replaces the very strong negative role observed in 1970-1980. However, this former negative role of ages 80+ among women is not absolutely reliable; it could be due partly to some improvement in registration of oldest female deaths.
In any case, this important change in the age components of mortality trends is closely related to the change in cause components. As the leading group of causes shifts from cardiovascular diseases to injury and poisoning, the impact shifts automatically from older to younger ages. For the same reason, the main part of the increase in life expectancy between 1980 and 1987 is among young adults.