Byrd, RS, 2002. The Epidemiology of Autism in California: A Comprehensive Pilot Study. Report to the Legislature on the Principal Findings.

NYT coverage of this study
NYT editorial
LA Times editorial

After a 1999 study by public health authorities had indicated that California was experiencing drastic increases in the number of autism cases, skeptics had attributed the increase to greater awareness of the condition and more systematic and better medical diagnosis, or perhaps simply to the growth of California's population. To determine whether the increase was real, and if so, what could be concluded about its origins, the State Legislature requested that the University of California’s Medical Investigation of Neurodevelopmental Disorders (M.I.N.D.) Institute conduct a comprehensive pilot study.

The new study, led by Dr. Robert S. Byrd, section chief of pediatrics at the Univ. of California at Davis Children's Hospital, issued its report on 17 August 2002. It concludes there is "no evidence that loosening in diagnostic criteria contributed to an increase in the number of children with autism, and that, therefore, "some, if not all, of the observed increase represents a true increase in cases of autism in California."

The earlier report, from the California Dept. of Developmental Services, had noted from 1987 to 1998 a 273% increase in autism cases from 2,778 to 10,360 reported cases. Their analysis indicated a sharp rise beginning for children after 1980 (figure below).

The new study by M.I.N.D. carefully dissected potential factors that might have misled the earlier report.

• First, the increase could not have been due solely to increases in California's population, because the number of children in California increased far less than the number of cases of autism.
• Another critique of the original study was that the criteria used to diagnose autism could have loosened, artificially inflating the number of new cases compared to old. To evaluate this possibility, M.I.N.D. carried out a detailed and independent analysis of two cohorts of autistic children, one born 1983-85, the second 1993-95. If criteria had changed, then this comparison should reveal that the second cohort included children with less severe autism than the first. M.I.N.D. found that there was no difference in the autistic characteristics of the two groups and that therefore changes in diagnosis criteria could not be responsible for the increasing frequency.
• A third potential source of error for the first report was mis-diagnosis. Perhaps a greater number of children with autism were present prior to 1980 but not diagnosed... or conversely that in years after 1980 children were classified as autistic when they should not have been. M.I.N.D.'s comparison of the 1983-85 and 1993-95 cohorts revealed that this was not the case.
• Another possible explanation for the increase that was suggested was that more families with autistic chidren were moving into California, perhaps attracted by a belief that health services in California were better for autistic chidren than elsewhere. M.I.N.D.'s analysis showed that in fact not only were the vast majority of autism cases of children born in California, the younger cohorts were more likely to have been born in the state than the older cohorts. This is the opposite of what would be required for this possible explanation. Hence differential immigration is not driving the changes.
• Perhaps changing demographics in California were responsible for the increase. M.I.N.D.'s study rejected this explanation after their detailed analysis of the 1983-85 and 1993-95 cohorts showed that there were no large differences in the diagnosed children's demographic characteristics.

All of the alternative explanations of the increase, —factors that might have made it seem like there was an increase in autism rates when there really wasn't— were rejected by M.I.N.D.'s study. This led to the inescapable conclusion that the rate of autism in children in California has increased. For whatever reason, children born in California since 1980 are more likely to develop autism than those born before.

M.I.N.D. attempted to assess what factors might be involved in causing the increase. The rate of increase is far too rapid for it to be based solely on genetic factors. Changes in gene frequency take generations, not years or even decades. Interactions between genes and environment, or environmental factors alone, then become the primary suspects.

M.I.N.D. did not have information about environmental exposures that would allow them to test the possible contributions of environmental contaminants. They did perform one analysis to test a proposed medical explanations for autism—mercury exposure via vaccination—but their sample size was too small to provide any sure answers.

California Department of Developmental Services. 1999. Changes in the Population of Persons with Autism and Pervasive Developmental Disorders in California’s Developmental Services System: 1987 through 1998. A Report to the Legislature March 1, 1999. Sacramento CA: California Health and Human Services Agency.