EXPOSURE ANALYSIS -CHAPTER 1 pptx

Ott Stanford University CONTENTS 1.1 Synopsis...3 1.2 Introduction...4 1.3 Full Risk Model...4 1.4 Total Human Exposure Concept ...6 1.4.1 Indirect Approach ...10 1.5 Receptor-Oriented A

Part I

Exposure Concepts

Receptor-Oriented Science

Wayne R Ott

Stanford University

CONTENTS

1.1 Synopsis 3

1.2 Introduction 4

1.3 Full Risk Model 4

1.4 Total Human Exposure Concept 6

1.4.1 Indirect Approach 10

1.5 Receptor-Oriented Approach 12

1.6 Indoor Air Quality 14

1.7 Activities That Increase Exposures 17

1.7.1 p-DCB in Residential Microenvironments 19

1.8 Source Apportionment of Exposure 21

1.9 Public Education 24

1.10 Exposurist as a Profession 26

1.11 Conclusions 27

1.12 Questions for Review 28

References 28

1.1 SYNOPSIS

Exposure science differs from past approaches in the environmental sciences, because, instead of beginning with an assumed source of pollution and then trying to find out who or what may be affected, exposure science works from both directions — not just from the source but also backward from the receptor In the past, environmental policy analysts have assumed some obvious source, such as a smokestack or a leaky drum, may be important and then tried to trace where the pollutant goes, often losing its complicated trail long before the much-diluted pollutant ever reaches any real

people, plants, or animals By following this older source-oriented approach, it has been difficult

to show in epidemiological studies that traditional sources actually affect anyone’s health or that these assumed sources cause adverse exposure of humans to harmful concentrations Exposure science, on the other hand, begins with the target itself, measuring the pollutant concentrations that actually reach people by using, for example, personal monitors worn by individuals Then the analyst works backward from the personal exposure to find the actual source By following this

receptor-oriented approach and measuring pollutant concentrations at the contact boundary of the person, new sources have been discovered, and some sources originally thought to be important were found to make negligible contributions to exposure Using direct measurements of exposure, many significant new indoor sources — consumer products, building materials, and personal activities — emerge as the main contributors to human exposure and thus are more likely to be

the causes of environment-related illness than traditional outdoor sources Activity pattern surveysshow that people spend on average more than 90% of their time indoors or in a vehicle, and efforts

to protect public health from environmental pollution and to reduce personal exposure must considerboth indoor air quality and outdoor air quality as well as other routes of exposure, such as dermalcontact, drinking water, nondietary contact, and food

1.2 INTRODUCTION

Historically, environmental risk problems often were discussed without identifying the target of

the risk As noted in definitions of the concepts of exposure and related terms (see Chapter 2), riskassessors must always answer a critically important question, “Whose risk is to be reduced?” Ifthe goal is to protect public health, then the target of the risk becomes the human being If the goal

is to protect an ecological system, then the target may become some organism within the system

— for example, an animal or a plant Failure to identify the target of the risk in environmentalproblems causes much confusion

In this book, we are concerned primarily with the risks of pollutants to public health, so thetarget of concern is the human being In particular, we are more interested in the risks caused by

environmental pollution to the general public than in the risks faced by workers in certain

occu-pational settings that are protected by occuoccu-pational health and safety laws Environmental protectionlaws, unlike occupational health laws, are intended to protect the general public The science ofexposure analysis has provided important insight by focusing more clearly than ever before onindividual members of the general public and the manner in which pollutants reach and adverselyaffect them Included, of course, are those persons in special high-risk occupations, but workerexposure constitutes a special subset of the general population

1.3 FULL RISK MODEL

To reduce the risks of environmental pollutants to human beings, a relationship between the sources

of pollution and their effects must be found If the risk is to be assessed accurately, then all thesources of the pollutant must be included The sources considered cannot be limited to just thetraditional or more obvious ones (smokestacks, sewage outfalls, hazardous waste sites, etc.) butneed to include the nontraditional sources as well (for example, building materials in indoor settings,consumer products, food, drinking water, and house dust)

Establishing the links between a particular target and a particular agent, or pollutant, requires

a knowledge of five fundamental components that may be viewed as links in a chain — from source

to effect — comprising the full risk model (Figure 1.1) Such a link sometimes is called a route

of exposure In this full risk model, each of the five components is linked to the others and isdependent on the one before it Thus, the pollutant output from one component is the input intothe next component Therefore, if information on one component of the model is lacking, then it

is not possible to fully characterize the relationship between the sources of pollutants and theresulting effects With a missing component, the effect of controlling a source on the exposure

received by the target cannot be determined The sources in this conceptual model are all sources

of a particular pollutant that may cause risk, regardless of whether the sources are found indoors,outdoors, or in-transit, or whether they are carried to the target by air, water, food, or dermal contact.Despite the importance of each of the five components in Figure 1.1 for determining the publichealth risk associated with environmental pollution, historically our scientific understanding of allfive components has not been equal Usually, environmental pollution has come to the attention ofpublic officials because traditional pollutant sources, such as smokestack plumes or leaking drums,have caused alarm because they were so obvious The obviousness of certain traditional sourceshas caused an overemphasis on this source category in the complete risk model Consequently, a

Exposure Analysis: A Receptor-Oriented Science 5

great body of knowledge exists today about source abatement and control of traditional sources,and many of the environmental laws deal with direct regulation of traditional sources, regardless

of whether their contribution to health risks is large or small By comparison, nontraditional sources,which release pollutants that reach people by nontraditional routes of exposure (for example,consumer products emitting pollutants in the home), have received relatively little attention eitherfrom regulators or the public

Once a traditional source of pollution is known and identified, environmental analysts usuallyfocus on the manner in which it moves from its source — sometimes called its fate and transport

— until it ultimately is converted into other harmless chemicals or reaches humans Like the sourcecomponent (first box in Figure 1.1), the pollutant movement component (second box) has receivedconsiderable research attention Meteorologists have developed atmospheric dispersion models, andwater quality and groundwater specialists have developed geophysical models for the movement

of pollutants in streams, groundwater, surface water, soil, and the food chain Research on themovement, or fate and transport, of pollutants has dealt primarily with traditional routes of exposure,tracing the movement of pollutants through geophysical carrier media on a large physical scale(1–100 km), while nontraditional routes of exposure (for example, microenvironments with sourceswithin 30 m of the person) usually have been overlooked

As with the first two components, the fifth component in Figure 1.1 — the effects of a pollutant

on the target — also has received considerable past research attention Numerous studies havelooked at the effects on animals and humans of specific concentration levels, often without knowingwhether suspected sources can create these same high exposure levels Because research on thehealth effects of each pollutant is difficult, much uncertainty still remains despite years of studies

of the effect of a given exposure level on rats and other animals in the laboratory, humans in clinicalexperimental settings, and community epidemiology studies

In contrast with the source, movement, and effect components of the full risk model, very littleknowledge has been available up until the last 20 years for the remaining two components —exposure and dose As the science advances, exposure and dose, because they have been neglected,are proving to be the two components of the full risk model for which major contributions toknowledge are possible Without accurate knowledge of human exposure (the middle box), it isoften impossible to determine which sources should be controlled and by how much, or the likelyeffect on public health of controlling a source The shaded zone in Figure 1.1 shows the main areas

of emphasis in exposure analysis

Filling the critical gaps in the full risk model is necessary for implementing a risk-basedapproach to environmental management Completing the full risk model allows an analyst todetermine if traditional source control efforts are actually reducing the risks to public health in themanner expected and to the degree needed If we do not know whether the right sources (or othercontributors to exposure and risk) are being controlled, or whether they are being controlled by

FIGURE 1.1 Five components of the conceptual full risk model To determine how much a change in a source

causes a corresponding change in an effect, it is necessary to know: (a) all five components and (b) the linkage between each pair of components The shaded zone indicates the primary emphasis of the science of exposure analysis (From Ott 1985 With permission.)

Source

Movement of Pollutants

Exposure

the correct amount, then our environmental programs will become ineffective and inefficient inreducing public health risks.

Fortunately, exposure research completed over two decades has demonstrated for a handful ofpollutants that the all-important missing exposure and dose data can be obtained and that the fullrisk model can be completed, thus making possible a true risk-based approach to environmentalmanagement This exposure research also has identified a variety of new nontraditional pollutantsources and other important contributors to exposure that environmental programs currently arenot addressing adequately, and the data show that many of these newly identified sources contributemore to public health risk than many traditional sources now subject to environmental laws andregulations As a consequence of these new findings, some observers believe that our environmentalpriorities and policies need to be reshaped and our laws revised

Although it is important to link sources to exposures to effects in the full risk model, evenlinking sources to exposures (but not necessarily to dose or effects) provides a great new body ofpractical knowledge that is important to regulators, decision makers, public health officials, andthe general public If an accurate source-exposure relationship can be established for a particularenvironmental pollutant, then it is possible to discover the most economical and efficient way toreduce risk by reducing exposures, with a consequent reduction in potential risk This benefit ispossible because it is reasonable to assume that a monotonic relationship exists between exposureand risk — that is, decreases in exposure lead to corresponding decreases in health risk, eventhough the exact form of the relationship may not be known Indeed, by completing even the partialrisk model between sources (first box) and exposure (middle box) — that is, finding the relationshipsfor just the first three boxes — we may discover that our mitigation activities are placing too muchemphasis on the wrong sources and that other overlooked nonregulatory approaches may be moreeffective in reducing exposures than the ones we have chosen Indeed, the evidence from someexposure research findings points toward many simple steps not usually associated with environ-mental laws that can be taken by individuals to reduce their own risks

1.4 TOTAL HUMAN EXPOSURE CONCEPT

The total human exposure concept seeks to provide the missing component in the full risk model:estimates of the total exposure of the population to an environmental pollutant with known accuracyand precision, usually averaged over a time period of interest, such as 24 hours Ideally, the exposureinformation needed is not for just one member of the population, nor is it the average exposure ofthe entire population Rather, what usually is sought is a frequency distribution of all members ofthe population, giving not only the population mean exposure but also the highest 1%, 5%, and10% of the exposures of the population.1 This measurement methodology has been completelydeveloped and demonstrated successfully for one major pollutant, carbon monoxide (CO; see

Chapter 6), and the total human exposure methodology also has been applied effectively to a number

of other toxic environmental pollutants

The total human exposure approach begins first by applying the conceptual definitions ofexposure that can be described quantitatively (Chapter 2) A target must be identified, and a contactboundary also must be identified, even though it might not be formally stated in a real-life exposureanalysis Any pollutant in any carrier medium that comes into contact with this conceptual contactboundary — either through the air, food, water, or dermal route — is considered to be an exposure

to that pollutant at that instant of time (Figure 1.2) The instantaneous exposure usually is expressed

as a pollutant concentration (for example, mass/volume) in a particular carrier medium at a particularinstant of time Some pollutants, such as CO, can reach humans through only one carrier medium,

1 These percentages are based on the quantiles of the distribution of population exposures For example, if we rank all the population exposures we have from highest to lowest, and 99% of the exposures are found to be below 65 µ g/m 3 of some pollutant, then can we say that 1% of the population of our sample have exposures at or above 65 µ g/m 3

Exposure Analysis: A Receptor-Oriented Science 7

the air inhaled Other pollutants, such as lead and chloroform, can reach humans through two ormore routes of exposure in multiple carrier media (for example, air, food, dermal, and drinkingwater) If multiple routes of exposure are involved, then the total human exposure approach seeks

to determine a person’s exposure (concentration in each carrier medium at a particular instant oftime) through all routes of exposure from all the sources of that pollutant (Figure 1.3)

Once applied, the total human exposure methodology seeks to provide information, with knownprecision and accuracy, about the exposures of the general public through all environmental media,regardless of whether the pathways of exposure are air, drinking water, food, or dermal contact Itseeks to provide reliable, quantitative data on the number of people exposed and their levels of

FIGURE 1.2 Exposure occurs when any pollutant makes direct contact with the human being through one

of four possible carrier media.

FIGURE 1.3 Full risk model when there is more than one source of the same agent Here the movement of

the pollutant from the source to the target (a human being) is shown as an “exposure pathway” for each source The entire journey of a pollutant from a source to a target is a “route of exposure.”

exposure, as well as the sources and other contributing factors responsible for these exposures,including the reasons for the individual person-to-person differences in exposure Ideally, theseexposure data are presented in an exposure distribution — a frequency distribution showing theproportion of the population exposed to different concentration level intervals over a specific timeperiod.

Placing the human being at the center of attention makes the total human exposure approachdifferent from the older and more traditional environmental analysis approaches — total humanexposure is a receptor-oriented approach, because it begins with the receptor, the human being.The total human exposure approach first considers all routes of exposure by which a pollutant mayreach a target Then, it focuses on those particular routes that are relevant for the pollutant ofconcern, developing measurements of the concentrations reaching the target Activity pattern infor-mation from diaries maintained by respondents can help identify those activities and microenvi-ronments that are of greatest importance, which, in many cases, helps to identify the contributing

sources Body burden, the amount of pollutant present in the person’s body — for example, blood

concentrations of lead (Pb) or breath concentrations of volatile organic compounds (VOCs) —often is included in exposure studies as an indication of previous exposure, thereby helping toconfirm and strengthen our understanding of the exposure measurements

The direct approach consists of direct measurements of exposures of the general population

to the pollutants of concern, such as the U.S Environmental Protection Agency’s (USEPA) TotalExposure Assessment Methodology (TEAM) studies In a TEAM study, a representative randomsample of the population is selected based on a carefully planned statistical design using theprobability sampling methods outlined in Chapter 3 Then, for the particular pollutant (or class ofpollutants) under study, the pollutant concentrations reaching the respondents selected according

to this statistical design are measured from all sources and for all relevant carrier media A sufficientnumber of people are sampled using statistical sampling techniques — sometimes called a multi-stage probability design — to permit inferences to be drawn, with known precision, about theexposures of the larger population from which the sample is drawn Often a stratified randomsample is used to select a greater number of those persons who are of special interest from anexposure standpoint (for example, long-distance commuters if a vehicular air pollutant is underinvestigation), and then sample weights are used in the subsequent data analysis to adjust for theover-sampling From statistical analyses of the diaries (activities and locations visited), it often ispossible to identify the likely sources, microenvironments, and human activities that contributemost to pollutant exposures, including both traditional and nontraditional sources Many of theselarge-scale exposure field studies, such as the TEAM studies (Table 1.1), have been made possiblebecause of new compact personal exposure monitors (PEMs) capable of measuring exposure tosome air pollutants with high precision (see Chapters 5– and Chapter 15) An exposure field study

of this kind typically samples 25 to 800 respondents and includes the following four basic elements:

1 Use of a representative probability sample of the population

2 Direct measurement of the pollutant concentrations reaching these people through allcarrier media (air, food, water, and dermal contact)

3 Direct measurement of body burden to infer dosage

4 Direct recording of each person’s daily activities through diaries

For volatile pollutants that are readily absorbed and given up by the blood, it is possible todetermine with reasonable precision the concentration of that pollutant present in the blood byasking respondents to exhale their deep-lung breath into a sampling bag and then to analyze thecontents of the bag This is basically the same method that is used by police officers who stopdrivers on the highway to measure the concentration of their blood alcohol using a breath analyzertest Breath measurement works satisfactorily for a number of other pollutants of interest besidesalcohol (ethanol) that are absorbed by the blood (for example, CO and benzene) The resulting

Exposure Analysis: A Receptor-Oriented Science 9

concentrations of the pollutant in the breath can be used to assess the concentration of the pollutant

in the blood and is a useful measure of body burden at that time Measurements of body burden

in some bodily fluids, such as breath or urine, have the advantage that they provide a useful indicator

of previous exposure and are noninvasive — they do not require a sample of blood or tissue fromthe person In addition, pharmacokinetic models have been developed that express the mathematicalrelationship between prior exposure and predicted blood and breath concentrations

Although the final sample sizes for some of the TEAM studies in Table 1.1 were not extremelylarge, the representative probability designs of the larger studies enable the analyst to makeinferences about the exposures of the much larger populations from which these samples weredrawn, such as the population of an entire community The TEAM studies were the first to showfor large populations that indoor sources of toxic chemicals not only greatly outnumber outdoor

Pesticides, House dust 1986–87 200

Springfield, MA (3 seasons) Pesticides 1987 100

Riverside, CA Particles, metals 1991 178

Toronto, Ontario 2 Particles, metals 1995–96 750

Indianapolis, IN 2 Particles, metals 1996 240

1 Study with personal monitors by Alyeska Pipeline Service Co through research

contractors.

2 Approach similar to TEAM by Ethyl Corp under a contract with Research Triangle

Institute, NC.

sources but also cause the bulk of human exposures, often two to five times the amount caused byoutdoor sources (Figure 1.4) For example, tetrachloroethylene (sometimes called perchloroethyl-ene, or PERC) comes from clothes recently brought home from the dry cleaners; chloroform isemitted by water boiled during cooking and released when showering; bathroom and kitchen

deodorizers and mothballs are sources of para-dichlorobenzene (p-DCB); solvents stored in the

home and cleaning fluids are sources of 1,1,1-trichloroethane and tetrachloroethylene; automobilesand gasoline stored in attached garages emit benzene; aerosol sprays release a variety of organic

compounds, including p-DCB; other organic pollutants arise from paints, glues, varnishes,

turpen-tine, and adhesives, as well as furniture polish, carpets, and building materials Finally, pesticidesand herbicides stored in homes or applied indoors can be important sources of personal exposure,

in addition to house dust tracked in from outdoors There is a greater likelihood of receiving a highexposure from sources that are nearby (usually indoors), while the lower probability of receiving

an elevated exposure from a traditional source suggests that the outdoor sources depicted in Figure1.4 should be shown in the far distance and should be smaller than they are now depicted.Indoor exposures are significant for a number of reasons A great variety and number ofchemicals are present indoors close to people, and the small physical scale (under 30 m) limitsdilution, causing relatively high concentrations Also, the infiltration of outdoor air into indoor

settings is relatively slow, causing a low air change rate and a long residence time — the time

required for the volume of air equal to the volume of the indoor setting to be fully replaced Airchange rates in homes typically range from 0.2 to 1 air change per hour, giving residence times(the reciprocal of the air change rate) of 1–5 hours These factors — intensity and variety of sourcescombined with low dilution and low air change rate — make indoor air pollutants especially

important contributors to the total exposure of occupants Finally, activity pattern, or time budget,

studies based on large representative probability samples in which participants completed 24-hourdiaries show that people spend most of their time indoors (Klepeis et al 2001), as discussed briefly

in the following sections All these factors combine to make indoor air pollution an importantcontributor to total exposure and to total risk

1.4.1 I NDIRECT A PPROACH

The direct approach described above is valuable for measuring directly the distribution of exposures

across a population by using personal exposure monitors to measure air exposure and other methods

to measure dermal, food, and drinking water exposure, making it possible to draw inferences from

a representative sample to a much larger population By contrast, the indirect approach seeks to

measure and understand the basic relationships between causative factors and the resulting

expo-sures in the small settings called microenvironments, such as homes or motor vehicles By combining

microenvironmental air pollutant concentrations with diary-based activity patterns in an exposuremodel, it is possible to predict population exposure distributions Similar approaches are applicable

to dermal exposure

Thus, the indirect approach is basically a modeling approach combining measurements fromsmall-scale microenvironmental experiments with large-scale activity pattern data in order to predictpopulation exposure distributions The resulting exposure model is intended to complement thedirect approach field studies by helping to translate their findings to other locales and to othersituations These models also are designed to allow the analyst to consider “what if” questions and

to predict how exposures might change in response to different policies and regulatory (andnonregulatory) actions Exposure models are not the same as the traditional outdoor pollutanttransport models for predicting outdoor concentrations in ambient air, surface water, or groundwater.Rather, they are designed to predict human exposure of a rather mobile human being (see Chapter

19 and Chapter 20) Thus, they require information on typical personal activities, locations visitedduring the day, and time budgets of people, as well as information on the likely concentration

Exposure Analysis: A Receptor

FIGURE 1.4 Examples of traditional (outdoor) and nontraditional (indoor) sources of exposure, based on findings from the total exposure measurement studies.

(From Ott 1990 With permission.)

Pesticide Application

Toxic Wastes, Chemicals

Fresh Dry Cleaning

Bathroom Cleaners, Room Deodorizers, Hair Products

Outdoor Indoor

Lighter Fluid, Solvents, Cleaning Fluids

Automobile Exhaust, Oil, Gasoline, Antifreeze

Insecticides, Herbicides, Pool Chemicals

Furniture Polish, Furniture Stuffing, Paneling

Cigarette Smoke

Paint, Glue, Varnish, Turpentine

distributions in the places the people spend their time (ordinary microenvironments such as homes,motor vehicles, stores, restaurants, schools, etc.).

An example of an early human activity pattern-exposure model was the Simulation of HumanActivities and Exposures (SHAPE) model, which was designed to predict the exposures of thepopulation to CO in urban areas (Ott 1984) The SHAPE model used the CO concentrationsmeasured in microenvironments in Denver and Washington, DC, to determine the contributionsfrom commuting, cooking, cigarette smoke, and other factors Once an exposure model is validated

by demonstrating that it accurately predicts exposure distributions measured in a total exposurefield study, then, in theory, it should be possible in a new city to make a reasonably accurateprediction of that population’s exposures using the new city’s data on human activities, travel habits,and outdoor concentrations USEPA has developed an exposure model for particulate matter that

is similar to the SHAPE model, the Stochastic Human Exposure and Dose Simulation PM) computer model (Burke, Zufall, and Özkaynak 2001) A SHEDS model also has been devel-oped for predicting the exposure of adults and children to pesticides Another modeling approach,called the Random Component Superposition (RCS) model, uses empirical exposure measurements

(SHEDS-to predict population exposure distributions in cities (Ott, Wallace, and Mage 2000) The RCSmodel is a statistical approach that describes the difference between indoor, outdoor, and personalexposures Other statistical approaches treat short-term and long-term exposure distributions(Wallace, Duan, and Ziegenfus 1994, Wallace and Williams 2005)

1.5 RECEPTOR-ORIENTED APPROACH

There is a fundamentally different conceptual approach between the new science of exposureanalysis and traditional approaches used historically in the environmental fields To illustrate thisdifference intuitively, consider an Agatha Christie murder mystery story At some point in the tale

— usually fairly early in the plot — we usually encounter a body, the victim of murder Theprincipal investigator — Monsieur Hercules Poirot — follows various clues and traces eventssurrounding the murder to try to discover the cause of the crime and to identify the murderer.Like Sherlock Holmes stories and like real crimes, each Agatha Christie story begins with thevictim — a dead person — and then works backward to find the cause of the crime (that is, thesource of the mortal wound) If the death occurred by gunshot, M Poirot may analyze bulletfragments to find out more details about the gun Samples may be collected from the victim’s bodyand at the crime scene to try to find the possible killer, the murder weapon, and any other factorsthat help explain the crime All these investigative efforts seek to find the facts surrounding thefiring of the gun, which is analogous to the emission of a pollutant from its source

The same approach that is used to investigate a crime — working backward from the victim

— applies to the exposure sciences By making appropriate measurements very close to the person,the exposure analyst tries to determine if the person is being exposed to a particular pollutant Bymaking measurements of body burden (breath, blood, etc.), the exposure analyst tries to determine

if a person has been exposed to a particular pollutant Similarly, Agatha Christie’s Monsieur Poirotfirst determines if the person has been exposed to a bullet and then seeks to find the source of thebullet Both fields — exposure analysis and criminology — use the same conceptual approach:they work backward from the point of contact to identify the source Thus, both techniques arereceptor-oriented

Many of the environmental regulatory approaches, by contrast, proceed in the opposite direction.They use a source-oriented rather than receptor-oriented approach Most environmental laws areconcerned with emissions or with effluents They do not, for example, begin with the people andattempt to see whether these pollutants actually are reaching any members of the general public.One probable reason is that the framers of these laws were overly impressed by the appearance ofleaking storage drums, smokestack emissions, effluents discharged into waterways, and othertraditional images However, as progress has been made in reducing source emissions, effluents,

Exposure Analysis: A Receptor-Oriented Science 13

leaking underground storage tanks, etc., it becomes extremely important to determine whetherpollutants actually still are reaching the people, by how much, and through what routes That is,

it becomes extremely important to measure the actual exposures that the population receives veryclose to the body and to answer the four basic questions mentioned above concerning: (1) thenumber of people exposed, (2) the level of their exposure, (3) the causes of their exposure, and (4)the manner in which exposure can be reduced most efficiently

A source-oriented crime approach — unlike the Agatha Christie and Sherlock Holmes stories

— would not originate with a dead body Rather, it would originate by discovering a gun and itsspent cartridges All we would know is that this gun had been fired, but we would not know wherethe bullets went or whether they actually hit anyone Even though there was no body — or evenany evidence that a bullet had injured anyone — the source-oriented approach would try to tracethe path of the bullets forward to find out if anyone was hit by a bullet Thus, we would investigateeven if we did not know if a crime had been committed Conducting source-oriented criminologywould be expensive to society, because the police would have to investigate carefully the firing ofevery gun, even if no person were known to be hit As readers of fiction, we probably all wouldfind novels based on a source-oriented crime approach to be rather boring — much ado aboutnothing Yet, it appears that much of the environmental regulatory community has embarked on asource-oriented approach Usually, the process of tracing the pollutant to its receptor in a source-oriented approach is so difficult that the investigator runs out of resources before discovering who

is exposed, if anyone

The public health community, like criminology and exposure analysis, also follows a oriented approach Many harmful viruses exist in the environment If a harmful virus has beenidentified, public health officials usually immediately ask, “Has anyone been exposed (i.e., comeinto contact) with the virus?” Simply knowing a virus is present somewhere in the environment isnot assumed automatically to cause a threat to public health Someone must be exposed first Once

receptor-it is known that an exposure has occurred, public health officials usually next ask, “How manypeople were exposed?” “What was the cause of their exposure?” “How can exposures be reduced?”Only by knowing the extent of the exposure and the causes of the exposures can public health andenvironmental officials intervene to reduce exposures in the most effective manner possible andthereby reduce public health risks

The importance of the receptor-oriented approach in all these fields is obvious It is not possible

for one to get sick from a virus unless an exposure has occurred Similarly, it is not possible to

die from a bullet unless the bullet makes contact with a person (that is, the victim must be exposed

to the bullet by being shot) Similarly, without exposure to an environmental pollutant, one cannotexperience adverse health effects More importantly, changing one’s exposure often may be achievedmore effectively through nonregulatory approaches than by reducing some sources directly throughregulations Often those sources that can be changed by making regulations contribute the least toone’s personal exposure Both regulatory and nonregulatory approaches should be considered forreducing human exposure to pollutants The public health practitioner uses public information, as

in the war against AIDS, to advise people to modify their activities and thereby to reduce their

exposures Once the pollutant enters the body, however, it becomes a dose, and little can be done

to reduce its effects For example, one cannot usually take a pill to reduce the effects of one’sexposure The best approach is to reduce one’s exposure first Of course, like Agatha Christiesolving a crime, it is not possible to reduce exposures unless we know they have occurred, sodetermining population exposures by direct measurement and attempting to answer the four ques-tions mentioned above are essential steps for protecting public health

An important scientific contribution to the field of exposure analysis was the development andsuccessful implementation of the TEAM studies (Wallace 1987; see Table 1.1), followed a decadelater by the National Human Exposure Assessment Survey (NHEXAS) studies (Whitmore et al.1999; Clayton et al 1999; Robertson et al 1999) These total human exposure field studies combineprobability sampling with direct measurement of exposure Conceptually, exposure is measured in

the food eaten, water drunk, skin surface contacted, and air breathed using small, portable personalexposure monitors (Chapter 5), and body burden is measured in breath and urine For the pollutantsthat are not present in food or water or soil, air usually is the main route of exposure By using agood statistical design (Chapter 3), inferences can be made from a carefully designed sample ofrespondents to larger populations of the community By using direct measurement of exposure,important discoveries are possible that cannot be made with predictive models alone, at least notwith the present state of the art of modeling.

1.6 INDOOR AIR QUALITY

Why is indoor air quality important in the exposure sciences? A receptor-oriented approach focuses

on humans and on measuring pollutants very close to the body The monitors must be located veryclose to the person — preferably carried by people Large-scale field studies in which people recalltheir locations and activities through diaries show that people usually are indoors — either at home,

at the office, or in a motor vehicle — with very little time spent outdoors An activity patterndatabase of 1,579 California adults (Jenkins et al 1992) — a representative sample of the state’spopulation — was analyzed using custom-designed computer programs to find where people arelocated over 24 hours (Figure 1.5) Examining the proportion of the California population located

in each of six microenvironments by time of day (beginning and ending at midnight with time onthe chart moving from top to bottom) shows that the most frequently occupied location forCalifornians is indoors at home, accounting for approximately 15 hours on the average The largeproportion of adults found inside a vehicle by time of day is striking when it is compared with theproportion of adults found outdoors For all but 2 hours of the day (10:00 A.M and 11:00 A.M.),the proportion of California adults inside a vehicle is greater than the proportion of adults outdoors.This proportion varies considerably with the time of day At 5:00 P.M., the maximum differencebetween these two locations occurs: 18% of the adults are in a vehicle but only 8% are outdoors

Comparing the time budgets for these three locations — indoors, outdoors, and inside a vehicle

— shows that outdoors accounts for the smallest amount of total time (Table 1.2) and is evensmaller than the average time spent in a motor vehicle Despite the widely held view that Califor-nians are outdoors much more than people living elsewhere in the United States, the differencesbetween California and the United States are extremely small (Jenkins et al 1992; Klepeis et al.2001) These diary studies show that Californians aged 12 and older were outdoors 6% of the day(86 minutes), while the U.S population was outdoors 7.6% of the day (109 minutes) On average,both the U.S and California population was indoors 87% of the day, and members of the U.S.population were indoors or in a vehicle more than 92% of the day

The graph showing the locations of Californians by time of day was plotted so that time over the24-hour period from midnight to midnight flows from top to bottom (Figure 1.5) An alternativeapproach is to plot time over 24 hours from left to right (Figure 1.6), and fortunately diary-basedactivity pattern data similar to that in California are available for the entire nation The NationalHuman Activity Pattern Survey (NHAPS) study was a probability study of a sample of 9,386 personsselected to represent the entire U.S population (Klepeis et al 2001) As with the California activitypattern study, more people were found to be indoors at home (top, lightly shaded labeled

“Residence–Indoors” in Figure 1.6) than in any of nine other locations throughout the day Figure1.5 and Figure 1.6 are essentially the same type of 24-hour diurnal graphs with the axes interchanged,except that Figure 1.5 is for California adults and Figure 1.6 is for the entire U.S population

In addition to spending a large amount of time indoors, especially in their homes, people breatheair pollutant concentrations in indoor settings that generally are higher than those measured out-doors The results from the TEAM studies of CO (Akland et al 1985; Hartwell et al 1984; Johnson1983; Ott et al 1988; Ott, Mage, and Thomas 1992; Wallace et al 1988b; Wallace and Ott 1982;see Chapter 6), volatile organic compounds (VOCs) (Pellizzari et al 1987a,b; Wallace and O’Neill

Exposure Analysis: A Receptor-Oriented Science 15

FIGURE 1.5 Proportion of the California adult population present in each of six locations by time of day,

based on a computer analysis of the diaries from a stratified representative random samples of 1579 persons aged 18 or older throughout the state (From Ott 1995a With permission.)

1987; Wallace 1986, 1991a,b, 1993a,b,c, 1997, 2001; Wallace et al 1982, 1984, 1985, 1986,1987a,b,c, 1988a, 1989; Zweidinger et al 1982; see Chapter 7), particles (Clayton et al 1993;Özkaynak et al 1996; Pellizzari et al 1993a,b; Thomas et al 1993b; Wallace 1996; Yakovleva,Hopke, and Wallace 1999; see Chapter 8), and pesticides (Whitmore et al 1994; see Chapter 15),

as well as various indoor air quality studies (Thomas et al 1993a) show that indoor concentrationsare generally higher than outdoor concentrations for a large number of pollutants (Table 1.3) Thistable lists all those pollutants for which the average concentration indoors has been found to behigher than the corresponding average concentration outdoors The main reason for the higherindoor concentration is the presence of sources indoors A number of these same pollutants arefound on California’s list of toxic air pollutants under Proposition 65, a referendum passed by thevoters that requires products containing toxic pollutants to be labeled To place these findings inperspective, the pollutants in Table 1.3 are grouped into several different categories: criteria airpollutants, pesticides, toxic air pollutants, and others

The 3-year Valdez Health Study was interesting because it focused on a single large pointsource, the Alyeska Marine Terminal (Yocum, Murray, and Mikkelsen 1991; Goldstein et al 1992).This large petroleum storage and loading terminal was approximately 3 miles from the 3600residents living in Valdez, AK As with the TEAM VOC studies, personal monitoring of the residentsshowed that the contribution of indoor sources and personal activities to benzene exposures wasmuch higher than the contribution of the marine terminal, accounting for 89% of the total exposure

to benzene in this community The Marine Terminal accounted for only 11% of the populationexposures, showing the importance of everyday sources that are close to people vs distant pointsources, even if their total emissions may be large

FIGURE 1.6 Stacked plot showing the weighted percentage of respondents in the U.S in each of 10 different

locations by time of day The original minute-by-minute diary data have been smoothed for clarification As with the California activity pattern data, the percentage of people outdoors (either at “Residence–Outdoors”

or “Near Vehicle (Outdoors)” or in “Other Outdoor” locations) is relatively small, though it varies in a diurnal pattern When people were outdoors, they were most likely to be outdoors at their residence (“Residence–Out- doors”), while the most likely location for finding anyone indoors was at their residence (“Residence–Indoors”), which varied from 96% of the time to about 36% over the 24-hour period (From Klepeis et al 2001 With permission.)

Near Vehicle (Outdoors) Mall/Store