HANFORD DOWNWINDERS INFORMATION SITE

Understanding Health Studies

HERE YOU'LL FIND...
What is Epidemiology?
What Has Epidemiology Accomplished?
Downwinder Perspective
Important Challenges for Studies
on Radiation Health Effects
Four Key Elements in
Understanding Epidemiology
The Hanford Experience
Summary
For Further Reading

Epidemiology is the study of disease in human populations. This branch of science has been essential in saving the lives of millions of people by discovering the causes of many diseases. By identifying what causes certain diseases, epidemiology has prompted advances in medicine and better ways of controlling and preventing disease.

Many people have called the Hanford Health Information Network over the last several years with questions about whether their exposure to Hanford radiation caused their health problems. As will be discussed later, epidemiology cannot presently answer those questions on an individual basis but only on a group or population basis. However, epidemiology can help to determine whether exposed populations have experienced more health problems than they would have under normal circumstances, that is, without the exposure.

Epidemiology in the nuclear arena is nearly always controversial. With regard to Hanford, epidemiologic studies are viewed by some as a way to prove that certain health problems were caused by Hanford's radiation releases. For other people (including some scientists), epidemiologic studies of populations downwind or downstream of Hanford are viewed as a waste of taxpayer funds. Still others feel distrustful of science because so many studies in the past have provided no link between radiation exposure and health problems. However, some people see in epidemiology a promise that at least some of their questions might be resolved.

It is the purpose of this report to explain the basics of epidemiology. Readers might find this information useful when reading news stories about study findings or more formal reports on epidemiologic studies, or when discussing studies with scientists and government agencies.

In this report, we will explore:

what epidemiology is and its importance,
the challenges in doing epidemiologic studies,
the significance of some of the key terms, and
the epidemiologic studies that have been done around Hanford.

This report is an introduction to epidemiology. It focuses on the possible health effects of environmental exposures to low-level radiation, especially as they relate to Hanford. It does not deal with exposures to workers nor exposures from toxic chemicals. Nor does it consider the study of health effects resulting from high-level radiation exposures.

What Is Epidemiology?

Epidemiology is the study of disease in human populations. Environmental epidemiology is the study of the ways things in the environment can be factors in causing disease. To study diseases and their causes, epidemiologic studies use scientific methods. Such studies are commonly referred to as disease studies or health studies.

Many exposed people have raised the question: "Did my exposure cause my health problems?" Science cannot currently answer that question. What epidemiology can try to answer is: "Are there more health problems among those exposed because of their exposure to radioactive materials?"

The basic approach of epidemiology is to compare groups of people. In doing so, it explores whether an association, or a link, exists between exposures and health effects. The comparisons are usually done by placing people into categories. There are two main categories: exposure and disease. A "cohort" study compares groups of people based on exposure. It tries to determine whether disease occurs more frequently or less frequently among a population (or group of people) which has been exposed than among those who have not been exposed.

A "case-control" study compares groups of people based on disease. It examines whether exposure occurred more frequently or less frequently in persons who have a particular disease than in persons who do not have that disease.

What Has Epidemiology Accomplished?

"Epidemiology's major contribution to the understanding of radiation health effects is the identification of 'late effects' of radiation exposure in human populations."¹ "Late effects" are health problems that are caused by an exposure but are not detected for many years, such as cancers or genetic effects.

New insights gained through epidemiologic studies have resulted in better ways to protect people from radiation exposure. For example, pregnant women used to undergo routine X-rays. But an epidemiologic study by Dr. Alice Stewart detected that these prenatal exposures were associated with a higher occurrence of leukemia among the children of those mothers.² Now when a pregnant woman receives medical X-rays, the fetus is protected from the X-rays.

downwinder perspective

Many callers to the Hanford Health Information Lines have questions and concerns about the release of radioactive materials to the Columbia River and possible effects on humans and on fish. Some downwinders have health problems that they believe are, or might be, related to Hanford. The following personal perspective is offered to help readers understand these experiences and concerns.

"Valuable health data can be gathered from ordinary people. Individuals notice when something is amiss. It may start with a group of friends sharing observations around a kitchen table. They eventually tally deaths or illnesses in their area over a period of time. This informal data gathering says, 'Hey, something drastic has happened to us, and we want to know what happened!'

"That concern prompted me to survey the health of my high school classmates who graduated from 1951 through 1954. Results showed something serious had happened. Similar "studies" were coincidentally being done by individuals in other communities. "Hanford Downwinders" have many health problems in common, in addition to cancer and thyroid disease. Unfortunately, full documentation of health effects for those now deceased is largely lost.

"The citizens affected by Hanford's emissions resemble "canaries in the mines." Their stories warn of serious human and environmental consequences for nuclear-oriented countries.

"Formal scientific studies will, of course, reveal important information. But troubling uncertainties exist. Records were made at a time when it was important for findings to support the efforts of the Cold War. And data based on earlier studies funded by the nuclear industry need very critical evaluation before being used.

"A more comprehensive health picture continues to be available from those still living. That information may rival "scientific" study and present a more accurate picture about what long-term, low-level radiation exposure does to people.

"Questions still remain for me. How do we get the most valid picture about all the health effects on citizens who lived even hundreds of miles from Hanford? And what is in store for future generations? From my point of view, those are the questions that studies should answer. Maybe these answers can be found in something other than epidemiological studies.

This perspective was contributed by a downwinder who lived in Northern Idaho during the time of the highest radioactive releases to the air - Name withheld upon request.

Important Challenges for Studies on Radiation Health Effects

There are five main challenges to consider when designing, conducting or evaluating studies of possible links between low-level radiation exposure and groups of exposed people with specific illnesses:

1. Other things can cause the same kinds of health effects that radiation exposure can cause. Although some studies have established a very strong link between radiation and certain types of cancer, radiation is not the only cause of cancer. Leukemia, for example, is one form of cancer that has been found to be associated with exposure to radiation. Yet, there are other things that can cause leukemia.

2. Usually there are no individual dose estimates. Thus far, there have been no epidemiologic studies around Hanford that have used individual dose estimates. The use of such estimates can greatly enhance the validity of a study's findings. Without accurate estimates, it can be much more difficult to assess whether any health problems detected are connected to the exposures.

3. People are exposed to other sources of radiation. The planet we live on is naturally radioactive. Long before scientists discovered radiation, all living things were being exposed to low levels of naturally occurring radiation. In addition to natural sources, people are also exposed to radiation from medical and dental procedures, consumer goods (such as tobacco products) and fallout from nuclear weapons testing. The current prevailing scientific view is that even the smallest exposure to radiation has the potential to cause a health effect.

4. People are different and can change. People move around, eat different foods, have different lifestyles and genetic backgrounds, and change their habits over time. All of these factors can directly or indirectly influence their health and the study of their health.

5. Health effects from low-level exposures cannot be detected immediately. There are long delays between the time of exposure and the time when a health effect occurs. This period of time is called the latency period. The length varies among diseases and among individuals. For leukemia, the latency period is as short as five years. For thyroid cancer, it will usually take at least five years or so before the cancer can grow large enough to be diagnosed as cancer. Most thyroid cancers would be expected to appear within 10 to 20 years following exposure. For some people, the delay could be much longer.

There are no known limits as to how long a latency period can be. It is possible that health effects can occur many decades following a harmful exposure. For genetic effects, it may be several generations before an effect shows up.

These five factors are significant challenges to scientists and citizens in trying to determine possible health effects from radiation exposure. Some scientists have the luxury of examining problems under a microscope in the controlled setting of a laboratory. The scientists doing environmental epidemiology have a harder job. The factors listed above help determine what an epidemiologist can and cannot do and why the results of epidemiologic studies are sometimes inconclusive or inconsistent. For citizens concerned about the same questions, it is important to understand the challenges in doing epidemiologic studies.

These different characteristics of the study population can sometimes work as "confounding factors." Confounding factors can mask an effect so that the relationship of the health problem and the exposure is not recognized. They can also make it appear as though there is an effect when, in fact, none exists.

Four Key Elements in Understanding Epidemiology

There are four key elements in an environmental epidemiologic study. These are by no means a complete list, but they are the most essential elements for understanding environmental epidemiology in the context of Hanford's relatively low-level radiation exposures.

1. The Nature of Epidemiologic Evidence

Many people who were exposed to radiation from Hanford want to know whether the health problems they, or their families or friends, have experienced were caused by radiation. Unfortunately, this kind of question cannot be answered with certainty. As mentioned above, the kinds of health effects that may be caused by low-level radiation have other causes as well. These other causes are usually unknown, and there is currently no way to determine what caused a particular case of disease. This is why epidemiologists must learn about the effects of radiation by finding out whether diseases happen more frequently among groups of exposed individuals than among unexposed groups.

Consider, for example, Dr. Alice Stewart's study of prenatal X-ray exposure and childhood leukemia mentioned above. Leukemia sometimes occurs among children who were not exposed to X-rays in the womb. Stewart showed that leukemia happened more often in children with prenatal X-ray exposures compared to similar children without such exposures. In other words, there were "extra" leukemias among the exposed children. Stewart had no way to identify which particular cases of leukemia were the extra ones: leukemias caused by radiation cannot now be distinguished from those having other causes. However, the evidence of extra leukemias among the children exposed to prenatal X-rays was strong enough for Stewart to conclude that the risk of leukemia was greater among the exposed children. In epidemiological terms, she found evidence that risk of leukemia was associated with, or related to, the radiation exposure.

2. Statistics

How was Stewart able to say that the evidence was "strong enough" to base her conclusions on it? Epidemiologists use statistics, a branch of mathematics, to analyze the information collected in a study. There are two aspects of statistics that are important in understanding epidemiologic studies: significance and power. Statistical power is evaluated during the design phase of a study. The calculations of statistical power help determine whether and how large a study should be done. Statistical significance, on the other hand, is assessed at the end of the study, after the results are known.

"Statistical significance" is the likelihood that the results found could not have occurred by chance alone. The association of disease risk with radiation exposure in a study is said to be statistically significant if the association is so strong that it is unlikely to have occurred simply by chance. If, on the other hand, the observed association is not that strong, then the association could have occurred by chance and therefore does not provide convincing evidence that risk was increased by the radiation exposure.

"Statistical power" measures the ability of a study to find an association between radiation and disease, when such an association actually exists. If exposure to radiation does indeed increase the risk of disease, then a study with high power will be very likely to find an association. However, if the study has low power, then it has little chance of finding an association even if there is an actual association (that radiation causes an increase in risk).

The power of a study is determined by several things, including the number of people who participate in the study and the sizes of the radiation doses they received. If the number of participants is small, or if the radiation doses are all low, then the study will have little, or low, power.

Achieving adequate power is an important goal in the planning of an epidemiological study. A study with low power is not worth doing. Even if the radiation exposure greatly increases the risk of the disease, the study will have too little chance of producing the correct finding of an association between exposure and disease.

The Hanford Thyroid Disease Study illustrates the role of statistical power in planning epidemiological studies. In the pilot study, the thyroid study estimated the radiation doses to the thyroid from Hanford's iodine-131 releases for over 800 people. Using this information about doses, study scientists then calculated that about 3,400 people will need to participate in the total study for it to have adequate statistical power.

Knowing the study's power also helps with interpreting the results of the study. If a study with high power finds no significant association between radiation exposure and disease, then it is unlikely that the exposure causes large increases in disease risk. However, if no significant association is found in a study with low power, then the results are inconclusive: there is no convincing evidence that an association exists, but the possibility that radiation increases risk cannot be ruled out.

3. Dose Response

Dose response is the term used to describe the part of an epidemiologic analysis that examines whether there is a relationship between the disease rate and the dose level. If a study finds that the higher the dose, the higher the rate of disease, then it is more likely the radiation dose caused the disease.

In order to do a dose-response analysis, scientists need dose estimates of those being studied. The importance of individual dose estimates has become more widely recognized in recent years. In 1991, Bernard Shleien and two officials of the Centers for Disease Control and Prevention (CDC) did a review of about 50 epidemiologic studies around nuclear facilities. They found that the "most serious problem...was the absence of quantitative estimates of radiation dose that could be used to assess dose-response relations."³ The studies used geographic location of people as a substitute for a dose estimate. Because a person's radiation dose is determined by many factors, geographic location alone is a weak indicator of dose. The solution to such problems, according to Shleien and his colleagues, is for researchers to focus much more attention on getting firm individual dose estimates. Moreover, they "suggest that there is little value in pursuing epidemiologic analyses of radiation-induced cancer without such estimates [of dose]."⁴

In spite of the potential value of individual dose estimates, the reliability of dose reconstruction is itself a controversial issue. Environmental scientist F. Owen Hoffman notes three problems with dose reconstruction: (1) it is an inexact science; (2) the circumstances, procedures and models used are complex; and (3) each dose reconstruction situation "requires the use of an extensive amount of judgment."⁵ For these reasons, Hoffman notes that the results of a dose reconstruction will differ from one investigator to another. To address these problems, Hoffman has proposed that especially important aspects of a dose reconstruction be independently approached by two or more teams of scientists. The particular aspects would vary from study to study.

4. Timing

Timing considers two aspects: (1) the possibility of an exposure causing a health effect, and (2) when to conduct an epidemiologic study.

First, the timing of the health effect in relation to the exposure is important. A health effect cannot be caused by an exposure if the health effect is detected before the exposure occurs. The health effect must occur after the suspected exposure and be within the latency period for such a health effect.

Second, a study must consider timing to account for "late effects." In order for a study to measure the "late effects" that might be related to a particular exposure, that study should take into account the latency period(s) involved. For example, if there were to be a thyroid cancer study of a population exposed to iodine-131, the scientists would not complete it before 10 to 20 years had passed following the exposure. This delay would allow for the latency period to pass so that the study would have a chance of measuring the effect from the iodine-131 exposure.

The Hanford Experience

In this section, we present a survey of the five environmental epidemiologic studies concerning Hanford that have been done to date. Each study is briefly described and discussed in light of the key elements presented above. It is important to note that none of the five studies used individual dose estimates in their analyses. This capability has only recently become available for the Hanford situation. The Hanford Thyroid Disease Study, which is currently underway, is using individual dose estimates.

Two Columbia River Cancer Death Studies

Two studies in the mid-1960s considered whether a link existed between radiation exposure from Hanford and rates of death from cancer. They compared cancer death rates in counties along the Columbia River and the Pacific Coast to those rates in counties not bordering those bodies of water.

Fadeley Study

Description: Robert C. Fadeley reported his findings in 1965.⁶ He looked at cancer death records for the years 1959 to 1964. Fadeley used geographic location as a substitute for exposure. In other words, he used the county (in which a person died) as an indicator of the exposure that person received from Hanford's operations. Fadeley found an increased rate of cancer deaths in Oregon counties bordering the Columbia River or the Pacific Ocean compared with the rates in counties not bordering the river or ocean. He believed that the increased cancer death rate was linked to exposure to Hanford contamination in the river and along the coast.

Discussion: Geographic location is not a strong indicator of exposure. Also, Fadeley did not allow for enough time to elapse for cancers to develop after the years of highest exposure from the Columbia River pathway (1957-1964). According to two scientists who conducted a similar study (see next study description), Fadeley did not account for the difference between cancer rates in people who live in cities and in those who live in rural areas.

Bailar and Young Study

Description: John C. Bailar III, M.D., and John L. Young, Jr., M.P.H., reported in 1966 on a study that was similar to Fadeley's work.⁷ Along with the Oregon counties, they included counties in Washington which border the Columbia River downstream from Hanford. Including more years than Fadeley did, Bailar and Young looked at cancer death rates from 1934 to 1963. They compared the rates for the years before Hanford started to the rates during the years Hanford was operating. Unlike Fadeley, they did not find any increase in rate of cancer deaths.

Discussion: Bailar and Young, like Fadeley, did not wait long enough (after the years of highest releases to the Columbia River) to allow for the latency period of even leukemia, about five years after exposure. They, too, used geographic location as a substitute for exposure.

Study of Birth Defects near Hanford

Description: Lowell E. Sever, Ph.D. and others conducted a study of birth defects in Washington's Benton and Franklin counties near Hanford.⁸ The researchers examined the number of cases of certain birth defects between 1968 and 1980. When the county rates were compared with rates from Washington, Oregon and Idaho, there were more neural tube defects than expected. However, cleft lip was reported less often in Benton and Franklin counties than in the three-state area.

Using information from a study of Hanford workers,⁹ the researchers concluded that the increase in neural tube defects was not explained by parental employment at Hanford or by occupational exposure to radiation. The researchers also concluded it was unlikely that exposure of the general public to radiation from Hanford operations caused the increase in neural tube defects. This conclusion was based on a dose estimate for the years 1974 to 1980.

Discussion: The study's dose estimate for the public only includes the years 1974 through 1980, during which there were limited Hanford operations. It does not include the years of highest releases of radioactive materials, 1944-1965. Also, the study was conducted before any individual dose estimates were available for the Hanford area.

Study of Childhood Leukemia Deaths

Description: John R. Goldsmith looked at childhood (ages 0-9 years) leukemia deaths in Benton and Franklin counties (and in two counties near Oak Ridge, Tennessee, another nuclear weapons facility).¹⁰ The time periods studied were three decades: 1950-59, 1960-69 and 1970-79. Goldsmith compared the actual deaths from childhood leukemia with the number of deaths that would have been expected. He based the expected number on national U.S. rates and the size of the local population.

Goldsmith found a significant excess of childhood leukemia deaths during 1950-59, a non-significant excess during 1960-69 and a deficiency during 1970-79. For Benton County in the 1950-59 time period, there were eight deaths from childhood leukemia when 6.65 would have been expected. For Franklin County in that same time period, there were four deaths when 2.01 would have been expected.

Discussion: Goldsmith did allow sufficient time to account for the latency period. However, as he himself admitted, the analysis by the decades was "arbitrary."¹¹ Goldsmith did not estimate exposures for the people being studied. He did not discuss the power of the study, although it was probably low because of the small numbers of deaths.

National Cancer Institute (NCI) Study

Description: The NCI study compared the deaths from cancer between two sets of counties around 62 nuclear facilities: the first set consisted of three "study" (or exposed) counties and the second set consisted of three "control" (or unexposed) counties. The cancer deaths included in the study occurred between 1950 and 1984. The main finding was that if there was an excess cancer risk present, "it was too small to be detected" with the methods used.¹² The NCI study was nationwide, but this publication describes only the Hanford component.

To analyze the Hanford site, the NCI team selected Benton, Franklin and Grant counties as the study set. Hanford covers parts of each county. Snohomish, Walla Walla and Whitman counties made up the control set.

Discussion: Acknowledging the inherent weakness in its own study, the NCI team provided its own critique: "It does not prove the absence of any effect."¹³ Of particular importance to the Hanford situation, the NCI team pointed out that since thyroid cancer is rarely fatal, it does not appear very often on death certificates.¹⁴

The NCI study did not compare those people actually exposed with those not exposed: "the exposures of individuals are not known.... Persons who lived in particular counties at the time of death may not have been long-term residents. Some residents will have moved elsewhere and died in another part of the country. Some residents of counties that have a nuclear facility may live far from the plant, not be at any risk, and their experience may dilute that of residents living closer to the plant."¹⁵ Additionally, two of the three control counties (Walla Walla and Whitman) are now considered to be downwind from Hanford and were subjected to contamination from the Hanford releases of iodine-131 and other airborne materials.

Information From Death Certificates

Death certificates are usually the main source of information used for analysis of causes of death. There are several problems with doing studies based on information from death certificates:

Death certificates record the number of deaths caused by a disease.
However, not all occurrences of a disease result in death.
Some diseases are rarely fatal, such as thyroid cancer.
So analyses of these diseases cannot be done based on deaths.
The information from death certificates is not always accurate or complete.
Death certificate information doesn't account for all
diseases (such as cancers present at death but not detected or diagnosed).
Some people might have lived most of their lives in other locations,
but the death certificate indicates only their last residence. This becomes a
problem if a study uses residence location as an indication of dose level.

SUMMARY

Epidemiologic studies can help us in trying to understand the possible connections between low-dose radiation and health effects. But the scientific field of epidemiology has limitations. Society cannot expect it to provide all of the answers to situations like the radiation releases from Hanford. It is an important tool, and we need to learn how to use it well.

FOR FURTHER READING

Beebe, Gilbert W. "A Methodologic Assessment of Radiation Epidemiology Studies." Health Physics. April 1984, pp. 745-762.

Legator, Marvin S., Barbara L. Harper and Michael J. Scott, eds. The Health Detective's Handbook: A Guide to the Investigation of Environmental Hazards by Nonprofessionals. Baltimore: John Hopkins University Press, 1985.

Shleien, Bernard, A. James Ruttenber and Michael Sage. "Epidemiologic Studies of Cancer in Populations Near Nuclear Facilities." Health Physics. December 1991 (61), pp. 699-713.

NOTES

1. - Steve Wing, Ph.D. "The Basics of Radiation Epidemiology" (Module 3). Radiation Health Effects: A Monograph Study of the Health Effects of Radiation and Information Concerning Radioactive Releases from the Hanford Site: 1944-1972 (published by HHIN and University of Washington); September 1994, p. 26. Wing is with the Dept. of Epidemiology at the University of North Carolina's School of Public Health.

2. - Dr. Stewart is an M.D. and works as an epidemiologist in the Department of Public Health and Epidemiology at the University of Birmingham's School of Medicine in England.

3. - Bernard Shleien, Pharm. D. et al. "Epidemiologic Studies of Cancer in Populations Near Nuclear Facilities." Health Physics. December 1991 (61), p. 710. Shleien is a health physicist and the president of Scinta, Inc. He also served as a member of the Technical Steering Panel (1988-1994) for the Hanford Environmental Dose Reconstruction Project.

4. - Shleien, p. 710.

5. - F. Owen Hoffman, Ph.D. "Environmental Dose Reconstruction, Approaches to an Inexact Science." It was presented at the Department of Health and Human Services Workshop on Energy-Related Epidemiologic Research Agenda, Atlanta, Georgia, Dec. 3-4, 1991. At the time, Hoffman worked under a Department of Energy contract at Oak Ridge. His field of expertise is in environmental pathways analysis for dose reconstruction projects.

6. - Robert C. Fadeley. "Oregon Malignancy Pattern Physiographically Related to Hanford Washington Radioisotope Storage." Journal of Environmental Health; Vol. 27, No. 6, May-June 1965; pp. 883-897. Fadeley was Director of Research of the Foundation for Environmental Research in Golden, Colorado.

7. - John C. Bailar III, M.D. and John L. Young, Jr., M.P.H. "Oregon Malignancy Pattern and Radioisotope Storage: A Reappraisal." Public Health Reports; Vol. 81, No. 4, April 1966; pp. 311-317. Both authors were with the Biometry Branch of the National Cancer Institute, Public Health Service.

8. - Lowell E. Sever, Ph.D. et al. "The Prevalence at Birth of Congenital Malformations in Communities near the Hanford Site." American Journal of Epidemiology; Vol. 127, No. 2, 1988, pp. 243-254. At the time of this study, Sever was with the Division of Birth Defects and Developmental Disabilities, Center for Environmental Health, Centers for Disease Control in Atlanta, Ga. The work was supported by the US Department of Energy.

9. - Lowell E. Sever, Ph.D. et al. "A Case-Control Study of Congenital Malformations and Occupational Exposure to Low-Level Ionizing Radiation." American Journal of Epidemiology; Vol. 127, No. 2, 1988, pp. 226-242.

10. - John R. Goldsmith. "Childhood Leukaemia Mortality Before 1970 Among Populations near Two US Nuclear Installations." The Lancet; April 8, 1989, p. 793. See also letter by Samuel Milham, Jr., and Goldsmith's reply in The Lancet; June 24, 1989, pp. 1443-1444. At the time of the article, Goldsmith was with the Epidemiology and Health Services Evaluation Unit, Occupational Epidemiology Section, Faculty of Health Sciences, Ben Gurion University of the Negev in Beer Sheva, Israel.

11. - Goldsmith, The Lancet; June 24, 1989, p. 1444.

12. - Seymour Jablon, M.A. et al. "Cancer in Populations Living Near Nuclear Facilities: A Survey of Mortality Nationwide and Incidence in Two States." Journal of the American Medical Association (JAMA); Vol. 265, No. 11, March 20, 1991, pp. 1403-1408. Jablon and his co-authors are with the Radiation Epidemiology Branch, Epidemiology and Biostatistics Program of the National Cancer Institute (NCI).

13. - Jablon, p. 1403.

14. - Seymour Jablon, M.A. et al. Cancer in Populations Living Near Nuclear Facilities. Bethesda, MD: Public Health Service, Department of Health and Human Services; 1990. National Institutes of Health publication 90-874, Vol. 1, p. 18.

15. - Jablon, JAMA; March 20, 1991, p. 1407.

TOP OF PAGE

SITE INDEX

Click Here to Search This Site

Hosted by www.Geocities.ws