Report to the Congress from the Presidential Commission on Catastrophic Nuclear Accidents

BNFL/UKAEA Agreement with Unions

As noted in the text, British Nuclear Fuels (formerly BNFL and UKAEA have entered into a Compensation Agreement with their respective unions covering cancer mortality and morbidity among radiation workers. The agreement provides a voluntary alternative to litigation for determining whether the disease or death should be deemed to have been linked to employment-related radiation exposure. When there is a sufficient likelihood that a cancer or cancer death should be attributed to radiation exposure in the course of employment, partial or full compensation is paid, depending upon the degree of likelihood (an example of proportionate liability). This approach provides an alternative to “all or nothing” compensation through the courts. The current agreement, signed in December 1987 and effective March 31, 1988, represents an extension of a 1982 agreement between BNFL and its unions that covered only cancer deaths, whereas the present agreement deals with disease as well.

About 1,200 cancer deaths among radiation workers employed by BNFL and its predecessors were known at the end of 1987; by the end of 1988, nearly all of them had been processed under the agreement and awards made in 17 cases. A 1989 paper authored by a member and retired member of the Health and Safety Directorate of BNFL1 sets out the background, framework, and philosophy of the Compensation Agreement, together with some indication of the factors determining the form and amount of the award, and some details regarding the 17 awards theretofore made.

The authors point out that U.K. regulations governing occupational radiation exposure take into account recommendations of the International Commission on Radiological Protection (ICRP) and National Radiological Protection Board advice and that these regulations are designed to ensure that the risk of any injury or disease caused by such exposure is very small.

In particular, there should be no acute non-stochastic injury, and there should only be a slight chance that any longterm stochastic injury, i.e., a cancer, will occur to the particular individual exposed.

Nevertheless, the philosophy is that, when cancer has occurred that could have been caused by radiation, compensation needs to be considered, and in those rare circumstances where the chance of induction of a particular cancer at a particular time by the occupational radiation exposure is greater than the chance of induction of that cancer at that time by all other causes, compensation should be paid. This philosophy of paying compensation on the basis of a balance of probabilities was developed in light of the following considerations:
  • one-third of the U.K. population develops cancer, and three-fourths of these are fatal;

  • occupational radiation exposure is assumed to pose some risk, however small, of cancer induction; and

  • it is rarely, if ever, possible to distinguish by medical examination between (a) cancers caused by toxic agents other than radiation or occurring spontaneously and (b) those caused by radiation.

The circumstances that permitted the creation of this contractual (but voluntary) alternative to litigation grew out of the continuing relationship between the employers and the employees' unions and the experience of both sides with lengthy, time consuming, and expensive litigation of five claims sponsored by the unions prior to the 1982 agreement. There was also precedent for such agreements, such as those with respect to occupational disease in the U.K. coal mining industry.

The objectives of the parties were to arrive at procedures that would (a) preserve the finality and binding nature of a settlement, once it was agreed to, to the same degree as a judicial award or settlement of litigation, (b) approximate the amounts of compensation that would be made in out-of-court settlements (lesser compensation in cases of weaker association between exposure and disease) rather than use the “all or nothing” approach of litigating to judgment, (c) achieve a quicker resolution than litigation, and (d) be fair to individuals or survivors, while reflecting the state of medical knowledge regarding the induction of cancers by radiation and other toxics.

The principal technical problem, according to the authors, was to sort out whether radiation exposure is a more probable cause than all other possible causes. The resolution settled upon is termed probability of causation (which the authors, as we have, short-form as “PC”2). The authors point out that

Quantitative assessment of PC for a particular case may involve many complex factors. These include the extent of each exposure to radiation and to other carcinogenic agents, arising from the individual's occupation, lifestyle, and medical diagnosis or therapeutic treatment, together with their contribution to the induction of the particular cancer at the time it developed. In addition, account should be taken of any other evidence, e.g. family history of the particular cancer, distinguishing the particular individual from the general population of the same age and sex in respect of published mortality and morbidity statistics.3

The authors report that the parties came to recognize that the task of devising a comprehensive framework for all cases is difficult, if not impossible, in light of the lack of quantitative data for some of the foregoing factors and the need to accommodate advances in scientific knowledge, among other considerations. To arrive at a workable system in light of these difficulties, the parties devised a tripartite categorization of claims. In the first category are cases where PC can be determined by computation because three agreed-upon tests are met:

  1. radiation exposure is the only significant factor that distinguishes the individual from the general population of the same age and sex;

  2. occupational exposure was received at a dose rate within or not far above4 that permitted by regulation; and

  3. the particular cancer is one whose association with radiation exposure is reasonably well-known.

In the second category are cases where only condition (c) is not met; here, a standing expert panel (about which more below) can determine PC (or the range of PC, as also explained below). And in the third category are those cases in which (a) or (b) or both are not satisfied; the standing expert panel, with such additional expert input as may be required, can determine PC.

The standing expert panels consist of independent scientific and medical experts agreed to by the patties. All of the significant facts to be considered are to be agreed to by the relevant union and the relevant employer. Such facts would include:

  1. personal data regarding the individual

  2. the particular cancer

  3. the extent of occupational exposure to radiation (both external and internal)

  4. the extent of other radiation exposure

  5. the extent of exposure to other carcinogens

  6. any significant differences between the individual and the general population of the same age and sex.5

In practice, Category I cases are those satisfying conditions (a) and (b) above, in which only external occupational radiation is significantly involved and in which the condition (c) disease is either a form of leukemia (other than chronic lymphatic leukemia) or bone cancer.

The risk function that underlies these PC values is an absolute risk model, in which the risk arising from the exposure at a particular time is independent of the subsequent natural risk of incurring leukemia. Latency is considered by a function in which the risk manifests itself after a minimum interval, rises though a peak, and is terminated after a defined period. The application of the risk function is somewhat biased in favor of cancer induction by radiation, in that the lifetime risk on average corresponds to two to three times that identified by the ICRP.6 An additional factor tending to favor compensation of the claimant is added to account for any non-conservatism that might arise from uncertainties in the data and models used.7

For Category II and III cases, additional screening criteria had to be adopted, in that about one-third of the U.K. population contracts cancer, while only 1 percent of the relevant workforce cancers fall into Category I. For Category II, a minimum preliminary PC level was agreed that is below that needed to qualify for the minimum award. The computation differs in that there is no cutoff for maximum latency period, and the risk factors used are even mom conservative, between three and four times ICRP 26 (1977).

Apparently the only special problem for Category III cases is the case of lung cancer, where an adjustment has to be made for smoking. Where there is no evidence that the individual abstained from smoking for any substantial period, the same screening technique is used as for Category II cases (i.e., smoking would be a significant competing risk of lung cancer). However, when there is evidence of any period of non-smoking, the general population risk (smokers and non-smokers) is used, which reflects a much reduced risk. Nevertheless, in the final determination, the expert panel (and any additional experts) do take into account the smoking habits of the individual.

Awards fall into five brackets: 100, 75, 50, 25, and 0 percent of the amount that would be recoverable in litigation. The 100 percent award applies where the PC range is above 50 percent, and lesser awards reflect lower PC values, though the authors do not specify the PC ranges corresponding to the partial awards, and these apparently have not been made public.8 The base amount of death awards varies with individual circumstances, including life expectancy given other medical history, earnings, and career prospects. Disease awards can be full (if death in the near future is the prognosis) or provisional. If provisional, they are based on known and foreseeable disabilities, further potential disabilities are identified, and the provisional award can be reopened if and as those additional disabilities become manifest.

As noted above, about 1,200 cancer death cases have been processed and 17 awards made. Of these, six were at the 100 percent level, two at 75 percent, four at 50 percent, and five at 25 percent. Three were Category I leukemia cases, five were Category II (one each multiple myeloma, and breast, liver, stomach, and thyroid cancers), and three Category III lung cancers, each involving smokers with a period of non-smoking.

The amounts of awards have been made public in 11 of the cases (the remainder being held confidential at the request of the beneficiary). The range of awards in these cases was from about $30,000 to about $200,000.9

  1. Mummery and Alderson,“The BNFL Compensation Agreement for Radiation Linked Disease”, 9 [U.K.] Journal of Radiological Protection 179 (1989).

  2. The approach is conceptually similar to that described in Chapter 4, though it employs its own set of values rather than the NIH radioepidemiological tables.

  3. Ibid., P. 180.

  4. As will be seen below, doses approximately within regulatory limits are reflected in tables which can be applied mechanically; greater doses call for the case-by-case judgment of the standing expert panel.

  5. Compare the practice of the Veteran's Administration prior to the adoption of the Radiation-Exposed Veterans Compensation Act. VA regulations in 38 C.F.R. §3.311b called for a case-by-case approach. The first step was to verify that a radiogenic disease had developed within the expected latency period following service-connected exposure. The following other factors we then considered:

    • the probable dose, in terms of type of radiation. doserate, and duration. taking into account any limitations in the dosimetry;
    • the relative sensitivity of the tissue involved to induction of the pathology by ionizing radiation;
    • the veteran's gender and pertinent family history;
    • age at exposure;
    • time-lapse between exposure and onset of disease; and the extent to which non-service-related exposure to radiation or other carcinogens may have contributed to development of the disease.

  6. International Commission on Radiological Protection. Recommendations of the International Commission on Radiological Protection. Publication 26. New York, 1977.

  7. The formulae used are set out in the paper. Ibid., pp. 181-82.

  8. One can speculate that the ranges corresponding to the award levels might be as follows: (100) greater than 50 percent; (75) 40-50 percent; (50) 30-40 percent; (25) 20-30 percent; and (0) less than 20 percent.

  9. Mummery and Alderson report the range as from 16,924 to 120,633 pounds sterling.

Appendix F « Index » Appendix H