OBJECTIVES: Using environmental concentrations and Geographic Information Systems (GIS) to model human exposure is an emerging tool in environmental epidemiology. To evaluate the usefulness of this, we investigated to what extent cadmium and lead concentrations in soil and moss could be used as potential determinants for cadmium and lead exposure for a population living near a battery plant. METHODS: Cadmium in urine and blood and lead in blood, as well as food, smoking habits, places of residence, occupations and health, were determined for 512 individuals living near a closed down battery factory in Sweden. Cadmium and lead contents in moss were available through previous assessments by the Geological Survey of Sweden. Soil samples had been collected at various distances from the factory, and analyzed for cadmium and lead contents in 1990. Kriging technique in GIS was used to create areas with different metal concentrations based on these measurements. By linking individual address coordinates to mapped concentration levels, each study subject home address received a value of cadmium and lead in moss and soil. RESULTS: We found a statistically significant association between lead in soil and lead in blood, for female subjects eating homegrown vegetables regularly. No significant association was found between cadmium in soil and cadmium in urine for either gender in the study population. No clear associations were found for either gender regarding lead and cadmium in moss and lead in blood or cadmium in urine or in blood. CONCLUSION: In general, environmental concentrations may not be useful surrogates for assessing human exposure to lead and cadmium, but concentrations of metals in soil around emitting point sources can be a complement for estimating the exposure in certain subgroups.