Characteristics and Clinical Outcomes of Hospitalized COVID-19 Patients in Bojanala, South Africa

1. INTRODUCTION

Severe Acute Respiratory Syndrome (SARS-CoV-2), the virus responsible for COVID-19, emerged as a global public health challenge, placing substantial pressure on healthcare systems worldwide [1]. First identified in Wuhan, China, in December 2019, SARS-CoV-2 belongs to the coronavirus family [1, 2]. As of October 2023, the global mortality burden from COVID-19 exceeded 6 million, with over 3 million fatalities in the United States of America alone [1, 3]. As of January 2020, the World Health Organization declared COVID-19 a global pandemic following its rapid spread [4]. It was widely believed that the virus originated through zoonotic transmission from animals, most likely bats, to humans [5]. Although two competing hypotheses remain, natural zoonotic spillover versus a laboratory-associated incident, no conclusive evidence has definitively supported either [5-6]. In Africa, the spread of the COVID-19 pandemic was marked by various factors, including transmission dynamics, public health responses, and regional healthcare capacity [7]. While SARS-CoV-2 spread across nearly all African countries within three months, the continent reported lower transmission and mortality rates compared to other regions [8, 9]. South Africa, however, had the most COVID-19 cases and fatalities on the continent [8, 10].

Globally, successive waves of COVID-19 were driven by emerging viral variants, raising concerns about diagnostic efficacy, treatment responses, vaccine effectiveness, and disease severity [11, 12]. In both developed and developing countries, delayed care-seeking, limited access to healthcare, and underlying comorbidities were identified as major contributors to adverse outcomes. Studies linked conditions, including hypertension, diabetes mellitus, cardiovascular disease, chronic obstructive pulmonary disease, malignancy, and smoking history, to poorer COVID-19 outcomes [13, 14]. These comorbidities are thought to worsen disease progression through immune dysregulation, systemic inflammation, and endothelial dysfunction.

Whereas international data is abundant, there remains a lack of region-specific evidence documenting the highest incidence of COVID-19 cases and fatalities across the continent concerning clinical characteristics and outcomes of hospitalized patients in environments with limited resources. Bojanala, a district in South Africa’s North West province, exemplifies such a context. The country’s dual healthcare system, characterized by marked inequalities in access, provides a unique environment for examining the clinical burden of COVID-19 [15]. Therefore, the objectives of this study were to assess demographic characteristics, comorbidity burden, treatment interventions, and clinical outcomes of hospitalized patients with confirmed COVID-19 infection.

2. METHODOLOGY

2.2. Study Setting

The study was conducted in five public health facilities in the Bojanala district, North West Province, in South Africa. Participating hospitals in the COVID-19 response effort included a regional hospital, three district hospitals, and a field hospital, which the district mandated be responsible for the admission of patients with confirmed diagnoses. The facilities were designated to treat patients who tested positive for COVID-19, each facility playing a pivotal role in managing the healthcare demands during the pandemic. (Fig. 1) shows the map of the Bojanala District where the data were collected. There are four districts in the North West Province, with Bojanala being the largest, having a population of 1.62 million people according to the Department of Statistics 2023 Census [16]. The area of Bojanala District is 18,333 km², and is made up of five regional governments, namely Kgetleng Rivier, Madibeng, Moses Kotane, Moretele, and Rustenburg, which have a total of 139 wards. It is situated within a mining area, along the N4 National Road, running from the Botswana border into the Mpumalanga Province. Most of the district's population is black Africans, making up 98% of the overall population, with whites making up 2% of the total population [17]. Most of the population (84%) relies on public health facilities, while roughly 16% is covered by medical schemes that provide them access to private healthcare services. A small proportion of individuals use both public and private facilities. All patients hospitalized at any of the five designated public health facilities in the Bojanala District who had a laboratory-confirmed diagnosis of COVID-19 were included in the study. The study employed a census sampling approach, capturing the entire eligible population during the enrolment period from May to July 2020, which corresponded with the peak of the first COVID-19 wave in the North West Province. This period was selected to ensure uniformity in treatment protocols and to reflect the healthcare system's response during the height of the pandemic. A total of 289 patients met the inclusion criteria and were enrolled. By including all eligible cases, the study aimed to maximize data completeness and avoid sampling bias within the defined setting. The sample size of this study was relatively small compared to studies conducted in similar regions. In South Africa, the rollout of electronic health record platforms is still under development, and public health facilities remain reliant on paper-based or partial documentation systems [18]. A recent qualitative study of the South African public health sector showed that, despite existing policy frameworks on rolling out a comprehensive national electronic health system [19]. Some barriers, including deficiencies in infrastructure, interoperability, and leadership, still exist. An integrated electronic repository for the reliable extraction of population-level clinical data is absent, and patient records are fragmented across hospitals, maintained in paper-based files, and not incorporated into any centralized system that facilitates large-scale epidemiological analysis.

3. RESULTS

3.5. Model Strength and Predictive Utility

The Receiver Operating Characteristic curve yielded an Area Under the Curve of 0.92, demonstrating excellent discriminatory performance. The ROC curve analysis for the final logistic regression model is provided in Fig. (2). The sensitivity was 91%, and the specificity was 78%, confirming the model's effectiveness in predicting adverse COVID-19 outcomes. The final logistic regression model exhibited high discriminatory ability with an AUC of 0.92. High sensitivity and specificity indicate clinical utility in flagging high-risk patients. The absence of significance for age and comorbidities in the adjusted model highlights the complex interaction between clinical factors, possibly due to multicollinearity or shared predictive weight with treatment indicators.

4. DISCUSSION

This study was conducted in five public healthcare facilities in the Bojanala district of the North West Province in South Africa to explore the demographic profile, multimorbidity, and outcomes of admitted patients with confirmed COVID-19.

Variations in patient demographics and outcomes from this study and other regions were observed. Studies conducted in KwaZulu-Natal and Cape Town, South Africa, showed that advanced age and comorbidities were independent risk factors associated with adverse COVID-19 severity and mortality [21, 23]. The findings of this study were similar to those of studies conducted in South Africa and other regions. In studies conducted in the USA and China, older patients, male gender, and presence of multimorbidity were associated with the risk of mortality [24, 25]. There are differences noted in findings of various COVID-19 studies, possibly influenced by the methodologies used in each study, the criteria for hospital admissions, patient health-seeking behaviour, and prevalence and severity of chronic conditions in different regions [26].

Healthcare systems and utilization were affected globally by the COVID-19 pandemic due to a combination of factors. In this study, patient health-seeking behaviour, financial constraints, chronic disease burden, and lockdown regulations likely affected some patients in seeking timely medical care. The delay of regular monitoring and treatment might lead to insufficient disease management, which would eventually lead to an increase in the risk of complications, long-term impairment, and a higher cost on the healthcare system in the future [26].

In a study conducted in China on COVID-19 severity and mortality, with a focus on gender variations, found that among older patients who died, more than 70% were male [27]. Although susceptibility and ages were comparable between males and females, males were more likely to experience adverse outcomes than females. The association between male gender and mortality was linked to male health-seeking behaviour and life expectancy in China and possibly globally.

Although our findings indicate that the mean age of patients who died was significantly higher than that of patients who recovered, age was not a predictive factor for mortality. Male patients represented a larger proportion of this study; however, there were no differences in mortality between male and female patients. Similar studies in South Africa and other regions reflect that male patients were associated with severe COVID-19 outcomes and mortality; our findings reflect no age or gender differences in adverse outcomes.

South Africa has a high prevalence of HIV, diabetes, hypertension, and arthritis compared to most regions on the continent and globally [28]. Often, older populations require regular screening, given the high prevalence of multimorbidity in this population.

We studied the correlation between multimorbidity and COVID-19 outcomes. The severity and management of chronic conditions were not established due to the retrospective nature of the study design, and public health facilities rely on paper-based systems, while the rollout of electronic systems in the country is developing.

Previous studies have shown that patients with comorbidities, including diabetes, hypertension, cardiovascular disease, and kidney disease, were associated with an increased risk of adverse COVID-19 outcomes [29, 30].

We report on the association of multimorbidity in patients admitted for COVID-19, where almost half of the patients experienced severe COVID-19 outcomes and mortality. The patients were 2.3 times more likely to experience adverse outcomes than patients with one or no chronic conditions. Multimorbidity was reported as a simple count and not graded for clinical severity due to how information was recorded in the patients' files at the healthcare facilities during data collection. PLHIV with low CD4 T cell count or immunological weakness, HIV-related immunosuppression may mitigate the hyperinflammatory response that is associated with severe COVID-19, potentially protecting against a cytokine storm [31].

The susceptibility for adverse COVID-19 outcomes may increase due to the combined effect of immunosuppression and significant comorbidity, particularly when HIV was untreated or chronic immunological activation [32, 33]. However, comorbidities, immunological dysfunction, and delayed treatment continued to be significant concerns for PLHIV during the epidemic [33]

Studies suggested that PLHIV hospitalized with COVID-19 were not at an increased risk of adverse outcomes [31, 34, 35]. However, PLHIV with detectable viral loads showed less severe respiratory disease than those who were virologically suppressed, and their inflammatory responses were comparable to those without HIV. The number of PLHIV in this study was lower than in other studies. The status of their chronic condition was not known, and only a few had multimorbidity. Most of these patients recovered, and their outcomes were comparable to those without HIV.

Compared to national surveillance statistics and estimates from other studies conducted in more severely affected regions, the mortality rate was lower in our region [36]. A variety of factors could have contributed to this trend, such as differences in healthcare facility capacities, proactive quarantine measures, and earlier diagnoses. In studies conducted in similar settings in South Africa, higher mortality rates were reported linked to delays at presentation and restrictive access to medical treatment due to the high volume of patients in need of medical care during the period of the pandemic [21, 35].

Patients, particularly those with multimorbidity, were at a great risk of severe COVID-19 outcomes and mortality with every increase in age, as a result of their predisposition to organ failure, compromised immunological responses, and reduced susceptibility to viral infections [37]. This combination of factors is associated with less favourable recovery outcomes and likely necessitates longer hospital stays and more intensive treatment. While the age was a considerable indicator for adverse effects in univariable analysis, the lack of significance in multivariable regression suggests that its impact was influenced by other factors, including the severity of comorbidities, requirements for treatment, and other factors.

A young age served as a protective factor against COVID-19 complications; this beneficial effect might also be mitigated by the high prevalence of uncontrolled HIV infection and non-communicable diseases in certain geographic regions [38].

In a review on the impact of comorbidities on COVID-19 patients, PLWHIV hospitalized due to COVID-19 were noted to be younger and had a higher hospital mortality than patients without HIV [39]. Patients without comorbidities were also reported to recover with minimal intervention, while patients with multimorbidity and comorbidities often required escalation to higher care and extended treatment durations. This was also possible due to the added risk of organ support complications and acquiring secondary infections due to prolonged ventilator support.

We also report that treatment with anticoagulants and antibiotics was associated with adverse COVID-19 outcomes after adjusting for confounders, including age and chronic conditions. However, this finding should be treated with caution, as factors in policy frameworks during the early pandemic and the severity of each patient’s chronic condition may affect outcomes.

In this study, a higher mean hospital stay was noted in patients who recovered than in those who died. This was due to the recommended isolation period of 14 days for individuals with a positive COVID-19 test to limit the spread, rather than an indication for treatment of a severe medical condition. Most countries recommended that patients with COVID-19 symptoms who were not able to self-isolate at home, to go to designated healthcare facilities regardless of the severity of their symptoms during the first waves of the pandemic [22]. Our multivariable model's higher AUC (0.92) shows its predictive power in early high-risk patient identification. In the adjusted model, age, comorbidities, and steroid usage lacked statistical significance, suggesting multicollinearity or shared predictive weight.

5. STRENGTHS AND LIMITATIONS OF THE STUDY

The study was conducted at public health facilities in the Bojanala District, reflecting the clinical conditions of resource-limited settings where more than 80% of South Africans seek medical care. Using the census sampling method, all 289 confirmed COVID-19 patients hospitalized during the first pandemic period were included. This improved internal validity ensured that the results accurately represented real-world clinical practice. To address prevalent limitations of retrospective studies, standardized data collecting methods, cross-referencing records within health facilities, and trained research assistants were used to improve reliability and consistency. Several comorbid conditions presented occasionally, constraining the capacity to analyze them individually or investigate condition-specific risks. Consequently, the comorbidity load was evaluated primarily as a count of chronic diseases, potentially underrepresenting the significance of high-risk severe comorbidities. The study used applicable non-parametric tests and multivariable logistic regression, in conjunction with diagnostic evaluations, to determine the relationships between clinical variables and outcomes. The resulting model demonstrated a high discriminative ability, achieving an area under the ROC curve of 0.92, thus affirming the strength of its prediction accuracy.

However, several limitations need to be recognized. The study's exclusion of private sector facilities limits the generalizability of its results to the wider South African population or alternative healthcare systems. The use of non-random sampling, although inclusive, may increase the probability of selection bias. The limited sample size constrains the detection of lower effect sizes and results in broader confidence ranges in adjusted models. Multicollinearity among clinical factors may have reduced individual impact estimates in the multivariable study. The national policy in early 2020 mandating facility isolation for patients unable to quarantine at home may have extended hospital stays for non-clinical reasons, especially in field hospitals, complicating the interpretation of length of stay as a measure of illness severity.

Despite measures used to verify records, the research did not systematically quantify missing data. A comprehensive case analysis method was used, including only patients with complete data for each model in the studies. This strategy guaranteed consistency but might have added bias if the missing data were associated with clinical severity or other unmeasured variables. Evolving clinical guidelines and limited healthcare resources in the early months of the epidemic may have influenced treatment choices and results, hence restricting comparison with subsequent stages of the pandemic or varying health system settings.

CONCLUSION

This study showed significant associations between comorbidity burden, age, and clinical outcomes in COVID-19 patients hospitalized in public facilities. Hypertension, diabetes mellitus, and HIV were the prevalent comorbidities. Patients with multiple comorbidities had an over 50% death rate, reflecting the cumulative effect of chronic illness clustering. Older age was a significant predictor, as it correlated with higher mortality; however, its predictive value decreased in multivariate models, showing confounding by treatment intensity or comorbidity burden. Patients receiving treatment, particularly antibiotics and anticoagulants, had an increased mortality risk, independent of adjustments for other clinical and demographic variables. This suggests that these treatments may have acted more as markers for illness severity than as direct factors contributing to adverse outcomes. During the early waves of the pandemic, all COVID-19-positive individuals unable to self-isolate at home were hospitalized for a 14-day isolation period, regardless of treatment need. This approach may have led to the extended duration of hospitalization in some patients who need no active treatment, as well as to the perceived gap between hospitalization and clinical severity in low-risk populations. Patients with one or no comorbidities often achieve recovery with minimal intervention, while the occurrence of multiple chronic diseases correlated with prolonged hospitalization and increased treatment need, indicating more severe disease trajectories and reduced physiological reserve. Emerging evidence suggested positive as well as negative impacts of HIV-related immunosuppression on COVID-19 outcomes. While studies show that people using ART may evade the severe inflammatory consequences of COVID-19, poorly managed HIV with immunological impairment continues to be a substantial risk factor. This intricate interaction needs more investigation in multiple settings. Although the mortality rates in this group are modest compared to national averages, the results highlighted age, multimorbidity, and treatment intensity as fundamental predictors of adverse outcomes. Individualized risk classification and proactive clinical care are essential, especially in resource-constrained environments.

This study emphasized the importance of early identification and stratification of high-risk COVID-19 patients, especially those with multiple comorbidities such as hypertension, diabetes, and HIV, in resource-constrained settings, where advanced care is limited. Clinical decisions should rely on accessible indicators such as age, comorbidity burden, and treatment needs. Clinical protocols should incorporate comorbidity screening and severity at admission, ensure access to essential treatments, and always keep chronic disease management to reduce preventable mortality.

REFERENCES