Predicting responders to prone positioning in mechanically ventilated patients with COVID-19 using machine learning.

BACKGROUND: For mechanically ventilated critically ill COVID-19 patients, prone positioning has quickly become an important treatment strategy, however, prone positioning is labor intensive and comes with potential adverse effects. Therefore, identifying which critically ill intubated COVID-19 patients will benefit may help allocate labor resources. METHODS: From the multi-center Dutch Data Warehouse of COVID-19 ICU patients from 25 hospitals, we selected all 3619 episodes of prone positioning in 1142 invasively mechanically ventilated patients. We excluded episodes longer than 24 h. Berlin ARDS criteria were not formally documented. We used supervised machine learning algorithms Logistic Regression, Random Forest, Naive Bayes, K-Nearest Neighbors, Support Vector Machine and Extreme Gradient Boosting on readily available and clinically relevant features to predict success of prone positioning after 4 h (window of 1 to 7 h) based on various possible outcomes. These outcomes were defined as improvements of at least 10% in PaO2/FiO2 ratio, ventilatory ratio, respiratory system compliance, or mechanical power. Separate models were created for each of these outcomes. Re-supination within 4 h after pronation was labeled as failure. We also developed models using a 20 mmHg improvement cut-off for PaO2/FiO2 ratio and using a combined outcome parameter. For all models, we evaluated feature importance expressed as contribution to predictive performance based on their relative ranking. RESULTS: The median duration of prone episodes was 17 h (11-20, median and IQR, N = 2632). Despite extensive modeling using a plethora of machine learning techniques and a large number of potentially clinically relevant features, discrimination between responders and non-responders remained poor with an area under the receiver operator characteristic curve of 0.62 for PaO2/FiO2 ratio using Logistic Regression, Random Forest and XGBoost. Feature importance was inconsistent between models for different outcomes. Notably, not even being a previous responder to prone positioning, or PEEP-levels before prone positioning, provided any meaningful contribution to predicting a successful next proning episode. CONCLUSIONS: In mechanically ventilated COVID-19 patients, predicting the success of prone positioning using clinically relevant and readily available parameters from electronic health records is currently not feasible. Given the current evidence base, a liberal approach to proning in all patients with severe COVID-19 ARDS is therefore justified and in particular regardless of previous results of proning.

Assess and validate predictive performance of models for in-hospital mortality in COVID-19 patients: A retrospective cohort study in the Netherlands comparing the value of registry data with high-granular electronic health records.

PURPOSE: To assess, validate and compare the predictive performance of models for in-hospital mortality of COVID-19 patients admitted to the intensive care unit (ICU) over two different waves of infections. Our models were built with high-granular Electronic Health Records (EHR) data versus less-granular registry data. METHODS: Observational study of all COVID-19 patients admitted to 19 Dutch ICUs participating in both the national quality registry National Intensive Care Evaluation (NICE) and the EHR-based Dutch Data Warehouse (hereafter EHR). Multiple models were developed on data from the first 24 h of ICU admissions from February to June 2020 (first COVID-19 wave) and validated on prospective patients admitted to the same ICUs between July and December 2020 (second COVID-19 wave). We assessed model discrimination, calibration, and the degree of relatedness between development and validation population. Coefficients were used to identify relevant risk factors. RESULTS: A total of 1533 patients from the EHR and 1563 from the registry were included. With high granular EHR data, the average AUROC was 0.69 (standard deviation of 0.05) for the internal validation, and the AUROC was 0.75 for the temporal validation. The registry model achieved an average AUROC of 0.76 (standard deviation of 0.05) in the internal validation and 0.77 in the temporal validation. In the EHR data, age, and respiratory-system related variables were the most important risk factors identified. In the NICE registry data, age and chronic respiratory insufficiency were the most important risk factors. CONCLUSION: In our study, prognostic models built on less-granular but readily-available registry data had similar performance to models built on high-granular EHR data and showed similar transportability to a prospective COVID-19 population. Future research is needed to verify whether this finding can be confirmed for upcoming waves.

New Physical, Mental, and Cognitive Problems 1 Year after ICU Admission: A Prospective Multicenter Study.

Rationale: Comprehensive studies addressing the incidence of physical, mental, and cognitive problems after ICU admission are lacking. With an increasing number of ICU survivors, an improved understanding of post-ICU problems is necessary. Objectives: To determine the occurrence and cooccurrence of new physical, mental, and cognitive problems among ICU survivors 1 year after ICU admission, their impact on daily functioning, and risk factors associated with 1-year outcomes. Methods: Prospective multicenter cohort study, including ICU patients ⩾16 years of age, admitted for ⩾12 hours between July 2016 and June 2019. Patients, or proxies, rated their health status before and 1 year after ICU admission using questionnaires. Measurements and Main Results: Validated questionnaires were used to measure frailty, fatigue, new physical symptoms, anxiety and depression, post-traumatic stress disorder, cognitive impairment, and quality of life. Of the 4,793 patients included, 2,345 completed the questionnaires both before and 1 year after ICU admission. New physical, mental, and/or cognitive problems 1 year after ICU admission were experienced by 58% of the medical patients, 64% of the urgent surgical patients, and 43% of the elective surgical patients. Urgent surgical patients experienced a significant deterioration in their physical and mental functioning, whereas elective surgical patients experienced a significant improvement. Medical patients experienced an increase in symptoms of depression. A significant decline in cognitive functioning was experienced by all types of patients. Pre-ICU health status was strongly associated with post-ICU health problems. Conclusions: Overall, 50% of ICU survivors suffer from new physical, mental, and/or cognitive problems. An improved insight into the specific health problems of ICU survivors would enable more personalized post-ICU care.

Physical, Mental, and Cognitive Health Status of ICU Survivors Before ICU Admission: A Cohort Study.

OBJECTIVES: Although patient's health status before ICU admission is the most important predictor for long-term outcomes, it is often not taken into account, potentially overestimating the attributable effects of critical illness. Studies that did assess the pre-ICU health status often included specific patient groups or assessed one specific health domain. Our aim was to explore patient's physical, mental, and cognitive functioning, as well as their quality of life before ICU admission. DESIGN: Baseline data were used from the longitudinal prospective MONITOR-IC cohort study. SETTING: ICUs of four Dutch hospitals. PATIENTS: Adult ICU survivors (n = 2,467) admitted between July 2016 and December 2018. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Patients, or their proxy, rated their level of frailty (Clinical Frailty Scale), fatigue (Checklist Individual Strength-8), anxiety and depression (Hospital Anxiety and Depression Scale), cognitive functioning (Cognitive Failure Questionnaire-14), and quality of life (Short Form-36) before ICU admission. Unplanned patients rated their pre-ICU health status retrospectively after ICU admission. Before ICU admission, 13% of all patients was frail, 65% suffered from fatigue, 28% and 26% from symptoms of anxiety and depression, respectively, and 6% from cognitive problems. Unplanned patients were significantly more frail and depressed. Patients with a poor pre-ICU health status were more often likely to be female, older, lower educated, divorced or widowed, living in a healthcare facility, and suffering from a chronic condition. CONCLUSIONS: In an era with increasing attention for health problems after ICU admission, the results of this study indicate that a part of the ICU survivors already experience serious impairments in their physical, mental, and cognitive functioning before ICU admission. Substantial differences were seen between patient subgroups. These findings underline the importance of accounting for pre-ICU health status when studying long-term outcomes.