Characterization of health effects of pesticide exposure in California: Estimation of the extent of under-reporting

This study investigated the information available about health effects related to pesticide exposure in California. It was designed to develop information to guide interpretation of statistical summaries of data from death certificates, hospital discharge summaries, poison control logs, and the California Pesticide Illness Surveillance Program (PISP).; Cases for review were selected based on diagnostic coding for toxicity or chemical injury that occurred from 1994 through 1996. We selected stratified random samples of death certificates, hospital discharges, and poison control records for review. All PISP records were converted electronically for inclusion.; We reviewed source documents to identify products involved and exposure circumstances. We avoided duplications by collecting individual identifiers from each source, which permitted probabilistic linkage confirmed by operator review. We interpreted the results of file linkage as a measure of the PISP's success in collecting information on health effects of pesticide exposure.; We found all sources to contain some misinformation. Toxicants other than pesticides were identified as solely responsible for roughly 15 percent of deaths and hospitalizations that received codes specific for pesticides. Adverse effects from antimicrobial pesticides were more likely to receive codes for caustics/corrosives than for disinfectants.; Most pesticide-related hospitalizations and poison control contacts involved residential exposures. Among fatalities, suicides predominated. The PISP, by contrast, collected primarily occupational exposures. Exposures of groups of people were nearly certain to be reported to the PISP, and exposures to agricultural-use pesticides were substantially more likely to be reported than exposure in other situations.; Misinterpretations and differences in definition distorted classification of pesticide-related health effects. Investigators should interpret statistical compilations conservatively, to allow for errors and misunderstandings.; Statistics drawn from the PISP may be expected to present a complete picture of group pesticide exposures, and a reasonable representation of exposures to agricultural-use pesticides. Characterization of residential exposures requires more attention.