Could Deficiencies in South African Data Be the Explanation for Its Early SARS-CoV-2 Peak?

TL;DR

This paper investigates whether data deficiencies in South Africa could explain its early SARS-CoV-2 peak, considering errors in infection estimates and pre-existing immunity, and finds these factors could significantly influence observed epidemic dynamics.

Contribution

It introduces a model analyzing how underreported infections and hidden immunity could account for early peaks, highlighting the importance of data accuracy in epidemic analysis.

Findings

Data deficiencies can produce perceived early peaks under certain conditions.

Significant underestimation of infections or immunity levels could alter epidemic threshold predictions.

Early peaks are associated with substantial changes in the basic reproduction number, $r_0$.

Abstract

The SARS-CoV-2 pandemic peaked very early in comparison to the thresholds predicted by an analysis of prior lockdown regimes. The most convenient explanation is that some, external factor changed the value of the basic reproduction number, $r_{\rm 0}$; and there certainly are arguments for this. Other factors could, nonetheless, have played a role. This research attempts to reconcile the observed peak with the thresholds predicted by lockdown regimes similar to the one in force at the time. It contemplates the effect of two, different, hypothetical errors in the data: The first is that the true level of infection has been underestimated by a multiplicative factor, while the second is that of an imperceptible, pre-existing, immune fraction of the population. While it is shown that it certainly is possible to manufacture the perception of an early peak as extreme as the one observed, solely by way of these two phenomena, the values need to be fairly high. The phenomena would not, by any measure, be insignificant. It also remains an inescapable fact that the early peak in infections coincided with a fairly profound change in $r_{\rm 0}$; in all the contemplated scenarios of data-deficiency.