HIV Testing Did Decline in High-Burden Settings During COVID-19 — But 'Nearly 25 Percent' Understates How Bad It Was

The claim holds that HIV testing declined by nearly 25 percent in high-burden settings, implying a serious but bounded disruption. The verdict is partially false: a decline of roughly that magnitude did occur in some aggregated data, but the actual figures in the hardest-hit countries were substantially worse, making '25%' a misleading summary of the crisis.

The strongest evidence against treating 25% as the headline number comes from the WHO's own survey of 106 countries, published in 2020. Among the most affected countries, the median reported decline in HIV tests during April through June 2020 — compared to the same period in 2019 — was approximately 41%. That is not an outlier; it is the median. PEPFAR's 2021 Annual Report to Congress, which covers the largest coordinated HIV program in the world, recorded a 22% drop across its supported programs in fiscal year 2020 versus fiscal year 2019. That 22% figure is the closest anchor for the '25%' claim, and it is a program-wide average that masks steep country-level variation.

The steelman version of the claim is real: PEPFAR's aggregate 22% decline is close to 'nearly 25%,' and UNAIDS's 2021 Global AIDS Update ('Confronting Inequalities') confirmed that testing services were severely disrupted in 2020. If someone cited PEPFAR's aggregate or a specific sub-period in a specific region, they could arrive at a figure near 25% without fabricating anything. The claim is not invented.

Where it breaks down is in the word 'high-burden.' The countries that carry the greatest HIV burden — concentrated in sub-Saharan Africa — are precisely the ones where declines were steepest, not closest to 25%. Dorward et al., writing in BMJ Global Health in 2021, documented a 33% decline in HIV testing in South Africa in April 2020 versus April 2019. South Africa has one of the largest HIV-positive populations on earth. Jewell et al. in The Lancet HIV modelled a 50% reduction scenario for sub-Saharan Africa as a realistic stress-test, not an extreme one. Using a 22% aggregate to characterize 'high-burden settings' applies a diluted average to the places where the damage was worst.

It is fair to concede that declines varied widely — some countries recovered testing volumes faster, and the 40–50% drops were concentrated in the April–June 2020 lockdown peak rather than the full fiscal year. The PEPFAR annual figure of 22% reflects a full-year average, which naturally smooths the worst months. Both facts are true simultaneously: the annual average was near 25%, and the peak disruption in the highest-burden countries was far above it.

The manipulation pattern here is the misleading denominator. Averaging across all PEPFAR-supported countries, across all twelve months of fiscal year 2020, produces a number that sounds alarming enough to generate concern but conservative enough to avoid conveying the actual scale of collapse in the worst-affected places. When a single headline statistic is applied to a category as diverse as 'high-burden settings,' check whether it is an annual average or a peak-period figure, and whether it covers the highest-burden countries specifically or a broader program aggregate. A 22% annual average and a 41% median peak decline are both accurate — they just describe different slices of the same crisis.