Evaluation of the effect values of risk factors by clustering method in patients who died due to COVID-19 disease

Abstract

Objectives: The aim of this study is to determine the factors that may be associated with mortality in patients who died due to COVID-19 and to determine the effect sizes of the factors that make a statistically significant difference. Methods: The patients who died due to COVID-19 between 01.03.2020 and 01.03.2021 in Bursa province were evaluated retrospectively. In addition to demographic information such as age, gender, nationality, existing chronic diseases of the patients, COVID- PCR test results, length of hospital stay, intensive care unit follow up times, intubation application times were recorded. The effect size of the variables on mortality were evaluated. Results: Total of 3,510 deaths due to COVID-19 were evaluated. Of these, 2,107 (60%) were male and 1,403 (40%) were female. Three thousand three hundred and seventy-four (96.12%) patients are 50 years or older. In both sexes, the highest number of deaths were in the age range of 70-79. The most common comorbidities were hypertension (HT) (n = 1,182; 34.16%) and diabetes mellitus (DM) (n =776; 22.43%). HT and DM had a strong effect value between the groups (p < 0.001 and p < 0.001, phi effect values: 0.661 and 0.681, respectively). Although there was a statistically significant difference for the age variable, it had an insignificant effect value (p = 0.008, 𝛿 = 0.074). Conclusions: Risk factors frequently reported for COVID-19 deaths but there are no studies showing the true effect values. In this study, HT and DM had a strong effect separately, gender and coronary artery disease (CAD) variables were moderate, chronic obstructive pulmonary disease (COPD), lung cancer and other chronic disease variables had weak effect values, age and non-lung cancers had insignificant effect.