A great number of studies have analysed the mortality disadvantage of Central and Eastern European countries compared to other European countries. Theories of mortality trend convergence and divergence are based primarily on analysis of situations in individual Central and Eastern European countries. The authors adopt a new perspective and method here, using truncated cross-sectional average length of life (TCAL) to see how different cohorts contribute to the observed mortality gaps. This method produces quite different results from life expectancy at a given time, while bringing to light histories that differ considerably by cohort and country.
1Improvements in mortality do not take place uniformly across regions. Generally, mortality is higher in places where the standard of living is lower. In Europe, it is well documented that the mortality levels of countries in Central and Eastern Europe (CEE) are far higher than those of more developed regions (WHO, 1995; Meslé, 1996, 2004; Mustard, 1996; Velkova et al., 1997; Meslé and Vallin, 2002; Andreev et al., 2003). For instance, in the early 1990s, the mortality gap stood at over 10 years between Eastern European countries with the lowest life expectancies and Western ones with the highest (Bobak and Marmot, 1996). Yet, despite the great survival improvements in Eastern European countries, mortality differences persist in Europe (Leon, 2011; Shkolnikov et al., 2013; Mackenbach, 2013; Meslé and Vallin, 2017).
2The mortality gap between Central and Eastern and Western Europe is due mostly to changes in health and disease patterns (Bobak and Marmot, 1996; Nolte et al., 2000; Andreev et al., 2003; Meslé et al., 2012), and it is usually attributed to differences in socioeconomic, environmental, and public-health investments (Watson, 1995; Bobak, 1996; Bobak and Marmot, 1996; Forster, 1996; Velkova et al., 1997; Vogt et al., 2017). From the end of World War II to the mid-1960s, the increasing use of antibiotics and immunization led to CEE countries achieving significant progress in survival from infectious diseases, particularly among the youngest ages (Meslé and Vallin, 2002; Vallin and Meslé, 2004). During that time, CEE countries converged towards lower mortality levels, and some almost succeeded in catching up with Northern and Western European countries (Meslé and Vallin, 2002; Meslé, 2004). Then, in places where mortality at the youngest ages had already reached low levels, a new challenge emerged in the form of improving longevity (Vallin and Meslé, 2004; Canudas-Romo, 2010; Bergeron-Boucher et al., 2015). Following the epidemiologic transition (Omran, 1971), more developed regions experienced great progress in survival from degenerative and man-made diseases. Conversely, in CEE countries, particularly those in the former USSR, where high mortality hit adults the hardest, from the mid-1960s to the mid-1980s, adult and old-age mortality increased for men and stagnated for women (Bobak and Marmot, 1996; Shkolnikov et al., 1997; Meslé and Vallin, 2002). As a result, mortality began to diverge between Central and Eastern and Western Europe (Shkolnikov, 2004; Vallin and Meslé, 2004).
3From the late 1980s, a new divergence in mortality trends began to emerge within the CEE countries, thus producing a clear gap between Central and Eastern Europe. In Central Europe, health improvements reduced mortality from cardiovascular disease, which in turn increased life expectancy (Meslé, 2004). In contrast, countries of the former USSR experienced a brief period of improvement (1985–1986), followed by a sharp increase in mortality due to the economic crises of the 1990s and low investments in public health (Leon et al., 1997; Gavrilova et al., 2001; Andreev et al., 2003; Meslé, 2004; Shkolnikov, 2004; Grigoriev et al., 2010).
4However, since the beginning of the 21st century, a new mortality trend has been observed in the former USSR countries, where mortality from cardiovascular disease and external causes at adult ages has started to decline (Grigoriev et al., 2010; Jasilionis et al., 2011; Shkolnikov et al., 2013; Grigoriev et al., 2014; Grigoriev and Andreev, 2015). In Russia, for instance, life expectancy at birth for both sexes increased by more than 5 years between 2004 and 2014, reaching 70.91 years in 2014, the highest level in the country’s recent history (Human Mortality Database, 2019). Another huge survival improvement took place in Belarus, where male life expectancy increased by about 2 years in a calendar year (2011–2012) (Grigoriev and Andreev, 2015). Despite this recent great mortality improvement in terms of health, countries of the former USSR still lag behind those of Western Europe.
5More developed countries, like those in Western Europe, have shown long-term improvements in health while also experiencing a sustainable decline in mortality over recent decades. From a cohort perspective, a continuous decrease in mortality leads to several generations experiencing the gradual benefits of health improvements, with younger cohorts benefiting more because their mortality experience begins at levels lower than those of older cohorts.
6In the case of discontinuous or short-term health progress, on the other hand, only a few cohorts can enjoy the benefits of health advances. The anti-alcohol campaign that the Soviet Union established in a specific period (1985– 1986) provides an example. The campaign reduced the mortality of several cohorts at adult ages during the mid-1980s, leading to a short period of increasing life expectancy (Shkolnikov and Nemtsov, 1997; Shkolnikov et al., 2004; Shkolnikov, 2012; Grigoriev and Andreev, 2015). Then, the economic crises of the 1990s hit certain age groups in countries of the former USSR harder than others, with some cohorts experiencing greater negative health effects (Shkolnikov, 2012).
7German reunification also exemplifies the importance of period changes in reducing mortality. After the fall of the Berlin Wall, mortality improved for all age groups in East Germany but at a differential pace among ages (Vogt, 2013; Vogt and Kluge, 2015, Vogt et al., 2017), which suggests that the mortality gap differs across birth cohorts. For instance, the Eastern European cohorts born in 1890, 1900, and 1910 converged to Western European mortality levels faster than the younger cohorts (born in 1920 and 1930) (Vogt and Missov, 2017). Thus, we can speculate that the mortality differences between East and West German older cohorts fell faster than those between younger cohorts. This reduction indicates that the contributions to the mortality gap during the pre-unification period differ across birth cohorts in Germany.
8Given the different political regimes, the fact that CEE socioeconomic and medical policies have differed from those of Western Europe over recent decades, and because changes have had major effects on some CEE birth cohorts but not on others, we wondered whether the mortality gap between CEE and Western European countries also varies across birth cohorts. Further, we hypothesized that the contribution of each birth cohort to the overall mortality gap in a given period can vary.
9Our aim here is to compare mortality between CEE and a group of high-longevity countries (HLCs) through a measure similar to period life expectancy but based on available cohort survival data. To accomplish this, we calculate and then decompose the ‘truncated cross-sectional average length of life’ (TCAL) measure (Canudas-Romo and Guillot, 2015). By decomposing the gap in TCAL by age and cohort, we can analyse the survival trajectories of CEE cohorts in comparison with their HLC counterparts. Moreover, such decomposition makes it possible to identify both the short- and long-lasting effects of the survival advantage/disadvantage at a given age for a certain CEE cohort. We complement previous studies by adding cohort mortality dynamics to help form a better understanding of the mortality gap between CEE and HLCs.
I – Methods
10TCAL is a cross-sectional measure that summarizes historical mortality information about all cohorts present at a given time, and it is not limited to populations for which complete cohort mortality data is available (Canudas-Romo and Guillot, 2015). It derives from the cross-sectional average length of life (CAL) measure, developed by Brouard (1986). CAL is the sum of proportions of survivors among the various cohorts present in the population at time t (Guillot, 2003). The difference between CAL and TCAL is that the latter can be calculated for populations without complete cohort mortality data. TCAL provides a novel way of comparing mortality and investigating survival disparities between populations by considering all the information available for all cohorts present at a given time, regardless of whether complete cohort data is available and whether the data come from a young or old cohort. The TCAL measure has an advantage over assessments of just a single year, which, from a period analysis perspective, combines pieces of mortality information from different cohorts. It also has an advantage over assessing, one by one, each individual cohort present at a given time while not knowing how they jointly contribute to the overall survival disparities between populations.
11To calculate TCAL, we define the year, t, for which we are interested in estimating the measure, and the earliest year for the available mortality series, Y1. Thus, TCAL for year t, truncated at year Y1, is computed as:
13where l(x, t, Y) is the survival function for cohorts reaching age x in year t, whose members were born in year t − x. In the Lexis diagram shown in Figure 1, TCAL(t, Y1) includes mortality rates from year Y1, which are located along diagonals that cross the age axis at time t.
14Some of the cohorts were born after year Y1, for which we have full cohort information. For cohorts born before year Y1, only partial cohort mortality data are available, so we assume a set of death rates for the years before year Y1. Since our interest is in the mortality gap between populations, the TCAL differences will be consistent if we use the same set of death rates for the years before Y1 in all the examined countries (Canudas-Romo and Guillot, 2015). To eliminate any confounding effects of death rates before the year Y1, we assume death rates equal to 0 for all the years before Y1, thereby focusing our comparisons solely on the cohort information available.
15To compare two populations at time t, both TCALs must be truncated at the same year (Y1), which means, in this case, that the mortality series for all CEE countries and for the group of HLCs must start at Y1. Thus, after comparing TCALs of each CEE country with the group of HLCs, we see which populations had higher mortality levels according to historical mortality data. Lower TCAL values correspond to populations that had higher cohort mortality levels.
Lexis diagram for the location of death rates used in TCAL(t, Y1)
Lexis diagram for the location of death rates used in TCAL(t, Y1)
16The difference in TCALs between the group of HLCs and each CEE country, i, is then:
18where the integral corresponds to the cohorts present at time t, aged 0 to ω, and both populations have the same set of age-specific death rates in year Y1. The cohort survival differences on the right side of Equation 2 allow us to identify the mortality contribution of each cohort present in year t. The difference between TCALs is comparable to the difference between life expectancies in that it shows the number of years one population lags behind another.
19We can rewrite Equation 2 using the definition of cohort survival as:
21where μHLC (a, t – x + a) and μi (a, t – x + a) are the forces of mortality at age a and time t − x + a for, respectively, the HLCs and population i. As TCAL condenses the available cohort mortality history into one measure, Equations 2 and 3 show that any differences between TCALs allow us to identify cohortspecific contributions to the mortality gap. Thus, the age–cohort contribution Δ (a, t – x,i) to the difference between the TCAL of the HLCs and that of the population i, TCALi can be estimated as:
23where l(x, t, Y1, i) and l(x, t, Y1, HLC) are the survival functions for the cohort aged x at time t in, respectively, the group of HLCs and population i; and 1pa(t – x, HLC) and 1pa(t – x, i) are the probabilities of surviving from age a to a + 1 for the cohort born in year t – x in, respectively, the HLCs and population i. Finally, instead of the integrals in Equations 2 and 3, the sum over cohorts and ages of the age–cohort contributions, Δ(a, t – x, i), returns the difference in TCALs:
25By means of such decomposition, we compare mortality between birth cohorts from different populations.
26The main limitation of the method is data availability. In principle, we would be interested in presenting as much cohort data as possible. However, this is not possible for many regions of the world, such as in CEE. Furthermore, despite constraints on data quantity, data quality improves over time. Thus, any measure with a cohort perspective will include some of the quality bias that exists in the older information.
II – Data
27From the Human Mortality Database (2019), we selected 11 CEE countries: Belarus, Bulgaria, Czechia, Estonia, Hungary, Latvia, Lithuania, Poland, Russia, Slovakia, and Ukraine. The HLCs included in the analysis are Australia, Austria, Belgium, Canada, Denmark, Germany, Finland, France, Iceland, Ireland, Italy, Japan, the Netherlands, Norway, New Zealand, Portugal, Spain, Sweden, Switzerland, the United States, and the United Kingdom. The same selection of HLCs has been used elsewhere (Ho and Preston, 2010; Canudas-Romo and Engelman, 2012; Canudas-Romo and Guillot, 2015) to represent the lowest mortality levels. For this group, we calculated period age-specific death rates by adding annual death counts and exposures from each country.
28We used mortality series from 1959 to 2013, except in the cases of Canada and Bulgaria, where mortality series were up to 2011 and 2010, respectively. To compare mortality levels between each CEE country and the group of HLCs, we truncated all series in 1959, the first year of available HMD data for most CEE countries; then, we calculated TCAL in 2013, truncated in 1959—or, as expressed in Equation 1, TCAL (2013, 1959).
29The HMD data provide detailed historical information on mortality for most industrialized countries. However, the quality of data for 1959–1969 is lower than in later years for Latvia, Russia, Ukraine, and Belarus (Grigoriev, 2015; Pyrozhkov et al., 2015; Shkolnikov and Jdanov, 2016; Jasilionis, 2017b). Also, data for 1959–1979 in Lithuania should be used with caution, while the quality of the data in Estonia for 2001–2009 is lower than in previous years (Jasilionis, 2017a, 2017c). Information from earlier years (for old ages) should also be treated with caution due to data quality issues in countries of the former USSR. In Central Europe, the quality of data in Slovakia from 1959 to 1961 and in Hungary for 1959 is lower than in later years (Mészáros and Jasilionis, 2015; Jasilionis and Radnóti, 2016).
30The HMD publishes cohort death rates only for cohorts that have at least 30 years of data. For instance, period death rates for Ukraine are currently available up to 2013, but cohort death rates are available only up to the cohort born in 1983. Therefore, to avoid an interruption in the cohort series, we used period death rates to reconstruct the diagonals. This was done consistently across all the countries analysed, for both CEE countries and HLCs.
31We further carried out a sensitivity analysis using data for countries that have enough cohort data to construct complete (a) cohort life tables, which we then compared with (b) cohort life tables, as estimated previously, based on period rates. Based on the HMD data and combining females and males for the 1900 cohort, disparities between life expectancies at birth for each country were minor: (a) 59.3 and (b) 59.2 years for Denmark, (a) 55.1 and (b) 55.0 years for the Netherlands; (a) 59.6 and (b) 59.4 years for Norway; and (a) 58.8 and (b) 58.7 years for Sweden. Both the sensitivity analysis and the consistency of the procedure for all countries reassured us that our results are not biased by the set of death rates selected.
III – Results
32Table 1 provides rankings of life expectancy at birth from highest to lowest. The table covers life expectancy at birth (e0) based on current mortality in 2013, and TCAL (2013, 1959), which captures the available mortality history from 1959 of all cohorts present in 2013, for each CEE country by sex. As expected, all CEE countries lag behind the group of HLCs. In comparing e0 with TCAL, Table 1 displays TCAL values that are lower than e0 for all countries. This is explained by TCAL taking into account the higher mortality levels in the historical mortality data, while they are not considered in e0.
33When CEE countries are ranked according to TCAL and e0, the Central European countries are clustered at the top, while the Eastern European countries are at the bottom. This suggests that both historical and current mortality in Eastern Europe are higher than those in Central Europe. For instance, Ukraine and Russia have the highest mortality levels according to both TCAL and e0, while Czechia and Poland are in the top three for both rankings. Despite this overall picture, we also identify some country-specific arrangements in Table 1. For instance, Estonian women move from the top of the life expectancy ranking (e0 = 81.3 years) to third position in TCAL (77.9 years); Estonian men also go down in the rankings when transiting from e0 to TCAL, from fourth (e0 = 72.7 years) to sixth (TCAL = 66.7 years). Since the beginning of the 21st century, Estonia has shown remarkable mortality improvements. From 2000 to 2010, Estonian life expectancy increased by 4.2 years for women and 5.2 years for men. Indeed, these recent increases in life expectancy were 3 times and 11 times higher than over the previous decade for women and men, respectively. This exemplifies the existing differential between current and cohort mortality, as depicted by life expectancy and TCAL.
34To show the performance of CEE countries in relation to the HLCs, Table 1 presents the differences in e0 and in TCALs between each CEE country and the group of HLCs in 2013. In both cases, the greatest differences are between the countries of the former USSR and the HLCs. For instance, the Russian male TCAL was 14.1 years lower than the TCAL in HLCs in 2013, while the gap in e0 is 13.2 years. In Czechia, however, the male gap in TCAL (Czechia vs. HLCs) is 3.8 years, about 10 years lower than the male TCAL difference between Russia and the HLCs. Considering this smaller difference in TCALs between Czechia and the HLCs, compared with the gap between Russia and the HLCs, one may conclude that the historical improvements in male survival have been much greater in Czechia than in Russia.
Central and Eastern European countries: life expectancy at birth (e0), TCAL and differences from the high longevity countries’ e0 and TCAL for males and females in 2013*,**,***
Central and Eastern European countries: life expectancy at birth (e0), TCAL and differences from the high longevity countries’ e0 and TCAL for males and females in 2013*,**,**** e0(2013) for HLCs is 78.3 years for males and 84.5 years for females.
** TCAL(2013,1959) for HLCs is 75.6 years for males and 81.7 years for females.
*** Mortality series from 1959 to 2010: e0(2010) for HLCs is 77.9 years for males and 83.2 years for females.
35The male gap in life expectancy (CEE vs. HLCs) for most CEE countries is smaller than the difference in TCALs (Table 1). In Hungary, the male gap in e0 is 29% lower than the gap in TCALs, while in Estonia the male difference in life expectancy is almost 60% lower than the difference in TCALs (Estonia vs. HLCs). In other words, the difference is higher in male historical mortality than in current mortality when comparing CEE countries with HLCs. The difference in mortality measured with e0, was reduced by recent improvements in the male mortality of CEE countries compared with HLCs. Among women, by contrast, Table 1 shows higher gaps in e0 (CEE vs. HLCs) than differences in TCALs (CEE vs. HLCs) for all countries except Czechia. For instance, Lithuania’s female life expectancy is lagging behind the HLCs by 5.1 years, while the difference in cohort mortality is 20% lower. In Ukraine, the female gap in e0 (Ukraine vs. HLCs) is more than 1 year higher than the difference in TCALs (Ukraine vs. HLCs). This may be explained by the slow progress over recent decades in the female mortality of CEE countries when compared with HLCs. The greater disparity in female life expectancies might be explained by some past mortality improvements among women that were captured by TCAL but not by life expectancy.
36To further understand the differences in cohort survival, Figures 2A (females) and 2B (males) for Central European countries and Figures 3A and 3B (females and males, respectively) for Eastern European countries show the age–cohort decomposition of the TCAL difference between each CEE country and the group of HLCs. Such decomposition allows us to investigate the contribution that cohorts present in 2013 make to the gap in TCALs between each CEE country and the HLCs. Figures 2 and 3 show the Lexis surfaces of the cumulative age and cohort contributions to the difference in TCALs. Each data point (age x and time t) in these figures represents the cumulative difference in cohort survival up to the specific age x and year t. Negative values are associated with higher survival in the HLCs.
37Figures 2A and 2B show lower mortality levels for most cohorts in the HLCs than in their counterparts in Central Europe for both sexes. With the exception of some Czech and Polish cohorts, all Central European cohorts present in 2013 contribute to the overall mortality disadvantage between each Central European country and the HLCs in 2013.
38In comparison, the overall picture of Czechia versus the HLCs stands out from those of the other Central European countries. Figures 2A and 3A show that Czech cohorts born in the late 1950s and during the 1960s, especially those born in the early 1960s, had a particular survival advantage over their counterparts in HLCs from birth until the age they reached in 2013. After World War II and up to the mid-1960s, mortality in Czechia greatly decreased due to the country’s extending health coverage to the entire population. At this time, Czech life expectancy at birth increased at the same rate as in France, and both countries achieved a similar mortality level (Rychtaříková, 2004). In addition to confirming the great mortality improvement during the 1960s, our decomposition reveals the long-lasting effect of lower mortality at younger ages for Czech cohorts compared with the HLCs. The low mortality in infancy and childhood lasts until 2013, and it is seen in Czechia’s higher cohort survival for cohorts born from 1959 to the early 1970s. These figures also suggest a more recent cohort development, i.e. lower infant/child mortality in Czechia than in the HLCs. These recent mortality improvements in Czechia may contribute to the slightly lower gap in e0 (Czechia vs. HLCs) than in TCALs.
Lexis surfaces for the cumulative age and cohort contributions to the difference in TCALs between CEE countries and other HLCs
Lexis surfaces for the cumulative age and cohort contributions to the difference in TCALs between CEE countries and other HLCsNote: Negative values correspond to higher survival in HLCs.
39In our window of observation, from 1959 to 2013, Figures 2A and 2B also reveal that Central European cohorts born from the 1920s to the late 1950s experienced lower mortality compared with the group of HLCs. This survival advantage is greater for females than for males. However, all these Central European cohorts have gradually lost their survival advantage compared with HLCs. Figure 2B shows that the male cohorts aged 60–80 in the 2000s are the ones that contribute the most to the mortality gap between Central European countries and HLCs in 2013. Among women, despite the great contribution that the cohorts aged 60–80 in the 2000s make to the differences in TCALs, their contribution is lower in comparison to that of men.
Lexis surfaces for the cumulative age and cohort contributions to the difference in TCALs between former USSR countries and other HLCs
Lexis surfaces for the cumulative age and cohort contributions to the difference in TCALs between former USSR countries and other HLCsNote: Negative values correspond to higher survival in HLCs.
40For the former USSR countries (Figures 3A and 3B), the overall picture is very similar: a great survival disadvantage exists for most cohorts. Moreover, survival disadvantage (Eastern European countries vs. HLCs) for all cohorts born between 1959 and 2000 increases as the cohorts get older. No Eastern European cohort has experienced lower mortality than HLCs in middle-aged adults. Indeed, the survival disadvantage of Eastern European cohorts greatly increases after age 30.
41With the exception of Russia, all Eastern European cohorts born during the 1960s and the early 1970s experience a small survival advantage at younger ages than do the HLC cohorts (Figures 3A and 3B). However, the survival advantage of these cohorts gradually disappears up to 2013 for both sexes. We also observe a longer-lasting effect of this survival advantage in infancy and childhood for females than for males. Male cohorts experience survival advantages at younger ages (i.e., up to age 20–25) when compared with HLCs, while female cohorts had this survival advantage until age 30–35. The low mortality levels until adulthood of Eastern European cohorts compared with HLCs was probably triggered by the anti-alcohol campaign oriented towards adults in 1985–1987, when cohorts born in the early 1960s had reached ages 20–25. Moreover, our results suggest different effects on males and females as a result of the campaign, which reduced the mortality gap between each former USSR country and the group of HLCs.
42From a cohort perspective, our results also reveal that the effect of the anti-alcohol campaign differs in the former USSR countries when compared with HLCs. In Lithuania, for instance, the survival advantage at younger ages of male cohorts born from the mid-1960s to the early 1970s did not last until age 20 (Figure 3B). By contrast, in Latvia, Ukraine, and Estonia, male cohorts that experienced lower infant and childhood mortality than HLCs retained their advantage up to ages 20–25. These aspects may indicate that, compared to changes in the mortality gap between Lithuania and the HLCs, cohort mortality improvements due to the anti-alcohol campaign more greatly narrowed the gaps between Latvia and the HLCs, Ukraine and the HLCs, and Estonia and the HLCs.
43Since the beginning of the 21st century, recent cohorts are experiencing infant and child survival disadvantage in Eastern European countries when compared with HLCs. Child mortality differences in Belarus and Estonia (Belarus vs. HLCs, and Estonia vs. HLCs) narrowed rapidly for cohorts born in the late 2000s, while child survival progressed slowly in Lithuania, Russia, and Ukraine. Another relevant aspect displayed in Figure 3B for Eastern European countries is the great contribution of male cohorts aged 40–80 in 2013 to the difference in TCALs. This result highlights the important contribution of mortality over age 40 to the mortality gap between Eastern European countries and more developed countries.
44Particular attention is paid to Russia, the only country that has not experienced any cohort survival advantage compared with HLCs. In addition to showing the high cohort mortality levels in Russia, Figures 3A and 3B reveal differences in the survival trajectories of Russian cohorts when compared with HLCs. At younger ages, Russian cohorts born from 1960 to the mid-1970s have experienced a lower survival disadvantage than Russian cohorts born between 1980 and 2000. The mortality disadvantage of Russian cohorts born between 1960 to the mid-1970s lasts until age 20 for males and up to age 30 for females. Again, the anti-alcohol campaign launched by Gorbachev in 1985–1987 may explain this long-lasting effect of lower mortality difference up to adulthood between Russia and HLCs. Among the former USSR countries, only Russian cohorts born during the 1960s have not experienced a survival advantage when compared with HLCs. Even if the positive effect during the period of the anti-alcohol campaign was greater in Russia and led to more important mortality improvements there than in other countries, it was not enough to compensate for previous lower survival.
45Infant and child mortality is much higher in all Russian cohorts born between 1959 and 2013 than in the HLCs. Despite the great mortality improvements at the youngest ages in Russia over recent decades, the country still lags far behind when compared with HLC infant and child mortality. However, women have more quickly progressed to lower levels of infant and child mortality than have men. When compared to HLCs, survival disadvantage is lower for females than for males at the youngest ages of Russian cohorts born during the 1980s and 1990s.
46This study takes a cohort perspective to investigate the mortality gap between CEE countries and a group of HLCs. We have revealed the contribution of cohort survival to the mortality difference between each CEE country and the group of HLCs in 2013. Our decomposition shows a great survival disadvantage for most CEE cohorts present in 2013 compared with their counterparts in HLCs. The age–cohort decomposition of difference in TCALs also reveals some survival advantages of particular CEE cohorts over HLCs, as is the case for Czech cohorts born in the late 1950s and during the 1960s. These Czech cohorts had a particular survival advantage over their HLC counterparts from birth until the age reached in 2013. The survival advantage of these Czech cohorts confirm the documented mortality decline in Czechia during the 1960s, when Czech life expectancy at birth was very similar to that of high-longevity countries (Rychtaříková, 2004). We complement this result by showing the long-lasting effect of a survival advantage at first ages of Czech cohorts born during the 1960s when compared with HLCs. Except for Russia, the age–cohort decomposition of the difference in TCALs also reveals a particular survival advantage of the former USSR cohorts born during the 1960s and early 1970s when comparing them with HLCs. Conversely, the survival advantage of these Eastern European cohorts gradually disappears by 2013 for both sexes. Our results show that the survival advantage at younger ages for these Eastern European cohorts lasted until adulthood (up to age 20 for men and age 30 for women). This effect was probably triggered by the anti-alcohol campaign oriented towards adults and which was launched by Gorbachev in the mid-1980s (Shkolnikov and Nemtsov, 1997), when cohorts born during the 1960s reached young-adult ages.
47The TCAL decomposition helps us to observe how cohorts’ contributions to longevity evolve over time and age. The cohort perspective has been emphasized here because we have aimed to understand how CEE populations arrived at current mortality levels. The methodology of decomposing TCAL is flexible and helps focus attention on the age-specific contributions accumulated across ages or periods. While outside the scope of the current study, taking each of the three perspectives together (age, period, and cohort) in the TCAL decomposition could complement and enrich the knowledge on a population’s mortality transition.
48Since the 1980s, the high mortality in Eastern European countries has been largely attributed to premature deaths in the middle-aged adult population, particularly among males born in the former USSR (Shkolnikov et al., 1997; Shkolnikov and Nemtsov, 1997; Meslé and Vallin, 2002; Meslé, 2004). In our window of observation, from 1959 to 2013, we show that the mortality disadvantages of the middle-aged adult population compared with other HLCs always existed between the former USSR and HLCs. Our results suggest that it is not only mortality among adults that contributes to the current disadvantage gap between the former USSR and HLCs, but that mortality at first ages still contributes to this mortality difference.
49The decomposition of the TCAL differences between CEE countries and HLCs highlights the potential for public health interventions to eliminate and control avoidable mortality gaps.
50Acknowledgements: Marília Nepomuceno acknowledges the support of the European Research Council (grant number 716323) and the Conselho Nacional de Desenvolvimento e Pesquisa (CNPq).