Dear SirDr Kalman correctly stresses the importance of responding to
the health and safety concerns of workers in the nuclear
industry. However, the statements that we dismissed comparisons
of cancer rates among Springfields workers to national rates,
and that such analyses provide `real information' of greater
importance to BNFL employees than internal analyses of cancer
rates, are perhaps misleading. In our editorial we simply
reported the fact that the paper by McGeoghegan and Binks
focused on internal analyses that compared cancer rates between
workers who accrued different levels of cumulative external
radiation dose [1].
Scientists, as well as worker and community groups, have
previously raised concerns about the problems of bias that can
occur when workers' death rates are compared to national rates
[2-5]. Briefly, the exclusion from employment of people
with medical problems, and the socioeconomic advantages of some
workers, may lead to the observation of relatively low death
rates in an occupational cohort and mask evidence of occupationally-induced disease.
Internal comparisons are often conducted, as noted by
McGeoghegan and Binks [6], in order to avoid the bias. In
their internal analyses, cancer rates among workers with high
levels of exposure were compared to cancer rates among workers
with low levels of exposure. Such analyses can provide insight
into whether cancer rates increase among workers with increasing
levels of exposure. The observation that, overall, cancer rates
are not substantially elevated among Springfields workers
compared to national rates, therefore, may be of less interest
to workers than the finding that cancer rates are higher among
workers with higher external radiation doses. In our editorial,
we suggested that this trend may at least partially reflect
confounding due to other occupational exposures (such as
internal exposure to ionizing radiation) which may have been
received at higher levels among workers who also received higher
external radiation doses. From an occupational health and
safety perspective, however, the observation of increasing
cancer rates associated with an occupational exposure (whether
independently, or in conjuction with other exposures) is of more
than `academic' interest.
Yours faithfully,