Machine learning prediction for COVID-19 disease severity at hospital admission.

IMPORTANCE: Early prognostication of patients hospitalized with COVID-19 who may require mechanical ventilation and have worse outcomes within 30 days of admission is useful for delivering appropriate clinical care and optimizing resource allocation. OBJECTIVE: To develop machine learning models to predict COVID-19 severity at the time of the hospital admission based on a single institution data. DESIGN, SETTING, AND PARTICIPANTS: We established a retrospective cohort of patients with COVID-19 from University of Texas Southwestern Medical Center from May 2020 to March 2022. Easily accessible objective markers including basic laboratory variables and initial respiratory status were assessed using Random Forest's feature importance score to create a predictive risk score. Twenty-five significant variables were identified to be used in classification models. The best predictive models were selected with repeated tenfold cross-validation methods. MAIN OUTCOMES AND MEASURES: Among patients with COVID-19 admitted to the hospital, severity was defined by 30-day mortality (30DM) rates and need for mechanical ventilation. RESULTS: This was a large, single institution COVID-19 cohort including total of 1795 patients. The average age was 59.7 years old with diverse heterogeneity. 236 (13%) required mechanical ventilation and 156 patients (8.6%) died within 30 days of hospitalization. Predictive accuracy of each predictive model was validated with the 10-CV method. Random Forest classifier for 30DM model had 192 sub-trees, and obtained 0.72 sensitivity and 0.78 specificity, and 0.82 AUC. The model used to predict MV has 64 sub-trees and returned obtained 0.75 sensitivity and 0.75 specificity, and 0.81 AUC. Our scoring tool can be accessed at https://faculty.tamuc.edu/mmete/covid-risk.html . CONCLUSIONS AND RELEVANCE: In this study, we developed a risk score based on objective variables of COVID-19 patients within six hours of admission to the hospital, therefore helping predict a patient's risk of developing critical illness secondary to COVID-19.

Trends in utilization of deceased donor kidneys based on hepatitis C virus status and impact of public health service labeling on discard.

BACKGROUND: Kidneys from deceased donors infected with hepatitis C virus (HCV) are underutilized. Most HCV virus-infected donors are designated as Public Health Service increased donors (PHS-IR). Impact of PHS and HCV designations on discard is not well studied. METHODS: We queried the UNOS data set for all deceased donor kidneys between January 2015 and December 2018. The final study cohort donors (n = 38 702) were stratified into three groups based on HCV antibody (Ab) and NAT status: (a) Ab-/NAT- (n = 35 861); (b) Ab+/NAT- (n = 973); and (c) Ab±/NAT+ (n = 1868). We analyzed utilization/discard rates of these organs, the impact of PHS-IR and HCV designations on discard using multivariable two-level hierarchical logistic regression models, forecasted number of HCV viremic donors/kidneys by 2023. RESULTS: During the study period, (a) the number of viremic donor kidneys increased 2 folds; (b) the multilevel mixed-effects logistic regression models showed that, overall, the PHS labeling (OR 1.20, CI 95% CI 1.15-1.29) and HCV designation (OR 2.29; 95% CI 2.15-2.43) were independently associated with increased risk of discard; (c) contrary to the general perception, PHS-IR kidneys across all HCV groups, compared to PHS-IR kidneys were more likely to be discarded; (d) we forecasted that the number of kidneys from HCV viremic donor kidneys might increase from 1376 in 2019 to 2092 in 2023. CONCLUSION: Hepatitis C virus viremic kidneys might represent 10%-15% of deceased donor organ pool soon with the current rate of the opioid epidemic. PHS labeling effect on discard requires further discussion of the utility of this classification.