Chickenpox is easing in school-age children worldwide, but a 2030 model keeps the gaps

Varicella — chickenpox — remains one of the most common infections of childhood, and a new modelling study offers a global ledger of where its burden sits among 5-to-9-year-olds and where it is heading. Published in Human Vaccines & Immunotherapeutics on 2 September 2026, the analysis drew on the Global Burden of Disease (GBD) 2021 study together with country-level varicella vaccination policy information, and trained an Extreme Gradient Boosting (XGBoost) machine-learning model to project the burden to 2030 [s1].

The headline trend is reassuring on one measure and flat on two others. From 1990 to 2021, the global incidence rate in this age group edged up from 1,028.46 to 1,048.24 per 100,000, and prevalence rose marginally from 30.81 to 31.07 per 100,000 [s1]. Over the same span, the disability-adjusted life year (DALY) rate — the measure that captures healthy time lost to the disease — fell substantially, from 21.09 to 12.36 per 100,000 [s1]. In other words, roughly as many children were catching chickenpox, but each case was, on average, costing less in health terms.

What the model projects

By 2030, the model projects all three metrics to fall: incidence to 897.01 per 100,000, prevalence to 28.96 per 100,000, and DALYs to 11.71 per 100,000 [s1]. A projected drop in incidence of roughly 150 per 100,000 from the 2021 figure is the most striking of the three, and it would mark a genuine break from the essentially flat incidence of the preceding three decades.

That projection deserves the caution the authors themselves attach to it. It is the output of a machine-learning model fitted to historical GBD estimates, not a measurement, and its reliability depends on the patterns of the past continuing to hold. The paper is candid that "the lack of vaccination coverage and cohort-level immunization data limits interpretation of vaccine effects," and calls for future work to incorporate both coverage and cohort exposure information [s1]. A forecast built without direct coverage data cannot cleanly credit any decline to vaccination.

The disparities the model expects to persist

The more durable finding is geographic. Across the full 1990-to-2030 window, Southeast and South Asia had the highest varicella incidence, Asian and Pacific Island nations had the highest prevalence, and sub-Saharan Africa and South Asia carried the highest DALYs [s1]. The projected global decline, in other words, is not expected to even out the map: the model indicates that "regional disparities will persist" even as the overall burden falls [s1].

To identify what drives the burden, the authors used SHAP (Shapley Additive exPlanations), a method that apportions a model's predictions across its input variables. Population was the strongest predictor, with a mean absolute SHAP value of 0.167 for incidence and 0.103 for prevalence, and a higher vaccination-programme tier "may be associated with lower DALYs" [s1]. That last association is stated tentatively for good reason — it is a correlation surfaced by the model, hedged by the same missing coverage data the authors flag, not a demonstrated causal effect of vaccination policy.

Why the DALY story matters most

For a disease that is rarely fatal in healthy children, the DALY measure is the one that best captures what varicella actually costs: days of illness, complications such as skin infection or, less often, neurological disease, and the knock-on burden on families and health services. A falling DALY rate against roughly stable incidence suggests that outcomes, not exposure, have improved over three decades — through some mix of better care, changing case mix, and, in countries that vaccinate, averted severe disease.

Chickenpox sits at the softer end of the vaccine-preventable spectrum, and many national immunisation programmes have prioritised other pathogens ahead of it; the reduced-burden trend and the modelled decline are consistent with that patchwork rather than with a coordinated global push. The value of a study like this is less its precise 2030 numbers than its map of where the load concentrates — information that sits within the wider debate over how far to extend childhood immunisation when budgets are finite.

What to watch

The figures that would test this model are straightforward to name and harder to collect: direct national varicella vaccination coverage, linked to cohort-level exposure, in the high-burden settings of South and Southeast Asia and sub-Saharan Africa. Until that evidence exists, the projected decline to 2030 is best read as a plausible trajectory with an explicit blind spot at its centre — the very data on vaccination that would explain why the numbers move.

This article is informational and is not medical advice.

Sources

• Global burden of varicella in children aged 5-9 y, relationship with vaccination policy indicators, and projected trends to 2030 — Human Vaccines & Immunotherapeutics (Zeng and colleagues), 2 September 2026