A retrospective cohort study following more than 2.5 million American children found no association between the timing of a child's first MMR vaccine dose before age 2 and a later autism diagnosis.
The study, published in the Pediatric Infectious Disease Journal, used the Cosmos electronic health record database covering children with linked birth-parent records from January 2015 through December 2024, with follow-up through age 8 or the last recorded visit. In the primary analysis of children who received a first dose between 11.5 and 24 months, the adjusted hazard ratio for autism was 0.97, with a 99% confidence interval of 0.91 to 1.03.
The distinction that matters for parents is what question this study asked. It did not compare vaccinated children to unvaccinated children, a question already addressed by decades of research. It asked whether receiving MMR earlier within the recommended window, rather than later, changed the likelihood of an autism diagnosis. The answer was no.
The Method Built to Address a Known Bias
Vaccine timing studies have a specific analytic problem. Families who bring children in on schedule differ from families who do not, and children with early developmental concerns may have different patterns of medical visits, which can distort results in either direction.
The researchers used prespecified age-based landmark analyses, a technique that anchors comparisons at fixed ages to reduce distortion from differing follow-up and healthcare use. Adjusted hazard ratios were estimated using Cox proportional hazards models with covariates selected through directed acyclic graphs.
They also ran a negative control exposure, an unrelated exposure that should show no true effect. That analysis produced a hazard ratio of 1.01 with a 99% confidence interval of 0.99 to 1.04. The authors argue that near-null results for both the vaccine and the negative control make it unlikely that hidden confounding explains the finding. Across all age-specific landmark cohorts, the estimates stayed near null.
What an Observational Study Can and Cannot Establish
This is observational research, not a randomized trial, and observational studies do not by themselves prove cause and effect.
That limitation cuts both ways. It means a study of this design cannot definitively prove that vaccination does not cause a rare outcome in a small subgroup. It also means the study is not being asked to carry that weight alone. Its value is that it is very large, prospectively recorded rather than recalled, drawn from routine clinical records, and consistent with prior work.
The comparison points matter. A Danish nationwide cohort study of 657,461 children published in Annals of Internal Medicine found no increased autism risk after MMR vaccination overall, in subgroups, or in defined time periods after vaccination. Earlier US and international cohorts reached the same conclusion. The new study adds the largest US sample to date and addresses the specific timing question directly.
Its own limits are worth stating. Electronic health record data can misclassify diagnoses, autism diagnosis ages vary by region and access to specialists, and follow-up ended at age 8, so diagnoses recorded later were not captured. The authors write that the results reinforce existing evidence supporting MMR safety, which is a more careful claim than proving safety outright. Coverage of the findings was published by CIDRAP.
The Measles Backdrop Making This Urgent
The study arrives in a record year for measles. CDC has reported 2,371 confirmed measles cases in 2026, the highest annual total since 1991, with 93% of infections occurring in unvaccinated people.
Measles is not a mild childhood illness in aggregate. It can cause pneumonia, encephalitis, and death, and it suppresses immune memory for months to years after infection, leaving children more vulnerable to other pathogens. Infants under 12 months are typically too young for a first dose and depend on coverage around them.
Delayed vaccination creates a window of susceptibility precisely when risk of severe disease is highest. Data from US electronic records show a post-2021 decline in timely MMR coverage and an increase in children with no MMR dose recorded by age 2. Because measles requires very high community coverage to prevent sustained transmission, small local declines in timely vaccination can produce outbreaks in a single school district while national figures still look adequate.
Questions Worth Bringing to Your Pediatrician
The recommended US schedule places the first MMR dose at 12 to 15 months and the second at 4 to 6 years. Children traveling internationally may be eligible for an early dose starting at 6 months, which does not count toward the two-dose series and requires the routine doses afterward.
If your child is behind schedule, ask about a catch-up plan rather than assuming the opportunity has passed. Catch-up schedules exist for every routine vaccine.
If you have questions or concerns about vaccination, bring them to a pediatrician or family physician who knows your child's history. That conversation is more useful than any single study, and clinicians expect these questions.
If your child has a weakened immune system, a history of severe allergic reaction to a vaccine component, or is undergoing certain treatments, discuss timing with a specialist, since MMR is a live vaccine and specific contraindications apply.
Parents seeking an autism evaluation should know that developmental screening is recommended at 18 and 24 months regardless of vaccination history, and that early intervention services are available in every state.
Uninsured and underinsured children can receive MMR at no cost through the federal Vaccines for Children program, available at participating pediatric offices, community health centers, and county health departments.
Frequently Asked Questions
What did the study measure? Whether receiving a first MMR dose earlier within the recommended window, between 11.5 and 24 months, was associated with a later autism diagnosis in 2.5 million US children.
What was the result? No association. The adjusted hazard ratio was 0.97 with a 99% confidence interval of 0.91 to 1.03, and estimates stayed near null across all age cohorts.
Does this prove vaccines do not cause autism? Observational studies cannot prove causation on their own. This study adds the largest US timing analysis to a body of evidence that consistently finds no link.
How is this different from earlier research? Most prior studies compared vaccinated with unvaccinated children. This one examined timing within the recommended schedule using landmark analyses and a negative control exposure.
What are the study's limits? Electronic record data can misclassify diagnoses, diagnosis age varies by access to specialists, and follow-up ended at age 8.
When is MMR recommended? The first dose at 12 to 15 months and the second at 4 to 6 years, with an optional early dose from 6 months for international travel that does not replace the routine doses.
Where can uninsured children get MMR? Through the federal Vaccines for Children program at participating pediatric offices, community health centers, and county health departments.
