Measles Deaths Counted Two Ways as NCHS Takes the Lead
Two health statisticians warn that counting measles deaths from death certificates rather than state case investigations produces different numbers — and that the gap will confuse the public.
In a STAT News opinion piece published Sept. 25, 2026, Denys T. Lau and Jennifer D. Schoendorf argue that Health and Human Services Secretary Robert F. Kennedy Jr.'s reliance on the National Center for Health Statistics to tally measles deaths conflicts with how state public health officials count them, writing that the discrepancy "will only increase public confusion."
There is more than one way to count a death, and during an outbreak the choice of method is not a technical footnote. It determines the number that appears on television, in a governor's briefing and in a federal press release — and it determines whether those three numbers agree.
That is the argument made by Denys T. Lau and Jennifer D. Schoendorf in an opinion piece published by STAT News. Writing about Health and Human Services Secretary Robert F. Kennedy Jr.'s use of the National Center for Health Statistics in reporting measles deaths, the authors put the problem plainly: "A discrepancy between how federal and state public health officials determine measles deaths will only increase public confusion."
Two counting systems that were never built to match
The National Center for Health Statistics, the federal government's official statistical agency for births and deaths, derives its mortality figures from death certificates. A physician, medical examiner or coroner records a chain of causes on the certificate; the certificate is registered with the state; the state transmits the record to the federal system; and the causes are coded into standard classification categories before the death is counted in a national total.
State and local public health departments work from a different starting point. Their measles counts come out of outbreak surveillance: a case is identified, often through laboratory confirmation, investigated by disease detectives who trace exposures and contacts, and then tracked through its outcome. If a person in that confirmed case set dies, the health department can attribute the death to the outbreak on the basis of the investigation, without waiting for a certificate to be signed, coded and transmitted.
Neither approach is wrong. They were designed for different jobs. Death certificate statistics exist to produce consistent, comparable, nationally standardized mortality data over long stretches of time. Outbreak surveillance exists to tell public health officials what is happening this week, in this county, in time to do something about it. Ask them the same question during an active outbreak and they will frequently give different answers.
Why the federal number tends to be the smaller one
Several mechanical features of vital statistics push the certificate-based count downward or later, and they compound during a fast-moving outbreak.
- Lag. Certification, registration, transmission and cause-of-death coding all take time. A death that a state health department can attribute within days may not surface in national mortality data for considerably longer.
- Pending cases. Where a medical examiner has not completed an investigation, the certificate may be filed as pending and amended later, which delays the death's appearance in the coded count.
- How the cause is written. Measles kills mostly through complications — pneumonia and swelling of the brain among them. If a certifier records the complication and does not record measles as the underlying condition in the causal chain, the coded record will not land in the measles category, even though the outbreak investigation would classify it as a measles death.
- Scope. Vital statistics count deaths as certified. Outbreak surveillance counts deaths among people already confirmed as part of the outbreak. The two sets overlap heavily but not perfectly.
The result is a predictable pattern: the state figure moves first and the federal figure catches up, sometimes converging and sometimes not. To a statistician, that is a known property of the data. To a member of the public watching two arms of government publish two totals for the same event, it looks like someone is wrong — or worse, like someone is hiding something.
The credibility cost of a contested number
Measles is the sharpest possible test case for this problem, because it is a disease whose public response turns almost entirely on trust. Measles is vaccine-preventable. Every argument about whether an outbreak warrants an urgent vaccination push rests on a shared understanding of how serious the outbreak is, and severity is communicated to the public overwhelmingly through a death count.
Measles is the sharpest possible test case for this problem, because it is a disease whose public response turns almost entirely on trust.
When the federal government's stated total is lower than the states' totals, the gap does not read as a methodological artifact to most people. It reads as a dispute. Parents deciding whether to get a child vaccinated, school administrators weighing exclusion policies, and local officials deciding whether to stand up a clinic are all working from the same handful of headline numbers. A visible split between the federal count and the state count gives every side of an already polarized vaccine debate a number to cite and a reason to distrust the other one.
That is the substance of Lau and Schoendorf's warning. Their objection, as stated, is not that death certificate data are unreliable — the National Center for Health Statistics has long been the reference source for US mortality — but that using a slower, differently constructed data stream as the authoritative real-time tally of an active outbreak sets the federal number against the state numbers for no epidemiological gain.
What to watch as the outbreak reporting continues
Several things will indicate whether the discrepancy narrows or hardens. The first is whether federal communications publish the counting method alongside the count — stating explicitly that the figure reflects coded death certificates as of a given date, and that state health departments may report a different total from case investigations. Labeling the method does not eliminate the gap, but it changes the gap from an apparent contradiction into an explained one.
The second is whether state health departments continue to publish their own outbreak-attributed death counts independently, or whether they defer to the federal figure. Independent state reporting preserves the faster signal that outbreak response depends on; deference would remove the discrepancy at the cost of timeliness.
The third is provisional versus final data. Mortality statistics are routinely published first as provisional and later revised as pending certificates are completed and amended records flow through. If provisional federal figures are presented without that caveat, later upward revisions will themselves become a story — and each revision will reopen the question of which number was ever right.
The underlying tension is not new to the pandemic era: the public wants a single authoritative number, and the statistical system produces several, each correct for its own purpose. What is different here is that the divergence sits inside a vaccine debate where every number is contested on arrival. Lau and Schoendorf's point is that choosing the slower instrument to answer a fast question does not settle that debate. It supplies it with fresh ammunition.
Key facts
- Publication: Opinion piece in STAT News, Sept. 25, 2026
- Authors: Denys T. Lau and Jennifer D. Schoendorf
- Federal method: National Center for Health Statistics counts from coded death certificates
- State method: Health departments attribute deaths through outbreak case investigation
Frequently asked questions
What is the National Center for Health Statistics?
The National Center for Health Statistics, or NCHS, is the US federal government's principal statistical agency for health data, including official births and deaths. Its mortality figures are built from death certificates filed by physicians, medical examiners and coroners, registered by states, then transmitted federally and coded into standard cause-of-death categories before publication.
Why would federal and state measles death counts differ?
They use different source data. States attribute deaths through outbreak case investigations of laboratory-confirmed patients, which produces a figure within days. Federal mortality statistics come from death certificates that must be certified, registered, transmitted and coded. Lag, pending medical examiner cases and how a certifier writes the cause of death can all produce divergent totals.
Who wrote the STAT News opinion piece?
Denys T. Lau and Jennifer D. Schoendorf. They argue that Health and Human Services Secretary Robert F. Kennedy Jr.'s use of the National Center for Health Statistics to report measles deaths conflicts with state public health methods, writing that the discrepancy between the two approaches "will only increase public confusion."
Does measles usually appear as the cause on a death certificate?
Not always directly. Measles typically kills through complications such as pneumonia or brain inflammation. If the certifier records only the complication and does not place measles in the causal chain, the coded record will not be classified as a measles death, even though an outbreak investigation would attribute it to the outbreak.
Is death certificate data unreliable?
No. Death certificate data are the long-standing reference source for US mortality and are designed for consistency and comparability over time. The criticism in the STAT News piece is about fit for purpose: certificate-based statistics are slower than outbreak surveillance, which makes them a poor instrument for reporting a real-time death toll during an active outbreak.
Why does the counting method matter to the public?
Because severity is communicated almost entirely through a death count, and measles response depends on public trust. When federal and state totals differ without explanation, the gap looks like a dispute rather than a methodological artifact, giving competing sides in the vaccine debate conflicting official figures to cite against each other.
