Development of an index predicts the in-patient mortality of COVID-19 patients presenting at a hospital in rural Michigan.

Abstract

Abstract Background Laboratory test results and chief complaints (CC) for patients hospitalized with COVID-19 can contribute to a better understanding of inpatient mortality risk. This study used a combination of lab test results on admission, demographic information, comorbidity data, and reported patient chief complaints to create a predictive model for inpatient mortality from COVID-19. Methods Clinical data were collected from a regional hospital (MI, USA). 1,093 COVID-19 patients were admitted. The CC, lab variables, and health comorbidities were inserted into a multiple binary logistic regression model alongside comorbidity information and the lab results, which was used to create a risk estimation tool for inpatient mortality in patients hospitalized with COVID-19. Results 1,088 cases were included in the analysis. 23.25% of the hospitalized COVID-19 patients (N = 253) died. The average age of patients who died was 77.14 years (+/- 13.99) vs 64.22 years (+/- 18.35) for those who did not die. 49.9% (N = 545) of patients were female. Mortality was higher in non-white patients [OR = 3.7 (95% CI: 1.14–12.1)], those older in age [OR = 1.1 (95% CI: 1.04–1.14)]; those with a prior myocardial infarction/coronary artery disease [OR = 2.7 (95% CI: 1.02–7.03)], those with hypertension [OR = 5.2 (95% CI: 1.14–12.1)] and those with higher WBC counts [OR = 1.2 (95% CI = 1.02–1.50)]. High total protein indicated decreased mortality [OR = 0.4 (95% CI = 0.20–0.84)]. Conclusions Multiple comorbidities are associated with greater mortality in those hospitalized with COVID-19. Understanding these risks will aid clinicians and healthcare systems in decision-making and allocation of resources to control disease burden.