 1358 studies from EUH and 915 studies from GMH (n = 2273). The dataset contained 1506 ultrasound studies with 528 (35.1%) studies positive for VTE, and 767 CT studies with 91 (11.9%) positive for VTE. When validated on the EUH dataset, ClotCatcher performed best (AUC = 0.980) when trained on both EUH and GMH dataset without paraphrasing. When validated on the GMH dataset, ClotCatcher performed best (AUC = 0.995) when trained on both EUH and GMH dataset with paraphrasing. CONCLUSION: ClotCatcher, a novel deep learning model with data augmentation rapidly and accurately adjudicated the presence of VTE from radiology reports. Applying ClotCatcher to large databases would allow for rapid and accurate adjudication of incident VTE. This would reduce misclassification bias and form the foundation for future studies to estimate individual risk for patient to develop incident VTE.

The Coronavirus Disease 2019 Pandemic Impacts Burnout Syndrome Differently Among Multiprofessional Critical Care Clinicians-A Longitudinal Survey Study.

OBJECTIVES: To determine the impact of coronavirus disease 2019 on burnout syndrome in the multiprofessional ICU team and to identify factors associated with burnout syndrome. DESIGN: Longitudinal, cross-sectional survey. SETTING: All adult ICUs within an academic health system. SUBJECTS: Critical care nurses, advanced practice providers, physicians, respiratory therapists, pharmacists, social workers, and spiritual health workers were surveyed on burnout in 2017 and during the coronavirus disease 2019 pandemic in 2020. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Burnout syndrome and contributing factors were measured using the Maslach Burnout Inventory of Health and Human Service and Areas of Worklife Survey. Response rates were 46.5% (572 respondents) in 2017 and 49.9% (710 respondents) in 2020. The prevalence of burnout increased from 59% to 69% (p < 0.001). Nurses were disproportionately impacted, with the highest increase during the pandemic (58-72%; p < 0.0001) with increases in emotional exhaustion and depersonalization, and personal achievement decreases. In contrast, although burnout was high before and during coronavirus disease 2019 in all specialties, most professions had similar or lower burnout in 2020 as they had in 2017. Physicians had the lowest rates of burnout, measured at 51% and 58%, respectively. There was no difference in burnout between clinicians working in ICUs who treated coronavirus disease 2019 than those who did not (71% vs 67%; p = 0.26). Burnout significantly increased in females (71% vs 60%; p = 0.001) and was higher than in males during the pandemic (71% vs 60%; p = 0.01). CONCLUSIONS: Burnout syndrome was common in all multiprofessional ICU team members prior to and increased substantially during the pandemic, independent of whether one treated coronavirus disease 2019 patients. Nurses had the highest prevalence of burnout during coronavirus disease 2019 and had the highest increase in burnout from the prepandemic baseline. Female clinicians were significantly more impacted by burnout than males. Different susceptibility to burnout syndrome may require profession-specific interventions as well as work system improvements.

Viral Micro-RNAs Are Detected in the Early Systemic Response to Injury and Are Associated With Outcomes in Polytrauma Patients.

OBJECTIVES: To evaluate early activation of latent viruses in polytrauma patients and consider prognostic value of viral micro-RNAs in these patients. DESIGN: This was a subset analysis from a prospectively collected multicenter trauma database. Blood samples were obtained upon admission to the trauma bay (T0), and trauma metrics and recovery data were collected. SETTING: Two civilian Level 1 Trauma Centers and one Military Treatment Facility. PATIENTS: Adult polytrauma patients with Injury Severity Scores greater than or equal to 16 and available T0 plasma samples were included in this study. Patients with ICU admission greater than 14 days, mechanical ventilation greater than 7 days, or mortality within 28 days were considered to have a complicated recovery. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Polytrauma patients (n = 180) were identified, and complicated recovery was noted in 33%. Plasma samples from T0 underwent reverse transcriptase-quantitative polymerase chain reaction analysis for Kaposi's sarcoma-associated herpesvirus micro-RNAs (miR-K12_10b and miRK-12-12) and Epstein-Barr virus-associated micro-RNA (miR-BHRF-1), as well as Luminex multiplex array analysis for established mediators of inflammation. Ninety-eight percent of polytrauma patients were found to have detectable Kaposi's sarcoma-associated herpesvirus and Epstein-Barr virus micro-RNAs at T0, whereas healthy controls demonstrated 0% and 100% detection rate for Kaposi's sarcoma-associated herpesvirus and Epstein-Barr virus, respectively. Univariate analysis revealed associations between viral micro-RNAs and polytrauma patients' age, race, and postinjury complications. Multivariate least absolute shrinkage and selection operator analysis of clinical variables and systemic biomarkers at T0 revealed that interleukin-10 was the strongest predictor of all viral micro-RNAs. Multivariate least absolute shrinkage and selection operator analysis of systemic biomarkers as predictors of complicated recovery at T0 demonstrated that miR-BHRF-1, miR-K12-12, monocyte chemoattractant protein-1, and hepatocyte growth factor were independent predictors of complicated recovery with a model complicated recovery prediction area under the curve of 0.81. CONCLUSIONS: Viral micro-RNAs were detected within hours of injury and correlated with poor outcomes in polytrauma patients. Our findings suggest that transcription of viral micro-RNAs occurs early in the response to trauma and may be associated with the biological processes involved in polytrauma-induced complicated recovery.

A Locally Optimized Data-Driven Tool to Predict Sepsis-Associated Vasopressor Use in the ICU.

OBJECTIVES: To train a model to predict vasopressor use in ICU patients with sepsis and optimize external performance across hospital systems using domain adaptation, a transfer learning approach. DESIGN: Observational cohort study. SETTING: Two academic medical centers from January 2014 to June 2017. PATIENTS: Data were analyzed from 14,512 patients (9,423 at the development site and 5,089 at the validation site) who were admitted to an ICU and met Center for Medicare and Medicaid Services definition of severe sepsis either before or during the ICU stay. Patients were excluded if they never developed sepsis, if the ICU length of stay was less than 8 hours or more than 20 days or if they developed shock up to the first 4 hours of ICU admission. MEASUREMENTS AND MAIN RESULTS: Forty retrospectively collected features from the electronic medical records of adult ICU patients at the development site (four hospitals) were used as inputs for a neural network Weibull-Cox survival model to derive a prediction tool for future need of vasopressors. Domain adaptation updated parameters to optimize model performance in the validation site (two hospitals), a different healthcare system over 2,000 miles away. The cohorts at both sites were randomly split into training and testing sets (80% and 20%, respectively). When applied to the test set in the development site, the model predicted vasopressor use 4-24 hours in advance with an area under the receiver operator characteristic curve, specificity, and positive predictive value ranging from 0.80 to 0.81, 56.2% to 61.8%, and 5.6% to 12.1%, respectively. Domain adaptation improved performance of the model to predict vasopressor use within 4 hours at the validation site (area under the receiver operator characteristic curve 0.81 [CI, 0.80-0.81] from 0.77 [CI, 0.76-0.77], p < 0.01; specificity 59.7% [CI, 58.9-62.5%] from 49.9% [CI, 49.5-50.7%], p < 0.01; positive predictive value 8.9% [CI, 8.5-9.4%] from 7.3 [7.1-7.4%], p < 0.01). CONCLUSIONS: Domain adaptation improved performance of a model predicting sepsis-associated vasopressor use during external validation.

Accelerating Coronavirus Disease 2019 Therapeutic Interventions and Vaccines-Selecting Compounds for Clinical Evaluation in Coronavirus Disease 2019 Clinical Trials.

Given the urgent need for coronavirus disease 2019 therapeutics, early in the pandemic the Accelerating Coronavirus Disease 2019 Therapeutic Interventions and Vaccines (ACTIV) public-private partnership rapidly designed a unique therapeutic agent intake and assessment process for candidate treatments of coronavirus disease 2019. These treatments included antivirals, immune modulators, severe acute respiratory syndrome coronavirus 2 neutralizing antibodies, and organ-supportive treatments at both the preclinical and clinical stages of development. The ACTIV Therapeutics-Clinical Working Group Agent Prioritization subgroup established a uniform data collection process required to perform an assessment of any agent type using review criteria that were identified and differentially weighted for each agent class. The ACTIV Therapeutics-Clinical Working Group evaluated over 750 therapeutic agents with potential application for coronavirus disease 2019 and prioritized promising candidates for testing within the master protocols conducted by ACTIV. In addition, promising agents among preclinical candidates were selected by ACTIV to be matched with laboratories that could assist in executing rigorous preclinical studies. Between April 14, 2020, and May 31, 2021, the Agent Prioritization subgroup advanced 20 agents into the Accelerating Coronavirus Disease 2019 Therapeutic Interventions and Vaccines master protocols and matched 25 agents with laboratories to assist with preclinical testing.

Sepsis Among Medicare Beneficiaries: 4. Precoronavirus Disease 2019 Update January 2012-February 2020.

OBJECTIVES: To provide updated information on the burdens of sepsis during acute inpatient admissions for Medicare beneficiaries. DESIGN: Analysis of paid Medicare claims via the Centers for Medicare and Medicaid Services DataLink Project. SETTING: All U.S. acute-care hospitals, excluding federally operated hospitals (Veterans Administration and Defense Health Agency). PATIENTS: All Medicare beneficiaries, January 2012-February 2020, with an explicit sepsis diagnostic code assigned during an inpatient admission. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: The count of Medicare Part A/B (fee-for-service) plus Medicare Advantage inpatient sepsis admissions rose from 981,027 (CY2012) to 1,700,433 (CY 2019). The proportion of total admissions with sepsis in the Medicare Advantage population rose from 21.43% to 35.39%, reflecting the increasing beneficiary proportion enrolled in Medicare Advantage. In CY2019, 6-month mortality rates in Medicare fee-for-service beneficiaries for sepsis continued to decline, but remained high: 59.9% for septic shock, 35.5% for severe sepsis, 30.8% for sepsis attributed to a specific organism, and 26.5% for unspecified sepsis. Total fee-for-service-only inpatient hospital costs rose from $17.79B (CY2012) to $22.98B (CY2019). We estimated that the aggregate cost of sepsis hospital care for the entire U.S. population was at least $57.47B in 2019. Inclusion of 14 months' (January 2019-February 2020) newer data exposed new trends: the cost per patient, number of admissions, and fraction of patients with sepsis labeled as present on admission inflected around November 2015, coincident with the change to International Classification of Diseases, 10th Edition, and introduction of the Severe Sepsis and Septic Shock Management Bundle (SEP-1) metric. CONCLUSIONS: Sepsis among Medicare beneficiaries precoronavirus disease 2019 imposed immense burdens upon patients, their families, and the taxpayers.

Quantifying resilience of humans and other animals.

All life requires the capacity to recover from challenges that are as inevitable as they are unpredictable. Understanding this resilience is essential for managing the health of humans and their livestock. It has long been difficult to quantify resilience directly, forcing practitioners to rely on indirect static indicators of health. However, measurements from wearable electronics and other sources now allow us to analyze the dynamics of physiology and behavior with unsurpassed resolution. The resulting flood of data coincides with the emergence of novel analytical tools for estimating resilience from the pattern of microrecoveries observed in natural time series. Such dynamic indicators of resilience may be used to monitor the risk of systemic failure across systems ranging from organs to entire organisms. These tools invite a fundamental rethinking of our approach to the adaptive management of health and resilience.

Effect of Vitamin C, Thiamine, and Hydrocortisone on Ventilator- and Vasopressor-Free Days in Patients With Sepsis: The VICTAS Randomized Clinical Trial.

Importance: Sepsis is a common syndrome with substantial morbidity and mortality. A combination of vitamin C, thiamine, and corticosteroids has been proposed as a potential treatment for patients with sepsis. Objective: To determine whether a combination of vitamin C, thiamine, and hydrocortisone every 6 hours increases ventilator- and vasopressor-free days compared with placebo in patients with sepsis. Design, Setting, and Participants: Multicenter, randomized, double-blind, adaptive-sample-size, placebo-controlled trial conducted in adult patients with sepsis-induced respiratory and/or cardiovascular dysfunction. Participants were enrolled in the emergency departments or intensive care units at 43 hospitals in the United States between August 2018 and July 2019. After enrollment of 501 participants, funding was withheld, leading to an administrative termination of the trial. All study-related follow-up was completed by January 2020. Interventions: Participants were randomized to receive intravenous vitamin C (1.5 g), thiamine (100 mg), and hydrocortisone (50 mg) every 6 hours (n = 252) or matching placebo (n = 249) for 96 hours or until discharge from the intensive care unit or death. Participants could be treated with open-label corticosteroids by the clinical team, with study hydrocortisone or matching placebo withheld if the total daily dose was greater or equal to the equivalent of 200 mg of hydrocortisone. Main Outcomes and Measures: The primary outcome was the number of consecutive ventilator- and vasopressor-free days in the first 30 days following the day of randomization. The key secondary outcome was 30-day mortality. Results: Among 501 participants randomized (median age, 62 [interquartile range {IQR}, 50-70] years; 46% female; 30% Black; median Acute Physiology and Chronic Health Evaluation II score, 27 [IQR, 20.8-33.0]; median Sequential Organ Failure Assessment score, 9 [IQR, 7-12]), all completed the trial. Open-label corticosteroids were prescribed to 33% and 32% of the intervention and control groups, respectively. Ventilator- and vasopressor-free days were a median of 25 days (IQR, 0-29 days) in the intervention group and 26 days (IQR, 0-28 days) in the placebo group, with a median difference of -1 day (95% CI, -4 to 2 days; P = .85). Thirty-day mortality was 22% in the intervention group and 24% in the placebo group. Conclusions and Relevance: Among critically ill patients with sepsis, treatment with vitamin C, thiamine, and hydrocortisone, compared with placebo, did not significantly increase ventilator- and vasopressor-free days within 30 days. However, the trial was terminated early for administrative reasons and may have been underpowered to detect a clinically important difference. Trial Registration: ClinicalTrials.gov Identifier: NCT03509350.

Sepsis Among Medicare Beneficiaries: 1. The Burdens of Sepsis, 2012-2018.

OBJECTIVES: To provide contemporary estimates of the burdens (costs and mortality) associated with acute inpatient Medicare beneficiary admissions for sepsis. DESIGN: Analysis of paid Medicare claims via the Centers for Medicare & Medicaid Services DataLink Project. SETTING: All U.S. acute care hospitals, excluding federally operated hospitals (Veterans Administration and Defense Health Agency). PATIENTS: All Medicare beneficiaries, 2012-2018, with an inpatient admission including one or more explicit sepsis codes. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Total inpatient hospital and skilled nursing facility admission counts, costs, and mortality over time. From calendar year (CY)2012-CY2018, the total number of Medicare Part A/B (fee-for-service) beneficiaries with an inpatient hospital admission associated with an explicit sepsis code rose from 811,644 to 1,136,889. The total cost of inpatient hospital admission including an explicit sepsis code for those beneficiaries in those calendar years rose from $17,792,657,303 to $22,439,794,212. The total cost of skilled nursing facility care in the 90 days subsequent to an inpatient hospital discharge that included an explicit sepsis code for Medicare Part A/B rose from $3,931,616,160 to $5,623,862,486 over that same interval. Precise costs are not available for Medicare Part C (Medicare Advantage) patients. Using available federal data sources, we estimated the aggregate cost of inpatient admissions and skilled nursing facility admissions for Medicare Advantage patients to have risen from $6.0 to $13.4 billion over the CY2012-CY2018 interval. Combining data for fee-for-service beneficiaries and estimates for Medicare Advantage beneficiaries, we estimate the total inpatient admission sepsis cost and any subsequent skilled nursing facility admission for all (fee-for-service and Medicare Advantage) Medicare patients to have risen from $27.7 to $41.5 billion. Contemporary 6-month mortality rates for Medicare fee-for-service beneficiaries with a sepsis inpatient admission remain high: for septic shock, approximately 60%; for severe sepsis, approximately 36%; for sepsis attributed to a specific organism, approximately 31%; and for unspecified sepsis, approximately 27%. CONCLUSION: Sepsis remains common, costly to treat, and presages significant mortality for Medicare beneficiaries.

Sepsis Among Medicare Beneficiaries: 2. The Trajectories of Sepsis, 2012-2018.

OBJECTIVES: To distinguish characteristics of Medicare beneficiaries who will have an acute inpatient admission for sepsis from those who have an inpatient admission without sepsis, and to describe their further trajectories during and subsequent to those inpatient admissions. DESIGN: Analysis of paid Medicare claims via the Centers for Medicare and Medicaid Services DataLink Project. SETTING: All U.S. acute care hospitals, excepting federal hospitals (Veterans Administration and Defense Health Agency). PATIENTS: Medicare beneficiaries, 2012-2018, with an inpatient hospital admission including one or more explicit sepsis codes. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Prevalent diagnoses in the year prior to the inpatient admission; healthcare contacts in the week prior to the inpatient admission; discharges, transfers, readmissions, and deaths (trajectories) for 6 months following discharge from the inpatient admission. Beneficiaries with no sepsis inpatient hospital admission for a year prior to an index hospital admission for sepsis were nearly indistinguishable by accumulated diagnostic codes from beneficiaries who had an index hospital admission without sepsis. Although the timing of healthcare services in the week prior to inpatient hospital admission was similar among beneficiaries who would be admitted for sepsis versus those whose inpatient admission did not include a sepsis code, the setting differed: beneficiaries destined for a sepsis admission were more likely to have received skilled nursing or unskilled nursing (e.g., nursing aide for activities of daily living) care. In contrast, comparing beneficiaries who had been free of any inpatient admission for an entire year and then required an inpatient admission, acute inpatient stays that included a sepsis code led to more than three times as many deaths within 1 week of discharge, with more admissions to skilled nursing facilities and fewer discharges to home. Comparing all beneficiaries who were admitted to a skilled nursing facility after an inpatient hospital admission, those who had sepsis coded during the index admission were more likely to die in the skilled nursing facility; more likely to be readmitted to an acute inpatient hospital and subsequently die in that setting; or if they survive to discharge from the skilled nursing facility, they are more likely to go next to a custodial nursing home. CONCLUSIONS: Although Medicare beneficiaries destined for an inpatient hospital admission with a sepsis code are nearly indistinguishable by other diagnostic codes from those whose admissions will not have a sepsis code, their healthcare trajectories following the admission are worse. This suggests that an inpatient stay that included a sepsis code not only identifies beneficiaries who were less resilient to infection but also signals increased risk for worsening health, for mortality, and for increased use of advanced healthcare services during and postdischarge along with an increased likelihood of an inpatient hospital readmission.

Sepsis Among Medicare Beneficiaries: 3. The Methods, Models, and Forecasts of Sepsis, 2012-2018.

OBJECTIVE: To evaluate the impact of sepsis, age, and comorbidities on death following an acute inpatient admission and to model and forecast inpatient and skilled nursing facility costs for Medicare beneficiaries during and subsequent to an acute inpatient sepsis admission. DESIGN: Analysis of paid Medicare claims via the Centers for Medicare & Medicaid Services DataLink Project (CMS) and leveraging the CMS-Hierarchical Condition Category risk adjustment model. SETTING: All U.S. acute care hospitals, excepting federal hospitals (Veterans Administration and Defense Health Agency). PATIENTS: All Part A/B (fee-for-service) Medicare beneficiaries with an acute inpatient admission in 2017 and who had no inpatient sepsis admission in the prior year. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Logistic regression models to determine covariate risk contribution to death following an acute inpatient admission; conventional regression to predict Medicare beneficiary sepsis costs. Using the Hierarchical Condition Category risk adjustment model to illuminate influence of illness on outcome of inpatient admissions, representative odds ratios (with 95% CIs) for death within 6 months of an admission (referenced to beneficiaries admitted but without the characteristic) are as follows: septic shock, 7.27 (7.19-7.35); metastatic cancer and acute leukemia (Hierarchical Condition Category 8), 6.76 (6.71-6.82); all sepsis, 2.63 (2.62-2.65); respiratory arrest (Hierarchical Condition Category 83), 2.55 (2.35-2.77); end-stage liver disease (Hierarchical Condition Category 27), 2.53 (2.49-2.56); and severe sepsis without shock, 2.48 (2.45-2.51). Models of the cost of sepsis care for Medicare beneficiaries forecast arise approximately 13% over 2 years owing the rising enrollments in Medicare offset by the cost of care per admission. CONCLUSIONS: A sepsis inpatient admission is associated with marked increase in risk of death that is comparable to the risks associated with inpatient admissions for other common and serious chronic illnesses. The aggregate costs of sepsis care for Medicare beneficiaries will continue to increase.

Correction to: Advanced Modeling to Predict Pneumonia in Combat Trauma Patients.

In the original article, the units indicated on the y-axes of Fig. 3 are incorrectly labelled. The correct label is pg/mL. Following is the corrected Fig. 3.

Advanced Modeling to Predict Pneumonia in Combat Trauma Patients.

BACKGROUND: Tools to assist clinicians in predicting pneumonia could lead to a significant decline in morbidity. Therefore, we sought to develop a model in combat trauma patients for identifying those at highest risk of pneumonia. METHODS: This was a retrospective study of 73 primarily blast-injured casualties with combat extremity wounds. Binary classification models for pneumonia prediction were developed with measurements of injury severity from the Abbreviated Injury Scale (AIS), transfusion blood products received before arrival at Walter Reed National Military Medical Center (WRNMMC), and serum protein levels. Predictive models were generated with leave-one-out-cross-validation using the variable selection method of backward elimination (BE) and the machine learning algorithms of random forests (RF) and logistic regression (LR). BE was attempted with two predictor sets: (1) all variables and (2) serum proteins alone. RESULTS: Incidence of pneumonia was 12% (n = 9). Different variable sets were produced by BE when considering all variables and just serum proteins alone. BE selected the variables ISS, AIS chest, and cryoprecipitate within the first 24 h following injury for the first predictor set 1 and FGF-basic, IL-2R, and IL-6 for predictor set 2. Using both variable sets, a RF was generated with AUCs of 0.95 and 0.87-both higher than LR algorithms. CONCLUSION: Advanced modeling allowed for the identification of clinical and biomarker data predictive of pneumonia in a cohort of predominantly blast-injured combat trauma patients. The generalizability of the models developed here will require an external validation dataset.

An Interpretable Machine Learning Model for Accurate Prediction of Sepsis in the ICU.

OBJECTIVES: Sepsis is among the leading causes of morbidity, mortality, and cost overruns in critically ill patients. Early intervention with antibiotics improves survival in septic patients. However, no clinically validated system exists for real-time prediction of sepsis onset. We aimed to develop and validate an Artificial Intelligence Sepsis Expert algorithm for early prediction of sepsis. DESIGN: Observational cohort study. SETTING: Academic medical center from January 2013 to December 2015. PATIENTS: Over 31,000 admissions to the ICUs at two Emory University hospitals (development cohort), in addition to over 52,000 ICU patients from the publicly available Medical Information Mart for Intensive Care-III ICU database (validation cohort). Patients who met the Third International Consensus Definitions for Sepsis (Sepsis-3) prior to or within 4 hours of their ICU admission were excluded, resulting in roughly 27,000 and 42,000 patients within our development and validation cohorts, respectively. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: High-resolution vital signs time series and electronic medical record data were extracted. A set of 65 features (variables) were calculated on hourly basis and passed to the Artificial Intelligence Sepsis Expert algorithm to predict onset of sepsis in the proceeding T hours (where T = 12, 8, 6, or 4). Artificial Intelligence Sepsis Expert was used to predict onset of sepsis in the proceeding T hours and to produce a list of the most significant contributing factors. For the 12-, 8-, 6-, and 4-hour ahead prediction of sepsis, Artificial Intelligence Sepsis Expert achieved area under the receiver operating characteristic in the range of 0.83-0.85. Performance of the Artificial Intelligence Sepsis Expert on the development and validation cohorts was indistinguishable. CONCLUSIONS: Using data available in the ICU in real-time, Artificial Intelligence Sepsis Expert can accurately predict the onset of sepsis in an ICU patient 4-12 hours prior to clinical recognition. A prospective study is necessary to determine the clinical utility of the proposed sepsis prediction model.

Serial Daily Organ Failure Assessment Beyond ICU Day 5 Does Not Independently Add Precision to ICU Risk-of-Death Prediction.

OBJECTIVES: To identify circumstances in which repeated measures of organ failure would improve mortality prediction in ICU patients. DESIGN: Retrospective cohort study, with external validation in a deidentified ICU database. SETTING: Eleven ICUs in three university hospitals within an academic healthcare system in 2014. PATIENTS: Adults (18 yr old or older) who satisfied the following criteria: 1) two of four systemic inflammatory response syndrome criteria plus an ordered blood culture, all within 24 hours of hospital admission; and 2) ICU admission for at least 2 calendar days, within 72 hours of emergency department presentation. INTERVENTION: NoneMEASUREMENTS AND MAIN RESULTS:: Data were collected until death, ICU discharge, or the seventh ICU day, whichever came first. The highest Sequential Organ Failure Assessment score from the ICU admission day (ICU day 1) was included in a multivariable model controlling for other covariates. The worst Sequential Organ Failure Assessment scores from the first 7 days after ICU admission were incrementally added and retained if they obtained statistical significance (p < 0.05). The cohort was divided into seven subcohorts to facilitate statistical comparison using the integrated discriminatory index. Of the 1,290 derivation cohort patients, 83 patients (6.4%) died in the ICU, compared with 949 of the 8,441 patients (11.2%) in the validation cohort. Incremental addition of Sequential Organ Failure Assessment data up to ICU day 5 improved the integrated discriminatory index in the validation cohort. Adding ICU day 6 or 7 Sequential Organ Failure Assessment data did not further improve model performance. CONCLUSIONS: Serial organ failure data improve prediction of ICU mortality, but a point exists after which further data no longer improve ICU mortality prediction of early sepsis.