1. Epidemiological studies. They analyze disease patterns in exposed versus unexposed populations.

2. Dose reconstruction studies. These attempt to determine the risk of disease by estimating the exposure to a disease-producing agent like radiation.

Epidemiological studies (or "epi" studies for short) may be done without reference to dose reconstruction studies by simply examining disease patterns in populations working in or living near facilities. Epi studies may also be done subsequent to dose reconstruction studies. Dose reconstruction can help epi studies group together exposed populations more precisely and thus reveal disease patterns that may remain hidden without such dose data.

Dose reconstruction studies are not dependent on epi studies. However, since there are often great uncertainties in exposure estimates, dose reconstruction studies by themselves sometimes do not settle the issue of whether the exposures resulted in demonstrable harm to the exposed populations. Sometimes it may be desirable to do an epi study after a dose reconstruction study. This is a difficult decision that depends on a judgment about the prospects of determining with reasonable certainty whether or not one or more of the exposed groups have suffered increased risk of one or more diseases.

There are a number of basic complications that arise in doing both epi and dose reconstruction studies on DOE facilities. They are:

• While some data on the principal radionuclides (such as uranium or plutonium isotopes) generally exist, there are often are no data all on non-radioactive toxic materials and even of many radioactive materials that were used, processed, or incidentally present.
• Populations are often exposed to more than one disease-producing agent, and the synergistic effects of these agents are essentially unknown. Sometimes people are exposed to the same disease producing agent from multiple sources.
• Non-cancer effects of toxic materials are, in many cases, only now beginning to be understood, and are all too often ignored due to lack of knowledge.
• For a variety of reasons, there are usually large uncertainties about the effects of low-levels of exposure.

Studies of off-site populations are generally more difficult than those of workers because there are no direct measurements of doses for such populations. Worker doses can be estimated from a variety of data such as film badge readings, urinalyses, and direct measurements of body burdens of radioactive materials via whole-body or lung counting instruments. For off-site populations such direct measurement data generally do not exist. Therefore, dose reconstruction studies for off-site populations just begin with examination of data regarding releases of harmful substances ("source term" data) as well as measurement of these substances in the environment ("environmental monitoring" data). Source term and environmental monitoring data are complementary and can provide checks on each other. They must be coupled with knowledge of the behavior of the harmful substances in the environment and the manner in which each pollutant reaches people ("pathway analysis").

As if the problems of doing dose reconstruction and epi studies were not complex enough, studies involving DOE weapons plants are typically plagued by additional complications. The lack of sufficient data, poor quality of data, poor record keeping practices, and even data fabrication in some cases, create severe complications. Classification of and restricted access to data frequently pose additional challenges.

What should community groups look for in health studies? And how can you tell what kinds of studies should be done? Sometimes the question is even more basic -- should the studies be done at all or would the money be better invested in medical monitoring, for instance?

There are no simple, general answers to these questions, but there are some guidelines for determining whether the studies are being carried out in scientifically sound manner.

Determining Exposures

It is essential that exposure estimates be based on the raw data, whether this is exposure data for workers, or some combination of environmental monitoring data and source term data for off-site populations. The expression "source-term" means the amount of a harmful substance released by a facility. The quality of the data and its adequacy in establishing meaningful estimates must be carefully assessed. For instance, it was not uncommon until about 1970 to monitor the air at specific locations off-site for only an hour or even less each month. Such monitoring would not be able to detect process releases that are expected to occur only over short periods of time, or accidental releases that by their nature usually occur over short periods. If, moreover, source term data are of poor quality, as they typically were in DOE installations that IEER has studied, then estimation of doses becomes very difficult. (See the accompanying article on the neighbors of the Fernald plant.)

In general, individual or small population doses are far more difficult to estimate than large population doses. In such circumstances it is less difficult and sometimes more pertinent to examine whether then-prevailing or current standards for exposure to radioactive or non-radioactive toxic materials may have been violated. For instance, the average air concentration at the plant boundary for uranium or some other radioactive material might be determined and compared to regulations.

In assessing doses to off-site populations, one should:

1. Examine environmental monitoring data for air, water, and soil as well as source term data over time.
2. Compare environmental monitoring data and derived environmental concentrations (from models) to then-prevailing and current environmental standards. These standards are derived from dose limits, so that it is possible to infer hypothetical doses to a person exposed to such concentrations by comparing measurements with the regulatory concentration limits.
3. Evaluate data for accuracy and completeness:
   • Have all the relevant radioactive materials been measured?
   • Have all the relevant non-radioactive materials been considered?
   • If data are not available for some materials, has there been an attempt to infer the potential magnitudes of the releases or potential air concentrations from other data? If not, has there been some way to assess the importance of the specific material in question to the exposure?
4. Determine whether all exposure pathways have been considered and the relative importance of each has been estimated. The pathways include direct exposure via air and drinking water, exposure in various ways via contaminated soil, exposure via contaminated food, and exposure after concentration in the food chain (as happens, for instance, with iodine-131 or some organic, non-radioactive pollutants).
5. Examine whether non-cancer and synergistic effects have been analyzed or even mentioned.
6. Examine whether the accidental releases from the plant were accounted for and if their effects on maximum exposure been considered.
7. Compare the risk factors that are used for exposure to low-levels of radiation or to non-radioactive toxic materials with the regulatory literature.
8. Look for a discussion of uncertainties both ways -- that is, ways in which doses and risks could be higher and lower than those calculated and whether these uncertainties have been quantified.

A community is often faced with a question of whether an epi study should be done following a dose reconstruction study. This question cannot be answered in general, except in the case of "small" populations, where there are not enough people for a statistically sound study to be carried out. An epi study is not unlike an opinion poll, only it is a poll of the body's functioning. Since there are considerable differences in they way different people respond to disease-producing agents, there must be sufficient numbers of people in an epi study to determine with a reasonable certainty if there is an increased risk.

Second, an epi study is not likely to yield a statistically significant result if there are great uncertainties in the estimates of doses in the exposed populations. If the exposed people cannot be grouped into appropriate dose ranges, then estimation of increase in risk becomes very difficult. This is especially the case if a small proportion of highly exposed people are mixed in with a far larger number of people with relatively low exposure. Studies must structured so as to provide reasonable assurance that exposure to materials for which data do not exist will not vitiate their results.

IEER's experience studying off-site populations by performing dose reconstruction studies indicates that at least for the U.S. nuclear weapons complex, the environmental monitoring records and source term estimates are often too poor to yield satisfactory estimates of doses to exposed populations. It is quite possible to estimate compliance with regulations, but such estimates cannot be satisfactory bases for epi studies. Thus, the prospects for epi studies of off-site populations yielding reliable results are in many cases not very good. This conclusion must be a tentative one since we have not studied the majority of DOE plants.

In sum, while dose reconstruction efforts to determine the broad outlines of potential damage seem worthwhile, epi studies of off-site populations should be approached with much more caution as they are likely to yield inconclusive results and may even yield falsely negative results. That is, they may indicate no harm, when some individuals have, in fact, been harmed. When there are large uncertainties created by lack of sound data, it is important to give the communities the benefit of the doubt.

Epi studies on workers should be done only after a thorough re-evaluation of exposures, including of internal exposures. In our study on Fernald workers, we found that worker dose records did not include doses that could be inferred from urine and lung-counting data that were available. Radon exposures were also not included. (See the accompanying article in this issue.) The DOE and its contractors need to update worker dose records so that estimates of internal doses are included in them Such corrected records would provide a far better basis for worker health studies and health protection.