EPIDEMIOLOGY

CANCER IN DECLINE?

SOURCE. V. M. Shkolnikov, M. McKee, J. Vallin, E. Aksel, D. Leon, L. Chenet and F. Mesle, Cancer Mortality in Russia and Ukraine: Validity, Competing Risks and Cohort Effects, International Journal of Epidemiology, Vol. 28, 1999, pp. 19-29. Full PDF version available online at http://ije.oupjournals.org/cgi/reprint/28/1/19.pdf

The death rate in Russia and Ukraine has increased dramatically since the late 1980s, but that does not mean that mortality from ALL causes has risen to an equal extent. In contrast to the sharp rise in deaths from such causes as cardiovascular disease, accidents and violence, mortality from cancer has actually fallen slightly.

In this paper, a team of researchers from the Center of Demography and Human Ecology of the Institute for Economic Forecasting in Moscow explore four factors that might account for this decline in cancer deaths. Their analysis of patterns and trends in the data leads them to conclude that all four factors may have contributed to the observed result, but with important variations by age group and sex.

First of all, the decline may to some extent be a statistical illusion, due to change in the way data are collected rather than to any real change in the incidence of cancer. Indeed, the official figures for cancer among the elderly are implausibly low, especially for rural areas, probably because many cases remain undiagnosed in areas where medical provision is poor. The ways in which cause of death is recorded on death certificates, and then coded for statistical purposes, have also shifted over time. For instance, is the immediate cause of death given or the underlying cause? Often cancer is the underlying cause, but the immediate cause is something else (e.g., a heart attack following a fall).

Secondly, there is the problem of birth-cohort effects arising from the turnover of generations whose cancer risks differ for reasons having nothing to do with recent changes in the social environment. These effects can explain some reduction in the cancer rate among men after early middle age and among women at all ages.

The third factor is the "competition" between mortality from cancer and mortality from other causes. Those who die early from heart disease or accident are thereby deprived of the chance of dying from cancer later on. This effect was found to explain some reduction in cancer deaths among middle-aged men.

Finally, might the decline in cancer mortality have something to do with improvement in health care? The authors are not sure, but tentatively suggest that improvements in treatment may have brought about a reduction in deaths from childhood leukemia.