OBJECTIVETo compare the results of observed individual means (OIM) model with beta binomial-normal (BBN) model and to apply the two models to assessment of long-term dietary lead exposure.

METHODSFood consumption data were obtained from the National Nutrition and Health Survey conducted in 2002 by 24-hour recall method. Contamination data were derived from the national food contamination monitoring program from 2000 to 2006 and from monitoring data of Customs exports for agricultural products between 2005 and 2006. By multiplying the average consumption of food with the average concentration of contaminant, the OIM model calculated dietary intake per day. By correcting the within-person variation and keeping the between-person variation, the BBN model built dietary intake in the long-term.Using the example of food lead data, the results of two models were compared.

RESULTSThe high-end percentile of OIM model was higher than the BBN model in various age groups.In the general population, the dietary intake of OIM model from 25th percentile to 99.9th percentile was between 1.167 and 7.313 µg×kg(-1)×d(-1), and the dietary intake of BBN model with the same percentile range was between 1.193 and 5.729 µg×kg(-1)×d(-1). The median of various groups was similar between the two models. The dietary intakes in the general population of two models were 1.543 and 1.579 µg×kg(-1)×d(-1).

CONCLUSIONThe high-end percentile of OIM model is more conservative than BBN model in the long-term dietary exposure assessment.