If you’ve been following the dispute between Pennsylvania Gov. Josh Shapiro and Health and Human Services Secretary Robert F. Kennedy Jr., you probably know that it centers on the deaths of two infants in Lancaster County in late August 2026.
And you may be aware that despite the county coroner’s confirmation that one infant died due to measles, while the other’s death was “measles-associated,” the measles dashboard on the Centers for Disease Control and Prevention website counts neither death as of Sept. 10.
It’s easy to dismiss the squabble as political posturing between public officials. But as an epidemiologist who studies vaccine-preventable illnesses and health communication, I worry that for people seeking accurate information about how to keep their families safe – in this case, families in Lancaster County – there’s an enormous cost when health authorities no longer speak in one voice.
In the confusion that such rifts create, the public has no way of knowing which disagreements are about evidence and which are driven by ideology and narrative control.
Deaths in dispute
The U.S. government issues a certificate for every death, confirming its underlying cause following a physician’s examination. The death certificate poses a narrow question: What single condition set in motion the chain of events that ended in this person’s death? This is recorded as the cause of death.
Because measles is what the CDC calls a notifiable condition, or a reportable disease, measles deaths get added to a national count. Contributing conditions go on a separate line in the document and are not counted.
Nobody disputes that both infants had measles when they died. The argument is which line on the death certificate “measles” belongs on. The Pennsylvania Department of Health has called both deaths measles-associated. This term is used when there’s evidence the person had measles, but there’s no settled official cause of death yet. It’s a holding category, not a verdict.
Irina Starikova/iStock via Getty Images Plus
The job of determining the official cause – that is, certifying a death – falls to the county coroner. Lancaster County Coroner Stephen Diamantoni concluded that for one of the infants, a newborn boy who died in mid-August, the cause was a ruptured spleen unrelated to the measles infection. For a 6-week-old girl who died Aug. 18, he found that measles was the cause.
These are not competing conclusions. They come from two different systems asking two different questions. County coroners’ decisions on cause of death are legal finding, made one case at a time, that feed national vital statistics.
State health departments, on the other hand, run disease surveillance. They track who got sick and who died during an outbreak so they can contain it. Surveillance moves in days because it has to; death certification moves in weeks because it should. The two rarely line up in real time.
A moving target
While the apparent divergence between county and state data was basically a technicality, the exchange between Shapiro and Kennedy was more of a brawl.
Shapiro announced the infants’ deaths on Aug. 25 and blamed anti-vaccine rhetoric for falling vaccination rates and increasing cases. The CDC posted condolences on the social media platform X, using Pennsylvania’s language calling the deaths measles-associated.
But in response, Kennedy suggested the deaths might have been fabricated by a member of the governor’s staff. Two days after Shapiro’s announcement, the CDC walked back its post, expressing appreciation for Kennedy’s efforts to clarify conflicting information from Pennsylvania officials about whether the two reported deaths were caused by measles.
Then, in an unusual move, on Aug. 30, the CDC changed the death count on its measles dashboard from 2 to an asterisk, with a footnote stating that the National Center for Health Statistics currently has no death records from 2026 that indicate measles is the underlying cause.
No new evidence arrived in those two days that would explain the CDC’s new position.
HHS now says it is working with a national network of disease tracking experts on a standardized case definition for measles deaths. Standard case definitions exist for measles infections but not for measles deaths, for the simple reason that until recently, dying from measles in the U.S. was an extremely rare occurrence: Between 2001 and 2022, the U.S. recorded roughly 4,000 cases and three deaths.
But writing a definition in the middle of a count – and applying it to a single state’s deaths – is a departure from protocol. Previously, during outbreaks in Texas and New Mexico in 2025, the CDC did count three measles-associated deaths while they were still being investigated. Declining to do so now is inconsistent and confusing, and breeds mistrust.
Unnamed officials told news outlets that CDC Director Erica Schwartz removed the deaths at Kennedy’s request, after CDC staff had accepted the state’s data. HHS disputes this, and Kennedy said in a post on X that the coroner’s finding – that the newborn boy’s death was caused by a ruptured spleen – vindicated federal caution.
A parent sifting through all this information has no way to gauge who is right. The usual shortcut – check the CDC numbers on the measles dashboard – now yields an asterisk leading to a footnote that contradicts statements from state officials, a county coroner and the agency’s own condolence post from 10 days earlier. There’s no neutral tiebreaker left, which means the decision falls back on which institution you already trusted most.
A broader split in public health guidance
While that argument played out, a second rift was widening on a different track.
On Sept. 2, four physician groups – the American Academy of Pediatrics, the American Academy of Family Physicians, the American College of Obstetricians and Gynecologists, and the Infectious Diseases Society of America – issued their own vaccine recommendations for the fall respiratory season.
The pediatricians association had already split from the CDC’s childhood immunization schedule in 2025, ending a 30-year practice of publishing one agreed-upon set of recommendations for families. What’s new is that three more societies have now joined this public health separatist movement, and that the split has extended to the shots most adults get.
On influenza, the physician groups’ advice matches the CDC’s: Everyone 6 months and older should get a flu vaccine. On COVID-19 they go further, recommending vaccination for children 6 to 23 months, whereas the CDC tells parents to consult a doctor.
For decades, the CDC has issued fall vaccine recommendations each year, offering a unified slate of federal advice for everyone across the U.S. to follow. Now, with the federal advisory committee that makes those recommendations sidelined and embroiled in a legal battle, the agency has issued no new federal vaccine guidance for this season. The four physician groups stepped forward to fill the gap.
Yet technically, the CDC’s previous seasonal recommendations – and its revised childhood immunization schedule – still stand, creating a confusing doubling of critical health guidance.
The cost of fracture
This confusion has a measurable cost.
Measles is among the most contagious diseases known. Roughly 1 to 3 in 1,000 children who catch it die, and in 2025, 11% of cases required a hospital stay. A 2025 study put the average cost at roughly US$43,000 per case.
The national case count reached 3,134 on Sept. 4, the highest since 1991. As of the same date, Pennsylvania had recorded 624 cases – 91 in the previous week. More than 40% of Pennsylvania’s cases have been in Lancaster County.
Parents must now reconcile a recently revised federal immunization schedule, a set of physician-society recommendations that departs from it, a state count of two deaths and a national count of none.
Each of those positions has an institution behind it. None of them explains how it relates to the others – and that is what’s most damaging. Public trust doesn’t only collapse when officials lie. It erodes when people are handed different information that can’t be squared, or misleading information, and are left to sort it out alone.
