On August 25, the Pennsylvania Department of Health announced two measles-associated deaths in Lancaster County, the state's first in 35 years. Later that day the Centers for Disease Control and Prevention (CDC) extended condolences to the families and said it was working closely with the state.
The next day, Health and Human Services (HHS) Secretary Robert F. Kennedy Jr. suggested on X that the deaths may have been "fabricated by one of the Governor's hopeful staffers," and called the announcement a "replay of the Democrats' COVID playbook." The day after that, the same CDC account thanked him for his "efforts to clarify conflicting information from Pennsylvania officials." Its statement invoked COVID-19 directly, recalling concerns that deaths had been attributed to the disease when people had died of other causes.
No finding had established that the Pennsylvania's classification was wrong. What changed was that the county coroner disputed one case in the press, Kennedy called both possibly fabricated, and CDC publicly thanked him. On August 30, the agency published its weekly measles update late, on a Sunday, with an asterisk in place of the two deaths, explaining that available information did not establish "whether measles caused or contributed to the deaths or whether the individuals died from other causes while infected with measles." The Washington Post and CNN reported that the decision came from the new CDC director, Erica Schwartz, MD, MPH, JD, who was sworn in last month.
States investigate outbreak deaths and the CDC compiles them. That is how the system has worked for decades. When New Mexico reported an unvaccinated adult last year who tested positive only after death, the CDC counted that death while the investigation was still open. Nothing in the CDC's statement identifies an error in Pennsylvania's data. All of it describes a dispute with Pennsylvania's governor.
States investigate outbreak deaths and the CDC compiles them. That is how the system has worked for decades.
The dispute is being argued in a vocabulary borrowed from the pandemic. The county commissioner whose post started it wrote that the coroner had "ZERO reported cases where measles is the immediate cause of death" and that one person "died WITH measles, but not FROM it." That phrase carried real weight during COVID, and I know because I used it myself at the time, on the wards and in public. It does not carry over to measles, and understanding why requires understanding why it was true in the first place.
Why ‘with, not from’ was a problem for COVID
Two things made incidental COVID-19 diagnoses common, and neither exists for measles.
The first was screening. For most of 2020 through 2022, many hospitals tested every admission for SARS-CoV-2 regardless of symptoms, to protect roommates and staff. In one four-hospital system, 51,187 asymptomatic admission tests returned a positivity rate of 1.8%. When Omicron arrived, the yield of asymptomatic screening more than doubled. A patient admitted for a hip fracture, a heart attack, or a delivery was swabbed on the way in, and a meaningful fraction came back positive.
The second was prevalence. In mid-January 2022, the country was reporting about 783,000 new cases a day. Combine universal testing with that much virus in circulation and the count of "COVID hospitalizations" fills with people whose COVID-19 infection had nothing to do with why they were in the hospital. During Omicron, a chart review of veterans' admissions found that only 59% of hospitalizations among boosted patients were for COVID-specific causes. The hospitalization metric had stopped measuring what it was built to measure.
Death certificates moved in the same direction. From 2020 through mid-2022, COVID-19 was the underlying cause on 88% of certificates that mentioned it. In January 2022 that figure was about 85%; by April it had fallen to 60% to 70% and stayed there, and for 2022 as a whole it was 76%. Those numbers cannot show incidental positives directly, because a condition listed anywhere on a death certificate is supposed to have contributed to the death.
What they show is the mix shifting, with COVID increasingly recorded as a contributor rather than the cause. What I saw on the wards went further than the certificates can record. Patients admitted for a stroke or a fall or end-stage cancer screened positive on arrival, died of the thing that brought them in, and had COVID written onto the paperwork anyway. Nobody can count those cases precisely, which is part of why the grievance was so hard to answer, and part of why it was legitimate.
So the distinction was real. It rested on two facts about COVID-19, and the argument works only where those facts hold.
Why it does not work for measles
Nobody screens for measles. There is no admission swab, no routine test before a delivery or a surgery, no reason a laboratory would look for measles virus in someone without signs of it.
CDC guidance for clinicians is to collect specimens from patients "with clinical features compatible with measles," and its clinical fact sheet walks a physician through fever, cough, coryza, conjunctivitis, and rash before it mentions a test. The PCR (polymerase chain reaction test) itself often runs at a state public health laboratory. In a country in which measles was eliminated, the CDC notes, most suspected cases turn out not to be measles. A positive measles result in a death investigation generally exists because there was a clinical or epidemiologic reason to look for it.
Prevalence points the same way. This is the worst measles year in the United States since 1991, and the national count through August 27 stands at 2,903 confirmed cases in a country of 340 million people. At the Omicron peak, the country reported more COVID cases in an hour than measles has produced all year.
The probability that an unrelated death coincides with an unrelated measles infection is not zero, but it is small enough that a positive test in someone who died should move a reasonable person's belief hard in one direction. During COVID, the base rate made incidental positives expected. Here it makes them a surprise.
Nobody screens for measles.
The newborn's case illustrates the difference. According to the family, who spoke to The Atlantic and NBC News, the mother became severely ill with measles after her other children caught it, and went into labor about a month early. The baby was born not breathing and could not be resuscitated. Measles virus was later found in his lung tissue.
That test was not a screen. It was ordered because a newborn died shortly after birth to a mother with active measles, and the literature is unambiguous about what that infection does. Measles in pregnancy raises the risk of pregnancy loss, preterm birth, and low birth weight, and pregnant women with measles are more likely to be hospitalized, develop pneumonia, and die than women who are not pregnant, according to a 2015 review in Obstetrics & Gynecology.
In a cohort of 24 pregnant women with measles during the 2017-2018 outbreak in Catania, Italy, 25% delivered preterm, against 5% to 15% in the general population. And maternal infection in the weeks before delivery can pass to the baby; congenital measles usually appears within the first 10 days of life and can be fatal.
The coroner has said the forensic pathologist found the spleen neither enlarged nor inflamed and attributed the fatal laceration to something other than measles. That deserves to be taken seriously, with two caveats.
Splenic size is difficult to judge after a rupture with massive blood loss, and the pathway that concerns me most—a mother ill enough with measles to deliver a month early—does not run through the spleen at all. Measles may still prove incidental here. The coroner's own office says the cause of the rupture is under investigation, and that, rather than anyone's confidence, should govern what gets said.
What a death certificate actually says
Much of the confusion turns on the structure of a death certificate, which most people have never had reason to learn. Part 1 lists the chain of events that led to death, with the immediate cause on the top line and beneath it the conditions that produced it, ending in the underlying cause that started the sequence. Part 2, in the words of the National Center for Health Statistics handbook, is for "other significant conditions contributing to death, but not resulting in the underlying cause given in Part I."
The CDC’s handbook written for medical examiners and coroners says the same thing. During the pandemic, CDC put it more plainly: A condition belongs in part 2 when the certifier believes it "was significant enough to contribute to the fatal outcome and was not an incidental finding."
Measles is listed in part 2 of the newborn's death certificate. The coroner has said he placed it there to offer a broader view of the child's health, and that he does not personally believe measles contributed. Those two statements are hard to reconcile, because the form does not have a line for context. It has a line for conditions that contributed. Whatever he intended, the document his office signed says measles was not incidental.
The commissioner's original standard, that a measles death must show measles as the immediate cause, would exclude many genuine measles deaths. Measles usually kills through what it causes, most often pneumonia, which the CDC identifies as the most common cause of death from measles in young children, along with encephalitis, secondary bacterial infection, and complications of pregnancy.
The immediate-cause line reads "pneumonia" or "encephalitis" or, here, "laceration of the spleen." The disease that set the chain in motion sits below it. Pennsylvania says it uses "measles-associated" when laboratory or epidemiologic evidence of measles is present, even if measles is not judged the immediate cause of death. It should have explained that distinction when it announced the deaths. But poor communication is not evidence of fabrication.
COVID is not a template
There is a broader habit at work, and it goes beyond one coroner's office. A lot of people learned during the pandemic to distrust official case counts, to suspect that hospitalizations and deaths were inflated, and to hear any public health warning as the opening move in a lockdown.
Some of that distrust was earned. I have spent much of the past two years arguing that the pandemic response got specific things wrong, and that admitting so is the only way to be believed about anything else.
But the lesson many people took was not "ask how the cases were found" or "read the case definition." It was that whatever was said about COVID was probably said about everything, and can be dismissed the same way. That is how a commissioner in Lancaster County reaches for "with, not from" about a disease nobody screens for, and how the HHS secretary describes two dead children as a COVID playbook.
Measles is not COVID, and the differences all point in the same direction. It is extraordinarily contagious, which is why it finds unvaccinated people so reliably. It is a stable virus with one serotype, so a vaccine designed in the 1960s still works, and two doses of the measles, mumps, and rubella (MMR) vaccine are about 97% effective and usually protect for life, in contrast to a coronavirus that keeps changing.
Its risk gradient looks nothing like COVID's; children under 5 are among the groups at highest risk of complications, rather than severe disease concentrating at the oldest ages. And of this year's 2,903 measles patients, more than 90% were unvaccinated or of unknown status. Nothing about that picture requires a pandemic-era interpretation, and everything about it resists one.
Measles is not COVID, and the differences all point in the same direction.
The part of the COVID experience that does transfer is the part being ignored. The way out of the "with, not from" problem in 2022 was not to stop counting. It was to look harder, through chart review, better outcome definitions, and separating primary from incidental admissions, and to publish the results so people could see the work.
Texas did a version of this last year, releasing ages and basic clinical details for its measles deaths without compromising anyone's privacy, and nobody accused Texas of fabrication. Pennsylvania should have done the same. The CDC, for its part, could have continued reviewing the cases while keeping the state-reported deaths in the national count and footnoting them as under review. Instead, it thanked Kennedy and pulled the number.
What the count is for
A death tally that can be altered by officials whom it embarrasses will not be trusted by anyone, including the officials who edited it. Every state epidemiologist reporting a death next week now knows the count can be revised at the federal level, and that risks changing what data states are willing to report. The measles death count has become a political number. The case count has not, and it tells the same story.
Measles kills roughly one to three of every thousand children who get it. At nearly 3,000 cases and climbing, additional deaths would not be surprising, and the families who suffer them deserve a federal count that includes them.
There is a family in Lancaster County that lost a newborn on August 14 and then heard the HHS secretary suggest their child's death might have been invented. Whatever the final cause of death turns out to be, they deserved better than that, and so does everyone who relies on the national numbers to know what is happening.
Dr. Scott is a clinical associate professor of infectious diseases at Stanford University School of Medicine, and a coauthor of "Updated evidence for COVID-19, RSV, and Influenza Vaccines for 2025-2026" in the New England Journal of Medicine.
The opinions voiced in CIDRAP Op-Ed pieces are the authors' own and do not necessarily represent the official position of CIDRAP.
