Rapid mortality surveillance using a national population register to monitor excess deaths during SARS-CoV-2 pandemic in South Africa

Abstract

This paper describes how an up-to-date national population register recording deaths by age and sex, whether deaths were due to natural or unnatural causes, and the offices at which the deaths were recorded can be used to monitor excess death during the SARS-CoV-2 pandemic, both nationally, and sub-nationally, in a country with a vital registration system that is neither up to date nor complete. Apart from suggesting an approach for estimating completeness of reporting at a sub-national level, the application produces estimates of the number of deaths in excess of those expected in the absence of the SARS-CoV-2 epidemic that are highly correlated with the confirmed number of COVID-19 deaths over time, but at a level 2.5 to 3 times higher than the official numbers of COVID-19 deaths. Apportioning the observed excess deaths more precisely to COVID, COVID-related and collateral deaths, and non-COVID deaths averted by interventions with reduced mobility and gatherings, etc., requires access to real-time cause-of-death information. It is suggested that the transition from ICD-10 to ICD-11 should be used as an opportunity to change from a paper-based system to electronic capture of the medical cause-of-death information.

Population perspectives and demographic methods to strengthen CRVS systems: introduction

“…Lima et al ( 2021 ) evaluate monthly death registration data in six Latin American countries during the COVID-19 pandemic to understand impact on life expectancy at birth. Similarly to Dorrington et al ( 2021 ), they contend with notable variability in underregistration of deaths, delayed registration issues, and variations in the place of occurrence and place of registration. They use P-scores to measure excess mortality in subnational areas of Brazil, Chile, Ecuador, Guatemala, Mexico, and Peru and examine its relationship with life-table entropy—i.e., the association between the relative challenges in life expectancy with changes in age-specific mortality rates.…”

“…The final two papers in this Genus thematic series, Dorrington et al ( 2021 ) and Lima et al ( 2021 ), document technical challenges and practical solutions to analyzing excess mortality during the COVID-19 pandemic in countries where death registration is incomplete—namely Brazil, Chile, Ecuador, Guatemala, Mexico, Peru, and South Africa. Such analyses have informed public understanding of the COVID-19 pandemic and shaped policy responses (Agrawal et al, 2021 ; Clarke et al, 2021 ; Goldstein & Lee, 2020 ; Helleringer & Queiroz, 2021 ; Polyakova et al, 2020 ; Viglione, 2020 ).…”

“…Dorrington et al ( 2021 ), noting the challenges of incomplete and delayed reporting in the national population register (NPR) and death register of South Africa, demonstrate how weekly excess mortality monitoring is possible when the NPR is systematically updated via the CRVS system. They make use of 10 years of recent data on registered deaths, and employ death distribution methods and comparative validation analysis to identify and adjust for underregistration of both adult and child deaths.…”

The impact of the SARS-CoV-2 epidemic on mortality in South Africa in 2020

“…During 2020, the RMS System was modified to track the weekly number of deaths and thereby provide critical insight into the impact of COVID-19 in near to real time. [7,8] The two key features of the impact of the SARS-CoV-2 epidemic on life expectancy in South Africa (SA) in 2020 are the seemingly modest impact relative to countries with similar levels of infection (particularly given the high level of comorbidities in SA [9] ) and the fact that the impact on life expectancy of females was greater than that on life expectancy of males. For example, only seven countries (all with lower prevalences of SARS-CoV-2) out of 29 considered by Aburto et al [1] experienced a lower change in life expectancy between 2019 and 2020, and only three showed a lower drop in male than female life expectancy; all of these were smaller than that found in SA.…”

“…Mortality is determined not only by the extent of the epidemic and the level of comorbidities in a country, but also by efforts made to intervene and manage the epidemic. In the case of SA, the very severe lockdown early in the epidemic, with a higher level of popular support and enforcement than the continued lockdown, based on less convincing reasoning, resulted in a near shut-down of the economy, the closure of schools, a sharp decline in unnatural mortality [7,13] and a reduction in non-COVID-19 communicable mortality, particularly in infants and young children.…”