Relative burdens of the COVID-19, malaria, tuberculosis and HIV/AIDS epidemics in sub-Saharan Africa

Introduction

COVID-19 has massively impacted life and society in sub-Saharan Africa, as elsewhere. Despite relatively low COVID-19 mortality rates in most countries of sub-Saharan Africa, aspects of the lockdown responses including business and school closures and restricted health service access introduced in the early days of the pandemic continue in various forms.¹ In South Africa, infection with the virus itself has significantly impacted health, with 28469 deaths attributed by the end of 2020,¹ but the younger aged populations to the north have recorded far lower mortality.² Public health interventions must be tailored to address such variation. This requires realistic metrics for disease burden that take the characteristics of the population, and the individual impacted by disease, into account.³

Metrics encompassing life years lost or disrupted are critical when comparing diseases that are strongly age-associated. A young child dying from pneumonia is clearly expected to lose more potential life years than an 80-year-old dying from the same. Malaria, with its mortality centred in these countries, imparts a disproportionate burden in disability adjusted life years (DALYs) lost, as most mortality occurs before age five.⁴ HIV/AIDS leads to long spells of severe ill-health and premature death primarily among young and middle-aged adults leading to significant life years lost and extended time lost due to less-than-optimal health. In contrast, COVID-19 is characterized by it’s strong association with advanced age, with a mean age of death similar to that for all-cause mortality in many countries.^5,6 However, reporting of the burden of COVID-19 has generally it centred on comparisons of mortality which ignore the influence of age. These mortality-based burden estimates can further mislead through the strong association of death from COVID-19 with pre-existing morbidities, which further reduce expected life years lost from most COVID-19-associated deaths.⁶ Time-based measures such as Disability-Adjusted Life Years (DALYs) that incorporate lost healthy life caused both by premature death and time lived in less-than-optimal health have potential to better represent the full disease impact – as opposed to mortality alone.

Low COVID-19-associated mortality among sub-Saharan African populations is at least partially predicted by their young age structure,^5,7 while lifestyle factors may be protective through a lower prevalence of major co-morbidities,^8-10 higher vitamin D levels and broad antigen exposure leading to non-specific T-cell prior immunity.^11-13 Comparisons of lockdown severity suggest that more restrictive measures have had limited additional impact on reducing COVID-19 mortality.^14-16 However, as with many public health responses, lockdown responses are not without cost. Predicted exacerbations of high burden diseases including malaria, HIV/AIDS and tuberculosis impact particularly children and younger adults.^17-19 Broader impacts of reduced food security and interruption of vaccination will have far-reaching health consequences,^20,21 while loss of family income and reductions in national gross domestic product will impede the capacity to respond.²²

While lockdowns may be easing, proposals for continent-wide mass vaccination under the COVAX mechanism will raise new costs and divert resources, and the urgency of developing good public health policy that appropriately prioritizes relative burdens including COVID-19 management are no less urgent.²³ We therefore compared the disease burdens of COVID-19 and the three pre-existing major infectious disease ‘epidemics’ of sub-Saharan African countries,²⁴ with and without South Africa and Lesotho, to estimate the relative burden of COVID-19 in relation to these other epidemics.

Methods

The health indicators used for the analyses were the number of deaths and Disability-Adjusted Life Years (DALYs) lost by age groups caused by COVID-19 and three major diseases: malaria, HIV/AIDS and tuberculosis. Published data were the main sources of the presented analyses. For the calculation of DALYs lost caused by these diseases at population level, the estimates for number of deaths by age and the number of non-fatal episodes of ill-health by age in 2020 were required. Estimates with and without South Africa and Lesotho were included as these two countries have very different burdens of the four diseases, and differing demographics.

As COVID-19 cases were reported from Africa in February 2020 and widely spread by late March, all recorded COVID-19 deaths were included up until 31 March 2021 to include slightly over 12 months of reporting. According to Africa CDC there had been 77,463 reported COVID-19 deaths in sub-Saharan Africa up to March 2021 and 24,302 COVID-19 deaths if South Africa and Lesotho were excluded.² These total number of reported COVID-19 deaths were allocated across age groups and also used to estimate the number of COVID-19 infections by applying the following method and assumptions. Using data on COVID-19 deaths and seroprevalence surveys from 45 mainly European countries, O’Driscoll et al estimated the infection fatality ratio by age.²⁵ Assuming that these estimated infection fatality ratios also represented the sub-Saharan African situation and further assuming a constant share of infection across age groups, the number of infections and deaths were inferred for a sub-Saharan Africa population size according to United Nations population estimate for 2020 and compatible with a total number of 77,463 (24,302) COVID-19 deaths (see Table S1 in the supplementary material).²⁶ These estimated numbers of deaths and infections by age group were inserted in the standard DALY formula used for calculating the burden of disease and with no discounting of future life years and without the age weighting function.²⁷ For the calculation of life years lost, the standard life expectancies by age from the Global Burden of Disease Study 2019 were used,²⁸ and the reference life table was downloaded from the Institute of Health Metrics and Evaluation website.²⁹ The non-fatal COVID-19 infections were assumed to be mild of a two-week duration and using a disability weight of 0.051 corresponding to the weight attached to a moderate to severe upper respiratory infection.²⁸

The number of deaths and DALYs lost for HIV/AIDS and tuberculosis by age group for 2019 were extracted from the Global Burden of Disease Study 2019 results.²⁹ These estimates were updated to 2020 by assuming a growth from 2019 to 2020 corresponding to the annual population growth rate in sub-Saharan Africa. Population growth rates were estimated by age group using population estimates from 2015 and 2020.²⁶ Cases of combined HIV/AIDS and tuberculosis are considered as HIV/AIDS only and not included in the tuberculosis burden calculations.

Total number of deaths and non-fatal illness episodes caused by malaria in sub-Saharan Africa were obtained from the WHO estimates for 2019,⁴ however, the published numbers were not available by age group. It was assumed that the deaths and illness spells followed the same distribution across age groups as malaria deaths as estimated by the Institute of Health Metrics and Evaluation.²⁹ The number of malaria deaths and illness episodes by age in sub-Saharan Africa (with and without South Africa and Lesotho) were translated into DALYs lost applying the same method as described above including the assumption that a non-fatal malaria infection lasted two weeks and with a disability weight of 0.051.

Predicted exacerbations of the three comparator diseases accrued from 2020 lockdown responses were derived from modelling published by the World Health Organization, StopTB Partnership and The Global Fund.^17-19

Results

Recorded COVID-19 mortality constituted 6.4%, 4.8% and 6.3% of the mortality of tuberculosis, HIV/AIDS and malaria respectively in sub-Saharan Africa north of South Africa and Lesotho. As DALYs lost, the COVID-19 burden amounted to 3.7%, 2.3% and 2.4% of that estimated for the three comparator diseases respectively (Figure 1). Tuberculosis mortality was never dominated by COVID-19 in any age-group, while HIV/AIDS is only dominated above 75 years of age and malaria equaled at 80 years (Figure 2).

• Download figure
• Open in new tab

Figure 1.

Comparison of baseline mortality and disease burdens (DALYs lost) predicted for the 12 months of 2020 for malaria, tuberculosis and HIV/AIDs (pre-lockdown impact), and up to 31 March 2021 for COVID-19, in sub-Saharan Africa. A1 Mortality for sub-Saharan countries north of South Africa and Lesotho. A2: Mortality for all sub-Saharan countries. B1 DALYs lost for sub-Saharan countries north of South Africa and Lesotho. B2: DALYs lost for all sub-Saharan countries.

• Download figure
• Open in new tab

Figure 2.

Mortality predicted for the 12 months of 2020 for malaria, tuberculosis and HIV/AIDs (pre-lockdown impact), and up to 31 March 2021 for COVID-19, in sub-Saharan Africa. A: Mortality for sub-Saharan countries north of South Africa and Lesotho. B: Mortality for all sub-Saharan countries.

Including South Africa and Lesotho, COVID-19 mortality and DALYs lost were 19.2%, 11.7% and 20.2%, and 11.1%, 5.5% and 7.5% of tuberculosis, HIV/AIDS and malaria respectively (Figure 1). All three comparator diseases individually dominated COVID-19 mortality until 65-69 years where COVID-19 dominated HIV/AIDS, and 70-74 years where malaria was also dominated (Figure 2).

Accrued COVID-19 DALYs lost up to 31 March 2021 were also lower than the lower-limit estimates of exacerbation of malaria, tuberculosis and HIV/AIDS (increase over baseline) predicted due to the COVID-19 public health response in sub-Saharan Africa north of South Africa, well below those for tuberculosis and HIV/AIDS for sub-Saharan Africa as a whole. If we assume significant under-reporting of COVID-19, arbitrarily multiplying by a factor of 10 only dominates the lower predicted exacerbations of malaria-burden, and remains below all disease burdens of tuberculosis and HIV/AIDs in sub-Saharan Africa as a whole, and all comparator diseases where South Africa and Lesotho are excluded (Table 1).

Discussion

These comparisons emphasize the relatively low disease burden that COVID-19 is exerting on sub-Saharan African populations, with the pre-existing ‘epidemics’ of malaria, tuberculosis and HIV/AIDS all greatly dominating in life-years impacted, with mortality dominating also in all except the elderly. The use of DALYs lost to assess relative disease burden follows well-accepted practice,^30,31 but is relatively unusual for assessment of COVID-19. It is unclear why this standard public health metric, or the related metric of Quality-Adjusted Life Years (QALYs), has not been widely applied to a disease with such a characteristic age-dependent impact. Appropriate assessment of burden is of extreme importance as decisions are made globally on resources to be allocated to vaccination as well as the imposition of recurrent lockdowns and other cost-bearing responses. Mass vaccination of the sub-Saharan African population against COVID-19, as is advocated in some quarters,²³ will draw essential resources from interventions aimed at health problems with far greater burden through diversion of financial resources and personnel on the ground. Such a wide-scale vaccine intervention has never been attempted before and the implications for already over-stretched health services will be significant. To ensure equity in health care, comprehensive economic evaluation comparing costs and effects of interventions against all four epidemics, including cost-effectiveness analysis, is urgently needed.

This data analysis has a number of limitations. COVID-19 mortality reporting in sub-Saharan Africa is doubtless incomplete, though low mortality is predicted by population age structure and lower prevalence of major co-morbidities including obesity,^5,7,32 while other lifestyle factors and prior immunity may also be protective.^10,33-35 Given lack of strong local data on age-related mortality, we assumed that mortality rates reflect those found elsewhere.²⁵ Lack of transmission appears an unlikely explanation for low recorded mortality as high seroprevalence has been recorded in various sub-Saharan African settings.^36-40 While the higher mortality of COVID-19 in South Africa could be partially explained by higher reporting rates, South Africa also has higher rates of known mortality risk factors.⁴¹ Evidence of very high asymptomatic infection,⁴⁰ and the level of testing taking place (868823 tests for 333 deaths in Uganda alone by 23 February 2021),⁴² suggests that the low recorded mortality in most countries reflects reality, in common with much of Asia.⁴³

The relative burden of COVID-19 in 2020 is also subject to the first cases only being reported in March in most of these populations,⁴³ and so it will not have been in all populations within the first few months of recording. However, as the total mortality rate is declining across the continent at time of writing,⁴³ the bulk of COVID-19 burden may well been accrued in the >12 months total recording time used here.

DALYs lost through COVID-19 morbidity as estimated in this paper do not take post-viral syndromes into account (e.g. ‘long-covid’). These have limited prevalence and may be lighter in younger (less severe illness) African populations, but this is still unclear and will add somewhat to COVID-19 burden. Conversely, the age-based nature of DALYs lost applied to COVID-19 does not take into account the high prevalence of life-shortening comorbidities associated with these cases,⁶ which will in turn lead to an over-estimation of the actual life years lost. Even assuming 90% under-estimation of COVID-19 mortality here, other single disease metrics still dominate COVID-19, as do most estimates of exacerbation of these through the COVID-19 public health response.

In comparing impact of COVID-19 and other health burdens, we considered just three diseases. The impact of broad malnutrition, and of reduced educational attainment (closed schools) and damage to local and national economies will have major long-term impacts on population and societal health.^20-22,44 As a greater proportion of the population achieves post-infection immunity,^{36-38,42,45,46} COVID-19 burden is likely to further reduce, and the cost-effectiveness of response interventions may then decrease further. It is therefore imperative that the cost-effectiveness analyses of further COVID-19 related interventions in these sub-Saharan African populations be tailored to local need, based on realistic metrics that reflect the relative impact of COVID-19, of other disease states, and of the proposed public health responses. Continued emphasis on COVID-19 responses including severe lockdowns or mass vaccination risk exacerbating other health priorities, increasing health inequities rather than reducing them.

Declaration of competing interests

No conflict of interest to declare.

Funding source

None

Availability of data

All data is in the public domain, and calculations are freely available from the authors.

Dissemination statement

Dissemination to patients is not applicable.

Author contributions

DB and KSH jointly conceived the study, contributed to the study design and conceptualization, and sourced the data. KSH performed the economic analyses, both authors participated in the drafting, revising and approval of the manuscript.

Declaration

The lead author (DB) affirms that the manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.

Patient and public involvement

There was no direct patient or public involvement in this study. All data used was in the public domain.

SUPPLEMENTAL FILE