National Academy of Sciences
National Academy of Engineering
Institute of Medicine
National Research Council
Office of News and Public Information
National Academy of Engineering
Back | Home
News from the National Academies
Date: Dec. 14, 1999
Contacts: William Kearney, Media Relations Associate
Megan O'Neill, Media Relations Assistant
(202) 334-2138; e-mail <news@nas.edu>

FOR IMMEDIATE RELEASE

Hanford Study Design Sound but Data Analysis, Communication Fell Short

WASHINGTON -- A study of people exposed to the radioactive isotope iodine-131 in the Richland, Wash., area was well-designed, but the study's researchers reported the findings as more conclusive than they were, according to a new report from the National Research Council of the National Academies. A number of other errors were made in communicating the findings of the study to the public, as well.

The Research Council report focuses on a draft report of a study conducted by the Fred Hutchinson Cancer Research Center, Seattle, and issued last January. The study examined the likelihood that releases of iodine-131 into the environment between 1944 and 1957 from the Hanford Atomic Products Operation -- a nuclear weapons fuel production facility -- resulted in an excess of thyroid-related illnesses among the local population.

The panel that wrote the Research Council report concluded that the Hanford Thyroid Disease Study investigators chose the most relevant population to study -- those in the most highly exposed areas who were young children at the time of the greatest iodine-131 releases. It also found that the clinical examinations and laboratory studies were performed with good-quality, scientifically valid methods. But shortcomings in the analytical and statistical methods used by the study's investigators overestimated the ability to detect radiation effects, which means the study results are less definitive than had been reported. The investigators assumed a higher degree of precision in the exposure estimates than could be discerned from the data collected. The Research Council panel recommended that, if possible, the investigators re-do the statistical calculations to take into account the uncertainty of some of the data, and interpret the study results in accordance with those limitations.

"We believe that the study's investigators incorrectly assumed that exposure estimates calculated for each person were more precise than they actually were," said panel chair Roy E. Shore, professor of environmental medicine, New York University School of Medicine, New York City. "In fact, the estimates were based on a computer model that had to take into consideration a range of factors, many of which were uncertain. For example, estimates were made with respect to how much iodine-131 was released from Hanford at a given time, which way the wind was blowing, how much of the iodine-131 adhered to vegetation that may have been consumed by cows and transferred to their milk, and how much contaminated milk was consumed by any particular child."

Children in the area were the focal point of the study because they are more sensitive to exposure to iodine-131, which would have occurred largely through consumption of contaminated milk. Exposure to the radioactive isotope has been linked in other studies to the occurrence of benign and malignant thyroid diseases.

In 1988 the Centers for Disease Control and Prevention (CDC), through its contractor, the Hutchinson Center, sponsored the Hanford Thyroid Disease Study -- a nine-year, $18-million effort to determine the health effects of exposure to radioactive iodine-131 releases from the plant between 1944 and 1957. The researchers set out to contact 5,199 people who had been born near Hanford during that time frame. The 3,441 subjects enrolled in the study underwent extensive medical examinations to look for evidence of thyroid disease.

Findings in the study's draft report were leaked to the news media last January, shortly before its intended public release. The draft report concluded that there was no evidence for linking radiation exposure from Hanford to the frequency of thyroid disease found in the study population. While presenting their findings to reporters, the public, and regional citizen groups, the investigators used strong, emphatic language in their comments and press releases, indicating that the results were statistically powerful and leaving little doubt about the correctness of the findings. For the final report, however, the Research Council panel urged the CDC to clearly spell out the imprecise nature of the exposure estimates. Because this information was not presented appropriately in the draft report, the panel felt that the study's investigators overstated the strength of their main finding of no radiation effect.

Usually, scientists provide both a best estimate and a range of estimates -- called "confidence intervals" -- to help interpret statistical results. But the Hanford study investigators, in their draft report and public statements, provided only their best estimates, not the confidence limits, for the extent of possible radiation effects. That made their findings seem more reliable than they actually were.

Radiation Exposure Estimates and Uncertainties Questioned

Most of the factors in the Hanford Environmental Dose Reconstruction (HEDR) computer model used to estimate thyroid exposures to radiation in study participants were reasonable, the panel said, and the predicted exposure levels appeared accurate within the limits often encountered in historical dose reconstruction studies. However, several experts have claimed that an important factor in the model -- the amount of iodine-131 released by Hanford -- was higher than calculated by the computer model's developers. This error would lead to an underestimation of exposure by about 30 percent.

Also likely is that the model factor that determines how much iodine-131 had been transferred from pasture grass to cow's milk was too high, and would have caused overestimation of exposure levels. In addition, some uncertainties in the doses were not acknowledged, particularly the uncertainty of an individual's milk consumption. Model developers should conduct a careful re-assessment of these factors, the panel said, and include all uncertainties.

The panel was critical of the investigators' exclusive use of the HEDR model's estimates of thyroid exposure levels for the data analysis, and suggested supplemental examinations that could help confirm or weaken the study's conclusions. The breakdowns of the radiation effect also were difficult to interpret, since important tables that would have enabled better interpretation of the findings were missing from the draft report.

Although the panel criticized the investigators' analysis of the data, it acknowledged that their efforts to trace and examine the population were excellent and it was unlikely that a more comprehensive study could have been conducted. Because so little information is available on exposures that are now 45 years old, the investigators had quite a challenge before them. "This carefully designed study, with sound follow-up and sound medical methods, has examined a large fraction of the most heavily exposed population and failed to find any obvious evidence of a radiation effect," the panel wrote. "The lack of evidence of a dose-response relationship for any type of thyroid disease in the study suggests, but does not prove, that the overall risks were not affected by Hanford fallout."

Results Poorly Communicated

The release of the Hutchinson Center draft report led to public concern. Many citizens in the region affected by Hanford were upset not only about the findings of the report, but also about the way in which investigators conveyed the study's results. Shortly after the release of the draft report, the CDC asked the Research Council to conduct an independent and comprehensive appraisal of the study methodology and how the results were interpreted and presented to the public.

The Research Council panel was impressed by investigators' efforts to have open communication with the public throughout the nine-year study and to establish a citizen advisory group for the study. All of those early efforts should have helped to build trust and credibility for the study. The panel thought that the communication plan in the draft report for its public release was well-developed and should have been moderately successful if implemented as planned. However, several factors prevented carrying out this communication plan, which focused on releasing a final report after the draft report had been through peer review.

When the draft report was released, a number of communication problems occurred, including the investigators' use of strong statements to characterize their findings. This was inappropriate because that document had not been reviewed outside of CDC and a number of uncertainties remained. Additionally, investigators paid insufficient attention to the community's health concerns and fears, the panel said. The investigators and CDC officials should have offered more balanced, and possibly alternative, interpretations of the findings and discussed their implications for individuals.

A new communication plan should be developed for the release of the final report of the study, the panel said. In the final report and all public documents related to it, any significant changes made from the draft report, and all remaining uncertainties, should be clearly outlined and explained. In addition, a strategy for an advance briefing should be developed to give citizen advisory groups that have participated in the study over the years adequate time to look at the report before its public release.

A small group of risk-communication experts, scientists, journalists, and members of the general public should be convened to consider how to more effectively release and discuss controversial draft reports that have not been through a peer-review process, as well as other issues that could affect the future release of important CDC reports, the panel said.

The panel's work was funded by the Centers for Disease Control and Prevention. The National Research Council is the principal operating arm of the National Academy of Sciences and the National Academy of Engineering. It is a private, nonprofit organization that provides advice on science and technology under a congressional charter. A panel roster follows.


Read the full text of Review of the Hanford Thyroid Disease Study Draft Final Report for free on the Web, as well as more than 1,800 other publications from the National Academies. Printed copies are available for purchase from the National Academy Press Web site or at the mailing address in the letterhead; tel. (202) 334-3313 or 1-800-624-6242. Reporters may obtain a pre-publication copy from the Office of News and Public Information at the letterhead address (contacts listed above).


NATIONAL RESEARCH COUNCIL
Commission on Life Sciences
Board on Radiation Effects Research

Committee on an Assessment of Centers for Disease Control and Prevention
Radiation Studies from Department of Energy Contractor Sites

Subcommittee to Review the Hanford Thyroid Disease Study Final Results and Draft Report

Roy E. Shore, Ph.D. (chair)
Professor, Environmental Medicine, and
Director, Epidemiology and Biostatistics Program
New York University School of Medicine
New York City

André Bouville, Ph.D.
Senior Radiation Physicist
Radiation Epidemiology Branch
National Cancer Institute
Bethesda, Md.

Bruce B. Boecker, Ph.D.
Scientist Emeritus
Lovelace Respiratory Research Institute
Albuquerque, N.M.

A. Bertrand Brill, M.D., Ph.D.
Research Professor of Radiology and Physics
Radiology Department
Vanderbilt University Medical School
Nashville, Tenn.

Patricia A.H. Buffler, Ph.D.*
Dean Emerita
School of Public Health
University of California
Berkeley

Sharon M. Friedman, M.A.
Iacocca Professor and Director
Science and Environmental Writing Program
Department of Journalism and Communication
Lehigh University
Bethlehem, Pa.

Susan E. Lederer, Ph.D.
Assistant Professor
Section of the History of Medicine
Yale University School of Medicine
New Haven, Conn.

Carl M. Mansfield, M.D.
Chairman
Department of Radiation Oncology
University of Maryland School of Medicine
Baltimore

Donald E. Myers, Ph.D.
Professor of Mathematics
Univesity of Arizona
Tucson

Daniel O. Stram, Ph.D.
Associate Professor
Department of Preventive Medicine
University of Southern California
Los Angeles

Robert G. Thomas, Ph.D.
Private Consultant
Folsom, Calif.

RESEARCH COUNCIL STAFF

Steve Simon, Ph.D.
Study Director

* Member, Institute of Medicine