The entire world has been greatly challenged by the sudden outbreak of COVID-19, as the human race has no remedial measures to combat the lethal impact of the disease. According to the World Health Organization (WHO), the global pandemic of COVID-19 is derived from SARS-CoV-2, a member of a large family of viruses, named coronaviruses; these viruses cause respiratory infections ranging from the common cold to high fever, leading to disease. This blue planet has witnessed many epidemics, such as that caused by severe acute respiratory syndrome coronavirus (SARS-CoV) from 2002 to 2003 and H1N1 influenza in 2009 [1], due to various pernicious viruses in the last two decades; however, COVID-19 is incomparable with previous epidemics because of the indomitable growth rate of the disease and its high fatality rate. China was the first country to experience high numbers of cases at the beginning of the pandemic; presently, Chinese authorities have “flattened the curve” with continuous testing and aggressive quarantine measures [2]. Outside China, South Korea was the country that had the largest initial outbreak; they managed to slow the spread of COVID-19 and flatten the curve without imposing lockdown in their country [3]. The only method used to slow and contain the outbreak in Korea was mass diagnostic testing and quarantining. The WHO declared that incorporating self-isolation, sanitizing, washing hands repeatedly and abstaining from touching the mouth, face, or nose to stop the spread of COVID-19 [4]. To combat transmission of the disease in India, the government declared a lockdown on March 25, 2020. However, the disease has spread rapidly across the entire country, and as of June 30, 202, there were 1,385,494 cases, with 32,096 deaths and 886,235 recoveries [5]. For a developing country such as India, the COVID-19 pandemic is a serious problem facing the nation, and the main sufferers are marginalized sections of society. Even after a strict lockdown was established nationwide, the number of cases increased and surpassed 450,000. However, the fatality rates later decreased, and several studies have shown that the lockdown did slow the rate of increase in a number of cases [6]. A positive point to be noted is that the number of recovered cases is slowly exceeding that of active cases.

In the study of dynamics of infectious diseases, compartmental models and the basic reproductive number (r₀) have been observed to be the mostly commonly used over the past year [7]. Basic mathematical models such as the Gompertz, exponential, and logistic growth models have shown to be quite effective in understanding the growth patterns of the disease [8]. One of the main demerits of the Indian database for COVID-19 is the underreporting of cases due to misreporting and the lower number of tests [9]. Amid the growing number of deaths due to COVID-19, researchers worldwide have associated these deaths with additional important cofactors, namely the effect on the older population and the impact of pollution and smoking as well as the development of acute respiratory distress syndrome [10,11]. In one study [12], a district-level analysis showed that 92 districts in India are in red zones of the disease. These red zones are mostly found in the states of Maharashtra and Gujarat; in another study [13], it was predicted by the autoregressive integrated moving average (ARIMA) model that the number of cases will increase alarmingly.

Studies have described the impact of lockdown, the transmission dynamics of the disease, and forecasts of the pandemic. The survival of patients, taking death as the event that varies by age group and sex, is noteworthy. The aim of this study was to conduct a survival analysis to establish the variability in survivorship by age group and sex at different levels, that is, the national, state, district, and patient levels. This quantitative analysis (with analysis of data from patients with COVID-19) is exceptional not only for its gravity and pertinence but also for its subtle nuances and penetrating approach.

Kaplan-Meier estimates were obtained initially to estimate the survival functions of patients in India with COVID-19 by sex, age group, and region. It can be seen from Figure 2 that the survival curves from the Kaplan-Meier estimator for male and female sex are almost the same. In Table 1, it can be observed that according to both the log rank test (P>.001) and Wilcoxon test (P>.001) for comparing the survival functions, the difference is not significant, indicating that there is no significant difference between the survival curves of male and female patients with COVID-19.

As shown in Figure 3, the survival curves from the Kaplan-Meier estimator by 5-year age group are significantly different. This result was further supplemented by the log rank test (P<.001) and Wilcoxon test (P<.001) for comparing the survival functions (Table 2); both tests gave highly significant results, indicating that there are significant differences among the survival curves of various age groups.

Figure 4 depicts that the survival curves from the Kaplan-Meier estimator for different regions of India are significantly different. From Table 3, it can be inferred that the log rank test (P<.001) and Wilcoxon test (P<.001) for comparing the survival functions are both highly significant, indicating that there are significant differences among the survival curves by region due to regional variations. Therefore, we can find that both age and region are significantly associated with the survival rate of COVID-19 without adjusting for other covariates. Figure 5 depicts a comparison of the survival curves among the states most affected by COVID-19, which shows the low survival rates in Maharashtra, Delhi, Gujarat, West Bengal, and Rajasthan.

Table 4 presents the results of the survival analysis using the Cox proportional hazard model and reiterates that male patients with COVID-19 have a 1.14 times higher risk of dying than female patients (hazard ratio [HR] 1.14; SE 0.11; 95% CI 0.93-1.38). Coming to the agewise comparison, we observed that patients aged 45-49 years to ≥65 years had a 5.83, 10.08, 15.31, 22.03, and 39.21 times higher risk of dying due to COVID-19 than those in the age group of 0-5 years, respectively. The highest risk of death from the disease was among people in the ≥65 years age group, with 39.2 times higher risk (P<.001) but a larger confidence interval (95% CI 9.73-157.97). When analyzing the survival curves by region, it can be seen that patients in the East, North East, and South Indian regions were at 59%, 14%, and 26% lower risk of dying from COVID-19 infection than those in Central India, respectively, whereas patients in the West Indian region had a 1.9 times higher risk of dying than those in Central India.

A multilevel survival model was applied at the overall national level (Weibull regression) and the state and district level (mixed effects Weibull regression), taking age group and sex as the covariates. The results of the multilevel survival analysis, that is, after controlling the variability due to clustering of the lower level at the higher level, can be seen in Table 5; the hazard ratio at the India level shows that male patients are at 1.27 times higher risk than female patients of dying from COVID-19 (HR 1.27; SE 0.13), which is almost the same at the state level (HR 1.32; SE 0.13) as well as the district level (HR 1.21; SE 0.13). Significant variability in survival was observed for the age groups older than 45 years. At all levels, we found that the hazard ratio increased with increasing age but decreased across each level. For instance, in the ≥65 years age group, the patient is at 39.3 times higher risk of dying overall than a patient in a younger age group (HR 39.3; SE 27.94); meanwhile, at the state level, the hazard ratio is 32.28, and that at the district level is 23.55. Now, from the variance of the errors of model, we can infer that the heterogeneity is greater at the district level (σ_e² 6.85; SE 1.35) than at the state level (σ_e² 2.28; SE 0.83).

Epitomizing the whole study, it is notable that in the stipulated time period, our observation clearly revealed that the survival rate was continually declining, and to date, that trend has not abated. It is worth mentioning that age, sex, and regional variability were important determinants at each step. Also, from this study, it is very clear that the male population in India is more vulnerable to COVID-19, likely due to prevalent comorbidities and the dominant presence of men outside the home (also, our data support the fact that the survival rate of the female population is higher). This study also traced a different pattern for India than for other countries, as the younger population is greater in our country than in most countries where the number of affected people is numerous.

In this study, we strived to identify reliable features associated with survival patterns, and we inevitably scanned the roles of sex, age, and regional variability as controlling factors of the survival rate. For the survival analysis, the study period was 5 months, with death being the event of interest in our analysis. As we evaluated the Kaplan-Meier survival function, we observed that the probability of survival continually declined during the study period of 5 months. During the study period, no stabilization could be observed. Female patients were found to have better survival rates compared to their male counterparts, as is evident from the Cox proportional hazard results, which may be due to sex differentials in cellular compositions and the immunological microenvironment of the lungs [14,15]. Although we only observed a miniscule difference in the survival curves of male and female patients, it was stated in earlier studies that men with COVID-19 are at higher risk of death and health outcomes, independent of age [16], as men have greater disease burden (diabetes, hypertension, or cardiovascular diseases); therefore, men have shown markedly increased risk of developing severe COVID-19 in comparison to women. Also, a greater proportion of the confirmed cases are male rather than female; this finding is expected in a country with a gender hegemony in which work participation, mobility, and migration are predominately higher for men than for women [17], which makes men more vulnerable to the infection.

Significant variability was observed among age groups, as is evident from all the survival estimates, which show that with increasing age, the risk of dying from COVID-19 increases [18]. It was reported in a study that among comorbid patients with COVID-19, nearly 21% had hypertension, 11% had diabetes, and 7% had cardiovascular disease, which increased their risk of mortality [19]. In contrast to data from other countries, in India, only 15% of confirmed cases are aged >60 years, and the majority of these patients are in the age bracket of 25-59 years; this is most probably because the older population is the most affected by this pandemic and India has a fairly young population, which may contribute to a lower case fatality rate [9]. Approximately 84% of the patients with COVID-19 were men, and 82% patients overall were above 40 years of age, as reported in an Indian Council of Medical Research study [20]. India is one of the largest countries in the world, and it is highly diverse in every respect. Every province posseses its own demographic features, typical climatic character, and above all, its own lifestyle. Needless to say, these factors play a pivotal role. Therefore, variation in survival rate is easily traced. Although Western and Central India show continually decreasing survival rates in the framed time period, Eastern, Northeastern and Southern India show slightly better results in terms of survival. Maharashtra, Gujarat, Delhi, Rajasthan, and West Bengal showed alrmingly low survival rates as well.

Finally, this study has depicted a grave scenario of continual degradation of the COVID-19 survival rate in various regions. In essence, we can safely conclude that critical appraisal of the survival rate and thorough analysis of patient data in this study equipped us to identify risk groups and perform comparative studies of various segments of the population in India.