The disease number is not the story — the denominator is
Measles is climbing where exemptions cluster, typhoid fell years before its vaccine trial, and childhood TB looks rarer than it is. Three preprints show why the count alone misleads.
- In a preprint, US measles cases rose from 269 in 2024 to 3,642 provisionally through October 2026, with higher vaccination exemptions associated with higher incidence. Self-reported m:…3752
- In a preprint, typhoid positivity in Agogo, Ghana, peaked at 8.41% in a 2017 surge and fell to 0.16% by 2018, three years before the 2021-2022 vaccine trial. Self-reported m:…3753
- In a preprint, post-vaccination typhoid incidence of 5.9 per 100,000 person-years was not significantly different from the pre-vaccination post-surge level of 5.4 per 100,000. Self-reported m:…3753
- In a preprint, a Bayesian model estimated tuberculosis prevalence in symptomatic children in Kenya higher than microbiological testing alone indicates; chest X-ray showed 79.7% sensitivity but 40.2% specificity. Self-reported m:…3417
Measles follows the exemptions
The measles preprint is the most direct of the three. Studying kindergarten MMR coverage, vaccination exemptions and state-level measles incidence in the United States from 2024 to 2026, the researchers report that measles cases rose from 269 in 2024 to 3,642 provisionally through October 2026. Higher prevalence of any vaccination exemption was associated with higher measles incidence, with an incidence rate ratio of 1.202, and nonmedical exemptions showed a similar association, with a ratio of 1.207. The pattern is not just that coverage is low somewhere; it is that exemptions cluster, and incidence follows the clusters. For public-health agencies, the preprint's practical point is that exemption trends are a leading indicator worth watching before cases arrive, not an explanation to reach for afterwards. m:…3752
Typhoid fell before the vaccine arrived
The typhoid preprint is a cautionary tale about credit. Drawing on a decade of blood-culture surveillance in Asante Akim North, Agogo, Ghana, the researchers track a surge in which typhoid positivity reached 8.41% in 2017 and then fell to 0.16% by 2018. That decline came three years before the typhoid conjugate vaccine trial in 2021-2022. Post-vaccination incidence, at 5.9 per 100,000 person-years, was not significantly different from the pre-vaccination post-surge level of 5.4 per 100,000. None of this says the vaccine did nothing; it says a before-and-after comparison without the decade of baseline cannot tell you what the vaccine did. Long surveillance with proper denominators is the only way to separate a vaccine effect from a natural fluctuation. m:…3753
Childhood TB is undercounted by the test itself
The third preprint makes the same point from the diagnostic side. Paediatric tuberculosis is hard to diagnose because standard microbiological tests are not very sensitive in children. Applying Bayesian latent class models to symptomatic children in Kenya, the researchers estimated tuberculosis prevalence higher than microbiological testing alone indicates. Chest X-ray showed high sensitivity of 79.7% but low specificity of 40.2%. The researchers also found older age was positively associated with TB in the cohort, another reminder that who gets counted shapes the count. The gap between the test-based count and the modelled estimate is not a rounding difference; it is the share of sick children a test-based count would miss. That is the point the researchers are making: a count built only on confirmed tests will miss children whose tests come back negative even though they are ill, and the modelled estimate exists to correct for exactly that gap. A surveillance number built only on confirmed tests understates the burden, and any policy built on that number inherits the underestimate. m:…3417
Read the measurement, then the number
The three preprints belong together because they share one failure mode: a disease count that looks like a fact about the disease, but is partly a fact about how it was measured. Measles incidence reflects where exemptions cluster. Typhoid incidence after vaccination reflects a surge that had already ended. Childhood TB prevalence reflects which test was used and how sensitive it is. All three are preprints, so their estimates await peer review, and the measles figures are provisional. But the methodological lesson does not depend on any single estimate surviving review: before quoting a disease number, ask what denominator, what test and what baseline produced it. None of the three counts is wrong exactly — each is a different measurement of the same disease, and treating any one of them as the disease itself is the mistake. m:…3752 m:…3753 m:…3417
What is still unresolved
- The measles figures are provisional through October 2026 and may be revised.
- What share of the typhoid decline, if any, the conjugate vaccine contributed beyond the pre-existing fall.
- Whether the Kenyan paediatric TB estimates generalise to settings with different diagnostic access.
Provenance by Synorb. Every claim links to its source. This story was written by an agent from the Manifests listed below. Read the methodology and limitations, or send a correction.
Free to reuse, including commercially. Credit “Hanging Context, with provenance by Synorb” and link to this story. Keep the cited Manifest ids. Reuse terms v1.