The goal of this research is to enhance the use of parental reports of exposures and job histories as epidemiologic tools in the investigation of environmental risk factors for adverse pregnancy outcome. Three basic methodologic problems will be addressed: 1) How to estimate and correct the bias in the odds ratio estimate arising from nondifferential exposure misclassification; 2) How to analyze exposure measures so as to minimize further exposure misclassification; and 3) How to improve the reliability of subject reports and exposure assessment. This research will apply current understanding of measurement error to existing interview exposure data from studies of adverse pregnancy outcome and to simulation models that will lead to new methods of data collection. Specifically, measurement error in exposure will be characterized for both job history reports and job-exposure profiles. Exposure misclassification-rates and the odds ratio adjusted for nondifferential misclassification will be estimated. Simulation models will be used to clarify the range of effects on exposure-response relationships from observed errors. Means of implementing new data collection methods directed to resolving these errors will be identified based on the reliability and simulation studies. Acute and chronic exposure models for specific agents and specific to gestational age groups will be developed, their distributions described, and the impact of the modelled exposures on odds ratio estimates will be examined. Protocols for the collection and reliability assessment of exposure data from interviews will be developed. Pregnancy outcome studies to obtain detailed interview exposure data and measures of risk adjusted for nondifferential misclassification will be designed and initiated. The primary data sources for the development of methodology will be: 1) job history reports from the Baltimore-Washington Infant Study, a population based case-control study of cardiovascular malformations, combined with job-exposure matrices to be developed; and 2) interview-derived job histories and job-exposure matrices from a cohort study of spontaneous abortions among semiconductor manufacturers with potential chemical exposures. New data from these sources and from other population based studies will be collected and analyzed with the newly developed methods. This proposal is therefore designed to apply current knowledge to the evaluation of measurement error in interview exposure data in pregnancy outcome studies, and to develop methods for the collection and analysis of interview exposure data directed at resolving these errors.
| Correa, Adolfo; Min, Yuan-I; Stewart, Patricia Ann et al. (2006) Inter-rater agreement of assessed prenatal maternal occupational exposures to lead. Birth Defects Res A Clin Mol Teratol 76:811-24 |
| Jackson, Leila W; Correa-Villasenor, Adolfo; Lees, Peter S J et al. (2004) Parental lead exposure and total anomalous pulmonary venous return. Birth Defects Res A Clin Mol Teratol 70:185-93 |
| Correa, A; Gray, R H; Cohen, R et al. (1996) Ethylene glycol ethers and risks of spontaneous abortion and subfertility. Am J Epidemiol 143:707-17 |
| Correa-Villasenor, A; Stewart, W F; Franco-Marina, F et al. (1995) Bias from nondifferential misclassification in case-control studies with three exposure levels. Epidemiology 6:276-81 |
| Correa-Villasenor, A; Ferencz, C; Neill, C A et al. (1994) Ebstein's malformation of the tricuspid valve: genetic and environmental factors. The Baltimore-Washington Infant Study Group. Teratology 50:137-47 |
