diminished such that their continuing decline would no longer result in important gains in life expectancy. Now, cardiovascular diseases and cancer ranked highest among all causes of death, much above infectious diseases. At the same time, economic and social changes resulted in an increase in such mortality-related factors as alcoholism, smoking, and traffic accidents. (See also the discussion of the epidemiological transition in the chapters by Kingkade and Arriaga and by Murray and Bobadilla in this volume.) A return to favorable life expectancy trends required both an end to the increase in these ''civilization ills" and success in a new fight against cardiovascular diseases and cancer. That was accomplished in France (Meslé and Vallin, 1993c), as well as in the United States, but not in Russia. This is why life expectancy changes within these countries since the late 1960s have again moved in opposite directions: steady improvement resumed in France and in the United States, while stagnation, or even deterioration for males, prevailed in Russia. These contrasting trends, which were pointed out by Jean Bourgeois-Pichat (1985) early in the 1980s, are still being confirmed.
Comparison with Japan is yet more illustrative, as shown in Figure 2-1. Until the 1960s, the Russian and Japanese situations were quite similar. Like Russia, Japan was far behind Western countries until World War II, but in the 1960s, thanks to very rapid progress, it reached the level achieved by France and the United States. The increasing disadvantage of Russia since 1965 is much more obvious by comparison with Japan than with France or the United States. The spectacular success of Japan in the transition from the victory over infectious diseases to the control of cardiovascular diseases contrasts dramatically with the failure of Russia.
The unexpected deterioration in life expectancy experienced by Russia since the mid-1960s has already been subject to much discussion in the literature. Taking into account some peculiarities of the age structure of mortality of the total Soviet Union, Anderson and Silver (1990) have proposed that at least a part of this deterioration is a statistical artifact due to improved data collection. Before discussing the trends themselves, it is necessary to look at these possible distortions due to the inaccuracy of the available data, and to examine ways of at least partially overcoming these deficiencies to draw meaningful conclusions; the first section of this chapter addresses these issues. Even if partially overestimated, no one contests the reality of the Soviet health crisis; the second section of the chapter attempts to derive some explanation for the crisis from analysis of the age and cause-of-death patterns of mortality in Russia. The third section examines comparative international trends in some specific causes of death. The chapter ends with a summary and conclusion.
In examining trends in life expectancy and causes of death in Russia, one