RESUMO
A large range of prognostic models for determining the risk of COVID-19 patient mortality exist, but these typically restrict the set of biomarkers considered to measurements available at patient admission. Additionally, many of these models are trained and tested on patient cohorts from a single hospital, raising questions about the generalisability of results. We used a Bayesian Markov model to analyse time series data of biomarker measurements taken throughout the duration of a COVID-19 patients hospitalisation for n = 1540 patients from two hospitals in New York: State University of New York (SUNY) Downstate Health Sciences University and Maimonides Medical Center. Our main focus was to quantify the mortality risk associated with both static (e.g. demographic and patient history variables) and dynamic factors (e.g. changes in biomarkers) throughout hospitalisation, by so doing, to explain the observed patterns of mortality. By using our model to make predictions across the hospitals, we assessed how predictive factors generalised between the two cohorts. The individual dynamics of the measurements and their associated mortality risk were remarkably consistent across the hospitals. The model accuracy in predicting patient outcome (death or discharge) was 72.3% (predicting SUNY; posterior median accuracy) and 71.3% (predicting Maimonides) respectively. Model sensitivity was higher for detecting patients who would go on to be discharged (78.7%) versus those who died (61.8%). Our results indicate the utility of including dynamic clinical measurements when assessing patient mortality risk but also highlight the difficulty of identifying high risk patients.
RESUMO
IntroductionAs COVID-19 hospitalization rates remain high, there is an urgent need to identify prognostic factors to improve treatment. Our analysis, to our knowledge, is one of the first to quantify the risk associated with dynamic clinical measurements taken throughout the course of hospitalization. MethodsWe collected data for 553 PCR-positive COVID-19 patients admitted to hospital whose eventual outcomes were known. The data collected for the patients included demographics, comorbidities and laboratory values taken at admission and throughout the course of hospitalization. We trained multivariate Markov prognostic models to identify high-risk patients at admission along with a dynamic measure of risk incorporating time-dependent changes in patients laboratory values. ResultsFrom the set of factors available upon admission, the Markov model determined that age >80 years, history of coronary artery disease and chronic obstructive pulmonary disease increased mortality risk. The lab values upon admission most associated with mortality included neutrophil percentage, RBC, RDW, protein levels, platelets count, albumin levels and MCHC. Incorporating dynamic changes in lab values throughout hospitalization lead to dramatic gains in the predictive accuracy of the model and indicated a catalogue of variables for determining high-risk patients including eosinophil percentage, WBC, platelets, pCO2, RDW, LUC count, alkaline phosphatase and albumin. ConclusionOur prognostic model highlights the nuance of determining risk for COVID-19 patients and indicates that, rather than a single variable, a range of factors (at different points in hospitalization) are needed for effective risk stratification.