You are here:
FEASIBILITY OF MATCHING STUDY PARTICIPANT RESIDENCE WITH SPECIFIC WATER UTILITY TRIHALOMETHANE (THM) DATA IN EPIDEMIOLOGIC STUDIES
Citation:
Luben, R. A., J. Nuckols, M. C. Lynberg, P. Langlois, J. Wolf, AND P Mendola. FEASIBILITY OF MATCHING STUDY PARTICIPANT RESIDENCE WITH SPECIFIC WATER UTILITY TRIHALOMETHANE (THM) DATA IN EPIDEMIOLOGIC STUDIES. Presented at International Society for Environmental Epidemiology, New York, New York, August 1-4, 2004.
Description:
Many epidemiologic studies concerning by-products of water disinfection use utility monitoring data to estimate exposure. Use of such data requires linkage of residence location to a specific water utility and associated monitoring data during a given exposure period. The inability to perform this linkage successfully can lead to significant exposure misclassification and/or reduced sample size. We evaluated the feasibility of using public record data to link trihalomethane (THM) data to 1,980 participant residences on public water supply in the Texas Center for the National Birth Defects Prevention Study (NBDPS). The Center provided geocoded residence locations during an exposure period defined as 3 months prior to date of conception (DOC) through date of birth for each participant. We also assessed potential sample bias when only the subset of the study population linked to THM data was used in an epidemiological analysis. We developed a geographic-based procedure for linking each geocoded residence to a specific utility using census data. We determined whether THM monitoring data were available for these utilities through state agency records. Geocoding success was 82% (n=1,623), 90% of the geocoded residences could be matched to a specific water utility, and 589 (40%) of residences were linked to a utility with at least one THM measurement during the exposure period. Similar results were achieved using residence at date of conception, a critical exposure period for birth defect studies. Linkage to THM data was highly dependent on utility service population, with 99% of residences in service areas with a population > 5,000 successfully linked, versus 19.5% for those with a population < 5,000. We found no bias for age, income, or education in the subset of the study population successfully linked to THM data. However, participants that were not linked to a THM measurement were more likely to be Hispanic (p<0.01). Overall, the results of this study, if validated, indicate it is very feasible to link study participants to a specific water utility using our geographic-based method. Ability to geocode a residential address is an important limiting factor. Using public records to link THM data to participants during a specific time period, such as gestation, is much more challenging. These results also indicate that sampling bias is likely to be limited to risks associated with living in areas with small utility companies and possibly Hispanic ethnicity.
This is an abstract of a proposed presentation and does not necessarily reflect EPA policy.