"The veterans have legitimate complaints about their radiation dose reconstructions," said John E. Till, chairman of the National Research Council committee that prepared the report.
"Doses were often underestimated, sometimes considerably," he said Thursday at a briefing accompanying release of the report. "Many key assumptions and methods used are not appropriate ... many calculations are illogical or not explained."
Updating the estimates might make a difference in some cases, Till said. However, he added that the committee does not believe it would significantly affect the number of awards made.
"Perhaps a few more veterans who filed claims in the past would have been compensated" if the estimates had been more credible, he said.
Veterans with any of 21 types of cancer are already automatically granted compensation, and the dose estimates are used to determine payment for veterans with some other illnesses and types of cancer.
If the program is continued, improvements need to be made, said Till, adding that Congress and the Defense Department need to take a hard look at if and how it should go forward.
"The program itself is based on some good science," Till said, but quality assurance has been very weak and the department has done a poor job of communicating with the affected veterans.
Joe Violante, national legislative director of Disabled American Veterans, said his group has serious concerns about trying to use estimates to determine if illness is service connected.
He favors assuming that a disease that may be connected to radiation is service-related in the case of atomic-exposed veterans. The money spent on making exposure estimates could have been better spent helping affected veterans, he said.
Fred Allingham of the National Association of Radiation Survivors said that currently most cancers are now assumed to be caused by the exposure, but there have been few studies of the potential effect of radiation on illnesses such as heart and autoimmune disease.
Between 1945 and 1962 the United States conducted about 200 atmospheric nuclear-weapons tests, often with military personnel observing from shipboard or in trenches or buildings some distance from the blast. Most were exposed to radiation from fallout rather than directly from the blast.
The report concluded that the Defense Department's Defense Threat Reduction Agency should re-evaluate the methods it uses to estimate these doses and urged establishment of an independent system to oversee the program.
The program has assessed about 4,000 exposures, but it takes a fairly high dose to be awarded a claim, said Till, president of Risk Assessment Corp. in Neeses, S.C. Only about 50 claims have been awarded, he said.
The committee randomly picked 99 cases for review, including military personnel who were either involved in nuclear-weapons tests in New Mexico, Nevada and the Pacific; prisoners of war in Japan when Hiroshima and Nagasaki were bombed; or personnel stationed in Japan after the war.
Film badges that record radiation were the main system to estimate doses of radiation, but many of the badges have been lost and not all participants wore them or kept them on at all times, the committee noted.
The defense agency estimates the possible range of radiation that a veteran making a claim might have been exposed to and then reports to the Veterans Administration an estimate set at 95 percent of the highest possible exposure, called the upper bound dose. That dose is then used to calculate the likelihood that the illness is related to the radiation.
In many cases the upper bound doses were not calculated correctly and are lower than they should be, Till said.
In addition, the report said, a lack of quality control over dose reconstructions made it difficult to determine how doses were calculated. Questionable assumptions about soldiers' location and duration of exposure also contributed to the lower estimates.
The council is an arm of the National Academy of Sciences, a private organization chartered by Congress to advise the government of scientific matters.