The validity and precision of research on health effects that allows for quantitative risk assessment for environmental chemicals depend heavily on the quality of exposure information. Previous studies have been conducted to collect data on exposure to specific compounds for the U.S. population using large, probability samples [1, 2]. These efforts have produced valuable information on group means and inter-individual variability in exposure levels. One lesson was that exposure to toxins is influenced by both micro-environmental levels of chemicals and by human activity patterns bringing persons in contact with exposures. Three types of human activity patterns are of particular interest for exposure assessment: food consumption, temporal-spatial activity patterns, and use of household products.
Food consumption serves as a source of exposures in several respects. First, it provides basic nutrient and caloric needs; second, it is a conduit for numerous chemicals that partition into the food during production that are not necessary for growth, development, or the maintenance of life systems; third, processing, packaging, preparation and cooking of food may introduce chemicals that were not present in the raw food. For some toxic compounds, food may be the primary route of exposure, e.g., to mercury, PCBs, some pesticides, dioxins, and acrylamide. While nutritional epidemiologists have focused on estimating nutrient and energy intake, food toxicology assesses contaminants and safe levels of essential minerals that are toxic at higher levels. A variety of methods are used for assessing dietary intake, reviewed by Kroes et al. , include food balance sheets, household market basket surveys, duplicate diet, food frequency, 24-hour recall, food records, and biomarkers. However, most large surveys providing extensive nutritional assessments, such as NHANES, are cross sectional providing limited insight into seasonal or long-term dietary behavior changes . Although environmental contaminants of food has been a topic of increasing interest in recent years, especially among children, to date no surveys have comprehensively addressed exposure to non-nutritive chemicals through food.
Temporal-spatial patterns of activity strongly influence exposure in several ways. First, the specific micro-environment where people spend time influences exposure levels due to varying concentrations of an agent in that microenvironment. For instance, newer homes are constructed differently than older ones, with materials containing different chemicals. Second, certain types of human activities can increase exposure concentrations, such as smoking or burning candles, which causes higher ambient levels of polycyclic aromatic hydrocarbons and particles. Third, activities can influence the contact rate with the contaminated environment, for example, physical activity increases the total intake of a toxin through inhalation by increasing breathing rates. Important for children's exposures is crawling and hand-to-mouth activity, including placing their thumbs or other objects in their mouths, resulting in increased exposures via dermal or non-dietary ingestion, respectively. The National Human Activity Pattern Survey (NHAPS)  further generated activity pattern data intended for the estimation of the prevalence and duration of population exposure. Data from this survey are included in the U.S. Environmental Protection Agency's Consolidated Human Activity Database. However, previous studies concentrated on collecting short-term activity data, and have not emphasized longitudinal or seasonal data collection that would allow evaluation of intra-individual variability over time.
Finally, household products are a significant source of exposures to a variety of chemicals [6, 7]. Products with direct dermal and inhalation exposures span a wide array of purposes: household cleaning products for furniture, toilet bowls, glass, carpets, etc.; pesticides (including those used on pets); dry cleaning fluids; pre-sale treatments of carpets and draperies for stain resistance and their backings with flame retardants; air fresheners, containing perfumes and aerosol propellants. Other household products relate to personal care including: cosmetics; dyes, permanents, hair straighteners, and shampoos; sunscreens; nail polishes; antiperspirants; perfumes and other fragrances used on the body or as room fresheners; contact lens solutions, ear wax removal products, and nasal sprays. Pesticides have been deemed of particular concern given that they may be formulated with the specific intent of having sufficient toxicity to kill certain species of living organisms. However, chemicals used for numerous other purposes also often have toxic properties. Very little research on use of household products has been conducted with the aim of understanding the potentially harmful human exposures that may result from their use.
A major gap in the current knowledge base concerning exposure assessment and human activity patterns derives from the cross-sectional or short-term nature of most population-based exposure datasets. The Study of Use of Products and Exposure-Related Behaviors (SUPERB) was developed in response to growing interest in data collection platforms that can be used for longitudinal assessments of exposure-related behaviors that may change over time, e.g., seasonally. The SUPERB platforms were designed to vary in the burden imposed on the participants and to be appropriate for several different age strata in population-based samples of households. Three age/demographic strata were of particular focus: young children, adults with families, and older adults. We conducted the surveys in northern and central California, the recruitment aimed to enroll an ethnically diverse sample. Data collected in a multi-tiered approach from SUPERB participants covered short-term, seasonal, and long-term changes in food consumption habits, temporal-spatial activity, and use of household and personal care products. It is expected that these data, as well as the methodological lessons learned, will be particularly helpful in refining protocols for studies such as The National Children's Study, the first comprehensive longitudinal assessment of a nationally representative sample of U.S. pregnant mothers and their children, aimed at examining exposures in early life and their health effects. The present report provides an overview of the SUPERB design and data collection methods and presents descriptive information on the study samples participating in various Tiers, each with a different platform or approach for data collection.