Investigation on the Risk Factores for Mortality of Patients with COVID-19 and Prioritization These Factores Using Neural Network in Some Southern Cities of Iran

Epidemics of human viruses began during the period of Neolithic around 12,000 years ago. Humans developed more densely population which allowed viruses to spread rapidly among communities. Also, plant and livestock viruses increased along with human viruses (2). At the January 2020, the coronavirus disease (COVID-19 7th human coronavirus) was discovered in Wuhan, Hubei province of China. COVID-19 virus caused six million deads in the world to date and cussed infection of more than seven million of cases in Iran (1). This infectious disease caused by the SARS-CoV-2 virus. This virus was contagious and fast-spread. Despite the aquarantine politics, SARS-CoV-2 virus caused many permanent economic and health damages in most countries. Coronaviruses are positive-sense, single- stranded enveloped RNA viruses with helical capsids that infect a wide range of hosts including humans, bats, other mammals, and birds (2). Coronaviruses are belonging to Nidovirales order, Coronaviridae family, Coronovirinae subfamily and four genera of alpha, beta, delta, and gamma. Alpha and beta coronaviruses are known as human infection agents. SARS-COV-2 virus abilities are including: high mortality number, short period of incubation, widespread transmission protocols, asymptomatic infection and affecting on most vital organs (heart, brain, lungs and …) which have attracted the health system attention and caused neglect to the other chronic and non-communicable diseases (4). Therefore, the disease incidence, prevalence and prioritization around the world may change in the future. From the beginning of COVID-19 pandemic, some symptoms and risk-factors have been introduced to the world as the increase elements of morbidity and mortality. Studies have shown that having any kind of underlying diseases and risk factors will be effective in the COVID-19 disease severity and mortality (6). Some of these important risk factors are including of chronic kidney disease, hypertension, age, gender, obesity, obstructive pulmonary diseases, diabetes, lung diseases, cardiovascular diseases, cancer, and liver disease. Also, each risk factors have different impact in different geographic areas (7). Some factors, such as different viral load kinetics in each individual person, epidemiological history, therapeutic or pharmacological effects and immune response have some major impacts on the laboratory diagnostic results. Due to the successive mutations of the SARS-CoV-2 virus and the high incidence disease, it seems that the vaccination alone cannot prevent the COVID-19 (9). On the other hand, the World Health Organization has warned about the vaccination as the only pandemic control protocol. Therefore, the prevalence of morbidity and mortality have become the public health concerns in the world since the beginning of the COVID-19 epidemic and the vaccination. Recognizing of the risk-factors and symptoms on COVID-19 in different geographic areas can be a helpful source to prevent the mortality. Understanding risk factors can help the world to control of the coronaviruses pandemic period and similar situations in the future. Therefore, the aim of this study was to determine the risk-factors of mortality of COVID-19 patients in three cities of Khuzestan province, Iran.

This research was an analytical cross-sectional study. Some details of 27963 COVID-19 patients such as clinical symptoms, individual characteristics and underlying diseases were gathered from hospitals in Abadan, Shadegan and Khorramshahr cities in Khuzestan province, Iran, from 20 February 2020 to November 2020. All the under-study population was previously investigated in terms of COVID-19 infection by the medical examinations and laboratory methods. This under-study population was categorized into three different groups such as hospitalized, outpatients and dead patients. Hospitalized patients have admitted in general or ICU (Intensive Care Unit) sector. Obtained database of COVID-19 patients was analyzed by IBM SPSS version 22.0 under regression, logistic model (univariable and multivariable logistic regression models) with 95 percent confidence level. Also, the neural network method was used for prioritizing of significant risk factors for mortality. At the end of the analysis, the models of multiple logistic regression and neural network were evaluated for Goodness of Fit. In this study, the anonymity principle and patient's preservation of personal information was considered during analytical method.

The mean of age was 40 years. The sex ratio was higher for men. That ratio for dead patients were approximately 63 years (from 62.7597 to 64.9854). The number of hospitalizations and deaths was occurred in May-July 2020 and the greatest number of deaths reports was belonged to Abadan city. The most recognized prevalent symptoms were cough, fever, hard breath and sickness which observed more within dead patients. Prevalent underlying diseases were diabetes, hypertension and blood diseases. The mortality risk factors of the multivariable logistic regression model were diabetes, age, blood diseases, cardiovascular diseases, neurological diseases, respiratory diseases and cancers. Also, variables such as diseases related to blood lipid, gender and thyroid diseases were removed from the model of multivariable logistic regression according to the model univariable logistic regression under 0.2 confidence level because they had no statistical significance for entering to the model of multiple regression. According to neural network analysis, age was the most important mortality risk factor. Other important risk factors were neurological diseases, blood diseases, respiratory diseases, cardiovascular diseases, diabetes and cancers respectively. The odds ratio of mortality increased with increasing number of underlying diseases among COVID-19 patients. The presence of at least one risk factor increased the odds of mortality approximately 8.3 times. Variables as kidney diseases, hypertension, Immunodeficiency and obesity were not recognized as mortality risk factors in the model of multivariable logistic regression.

This study investigated the effective risk factors for mortality among patients with COVID-19 in southern of three cities of Abadan, Khorramshahr and Shadegan of Khuzestan province, Iran. According to the prioritizing in neural network and risk factors of mortality, we recommend that after attention to age as the most important risk factor for mortality, COVID-19 patients with blood and neurological disease history should receive more public health care services than the others till the end of COVID-19 pandemic period. Although, antigenic mutations of COVID-19 virus have reduced the effectiveness of recent vaccines. However, risk factors for mortality of COVID-19 patients as the important disease prevention levels are need more attention from health politicians. Understanding risk factors of mortality, can be useful for future researches and similar epidemic or pandemic of any coronaviruses.