included detailed information about a range of possible sources of exposures to environmental tobacco smoke. With the refined models developed in the study, a more-precise relation was identified between lung function and environmental tobacco smoke. This work illustrates the efficacy of using a population subset with improved exposure assessment in order to improve the sensitivity of the investigation.
Peters (1991) has conducted an epidemiologic investigation to identify the chronic effects of ambient air pollutants in southern California. The design for this 10-year study emphasizes exposure assessment. This emphasis is particularly warranted because of the magnitude of the health concerns and the potential cost of controls. The study is in 3 phases. Phase I considers the resources available to determine spatial and temporal patterns of pollutants, identify the locations that allow discrimination of pollutants, develop a sampling strategy, and determine cost-effective methods for estimating personal exposure. Phase II will consist of a cross-sectional study of health outcomes that can be related to personal exposure data. Five comparison groups are to be compared by their exposure to ozone, acid, and particles in varied exposure categories. Children will be the subject population. In phase III, the investigators will follow successive cohorts through either the duration of the study or high-school graduation. Changes in pulmonary function and incidence of disease will be compared with the exposures that have been assessed prospectively. This study will have a detailed exposure assessment for use in the evaluation of health effects associated with air-pollutant exposure. It will permit some estimation of the value of extensive exposure characterization in terms of both cost-resource use and the ability to define causal factors and dose-response relations in air-pollution studies.
Lioy et al. (1992) and Stern et al. (1992) have investigated population exposure to chromium waste, with particular reference to residential exposure. Lioy et al. sought to identify microenvironments that can lead to important chromium exposure. Chromium levels were determined for indoor air, outdoor air, and house dust. Surface dust was identified as the best index of potential chromium exposure. This study illustrates the role of total-exposure monitoring in the selection of exposure media for an epidemiologic study. Stern et al. (1992) compared chromium levels in urine with exposures estimated by environmental monitoring. The authors found an association between chromium in household dust and urine levels of chromium that was consistent with residential exposure to chromate-production waste. These 2 papers are excellent models of a systematic approach to exposure assessment.
Investigators from the National Cancer Institute have investigated methodologic issues in exposure assessment for case-control studies of cancer and herbicides-pesticides (Blair and Zahm, 1990a,b, 1992; Brown et