May 23, 2026
In November of 2024, I critiqued a JAMA publication for stating autism prevalence has changed due to diagnostic changes in criteria in spite of two CDC’s ADDM publications I was aware of at the time that found higher case counts for autism using DSM-IV-TR criteria than DSM-V, the researchers seeming contradictory statement that the transition from ICD-9 to ICD-10 codes in 2015 was found to have little to no impact on autism diagnosis rates -a contradiction to the statement that changes in diagnostic criteria account for increases in autism rates-, interpretations of adult autism rates without accounting for ‘age at diagnosis,’ and severity of autism symptoms at early life impacting age at diagnosis where profound autism cases may be more likely to have an early life diagnosis [1].
In June of 2025, I critiqued various WHO publications on autism for similar reasons, criticizing arguments that changes to diagnostic criteria accounted for increases in autism incidence rates, at the time identifying a total of 5 publications from the CDC’s own ADDM that found higher case counts using older criteria for autism, and critiqued contradictory statements by the WHO that individuals with developmental disabilities are living longer but are also subject to increase mortality, and critiqued interpretations of adult prevalence rates of autism without factoring age at diagnosis [2].
I desire to at this time also discuss the limitation of ‘age at diagnosis’ for as a variable it is not wholly indicative of ‘age at regression’ if such a factor impacts an autism diagnosis, nor is ‘age at diagnosis’ indicative of ‘failure to meet developmental milestones’ nor ‘start of autism symptoms’ in cases of repetitive behaviors meeting criteria.
For example, perhaps a case child is developing normally for the first year of life, however, the parents of the child begin to note a regression/loss-of-milestones-met around one year and six months; yet, a formal diagnosis of autism is not provided until 3 years of age in spite of the child beginning to receive early intervention services at age 2. Thus, two separate variables can be used in this case: regression at age 1 year 6 months and age of diagnosis at age 3.
As another example, perhaps a child does not meet developmental milestones at 1 year of life, yet for some reason they are not diagnosed with autism until age 4. Once again, two separate variables can be used in this case: failure to meet 1 year milestones and age of diagnosis at age 4.
As another example, ‘start of autism symptoms’ may be a case for a young female child of autism who exhibits repetitive/restrictive behaviors since age 2, but perhaps due to her high communication skills she is not diagnosed with autism until age 7. In this case, if the child has exhibited other symptoms of autism since a younger age but did not receive a formal diagnosis until age 7, then those two variables could be used by researchers.
I realize that autism diagnosis is used by researchers to set firm boundaries between what is or isn’t classified as autism and this assists in data interpretation. However, worth remembering is that some researchers also use autism-symptom scales and that these can be used despite a lacking autism diagnosis in a dataset; and that the interpretations may be more broad than just an autism diagnosis but to those exhibiting symptoms of autism (as defined in the autism-symptom scale used) without a formal diagnosis.
Of course, this brings us full circle into the use of behavioral definitions of autism and the limitations thereof, where results without a formal autism diagnosis may be argued to be too broadly interpreted to individuals/children with symptoms but perhaps not severe enough to meet/require diagnosis; and the same could be true for the autism biomarker literature, where results without a formal autism diagnosis may be argued to be too broadly interpreted.
Yet one must ask: is the presence of a biomarker such as oxidative stress not cause for concern despite not meeting criteria?
If good health is the absence of disease, and the absence of disease is impacted by the absence of oxidative stress, then perhaps it could be argued that despite not meeting criteria for autism diagnosis, the presence of a biomarker is cause for concern in the name of good health.
In April of this year 2026 I hypothesized a preliminary autism-biomarker equation [3]. My ongoing refining of that equation shall include reviewing early-life susceptibility to toxins that increase the probability of biomarkers impacting autism diagnosis. However, I understand autism-biomarkers are not limited to early-life exposures, but may also present themselves across the lifespan due to cumulative exposure to toxins and those same biomarkers may yet be connected to other disease-conditions.
I believe that’s the complicated nature of this dissertation that does not wish to exclude relevant context, yet definitions become blurry and boundaries between one disease and another, or biomarkers for one disease and another, become less differentiated.
Nevertheless, I believe this inclusive analysis will assist in yielding answers we are desperate for.
References
- Autism Librarian (2024, November 4). Scientific Negligence in Latest Autism Incidence Publication: Are JAMA Researchers Incapable of a Literature Review? https://thecausesofautism.com/2024/11/04/scientific-negligence-in-latest-autism-incidence-publication-are-jama-researchers-incapable-of-a-literature-review/
- Autism Librarian (2025, June 16). Trump’s Executive Order: Is Exiting the WHO the Right Thing to Do (for Autism)? https://thecausesofautism.com/2025/06/16/trumps-executive-order-is-exiting-the-who-the-right-thing-to-do-for-autism/
- Autism Librarian (2026, April 28). Hypothesizing More Equations: Did Aristotle Miss a Linguistic Fallacy? https://thecausesofautism.com/2026/04/28/hypothesizing-more-equations-did-aristotle-miss-a-linguistic-fallacy/
The Causes of Autism 