Over-reliance on modelled estimates
The reliance on extrapolated estimates from the B3 model rather than empirically supported, smoothed data is due to the limited availability of recent data at the period’s endpoints. Only 5% of the under-five mortality estimate for 2022 is derived from countries with actual data for that year [5]. When considering countries with data from either 2021 or 2022, the percentage increases to 8%. Given this limited empirical basis, independent validation using alternative data sources—such as real-time health surveillance or administrative records—would be essential for assessing the robustness of these estimates.
As shown in Table 1, the lack of data from specific regions and countries compounds the sparse data problem. Data from Sub-Saharan Africa (SSA), for example, is a critical region because of its disproportionate child mortality, accounting for 58.3% of the total estimated child deaths in 2022. A little more than half of all estimated SSA child deaths in 2022 occurred in just four countries (10.6% from the Democratic Republic of Congo; 6.2% from Ethiopia; 4.6% from Niger; and 29.3% from Nigeria). The most recent empirical data for the Democratic Republic of Congo was from 2010; for Ethiopia, it was from 2016; and for Nigeria, the country with the most significant contribution to child mortality in the world, it was from 2015. Only Niger provided relatively contemporaneous estimates to its mortality estimate—2020.
Table 1 Analysis of data* availability by country by year (2018–2022) for the top mortality countries in sub-Saharan Africa (50.5% of deaths) and South Asia (84.9% of deaths)Temporal discontinuity
The underlying assumption of any temporal extrapolation is that the model developed on data from one period of time will hold true for future periods. The failure of this crucial assumption underpins analytic techniques that rely on temporal discontinuities to demonstrate exogenous causal effects. the UN IGME did not adjust their estimates post-2020 to account for COVID-19, a major temporal discontinuity, despite the pandemic impacting mortality differently across age groups and countries. Globally, the mortality from COVID was heavily skewed towards older people, and mortality in children under 5 was relatively rare [6]. Nonetheless, there are reasons to anticipate that the pandemic may have indirectly affected child survival due to curtailed health services during shutdowns [7]. Particularly in sub-Saharan Africa, only nine countries provided high-quality child mortality data during 2020–2022, representing just 17% of the region’s child deaths. Therefore, the majority of estimates rely on extrapolation without considering potential pandemic-related changes. Although the UN IGME analyzed other sources and reported no widespread excess child mortality, they acknowledged data limitations and urged caution. Nonetheless, their assumption of mortality stability despite data gaps introduces a significant limitation, as it overlooks possible indirect effects of the pandemic on child survival.
Inconsistency with related data
Data related to factors that are known to affect child mortality have not been considered in the updated UN IGME estimates. It is well-established that childhood vaccinations prevent deaths, and conversely, the failure to vaccinate endangers child lives [8]. First, an unvaccinated child is at greater risk of death due to the vaccine-targeted pathogen directly and indirectly to other pathogens through heterologous effects associated with some live vaccines. Second, any cohort immunity effect is jeopardized when vaccine coverage decreases, placing all children at greater risk of death due to the vaccine-targeted pathogen. Finally, health services for treating unvaccinated children who become ill expend human, financial, and commodity resources unnecessarily, which could contribute to saving other lives. Furthermore, any impact of systemic vaccination failures is likely to be lagged. A child does not die when they miss a vaccination session. Rather, they could die sometime later when or if they contract the otherwise preventable disease, which could be a year or more later. When the pandemic started in 2020, routine immunization services were halted or curtailed in many countries. EPI provides the basic regimen of childhood vaccination to prevent diphtheria, tetanus, pertussis, pneumococcal pneumonia, rotavirus, and measles, among others. Despite evidence of declining vaccination rates in 2020 and 2021, the models do not incorporate these lag effects, nor do they reconcile discrepancies between reported vaccination coverage and actual delivery failures. Additionally, other risk factors like undernutrition [9] and limited access to antenatal care [10], affected by COVID-19, remain unconsidered in the updated estimates, risking an overly optimistic portrayal of child mortality trends.