A new Pediatric Infectious Disease Journal study is being promoted by the pharma-captured mass media apparatus as a massive 2.56-million-child analysis disproving an MMR–autism association, but major methodological problems undermine the entire paper.
They did not use a truly unvaccinated control group, did not clinically adjudicate autism, did not interview parents, did not independently verify vaccine histories, and used a narrow ICD-coded autism definition.
No True Unvaccinated Control Group
Most importantly, there was no true unvaccinated control group. The comparison children were only those without a recorded MMR vaccination during the relevant analysis. They were not required to be vaccine-naive and could have received other routine childhood vaccines. The study itself reports that 62.1% of the overall cohort received a PCV booster.
Even their MMR status was not independently verified. Vaccination exposure came from Epic records, including historically documented doses, with no reported state immunization-registry confirmation, parent interview, or outside-record review. That means some children classified as “unvaccinated” in the dataset could theoretically have received MMR outside the participating Epic/Cosmos system without that dose being captured.
Autism Was Based on ICD Codes, Not Clinical Adjudication
The study relied entirely on electronic health records. Autism was defined by an ICD-10 F84.0 encounter code, not by independent clinical adjudication, ADOS testing, chart review, or parent interview. Nearly 9% of autism cases had only a single encounter carrying the diagnosis. Vaccine status was likewise taken from Epic records, with no reported state-registry verification, parent confirmation, or external record review.
The Study Did Not Measure the Full Autism Spectrum
The investigators counted only ICD-10 F84.0 “childhood autism,” a relatively narrow diagnostic code, rather than the broader range of autism-spectrum diagnoses captured under F84.. Yet the 3.2% national benchmark they cite reflects that broader autism-spectrum definition. This means their 2.55% autism prevalence is not directly comparable to the national 3.2% figure and may miss children coded under other autism-spectrum categories. The authors themselves acknowledge this limitation and state that future studies should examine the full F84. spectrum.
Hellbound and Down 