Three categories of radiation data were collected for workers at Fernald:

1. Direct measurements of worker external radiation doses. (Collected using film badges worn by workers, for instance.)
2. Measurements of radioactive materials inside workers' bodies. The methods included analyzing urine samples and measuring gamma radiation emanating from radionuclides trapped in workers' lungs (called "lung counting").
3. Measurements of radioactivity in the workplace environment. These are made by sampling the air in the general area where the work is done and in the "breathing zones" close to workers' faces.

The third category is not a direct measurement of dose but provides an indication of working conditions leading to exposure. Standards are set limiting the concentrations of radionuclides in the air so that the radiation doses to workers might be kept below allowable maximum limits. IEER performed an independent assessment of radiation exposure to workers as part of a class action lawsuit filed by the plant's workers against National Lead of Ohio, DOE's contractor until 1985.

Working conditions at the Fernald uranium processing plant near Cincinnati were appalling, especially in the 1950s and early 1960s. They were typified by high air concentrations of uranium in many areas of the plant which often exceeded the Maximum Allowable Concentration (MAC) by tens of times, hundreds of times, and even thousands of times. One 1960 plant document lists the air dust concentration in the breathing zone of an operator cleaning under a burnout conveyor as 97,000 times the MAC.²

Work procedures also contributed to the high air dust concentrations in the plants. For example, a 1968 plant document described the procedure for emptying a dust collector:

In many plant situations, proper respiratory protection to prevent inhalation of this radioactive dust was not available. IEER's review revealed that workers were not properly trained regarding when to use respirators, and consequently did not wear them in many situations when air dust concentrations were high. In fact, in the early years of plant operation, workers were not even issued respirators as long as air concentrations of radioactivity remained less than ten times the MAC. In addition, a significant number of respirators cleaned for reissue remained contaminated. In some cases, the insides of respirators were contaminated. A plant doctor on an impromptu plant tour characterized some of the respirators as "the epitome of filth."⁴

Internal exposure estimates

Fernald worker dose records are highly misleading because they contain no mention of radiation doses due to the uranium that workers inhaled which then irradiated their bodies, notably their lungs. These doses were not included in worker records despite the urine sampling that was done throughout the plant's history and the lung counting that was done after 1968. Thus, when workers requested dose records, they were only given information on external doses (see below).

The urinalysis program used at the Fernald plant had several shortcomings. Twenty-four hour urine samples provide a good indication of how much uranium is in a person's body. However, 24-hour samples were not regularly taken at Fernald. Instead, the program relied on "Monday morning" single samples. It was not recorded which workers drank coffee and therefore possibly diluted their urine samples.

Another problem with the program was the infrequency of the samples, especially in the early years of plant operation. After uranium is inhaled, it is excreted from the body in diminishing amounts over a period of time. The amount of time it takes for an inhaled material to be excreted depends on its chemical form. When samples are taken only every few months or even just once a year, as they were in early years of Fernald operations, it is possible for large exposures to go undetected. As a result, infrequent monitoring makes it impossible to accurately determine the magnitude of the exposure.

IEER developed a method to estimate radiation doses to the lung from urine data by calibrating that data to the direct lung count data that was available after 1968. The concept was developed by Bernd Franke in collaboration with an IEER consultant, Mike Throne, who also created the mathematical formulation of the method. Kevin Gurney wrote the computer program to manage the enormous volume of data and run it through the mathematical model.

IEER's conclusions were that doses due to uranium inhaled by workers were above then-allowable limits (15 rem per year) in more than fifty percent of the cases in every year but one between 1952 and 1962. Significant proportions of workers continued to suffer overexposure after that. A chart of the proportion of workers exposed to more than the allowable limits due to lung burdens of uranium is shown below.

Source:B. Franke and K.R. Gurney, "Estimates of Lung Burdens for Workers at the Feed Materials Production Center, Fernald, Ohio," (Takoma Park: IEER, 1994), p. 8.

The presence of large and variable amounts of radon during lung counting appears to have created measurement errors in the records of many workers. Fernald's procedure for lung counting included measurement of subtraction of ambient external radiation readings, including radiation from radon and its decay products. However, differences in radon levels between the time that background measurements were taken and the time that the lung counting was done could mean that the actual lung burden may have been higher or lower than reported. Such fluctuations would tend to cancel out in population dose estimates, such as the ones that IEER made, which are presented in this article. Further, the result of subtracting high background readings resulted in many negative estimates of lung burden, which must necessarily be rejected as false, as well as a large number of low values below 5 milligrams, which IEER considered to be too unreliable to use. IEER's work took these statistical problems in lung count data into account by omitting all lung burden estimates below 5 milligrams. Worker doses from radon and its decay products would be in addition to those from uranium lung burdens discussed above. These remain to be estimated.

External exposures

In general, external exposures were also not carefully monitored at the Fernald plant. For instance, there was a high potential for some workers to experience significant external exposures, especially to their hands. Hand exposures were not calculated at all until 1970, when some workers began to wear wrist dosimeters. In many cases, external dose records indicate readings of zero. Without further records and investigation, its is not possible to assess whether these meant that there was no significant reading above background, whether there were other problems with the data, or even whether some of the data were fabricated (as was the case with some of the uranium release data).

One example of the problems of the external radiation dose record dates from the early 1980s, when thermoluminiscent dosimeters (TLDs) were introduced in place of film badges. Contamination of the TLDs by uranium prevented accurate readings and so a "correction factor' was introduced to the raw dose reading. However, NLO used the same correction factor for all workers regardless of the working conditions and duration of exposure during the month. The result was that some workers were, after "correction," estimated to have negative radiation doses. These records were apparently referred to the Health Physics section for further action, but it is still not known what was done with the estimates. One possible outcome is that a zero was entered into the dose record.

Concluding observations

Just after the presentation of IEER's findings in court in 1994, the DOE settled the lawsuit on behalf of NLO, providing workers with lifetime medical monitoring, and other benefits. But the DOE has still not acknowledged that worker dose records are severely flawed and incomplete. So far as IEER has been able to determine, DOE and its contractors still routinely fail to include estimates of internal doses in worker dose records. Therefore, in the nuclear weapons plants were workers have been exposed to conditions that might cause internal exposures, the dose records would be systematically incomplete and underestimate worker exposure.

In many epidemiological studies, or at least in some cases, the general assumption that film badge data are a useful proxy for actual total exposure may not be valid. Inaccurate external dose records, lack of dose records for many high internal exposures, and the highly variable conditions of uranium dust to which workers were exposed make film badge data suspect. Finally, worker records contain almost no information about exposures to non-radioactive toxic materials, such as acids, metals, and solvents, which are routinely used in large quantities.

Nuclear weapons production in the U.S. has involved 600,000 workers, many of whom worked in uranium and plutonium processing facilities, where there were risks of internal exposures. Identifying those most at risk by estimating internal exposures is a matter of elementary justice and health protection. Efforts must also begin to find groups at high risk due to chemical exposure. Such evaluations can lead to identifying high risk worker groups. Medical monitoring may provide early detection for such workers who may otherwise not suspect that they are at risk until it is too late for them.

4. Memo from J.A. Quigley to C. Dees, National Lead Company of Ohio, October 12, 1953, p. 3