Environmental Research on Autism Based on Administrative Data Linkage

Many of our concerns regarding autism prevalence and epidemiology have arisen from and been addressed through analyses of administrative data, particularly records of social service and educational interventions for children with autism and their families. The linkage of such records to birth certificate files has enabled the estimation of cohort prevalences and provided insight into their temporal trends. The linkage of such records to data describing environmental hazards that may be related to autism can be seen as an extension to this process, yielding information complimentary to clinic-based studies, animal studies, or survey research.

Reliance upon administrative data constrains our ability to verify diagnoses, specify phenotypes, and consider a variety of covariates; it also limits our attention to those environmental hazards for which we have spatially and temporally resolved data. In spite of these limitations, however, advances in Geographic Information Systems (GIS), the spatial and temporal modeling of pollutants, and pharmacokinetics suggest that the validity of data linkages can be expected to increase. With these advances it becomes more likely that environmental epidemiological studies based on data linkage will make contributions to our understanding of autism, particularly in the form of hypothesis generation.