10. Age pattern of mortality decline

10.1. The elderly in 12 selected countries

In order to provide a better perspective on the age pattern of recent mortality changes among the oldest-old, we give in the following first an overview of the changes observed among the elderly in general from age 60 up. The data for the age bracket 60-79 have been drawn from various editions of the United Nations Demographic Yearbook and refer approximately to the periods 1955-59 and 1985-89 (the death rates in certain years in some countries were not published). From them, percentage decline was calculated for quinquennial age groups in.the interval 60-84 years in twelve selected countries and the series was extended to older quinquennial age groups using data from our database and verification showed that the two sources agreed well in the pivotal age group 80-84. The results are given in Annex Table 6 and Figure 5. Because of small numbers the ratios for centenarians are given in the table only for the largest countries and, being rather erratic, are not reproduced in the figure.

        The results for males reveal striking differences between countries. The first frame shows considerable but differential progress in four large countries. The very deep decline in both Japan and France has been strongly age-selective and the younger elderly have made much greater gains than the oldest. The more moderate gains in West Germany and Italy do not adhere to this pattern beyond age 80: among the very oldest they are nearly as large as among sexagenarians.

        In the Netherlands (not shown in the figure) the age pattern has a shape similar to that in West Germany and Italy but the curve moves on a higher level and the death rates of the 70-79-year-old have actually risen.

        The Nordic countries show even greater divergence. In Finland a very substantial progress has followed the age pattern of France while the gains in Sweden have been almost non-selective. In Denmark and Norway the mortality of the younger elderly has increased and only the oldest-old show declining rates.

In the Central European countries of the former East bloc the development has been quite alarming. The mortality of the younger elderly men has increased sharply and only the 85-and-over have seen any improvement at all.

        It is perhaps expected that mortality can be more easily reduced among relatively younger people than the very oldest. This expectation is met in the data for men in Japan, France and Finland. The data for West Germany, Italy and Sweden are roughly neutral in this respect showing no clear dependency on age.

        In the Netherlands gains are recorded among the youngest elderly and again among the oldest. Finally, Norway, Denmark and the former East bloc countries show a contrasting picture where the younger elderly have not benefited at all and the gains have been largest among the oldest.

        The data on women present an entirely different, more regular and uniform picture. Although the size of the gains does vary considerably between countries, they are in all cases gradually reduced as age increases. The largest gains, though, have not taken place at the youngest ages but in the upper sixties or the seventies. After this, the gains have narrowed down with age in all twelve countries.


10.2 The oldest-old in 28 countries

In the narrower age band of the oldest-old the Odense database allows relatively accurate description of the changes in a number of countries as brought together in Table 12. The figures for some of them should, however, be regarded with caution. Besides the weaknesses in the basic data in a few countries as pointed out before, the rates for Iceland and Luxemburg are in this table affected by the small numbers involved. Steadier time series by age group are given in Annex Table 7 for aggregates of countries in quinquennial periods.

        The existence, among men, of sharp contrasts in the age pattern of mortality decline, noted above between 12 countries, is confirmed on a broader front. The expected diminishing gains with advancing age can be seen, in addition to Japan, France and Finland, unequivocally also in Australia, Austria, Portugal and Switzerland and somewhat less clearly in Ireland and Scotland. It deserves to be noted that all the countries where the overall gains have been the largest - Japan, France, Switzerland and Finland - share this same age pattern of change.

        Roughly equal mortality decline at all ages can be seen in Belgium, England and Wales, Germany (West), Italy, New Zealand and Sweden, and the same is suggested by the data for Spain.

        In sharp contrast to all the above, the men in the Netherlands and Norway have registered only small, if any, gains around age 80 (at ages 70-79 the mortality increased) and substantial benefits have accrued only to the very oldest. It will be recalled that in these two countries the mortality of old men has been traditionally low. The shape of the curve is basically similar in Denmark. The modest gains made in the former East bloc also tend to concentrate in oldest ages.

        Looking at the figures for the aggregates at the bottom of Table 12 we note that the gains increase with age in the low-mortality group, decline with age in the rapid decline group, the two tendencies are cancelled out in the medium-mortality group and are independent of age in the high-mortality area as well as in Western Europe as a whole.

        Female mortality has developed along very different lines and conforms fully to the expected law of diminishing return with advancing age. Prime examples of strong age-selectivity in this direction are Japan, France, Switzerland and all Nordic countries. The tendency is weak in Portugal and Scotland but there is no significant exception to it among the 27 countries examined. The pattern is confirmed in all aggregates.


10.3 Discussion

A satisfactory explanation of the varying age patterns of mortality decline would require an analysis of the causes of death. Some general features observed in the data at hand do, however, permit a few tentative suggestions. As age increases, degenerative diseases such as most heart diseases, arteriosclerosis and the various forms of senile dementia become gradually the prevalent causes of death. Aging itself becomes a prime determinant of mortality. Strong evidence has been published to the effect that wide-spread changes in lifestyle, including diet, smoking and exercise, probably combined with blood pressure control, have been the main factor in reducing mortality from degenerative diseases in industrialized countries. If this is the case and if the aging process itself is delayed, it is natural that its benefits are felt more at ages where it is not yet very advanced. Hence mortality should decline more, say, at age 80 than 90 or 100. Among the still younger elderly, on the other hand, other causes of death with different etiology predominate and even if they depend on lifestyle, they have often been less responsive to recent trends in it.

        It appears that in low-mortality societies women live closer to their physical endowment and die more in conformity with the aging process and this results in more regular and more uniform mortality patterns. Men are more prone to death by external events and through morbid processes caused by a multitude of interfering factors. Whether these depend on occupation, lifestyle or other influences, they are more diversified in space and over time and this leads to less uniform mortality patterns. It also follows that findings in studies like the present one are valid only for the kind of populations they represent and in the specified period. They are also liable to change at any time.


Updated by V. Castanova, 1 March 1999