Coordinated by Roberta PACE, Alain PARANT

 

Diverging tendencies by age in sex differentials

in mortality in Italy

Carlo MACCHERONI

Carlo F. Dondena Centre for Research on Social Dynamics,

Università Commerciale L. Bocconi, Milan

 

Abstract: Chronological and comparative analyses are based on the most recent available Eurostat and OECD data. These analyses shed some lights on countries’ commonalities and diversities, as observed in marriage patterns and family policies over time. In Italy the gap among male and female life expectancy at birth has progressively narrowed since the beginning of 1980s; currently women live 5.1 years longer, so Italy results in median position in comparison with the other countries of Western Europe. Between 1979 and 2009 life expectancy among men has grown by 8.5 years, whereas women gained only 6.8. This does not mean that male over-mortality is reduced at all ages; actually this regards only adult and central ages, while from the threshold of elderly ages over-mortality among men shows an increasing trend. In Italy the trend in gender differences in mortality is in line with the overall death postponement which has been registered in the last decades; this process has changed the role of gender differences in mortality at various ages in the present gap, with a heavier weight on those in the highest age groups due to cohort factors such as involvement in wars, plagues and economic crisis in the first half of the twentieth century, and to the concept of the elderly condition according to which old age was only seen as a period of bio-physiological and psychological regression.

Keywords: Mortality, Gender differences, Male excess mortality, Life-table Aging Rate.

1.                        INTRODUCTION

 

Compared to other European countries, Italy shows very low mortality rates and relatively low differential mortality by gender. The analysis of this differential through the (Fe0 – Me0) life expectancy gap at birth - between female (Fe0) and male (Me0) - highlights remarkable differences at regional level (see Table 1): in 2009 they ranged between 3.3 years in Iceland and 11.1 years in Lithuania. Italy placed in a nearly median position (5.2 years). In addition, from the joint analysis of the (Fe0 – Me0) gap and the correspondence male life expectancy (Me0), a strong negative correlation (-0.908) between them can be clearly observed (see Figure 1). This means that by higher male mortality levels the gap between female (Fe0) and male (Me0) life expectancy at birth increases. This statistical relation is instead much weaker if the (Fe0 – Me0) gap and the corresponding female life expectancy (Fe0) are jointly analysed.

The results shown in Figure 1 certainly reflect deep territorial differences in the past mortality trend. In the graph three clusters can be seen: the first, left on the top, is constituted by three Baltic countries from former Soviet Union, the second by neighbouring countries of former Soviet Union in that period, i.e. from Poland to Romania, and the third by the rest of European countries, the so called western bloc. As is known, countries belonging to the first two clusters were involved, although at different level, in the mortality crisis following the fall of communist regimes at the beginning of the 1990s[i]; this generated a net gap compared with the rest of Europe which is shown in Table 1 in terms of life expectancy and which will require still a long time to be filled. The comparison related to Italy is therefore to be made within the last cluster, in which it places in a nearly median position. It should be noticed, however, that notwithstanding the higher social-economic and cultural homogeneity of the area under study, differences in mortality measured by (Fe0 – Me0) still show a non-negligible variability: in France and Finland the life expectancy gap between sexes is relatively higher, though the overall mortality rates are among the lowest in the world.

 

Table 1 - Female-Male differences in life expectancy at birth in 2009 – Selected European countries

Countries

Male

Female

Difference

Countries

Male

Female

Difference

 

 

 

 

 

 

 

 

Iceland

80.2

83.5

3.3

Belgium

77.2

82.7

5.5

Malta

78.1

82.2

4.1

Czech Republic

74.2

80.3

6.1

Netherlands

78.4

82.6

4.2

Spain

78.5

84.6

6.1

Sweden

79.2

83.4

4.2

Portugal

76.3

82.4

6.1

United

Kingdom

78.0

82.2

4.2

Finland

76.6

83.2

6.6

Denmark

76.9

81.2

4.3

Slovenia

75.6

82.2

6.6

Norway

78.7

83.1

4.4

France

77.8

84.8

7.0

Switzerland

79.9

84.4

4.5

Bulgaria

70.0

77.2

7.2

Cyprus

78.2

83.0

4.8

Romania

69.7

77.3

7.6

Germany

77.7

82.7

5.0

Slovakia

71.3

78.9

7.6

Ireland

77.0

82.0

5.0

Hungary

70.1

78.1

8.0

Luxembourg

77.9

83.1

5.2

Poland

71.5

79.9

8.4

Italy

79.2

84.4

5.2

Estonia

69.8

80.1

10.3

Greece

77.6

82.9

5.3

Latvia

66.9

77.4

10.5

Austria

77.5

82.9

5.4

Lithuania

67.5

78.6

11.1

Source: World Health Statistics, 2013.

 

This male over-mortality registered at all ages is determined both by genetic and biologic differences and by gender differences; the variability observed derives from the fact that these factors interacted differently with psycho-social health determinants due to different environmental situations and health systems. As to the first aspect, an important study current focused on the role of genetic and biologic differences had emerged already in the past[ii]; continuous progress of research in the bio-medical field has gradually increased knowledge of possible risk factors determining the present male excess in mortality. One of the most recent studies[iii]  shows that the lower incidence of cardiovascular diseases in women, and consequent lower mortality, is due to the positive action of female sexual hormones on blood fat levels, while men high testosterone levels prove to have a negative effect on cholesterol. Other biological factors contribute to increase cardiovascular diseases among men; however, it is difficult, if not impossible, to separate the impact of biological factors from social and environmental factors that are also indicated as main mortality causes.

Mortality differences are, indeed, not only determined by sex but are also a matter of gender. The introduction of the social field has led to a global approach of the problem since risk factors in habits and behaviour cannot be separated from familiar, educational and working environment[iv]; with regard to this latter point, it should be highlighted that the hardest and/or most dangerous working activities are usually performed by men. Moreover, some behavioural factors such as scarce attention to road safety, use of drugs and smoking habit, all of which have an heavy impact on male over-mortality, are important features of life choices and of the different attention paid to one’s health status by genders, especially in the past.

For further information about these issues see the related bibliography, also as far as international comparisons are concerned; the target of the present study is to focus on the more strictly demographic aspects of sex differences in mortality and on the more recent trends in Italy by analysing 1975-2010 life table time series[v]  produced by Istat[vi]. In this period Italy, similarly to many other countries, experienced an overall mortality reduction, an important signal of a general improvement of health conditions; in more recent years this event was also accompanied by a progressive reduction of the differences (Fe0 – Me0). This applies to all age groups from birth to adult age but is not registered in subsequent age classes; once individuals have reached old age, male over-mortality starts to increase with a trend that introduces uncertainty margins on the expectations of further reductions of present differentials in the next years. In order to grasp these issues, an effort was made to try and focus on the general framework of the most recent trends, especially as far as the oldest age groups are concerned. Survival until adult age is today almost certain for women and much probable for men. The evolution of mortality in the last stage of life plays therefore an important role in the further increase of life expectancy and in the dynamics of relative differences by gender. In the following paragraphs it will be shown that both gender- and cohort-related factors play an important role in this evolution.

 

Figure 1 – Correlation between female-male differences in life expectancy at birth and male life expectancy at birth for selected European countries - 2009


Source: Data processed from World Health Statistics, 2013.

2.                       EVOLUTION OF SEX DIFFERENTIAL IN LIFE EXPECTANCY

 

According to Istat, gender difference in life expectancy at birth today (2010) is 5.06 years (Fe0 = 84.45 and Me0 = 79.39); compared to 1975, men gained 1.3 years more than women (Fe0 = 75.75 and Me0 = 69.4), thus diminishing the previous gap; especially for recent years, this more rapid growth of Me0 is shown clearly in Figure 2, which stresses the probable intersection of Me0 and Fe0 trends in the very long term. In the period under analysis, male disadvantage measured in terms of (Fe0 – Me0) has increased up to 1979 (it was nearly 7 years), then has started to diminish although with some pauses or breaks (see Figure 3) in 1992 and 2001, while showing at the same time a progressive, net difference depending on age.

 

 

Figure 2 – Life expectancy at birth by sex – Italy, 1975-2010


Source: Data processed from Istat life tables, 2014.

 

Figure 3 illustrates the juxtaposition between the residual life gap trends (Fex  - Mex) (x = 0, 1, 2,…) very clearly: on the one hand, those decreasing from birth up to about the age of 60 are visible, on the other, those increasing at following ages are shown. The second half of 2000s, however, seems to represent the starting point for a new stage with new distinctions among "Young Old" (aged 65-74),"Old" (aged 75-84) and "Oldest-Old" (aged 85+). As a matter of fact, by initially decreasing width of the gap for these age groups, an accelerating trend in the reduction is observed, while in cases where the gap profile had previously registered an increase, at the end of the period a reverse trend or, at least a break, take place, such in the case of those aged over 80, for whom the different residual life expectancy of men and women do not seem to decrease (see Figure 3). The available information, however, is not yet sufficient to understand if this is a casual event or rather a new break with the past.

The 1975-2010 time series considered[vii] allow nonetheless to understand the consequences of an important stage in the evolution of mortality in Italy from the beginning. After the success obtained in the previous years in the fight against infant mortality and infectious diseases, greater attention had been paid to health protection of middle-aged and elderly persons. These age groups started to benefit from the attention paid to them around the second half of the 1970s, when a reduction of mortality caused by cardiovascular diseases – still constituting one of the main causes of death – started.

 

  Figure 3 – Female-Male differences in life expectancy at selected ages – Italy, 1975-2010


Source: Data processed from Istat life tables, 2014.

 

A study by Istat[viii] on the relationship between the increase in male and female life expectancy and the decrease of deaths by cause between 1970 and 1990 shows the differences and delays in receiving new therapies, new life styles and prevention first in the central age group and then in subsequent age classes. Between 1970 and 1980 women aged 55-74 and even more women aged 75 and over had started to benefit from the fall in mortality due to cardiovascular diseases to a larger extent compared to men; in the following decade women benefited even further from it and always much more than men in both these age classes. Conversely, between 1970 and 1980 the effects of the fall in mortality due to this cause were limited to men aged 55 and 74, while remarkable results were registered for the subsequent age class only between 1980 and 1990. Both the differential mortality profile increasing until 1979 and the immediately subsequent variations are to be seen within this framework, as shown in Figure 3. 

Genders certainly do not pay - nor have paid in the past - the same attention to their health, and the activities of health protection and prevention by the National Health Service would be certainly more efficient if these took into account gender-related issues. This is clearly shown by studies on avoidable mortality[ix], i.e. on those deaths which could be avoided according to medical science. This category comprises[x] deaths occurred before 75 and due to lung and respiratory system cancers, ischemic heart diseases and external/accidental causes  (i.e. events which could be avoided or significantly diminished through radical actions on individual behaviours thanks to primary prevention or through precocious diagnosis and treatment in addition to improvement of health services).

Table 2 shows some results of research in this field related to 2002 Italian experience, which quantifies the possible increase in life expectancy from this point of view in 4.5 and 2.5 years for males and females, respectively; in relative terms, the increase for males is nearly double (5.8%) compared to that for women (3%).

 

Table 2 – Life expectancy at birth net of avoidable mortality - Italy, 2002

Life expectancy at birth and difference

Males

Females

Difference

       

Total

77.1

83.0

-5.8

Net of avoidable mortality

81.6

85.5

-3.8

Difference

-4.5

-2.5

-2.0

Source: Atlante ERA, 2007.

 

 

It should also be taken into account that the evolution of mortality deeply modified the role of the various age classes in determining male over-mortality. The methodology used to analyse these changes is that proposed by Arriaga[xi] and the study was carried out following the age group classes mentioned before (see Figure 3). In 1979 the difference between eF0 and eM0 was equal to about 6.9 years; table 3 shows clearly that the major contribution to this gap – equal to 0.322 decimal year, i.e. 4 months – was due to male higher mortality rates at birth. Subsequent evolution, however, highlights a deep change since the role of differences in elderly and old age group classes became increasingly important, while that of differences in mortality at birth reduced remarkably; the latter contributed in 2010 with only 0.047 decimal year (corresponding to about 18 days) to the gap of 5.06 years between eF0  and eM0. In addition, as can be seen in Table 3, the age classes that contributed more in determining male disadvantage moved increasingly forward: the peak in 1979 was in the 65-69 age class but it moved between age class 75-79 in 2010.

  

Table 3 – Contribution of age groups to differences in female-male life expectancy at birth - Italy, selected years

Age

1979

1992

2001

2010

1979

1992

2001

2010

Absolute values

Percentages

0

0.32

0.15

0.05

0.05

4.69%

2.21%

0.91%

0.93%

1-4

0.04

0.01

0.00

0.01

0.51%

0.14%

0.06%

0.12%

5-9

0.04

0.02

0.01

0.00

0.59%

0.36%

0.15%

0.08%

10-14

0.05

0.03

0.02

0.02

0.74%

0.50%

0.31%

0.35%

15-19

0.16

0.18

0.11

0.08

2.35%

2.71%

1.93%

1.53%

20-24

0.18

0.24

0.19

0.11

2.58%

3.62%

3.23%

2.16%

25-29

0.13

0.24

0.16

0.11

1.93%

3.64%

2.74%

2.12%

30-34

0.11

0.22

0.14

0.10

1.55%

3.29%

2.38%

2.00%

35-39

0.15

0.17

0.13

0.09

2.17%

2.57%

2.26%

1.85%

40-44

0.22

0.18

0.14

0.11

3.27%

2.77%

2.44%

2.26%

45-49

0.38

0.25

0.19

0.15

5.59%

3.83%

3.29%

2.86%

50-54

0.56

0.37

0.28

0.21

8.21%

5.66%

4.77%

4.07%

55-59

0.71

0.55

0.43

0.33

10.39%

8.30%

7.38%

6.48%

60-64

0.88

0.78

0.56

0.47

12.81%

11.81%

9.66%

9.18%

65-69

0.89

0.89

0.72

0.60

12.92%

13.56%

12.38%

11.79%

70-74

0.85

0.83

0.83

0.68

12.34%

12.63%

14.30%

13.38%

75-79

0.63

0.74

0.79

0.70

9.13%

11.22%

13.55%

13.72%

80-84

0.38

0.46

0.57

0.66

5.47%

6.96%

9.86%

12.95%

85 e +

0.19

0.28

0.49

0.62

2.77%

4.23%

8.40%

12.18%

Total

6.86

6.59

5.81

5.06

100.0%

100.0%

100.0%

100.0%

Source: Data processed from Istat life tables, 2014.

 

These results confirm that health problems and different health conditions tend to concentrate among elderly age groups. As a matter of fact, the growing trend of life expectancy shown by the evolution of Fe0 and Me0  in Figure 2 took place through a slow, more and more consolidated death postponement, resulting in a concentration of deaths in the highest age groups and within a short interval. With reference to the life tables analysed in this study, between 1975 and 2010 the share of deaths among women aged 80 and over – nearly 75% of total deaths – increased by 50%, while deaths among men – 61.5% of total deaths – doubled.

This fact underlines that the threshold to old-age cannot be schematised on the basis of a chronological age which is still calculated according to a procedure used in demography corresponding to the loss of the economic role and generically indicated as the age of 65. Age-related thresholds that are assumed to be valid at a certain moment in time are no longer so by successive health condition improvements. In Italy, reaching the age of 65 is today almost certain for women and much likely for men: according to 2010 life table, the probability for a new-born female to reach 65 years of age is 0.933, while a new-born male has a probability of 0.879.

For this reason, special attention is paid to the oldest age groups in the following pages.

 

3.         THE TREND OF SEX MORTALITY RATIOS STARTING FROM ADULT AGES

 

The gaps between sex differentials in life expectancy analysed in the previous paragraph can be explained by the dynamics of gender death risks. The first aspect is related to the different speed of the decreasing rate in death probability at different ages. The period considered is once again that between 1979 and 2010 for ages until 95: in Istat life tables, death probability at successive ages is graduated according to a logistic model[xii].

 

 

Figure 4 – Average annual decreasing rate in death probability between 1979-2010 by sex, Italy


Source: Data processed from Istat life tables, 2014.

 

Death probability at birth registered the greatest reduction for both sexes (see Figure 4), with an average annual decreasing rate of 5%; the rate was slightly higher for males. Passing on to other age groups, the greatest reduction, as far as men are concerned, was registered for infants and young people up until 20 and between 40 and 70[xiii], with a most substantial reduction, 50% higher than for females, between 50 and 55 (see Figure 4). In this case, the risk of death fell by 3% annually and this represents an important success because it is referred to a period in the life where men are still active and therefore more exposed to risks related to their working activity.

As far as women are concerned, mortality has decreased at a higher rate than among men aged between 21 and 39 and older than seventy (see Figure 4); what happened in the former age group has contributed less to the increase of female life expectancy, because the death probability is usually very low in that life span. On the contrary, in the latter age group, where the death probability is usually very high, there was a remarkable decrease (over 2%). Moreover, even in comparison to men, the decrease was more rapid also among women aged over 90.

 For a better understanding of the changes occurring in the last stage of life, differentials were measured in terms of average variation rates (see Figure 4) and also by the so-called M/F mortality ratio calculated on mortality rates (see Figure 5). As is known, this is an indicator constructed by comparing correspondent specific death rates at different ages (mx) or by comparing male death probability (qx ) with females’[xiv], thus obtaining a relative measure of the male excess mortality, independent of the death rate level which characterise them. As previously done, the 1975-2010 abridged life tables are also referred to in the case of M/F mortality ratio. In this case adult ages are examined.

  

Figure 5 – Sex mortality ratio: adult and old ages – Italy, 1975-2010


Source: Data processed from Istat life tables, 2014.

 

From several points of view the results obtained (see Figure 5) are in line with what already shown in Figure 3: before and after the age of 75 reverse trends are registered; in more detail, in all the age groups included between 40 and 59 a decreasing M/F mortality ratio can be observed due to the more rapid reduction of male death risk mentioned above. At subsequent ages the initially growing trend does not keep the same profile for the entire period considered; the trend reverts around the half of the 1990s in the age group 60-69 and at the beginning of 2000 for 70-74 year olds, whereby the M/F mortality ratio falls abruptly and rapidly. In the age group of people of 75 and over there has been sometimes a considerable increase of this rate, as in the age group 75-79, although this trend slowed down after 2005. Apparently in these age groups the variations are the more recent the older the age class considered. These variations could reflect cohort factors, as will be better described in the following paragraph; nonetheless the residual period of time until 2010 is still too short to evaluate these trends (see Figure 5).

Figure 5 also shows that by people aged 70 and over the levels of the M/F mortality ratio are always smaller by increasing age, because female death probability at very old age converges towards males’. Notwithstanding this narrowing of the gap, it should be highlighted that at very old age death probability is very high and, for this reason, even little variations produce remarkable effects, as will be explained by using Keyfiz approach[xv].

 

Table 4 – Excess of male over female death rates at all ages accounting for the sex differences in remaining life expectancy – Italy (percentage values)

Age

1975    1980    1985    1990    1995    2000    2005       2010

 

60

65

70

75

80

 

60

65

70

75

80

Excess

 55.9%      60.3%     61.0%   60.9%    61.2%    58.9%    59.4%       55.3%

 50.3%      54.7%   55.2%    55.6%    56.9%    55.9%    56.8%       53.1%

 42.5%      47.5%     48.2%   48.7%    50.4%    50.9%    53.2%       49.6%

 34.0%      40.3%     39.6%   39.9%    42.5%    43.7%    47.3%       44.7%

 26.7%      32.7%     29.9%   31.0%    34.2%    35.5%    39.5%       39.4%

Differences

-24.0%    -25.9%   -24.7%    -23.9%    -23.0%    -21.0%    -19.9%      -17.7%

-24.7%    -27.0%   -25.8%    -25.1%    -24.7%    -23.1%    -21.9%      -19.5%

-23.9%    -27.0%   -25.6%    -25.3%    -25.2%    -24.4%    -24.1%      -21.2%

-21.8%    -26.0%   -23.9%    -23.5%    -24.5%    -24.4%    -25.2%      -22.7%

-18.6%    -23.2%   -20.0%    -20.4%    -22.1%    -22.9%    -24.5%      -23.7%

 

Source: Data processed from Istat life tables, 2014.

 

Figure 5 also suggests that M/F mortality ratio changed remarkably in relation to both reference period and age group. If it were observed during a person’s life span, in Italy the absolute minimum would be registered at infant ages and the maximum at the age of 25; the latter is the age group where male mortality is more than triple compared to female’s, especially due to injures following road accidents. The ratio would therefore diminish, reaching its relative minimum around the age of 45[xvi] then it would grow progressively till the relative maximum around the age of 65, when male mortality level would be more than double; this second peak is especially due to sex differences in mortality incidence for cardiovascular diseases, cancers and diseases of the respiratory system, responsible in 2009 of about 38%, 30% and 7% of all deaths, respectively. After this second peak, the ratio falls again to then tend to 1 at older ages.

This great variability during the whole life span of a person can on the one hand barely be synthesised by an indicator structured as an average, expressing how male death risk compared to female’s is overall higher from a certain age onwards, i.e. in an interval covering his residual life. On the other hand, it is not easy to assume how to build this average since – as shown before – the weight of mortality varies at different ages when determining the general mortality rate.

In 1985 Keyfitz[xvii] suggested a solution from the analytical point of view: he found a relation between life expectancy relative variation and uniform death risk relative variation rate. In more detail, once a population shows a lower mortality rate than another (as in the case of female versus male component), Keyfitz proved that the mortality uniform increase rate at all ages δ (0< δ <1), expressing the relative variation between the lowest and the highest life expectancy (eM0, t and eF0, t respectively) is obtained from the following equation

                  eM0, t

              ———    =  1 - δ H(t)                                                                                              [1]           

                  eF0, t

where H(t) (0 ≤ H(t) ≤ 1) is entropy at time t of the survival function and is calculated starting from the function itself supplied in the life tables. Entropy is equal to 0 when all deaths occur at the highest age, a situation corresponding to the limit case of the present mortality postponement process in increasingly higher age groups mentioned at the end of the previous paragraph; this evolution is known in demography as rectangularisation of the survival curve. Entropy is equal to 1 when the mortality rate is constant in all age groups.

Equation [1] refers to life expectancy at birth at time t, however the ratio can be applied also to residual life expectancy at subsequent ages. In more detail, attention was focussed on elderly ages, whereby the M/F mortality ratio still shows a certain amount of variability in the different age groups, though remaining lower than that referring to a person’s entire life span. Passing on to the results for δ [1] – obtained by determining H(t) through a first approximation – Table 4 shows that, in relative terms, the average level of male over-mortality is not negligible at all and doesn’t follow the decreasing trend registered especially in the highest age groups since the second half of the 1990s, before the so called “young olds”.

 

Table 5 – Fraction added to life expectancy deriving from a proportional reduction in mortality rates at all ages – Italy 2010

Age

 

                                  Percentage reduction

0.50%   0.75%  1.00%   1.25%   1.50%   1.75%   2.00%   2.50%

 

 

 

45

50

55

60

65

70

75

80

85

 

 

 

45

50

55

60

65

70

75

80

85

 

 

Males

 

0.046    0.069   0.092    0.115    0.138    0.161    0.184    0.231

0.044    0.067   0.089    0.112    0.134    0.157    0.179    0.225

0.043    0.064   0.086    0.107    0.129    0.151    0.173    0.216

0.040    0.061   0.081    0.102    0.122    0.143    0.163    0.205

0.038    0.057   0.075    0.094    0.114    0.133    0.152    0.190

0.034    0.051   0.068    0.086    0.103    0.120    0.138    0.173

0.030    0.045   0.061    0.076    0.091    0.107    0.122    0.153

0.026    0.039   0.052    0.065    0.079    0.092    0.105    0.132

0.021    0.032   0.043    0.054    0.065    0.076    0.087    0.109

 

Females                                           

0.042    0.063   0.084    0.105    0.126    0.147    0.168    0.211

0.041    0.061   0.082    0.102    0.123    0.144    0.164    0.206

0.039    0.059   0.079    0.099    0.119    0.139    0.159    0.200

0.038    0.057   0.076    0.095    0.114    0.134    0.153    0.192

0.036    0.054   0.072    0.090    0.109    0.127    0.145    0.182

0.034    0.050   0.067    0.084    0.101    0.118    0.136    0.170

0.031    0.046   0.062    0.077    0.093    0.109    0.124    0.156

0.027    0.041   0.055    0.068    0.082    0.096    0.110    0.138

0.023    0.035   0.046    0.058    0.070    0.081    0.093    0.117

 

Source: Data processed from 2010 Istat life table, 2014.

 

These trends, as well as those surveyed with other indicators, raise the doubt that gender problems have not been taken into account by the National Health System when studying health conditions among elderly age groups. This is confirmed even on the basis of pilot tests conducted under the “neutral” assumption of a uniform mortality reduction for both sexes[xviii]: using some representative average annual rates among those registered in the period between 1979-2010 (see Figure 4), in the present mortality structure after the age of 75, a higher increase (in absolute terms) of women’s residual life expectancy compared to men’s can always be observed; in the previous ages the contrary is registered (see Table 5).

 

2.                  A HINT TO THE INTERACTION BETWEEN COHORT AND PERIOD-RELATED FACTORS IN THE ELDERLY AGES

 

In order to grasp the role played by cohort factors on mortality, it is necessary to use longitudinal rather than cross-sectional analysis. However, the “cohort effect” by gender in the age group comprised between the “young olds” and the “oldest olds” can be highlighted also by the accelerating-decelerating pattern of mortality using the approach of Horiuchi and Coale[xix].

These authors agree with Gompertz (1825), according to whom the deteriorating process of a person’s life funtions increases by aging and consequently his resistance against death decreases at an approximately constant rate; on the other hand, they also take into account that the population is composed of individuals with different personal frailties against death, a characteristic known as population heterogeneity. By increasing age the most selected individuals remain alive, i.e. those who can better resist the elimination process; for this reason, starting from a certain age, usually comprised between 55 and around 90 – the analysis of mortality will stress a process first characterised by a rising and then declining profile.

The existance of this process should be inferred from the so-called Life-Table Aging Rate (LAR), indicated by the authors with k(x) and obtained from the relative derivative of the instant mortality rate μ(x)[xx], i.e.:

 

                1       d μ(x)           dlog[μ(x)]

k(x)  =  ——  ———   =  ————                                                    [2]

              μ(x)      dx                    dx

 

In the present case the estimate of μ(x) at age x, and indicated with μx, is obtained by using the survival function lx supplied in the life tables. The equation used for μx  among the proposed equations is the following μx 

 

         8(  lx+1  –  lx–1 ) – ( lx–-2  –  lx+2 )

μx  ————————————                                                         [3]

                           12  lx

 

On the basis of the previous considerations, the k(x) age pattern by different x will be initially increasing and then decreasing, thus showing a bell-shaped profile as observed in Figures 6 and 7 for women and only in Figure 7 for men.

 

 

Figure 6 – Life-Table Aging Rate (LAR) by sex – Italy, 1974


Source: Data processed from 1974 Istat life table, 2014.

 

It should be mentioned that Horiuchi and Coale studies stopped in 1974 and took in exam different countries, with the exception of Italy. It is therefore certainly interesting to find in the present study’s results, constructed specifically for that year, strong similarities with Horiuchi and Coale findings for countries like France and the Federal Republic of Germany. According to the authors, the different profiles for males and females of k(x) between the end of the 1960s and the beginning of the 1970s are the consequence of First World War long-term effects on the health of male survivors in the cohorts which took part in the conflict (1915-18 for Italy); in the case of females, they are due to the interaction of female population heterogeneity with the death selection process mentioned above. Figure 6 shows indeed that the k(x) male pattern related to 1974 is very unstable, while that of women is shaped in the form of a bell; in more detail, their k(x), after reaching a minimum around the age of 60 (see Figure 6), reaches its maximum around the age of 75, equal to nearly 0.128, only slightly higher than that registered in 1974 in the previously mentioned countries. In the immediately subsequent years these profiles can be found again but, as time goes by and the cohorts that experienced more than others a particularly difficult historical period, such as the first half of 1900, disappear, male LAR assumes an increasingly similar profile to female’s. Towards the end of the observation period (2010) both k(x) profiles, though showing different levels, are somehow similar (see Figure 7).


 Figure 7 – Life-Table Aging Rate (LAR) by sex – Italy, 2010


Source: Data processed from 2010 Istat life table, 2014.

 

Comparing the case of women in 1974 and 2010, it can be noticed that, by increasing life expectancy, mortality deceleration has been postponed, because by reducing mortality also the selection process has been slowed down. Comparisons between males and females for 2010 highlights two processes characterised by higher speed in the case of women and decidedly lower speed for men (see Figure 7); the latter show a lower acceleration and deceleration pattern because they underwent a greater selection process, as seen in the previous paragraph. 

It is well known that among persons aged 65 and over women often have a worse perception of their health conditions. However, this fact does not seem to affect differences in mortality. Up until recently (2010), although in that age group women residual life expectancy was still very much longer than men’s, the number of years lived in good or very good health conditions (5.2) was little lower than among men (5.6)[xxi]. Within this life span, the main health problem is represented by the increase of chronic diseases and disabilities, which most likely found the National Health System unprepared. The reduction of the gap among male and female life expectancy is therefore also dependent on the effectiveness of the actions carried out that cannot be simply generically valid, as shown at the end of the previous paragraph.

 

3.              CONCLUSIONS

 

As illustrated in this paper, the evolution of differences in mortality is increasingly affected by mortality postponement with deaths concentrating in the highest ages, where aging and the quality of old age are the result of slowly changing life paths from one cohort to the next. The cohorts born between the end of the 19th and the beginning of the 20th century, who are to be found in the period life tables used for the present study, experienced catastrophic events such as wars, epidemics and economic crisis; particularly men, especially those employed in the industrial sector, carried out hard and stressful tasks, often in noxious working environments, whose harmful effects on health in the long term were aknowledged in many cases only much later thanks to the progress of medical sciences. In addition, those who were born and had lived for most of or all their lives under these conditions experienced different socialisation patterns between genders and became old differently from today, also because they absorbed those socio-cultural rules which considered old-age as an irreversible declining stage, the descendent parable of the individual’s life after the maximum peak represented by youth.

On the one hand, it is surely not possible to deny bio-physiological aging and its rules; on the other, recent observations showed that individual aging is also determined by socio-cultural rules and beside biologic aging there is also an elderly condition created by society. The present differences in mortality between “olds” and “oldest olds” is also a mirror of this concept of old age rooted very far away in time and well expressed by P. Terenzio Afro words: senectus ipsa est morbus (old age is a disease in itself). A trace of this vision of elderly persons as individuals wearing biologically due to the effect of time passing can be also found in one of the first drawing up of demographic models, the so-called survival law by Gompertz (1825) which is still today a reference for comparative analysis. Moreover, considering individual aging a disease led to look at many diseases typical of this stage of life as intrinsically progressive, irreversible, to be almost accepted fatalistically; for this reason, elderly persons of that time were the survivals of a strong selection process and longevity was a target reached only by some.

Later on, the population demographic aging process, going on in Italy for nearly 30 years, has increased the share of elderly persons now representing an important part of population with own economic and political weight. In addition, this fact contributed to stimulate the spreading of a proactive action within National Health Services as their health protection and longevity itself has become an increasingly studied event. Women benefited much more than men from these changing attitudes and one of the results of gender differences in mortality is represented by their dominant presence in the elderly and old age groups. In Italy in 2011 women represented 57.7% of population aged 65 and over and reached 69.7% of population aged over 85. The study of women life paths who had reached these ages led to identify a further explanation of women higher longevity in their psychology, which leads them to build social and solidarity networks in a much more efficient way than men. Thanks to their different way to relate to the external world, women showed higher capacity to adapt to life changes and life stages[xxii] and to different human relations. It is still unclear if genetic factors may play some role in the over than double life expectancy gender gap of persons aged over 75 (see Figure 3).

An acceleration in the change of gender relations took place in the years subsequent to 1968 which gave further impetus to women emancipation in Italy, promoting the homogenisation of male and female socialisation processes among the recent cohorts. On the one hand these changes are favouring the disappearance of social habits not in line with gender balance; on the other, through emancipation women have assumed typically male life styles which are noxious for their health.

The new sharing of equal life conditions by gender, as well as of the same socio-cultural environment, will certainly diminish the present imbalance in terms of survival; their abrupt reduction, however, will require, especially by the side of men, the choice of a life style as much as possible free from risk behaviours; presumably, the results of this process will be visible only in the long term.

 

 

 

Bibliography

 

ARRIAGA Eduardo, “Measuring and Explaining the Change in Life Expectancies.” Demography, 1984, 21, pp. 83–96.

Atlante ERA 2007. Mortalità evitabile e contesto demografico per Usl. Era epidemiologia e ricerca applicata, Universitalia, 2012.

BOCK, Gisela, “Women’s History and Gender History: Aspects of an International Debate”, in Gender and History, Volume 1, Issue 1, 1989, pp.7-30.

CASTELLI, Adriana, NIZALOVA Olena, “Avoidable Mortality: What it Means and How it is Measured”, CHE Research Paper 63, 2011.  www.york.ac.uk/che/publications/

CESA-BIANCHI, Marcello, Psicologia dell’invecchiamento, Carocci, Roma 2000.

CONNELL, Raewyn, Gender, Polity Press, Cambridge, 2002.

CRIMMINS, Eileen, PRESTON Samuel, BARNEY, Cohen (eds.), International Differences in Mortality at Older Ages: Dimensions and Sources, The National Academies Press, Washington, DC, 2011.

EUROPEAN HEALTH AND LIFE EXPECTANCY INFORMATION SYSTEM (EHLEIS), Country Reports, Issue 6, April 2013.

FEDERICI, Nora, “La mortalità differenziale dei due sessi e le sue possibili cause”, Statistica, 1950, 10, 3, pp. 274–320.

FRIEDAN, Betty, The fountain of age, Simon and Schuster, New York, 1993.

GAY, Juan, et al., “Mortality Amenable to Health Care in 31 OECD Countries. Estimates and Methodological Issues”, OECD Health Working Papers, n°55, OECD Publishing, 2011. 

HORIUCHI, Shiro, COALE, Ansley, “Age patterns of mortality for older women: an analysis using the age-specific rate of mortality change with age”, Mathematical Population Studies, 1990, Vol. 2 (4), pp. 245-267.

HORIUCHI, Shiro,  WILMOTH,  John, “Age  Patterns of the Life Table Aging Rate for Major  Causes of  Death in Japan, 1951-1990”, Journal of Gerontology: Biological Sciences, 1997, Vol. 52A, B67-B77.

KEYFITZ, Nathan, Applied Mathematical Demography, Springer Verlag, New York, 1985.

ISTAT-ISS, La mortalità in Italia nel periodo 1970-1992: evoluzione e geografia, Istat, Roma,1999.

ISTAT, “Tavole di mortalità della popolazione italiana per provincia e regione di residenza”. Anno 1998, Informazioni, Roma, 2001.

LIU, Yan et al., “Gender gaps in life expectancy: generalized trends and negative associations with development indices in OECD countries”, European Journal of Public Health, 2013, 23 (4), pp. 563-568.

MACCHERONI, Carlo, “La speranza di vita ed il rischio longevità nell’offerta di rendite”, in PACI, Sergio (ed.), Il sistema delle regole per le rendite di previdenza complementare, Milano, Egea, pp. 69–110, 2012.

MACCHERONI, Carlo, “Le differenze di genere alla morte in Italia: evoluzione recente e tendenze in atto”, Carefin Occasional  paper, Carefin, Università Bocconi, 2013.

MADIGAN, Francis, “Are sex mortality differentials biologically caused?” Millbank Memorial Fund Quarterly, 1957, 25 pp. 202-223.

OKSUZYAN, Anna, JUEL, Knud, VAUPEL, James, CHRISTENSEN, Kaare, “Men: good health and high mortality. Sex differences in health and aging”. Aging Clinical and Experimental Research, 2008, 20 (2), pp. 91-102.

PAMPEL, Fred, “Cigarette Use and the Narrowing Sex Differential in Mortality”, Population and Development Review, 2002, vol. 28, 1, pp. 77-104.

ROGERS, Richard et al.,Social Behavioral, And Biological Factors, And Sex Differences in Mortality”, Demography, Volume 47-Number 3, August 2010, pp. 555–578.

SUARDI, Teresa, Invecchiare al femminile, La Nuova Italia Scientifica, Roma, 1993.

SULLEROT, Évelyne, L'age de travailler, Fayard, Paris, 1986.

VALLIN, Jacques, “Social change and mortality decline: women's advantage regained or achieved?”,  in Nora FEDERICI, Karen OPPENHEIM-MASON, Sølvi SOGNER (eds), Women's position and  demographic change in the course of development, pp. 190–212., Clarendon Press, Oxford, 1993.

VALLIN, Jacques, MESLÉ, France, VALKONEN, Tapani, Trends in mortality and differential mortality, Council of Europe, Strasbourg, 2001.

WALDRON Ingrid, “Contributions of biological and behavioural factors to changing sex differences in ischaemic heart disease mortality”, in LOPEZ, Alan, CASELLI, Graziella, VALKONEN, Tapani (eds.), Adult mortality in developed countries. from description to explanation, pp. 161-178, Clarendon Press, Oxford, 1995.

WUNSCH, Guillaume, TERMOTE, Marc, Introduction to Demographic Analysis, Plenum Press, New York, 1978.

 

 



[i] Jacques VALLIN, France MESLÉ, Tapani VALKONEN, Trends in mortality and differential mortality, Council of Europe, Strasbourg, 2001. Yan LIU et al., “Gender gaps in life expectancy: generalized trends and negative associations with development indices in OECD countries”,  European Journal of Public Health, 2013, 23 (4), pp. 563-568.

[ii] Nora FEDERICI, “La mortalità differenziale dei due sessi e le sue possibili cause”, Statistica, 1950, 10, 3, pp. 274–320.

Francis, MADIGAN, “Are sex mortality differentials biologically caused?” Millbank Memorial Fund Quarterly, 1957, 25, pp. 202-223.

Jacques VALLIN, “Social change and mortality decline: women's advantage regained or achieved?”,  in Nora FEDERICI, Karen OPPENHEIM-MASON, Sølvi SOGNER (eds), Women's position and  demographic change in the course of development, pp. 190–212., Clarendon Press, Oxford, 1993.

Ingrid WALDRON, “Contributions of biological and behavioural factors to changing sex differences in ischemic heart disease mortality”, in LOPEZ, Alan, CASELLI,  Graziella, VALKONEN, Tapani (eds.), Adult mortality in developed countries : from description to explanation, pp. 161-178, Clarendon Press, Oxford, 1995.

[iii] Richard ROGERS, et al., “Social Behavioral, And Biological Factors, And Sex Differences in Mortality”, Demography, Volume 47-Number 3, August 2010: 555–578.

[iv] Gisela BOCK, “Women’s History and Gender History: Aspects of an International Debate”, in Gender and History, Volume 1, Issue 1,1989, pp.7-30.Teresa, SUARDI, Invecchiare al femminile, La Nuova Italia Scientifica, Roma, 1993.Raewyn, CONNELL, Gender, Polity Press, Cambridge, 2002.Betty, FRIEDAN, The fountain of age, Simon and Schuster, New York, 1993.

[v] Guillaume WUNSCH, Marc, TERMOTE, Introduction to Demographic  Analysis, Plenum Press, New York, 1978.

[vi] ISTAT, “Tavole di mortalità della popolazione italiana per provincia e regione di residenza”. Anno 1998, Informazioni, Roma, 2001.

[vii] ISTAT, “Tavole di mortalità della popolazione italiana per provincia...cit.”.

[viii] ISTAT-ISS, La mortalità in Italia nel periodo 1970-1992: evoluzione e geografia, Istat, Roma,1999.

[ix] Adriana CASTELLI, Olena, NIZALOVA, “Avoidable Mortality: What it Means and How it is Measured”, CHE Research Paper 63, 2011.  www.york.ac.uk/che/publications/

Juan, GAY, et al., “Mortality Amenable to Health Care in 31 OECD Countries. Estimates and Methodological Issues”, OECD Health Working Papers, n°55, OECD Publishing, 2011.

[x] Atlante ERA 2007. Mortalità evitabile e contesto demografico per Usl. Era epidemiologia e ricerca applicata, Universitalia, 2012.

[xi] Eduardo ARRIAGA, “Measuring and Explaining the Change in Life Expectancies” Demography, 1984, 21, pp. 83–96.

[xii] ISTAT, “Tavole di mortalità della popolazione italiana per provincia...cit.”.

[xiii] Carlo MACCHERONI, “La speranza di vita ed il rischio longevità nell’offerta di rendite”, in PACI, Sergio  (ed.), Il sistema delle regole per le rendite di previdenza complementare, Milano, Egea, 2012, , pp. 69–110.

[xiv] Guillaume WUNSCH, Marc, TERMOTE, Introduction to Demographic  Analysis...cit.

[xv] Nathan KEYFITZ, Applied Mathematical Demography, Springer Verlag, New York, 1985.

[xvi] Carlo MACCHERONI, “Le differenze di genere alla morte in Italia: evoluzione recente e tendenze in atto”, Carefin Occasional paper, Carefin, Università Bocconi, 2013.

[xvii] Nathan KEYFITZ, Applied Mathematical Demography...cit.

[xviii] Carlo MACCHERONI, “La speranza di vita ed il rischio longevità ...cit.”.

[xix] Shiro HORIUCHI, Ansley COALE, “Age patterns of mortality for older women: an analysis using the age-specific rate of mortality change with age”, Mathematical Population Studies, 1990, Vol. 2 (4), pp. 245-267.

[xx] Nathan KEYFITZ, Applied Mathematical Demography...cit.

[xxi] Anna OKSUZYAN, Knud JUEL, James VAUPEL, Kaare CHRISTENSEN, “Men: good health and high mortality. Sex differences in health and aging”. Aging Clinical and Experimental Research, 2008, 20 (2), pp. 91-102.

EUROPEAN HEALTH AND LIFE EXPECTANCY INFORMATION SYSTEM (EHLEIS), Country Reports, Issue 6, April 2013.

[xxii] CESA-BIANCHI, Marcello, Psicologia dell’invecchiamento, Carocci, Roma 2000.