MAXIMIZING ACCURACY AND PRECISION USING INDIVIDUAL AND GROUPED EXPOSURE ASSESSMENTS

Objectives Random errors in exposure data were explored to determine t heir effect on exposure-response relationships using individual, group ed, or combined (grouped and individual) exposure assessment methods. Methods Monte Carlo simulations were conducted by generating small ''s tudies'' of one hundred subjects divided into four exposure groups. Ob served exposure data were generated for each individual using assumed inter- and intraindividual variances and a lognormal distribution. The data were used to calculate the following three estimates of exposure : an individual mean, a group mean, and a hybrid estimate using the Ja mes-Stein shrinkage estimator. The exposure estimates were regressed o n generated (continuous) ''health outcomes,'' and the regression resul ts were stored and analyzed. Results Random errors in exposure data re sulted in attenuation of the exposure-response relationship when the i ndividual estimates were used, especially when the within-subject vari ability was high. The attenuation was substantially controlled by the group mean estimate, however, at a cost of decreased precision. The hy brid estimator simultaneously controlled both bias and imprecision in the observed exposure-response function. Conclusions While estimates o f exposure based on individual means may result in attenuation of the exposure-response relationship, grouped estimates may control bias whi le decreasing precision. Combining individual and group estimates can simultaneously control both types of error. However, further research is required to determine how robust these findings are to different er ror structures, grouping strategies, exposure-response models, and exp osure assessment methods.