Maternal risk factors vary between subpopulations of children with autism spectrum disorder.

This was an insurance claims study, not a clinical assessment. Autism and pregnancy variables came from diagnosis codes, and some codes are vague because clinicians may choose broad or detailed codes. The data included only commercially insured and Medicare Advantage enrollees, not Medicaid, so families with lower income or different coverage may be missing. The study assumed people used their medical coverage, so it may miss pregnancies affected by barriers to care. It could only look at events during pregnancy, not before or after, and it could not measure many environmental or family factors. It could not tell whether developmental delay or autism was diagnosed first. It could not tell what type of prenatal vitamins or vaccinations were involved, whether cesarean delivery was elective, or whether twins were identical or fraternal. The third trimester finding may be partly due to more medical visits near birth, not a real difference in timing of risk. Sociodemographic differences may have influenced results. For example, higher education and older age are related, and the study could not fully separate them. Because the study is observational, it shows associations between maternal pregnancy claims and autism clusters, not causes. Variables with fewer than 11 claims were removed, and the non-autism group was sampled at 10%, which may limit some comparisons. Some subgroup comparisons, such as mothers older than 40, had too few cases for reliable testing.

No author conflict-of-interest declaration is included in the supplied article text. The provided publication types list NIH extramural and non-U.S. government research support.

Do not read these prenatal claim associations as causes of autism, especially the vaccination finding. The study used insurance claims, could not identify vaccine type or vaccinations given outside insurance records, and cannot show cause.