Model projects dropping the hepatitis B birth dose would raise childhood infections

The United States has given hepatitis B vaccine at birth, to every newborn, since the early 1990s. In December 2025 federal advisers voted to step back from that universal policy [s2]. A modelling study published in Annals of Internal Medicine on October 20 asks what would happen to the number of children living with chronic hepatitis B if the proposed changes were carried through [s1]. It is a projection, not a measurement — but it attaches numbers to a debate that has mostly been conducted in principle.

What the model does

The researchers built a Markov cohort state-transition model of perinatal and childhood hepatitis B virus infection [s1]. The model follows a single one-year birth cohort of U.S.-born infants — 3.6 million of them — over a 20-year horizon [s1]. It is fed with published data: maternal hepatitis B prevalence of 0.4%, a hepatitis B surface antigen test sensitivity of 99.5%, and age-stratified risks of an infection becoming chronic [s1].

The key lever is vaccine completion under three strategies [s1]. A universal birth-dose strategy assumes 91.1% of children complete vaccination [s1]. A deferred strategy, giving the first dose at two months rather than at birth, assumes 87.1% [s1]. A shared-decision-making strategy, under which the birth dose is offered rather than recommended, assumes just 50.0% completion [s1]. Those completion figures do much of the work in what follows.

The projected numbers

Under universal birth-dose vaccination, the model projects 255 childhood chronic hepatitis B infections in the 2026 birth cohort, with a 95% uncertainty interval running from 130 to 530 [s1].

Moving to the deferred strategy raises that by 19%, to 303 infections (95% uncertainty interval, 155 to 610) [s1]. Moving to shared decision-making raises it by 83%, to 467 infections (209 to 1,032) [s1]. The shared-decision-making figure is the one driven hardest by the assumed halving of completion.

The model also tracks how many children remain susceptible to hepatitis B at the 20-year mark — a measure of how much of the cohort is left unprotected into adolescence and early adulthood. Against 232,000 susceptible under the universal strategy, the deferred strategy leaves 474,000 and the shared-decision-making strategy 1,806,000 [s1].

Why the birth dose matters

The logic the model encodes rests on timing. Hepatitis B transmitted from mother to baby around birth, or in early childhood, is far more likely to become a lifelong chronic infection than infection acquired as an adult. The birth dose exists to close that early window, including for infants whose mothers' infection status is unknown or missed despite screening. Deferring the first dose to two months, or making it optional, reopens part of that window — which is why even a modest drop in coverage shows up as more chronic infections in the model. The gap between a 4-percentage-point fall in completion under deferral and a 41-point fall under shared decision-making is what separates the 19% and 83% projected increases [s1].

How much weight to give it

This is a model, and its outputs are only as good as its inputs. The headline differences between strategies are driven largely by the assumed completion rates — 91.1% versus 87.1% versus 50.0% [s1] — and the real-world figures under a changed policy are not yet known. The 50% assumption for shared decision-making is a projection of how many families would opt in, not an observed number. The model also draws on a fixed maternal prevalence and test sensitivity [s1]; the authors report sensitivity analyses, including a lower-coverage scenario, to probe how far the conclusions hold.

What a model of this kind can and cannot do is worth stating plainly. It cannot tell you that a specific child will or will not be infected. It can show the direction and rough magnitude of a policy change under stated assumptions, and here the direction is consistent across strategies: less birth-dose coverage, more projected chronic infections and more susceptible children. The absolute numbers — a few hundred infections per cohort — are small against 3.6 million births, but chronic hepatitis B is a lifelong condition that can lead to cirrhosis and liver cancer, which is the reason the stakes are framed the way they are.

What to watch

Whether actual vaccine-completion rates under any revised schedule match the model's assumptions; whether the projected rise in susceptible adolescents prompts catch-up vaccination planning; and whether other groups reach similar numbers with independent models.

This article describes a single modelling study and is informational only. It is not medical advice, and nothing here should be used to make decisions about any individual child's vaccination.