Abstract
For over half a century, life expectancy in Eastern European (former communist) countries has been appreciably lower than in Western Europe, although this difference has been narrowing since 2000. We investigated the impact of the COVID-19 pandemic on these differences. The pandemic reversed the recent convergence and widened the gap to levels observed more than two decades ago (7.9 years for males and 4.9 for females in 2021). Moreover, the trajectory of excess mortality in the pandemic differed between East and West, with the first major peaks in Eastern Europe occurring on average six months after the first peaks seen in Western countries. Despite this, the East suffered greater losses in life expectancy, especially in 2021. This was due to larger relative mortality increases in the East rather than greater frailty of the Eastern European populations as indexed by higher pre-pandemic mortality levels. East-West differences in life expectancy losses in 2021 were substantially explained by COVID-19 vaccination, which together with trust in government accounted for half the gap. We conclude that the East-West differences in life expectancy losses are associated with structural and psychosocial traits that have their roots in the communist era. This includes differences in the connectivity of populations (which drives the differences in timing), as well as profound contrasts in levels of trust in science, authorities, and their capacity to enforce lockdowns and other regulatory measures (driving the huge differences in excess mortality from autumn 2020 onwards).
Introduction
Since the late 1960s, an East-West divide in European life expectancy has existed, with the former communist countries of the East lagging behind those of the West. Whereas countries in Western Europe have shown a steady upward trajectory, Eastern European improvements were small or non-existent up until the late 1980s when the fall of the Berlin Wall was followed by a rapid improvement in the countries of Central and Eastern Europe such as Poland and the Czech Republic. In contrast, the massive societal shock induced by the collapse of the Soviet Union in the early 1990s was accompanied by a sharp fall in life expectancy in nearly all former USSR countries. This negative tendency was transient and at various points between the mid-1990s (e.g. Estonia in 1995) and the mid-2000s (e.g. Russia in 2005) pronounced improvements were seen in all of these countries, with a consequent narrowing of the East-West gap in life expectancy.(1) However, the COVID-19 pandemic appears to have led to, once again, a widening of the East-West gap. This paper focuses on quantifying and understanding the differential impact of the pandemic on the two parts of Europe in terms of temporal dynamics in excess mortality as well as the overall impact on excess deaths and life expectancy losses up to the end of 2021.
East-West differences in the impact of COVID-19 on mortality have already been noted, with countries of Central and Eastern Europe (CEE) being seen to have had a relatively mild first phase of the pandemic but having some of the highest excess death rates from autumn 2020 onwards.(2) In contrast, many Western countries were hit hard in spring 2020 but managed to avoid a large increase in autumn/winter of 2020 and 2021.(3, 4)
The generally lower vaccination coverage in the East compared to the West in 2021 has been suggested as one of the factors contributing to these differences.(4) A systematic review(5) found that lower trust in government and the medical system in Eastern Europe could also have played an important role. Others have reported that excess death rates in the pandemic across 50 countries of the WHO European Region were strongly and inversely correlated with the effectiveness of regulatory enforcement and vaccine coverage.(6)
In this paper, we examine life expectancy losses due to the COVID-19 pandemic comparing the higher losses in the group of 11 Eastern European countries of the former communist bloc with the lower losses in 17 Western European countries (Supplementary Table S1). We investigate the components and possible determinants of this difference and carry out exploratory analyses to test the potential role of a range of factors for the East-West contrast in life expectancy losses in 2021. These include vaccination levels, the stringency of non-pharmaceutical interventions, trust in government/science and the effectiveness of regulatory enforcement in general. We also juxtapose these cross-sectional population characteristics in the context of post-war historical trajectories that may reflect persistent differences between East and West.
Results
Life expectancy trends since 2000 and abrupt changes in 2020-21
The long-term life expectancy disadvantage of the East compared to the West began to reduce in the late 1990s in the new EU member states of Central and Eastern Europe and in the mid-2000s in Russia (Figure 1). Between 2000 and 2019, the difference between the East and West life expectancy means decreased from 7.7 years to 6.3 years for males, and from 4.4 years to 3.2 years for females. This convergence was reversed by the COVID-19 pandemic: in 2020 the East-West difference increased to 6.5 for males and 3.5 years for females, while in 2021 the gap increased further to 7.9 and 4.9 years respectively. This widening was caused by a slightly larger life expectancy fall in the East than in the West in 2020: males — 0.9 in East vs 0.7 years in West, and females — 0.8 years in East vs 0.5 years in West. In 2021, the East saw a continuing fall of 1.3 years for both males and females, while the West experienced a small rebound with a gain of 0.1 years in both sexes. Importantly, the declines in life expectancy in all Eastern European countries were larger in 2021 than in 2020.
The figure shows trends in life expectancy across Eastern and Western European countries together with their the average life expectancy for East (thick Blue) and West (thick Red). The long-lasting longevity gap between East and West was shrinking in 2000-19 due to steeper improvements in the East. This convergence was substantially reversed by the COVID-19 pandemic.
Data shown in this Figure is provided at https://github.com/VMSdemo/East-West-contrast-in-life-expectancy-losses-in-2020-21
Temporal and geographic patterns of excess mortality
The weekly excess crude death rates (ECDR) in 2020-21 by country are shown in Figure 2 together with the Eastern and Western mean ECDR. The countries that showed the highest ECDR peaks in the first pandemic wave (March-May 2020) were in Western Europe. In this initial phase, the pandemic had little impact on mortality in countries of the East. This changed in October 2020 when most Eastern countries showed marked increases in weekly excess deaths. From November 2020 to January 2021, the ECDR peak was higher and wider in the East than in the West. Most importantly, countries of the East experienced further massive elevations of weekly ECDRs throughout 2021.
Western European countries showed the highest mortality peaks in the first pandemic wave (March-May 2020), while Eastern European countries showed marked increases in weekly excess deaths only in October 2020 - January 2021 and further experienced large elevations of excess mortality throughout 2021.
Data shown in this Figure is provided at https://github.com/VMSdemo/East-West-contrast-in-life-expectancy-losses-in-2020-21
Figure 3a shows that the first major peak in ECDR in each country tended to move over time from the southwest to the northeast of Europe. The earliest major excess mortality waves in the spring of 2020 occurred in Italy, Spain, England and Wales, Scotland, Northern Ireland, Belgium, Netherlands, France, and Sweden. Most other Western countries and nearly all Eastern ones experienced later initial peaks between November 2020 and January 2021.
In general, the first major excess mortality peak moves from the southwest to the northeast of Europe. Despite the later onset of excess mortality in Eastern compared to many Western European countries, the East suffered much higher total excess death rates in 2020-21.
Data shown in this Figure is provided at https://github.com/VMSdemo/East-West-contrast-in-life-expectancy-losses-in-2020-21
Despite the later onset of the first excess mortality peaks in Eastern compared to many Western European countries, Figure 3b shows that countries of Eastern Europe suffered much higher excess mortality in 2020-21 than those of Western Europe, with the sole exception of Slovenia. It is particularly striking that while the Western countries situated in the center of Europe such as Germany, Austria, and Switzerland had a similar timing of the first pandemic peak as neighboring countries of East (Poland, Czechia, Hungary, and others), their overall rate of excess mortality in 2020-21 was much lower. Similarly, Latvia and Estonia were similar to Norway and Finland in terms of the late timing of their first peak but experienced much higher total excess mortality.
Life expectancy losses in 2020 vs. 2021: reemergence of the East-West divide
Life expectancy losses express the years of life lost due to excess mortality compared to the counterfactual estimates that might have been observed in the absence of the pandemic. As described in Methods this measure is equal to the difference between life expectancies at birth observed in 2020 and 2021, and life expectancies predicted for 2020 and 2021 from age-specific mortality trends over the period 2005-2019.
Life expectancy losses in 2020 overlapped between East and West, despite a clear East-West difference in period life expectancy (Table 1). However, in 2021 the losses were substantially higher in East than West (Supplementary Figure S1). It is notable that the mean life expectancy losses in Western Europe in 2021 were below the minimum country-specific losses in Eastern countries for both sexes.
While in 2020 there was already a tendency for higher life expectancy losses in Eastern countries, this tendency was more pronounced in 2021 (Figure 4). The upper panels of Figure 4 shows that countries with the largest life expectancy losses in 2020 included both Western (Italy, Spain, parts of the UK, and Belgium) and Eastern (Russia, Lithuania, Bulgaria, Poland, Czechia) countries. However, the lower panels of Figure 4 demonstrate that in 2021, the ten countries with the highest life expectancy losses were all from Eastern Europe.
The figure shows the ranking of Eastern (yellow-brown colour) and Western (green) European countries by life expectancy losses in 2020-21 as well as contribution of mortality changes below and above age 65 to the loss. The contribution of younger ages was generally higher in males, in 2021 and in the East.
Data shown in this Figure is provided at https://github.com/VMSdemo/East-West-contrast-in-life-expectancy-losses-in-2020-21
Figure 4 also shows the contribution of different age groups to overall losses for each country. Overall in 2020 the contribution of deaths occurring under the age of 65 years to total life expectancy losses were 31.5% for males (25.3% in the West vs. 39.5% in the East) and 23.2% for females (19.3% in the West vs. 28.5% in the East). In 2021, deaths under 65 years made up 44.1% of the total losses of life expectancy for males (40.3% in the West vs. 49.7% in the East) and 35% for females (34.8% in the West vs. 35.2% in the East). Importantly, over half of the 2021 male losses were due to these younger deaths in Bulgaria, Russia, Slovakia, Latvia, Estonia, Lithuania, Hungary, and Scotland.
East-West differences: higher relative increase in mortality or higher baseline mortality levels?
The East-West differences in the impact of the COVID-19 pandemic on excess mortality and life expectancy losses can be thought of as having two potential components: those resulting from larger proportional increases in mortality in the East compared to the West and those due to the higher baseline mortality in the East reflecting the poorer health status of Eastern European populations.
In the upper panel of Figure 5, an exponential increase with age in death rates is seen in both Eastern and Western Europe for both sexes with higher death rates in the East at all ages. In contrast, the lower panel shows approximately constant relative excess death rates (observed death rate / expected death rate) in both Eastern and Western Europe. However, the extent of these relative increases was much larger in Eastern than Western Europe. For males, an excess of about 30%-40% for the East and 10%-15% for the West are seen between the ages of 35 and 75 years. For females, the equivalent plateaus are at 40%-45% for the East and 5%-15% for the West. Thus despite exponential increases in baseline death rates with age, the relative excesses did not show any systematic association with age.
Upper panels show baseline age-specific death rates for separate countries (thin lines) as well as East and West means (thick lines). Lower panels show the ratios of the observed to the baseline age-specific death rates. While, the East-West divide in the baseline reflects long-term levels of population health, the relative excess reflects elevation of mortality caused by the pandemic.
Data shown in this Figure is provided at https://github.com/VMSdemo/East-West-contrast-in-life-expectancy-losses-in-2020-21
Further insights into the respective roles of East-West differences in relative mortality increases during the pandemic compared to baseline mortality differences are provided in Supplementary Table S2. This shows that the difference in the life expectancy losses in 2021 between the East and West was 2.03 years for males and 1.97 years for females. The contributions of larger relative mortality excess in the East to the gap in the losses were 88% for males and 94% for females. From this we conclude that it is the larger relative increases in mortality that is the primary driver of the life expectancy losses, rather than higher levels of baseline mortality in the East compared to the West.
Associations of explanatory factors with East-West differences in life expectancy losses in 2021
In this paper, we have already established that the contrast in life expectancy losses in Eastern Europe vs. Western Europe was particularly clear in 2021. In this section, we therefore focus on how country-level characteristics may potentially contribute to the East-West differences in life expectancy losses in 2021 alone. The analysis was guided by a simple conceptual framework (Figure 6) relating various factors to 2021 losses. Proxy data for a number of these factors (marked in green) were used in a meta-regression framework.
The green color of the boxes indicates the availability of numerical proxy variables.
Table 2 shows the effect of adjustment for a series of the country-level characteristics on the East-West difference in life expectancy losses in 2021 while the Supplementary Figure S2 shows associations of life expectancy losses in 2021 with each of them. Without adjustment for any factors, the losses of life expectancy were almost 2 years larger in Eastern compared to Western Europe. Vaccination levels in 2021 led to the largest attenuation of this effect of any single variable. Regulatory enforcement and trust in government had the second and third largest impacts on the East-West differences respectively. It was striking that adjustment for the stringency index of non-pharmaceutical interventions did not attenuate the association at all. Vaccination in combination with trust in government showed the largest reduction of any model, resulting in a halving of the East-West differences in life expectancy losses. Adjustment for vaccination together with regulatory enforcement had a similar effect for females, but a slightly smaller attenuating effect for males. The two-factor models had a marginally better fit than the single-factor (vaccination) models.
Discussion
In this paper, we have described in detail two distinctive features of the differential impacts of the COVID-19 pandemic on excess mortality and life expectancy losses in Eastern compared to Western European countries. The first of these is the absence in nearly all Eastern European countries of a pronounced and substantial mortality peak in the first few months of the pandemic unlike in most Western European countries. The second is that from October 2020 to the end of 2021, most Eastern European countries tended to experience much more persistent and high levels of excess mortality compared to Western European countries. Despite the fact that the major impacts of COVID-19 on mortality arrived later in Eastern than in Western Europe, the overall impact of the pandemic 2020-21 on death rates and life expectancy losses was much greater in Eastern compared to Western Europe.
In 2021 the relative increases in the excess death rates above the expected level were appreciably higher in Eastern Europe and were the key driver of East-West differences in life expectancy losses rather than the poorer health of Eastern European populations as measured by lower baseline life expectancy. Our regression analyses confirm the important role of differences in vaccination coverage in explaining the greater 2021 life expectancy losses in the East compared to the West, although factors such as trust in government and enforcement of regulations also appear to be important, which may be due to their moderating impact on high-risk behavior at an individual level.
In developing a comprehensive explanation of East-West differences in excess deaths and life expectancy losses from COVID-19 it is useful to distinguish between proximal and distal influences. Proximal ones are those that influence the spread of the pandemic in each country such as promulgation of and population compliance with non-pharmaceutical innervations (NPIs) and vaccine uptake including hesitancy and availability, and prevalence of infection at the time of initial lockdown. Distal influences are features that may differ between Eastern and Western European countries that originate in the different post-war histories of the two blocs which may have influenced how governments and individuals responded to the pandemic. These include differences in public trust in government and science leading to a varying degree of compliance with government advice or regulation aimed at reducing levels of transmission as well as affecting differences in levels of vaccine hesitancy. We discuss each of these in turn below.
Most Western European countries experienced very sharp increases in infections and consequent deaths that took off in March and peaked around the end of April 2020. In contrast, no such peaks were seen in any Eastern European country. The only plausible explanation for this is differences in the prevalence of SARS-CoV-2 infection at the time of the introduction of lockdowns and related measures to reduce community transmission.(7, 8) These differences in prevalence in early March 2020 in turn are likely to be a reflection of the relatively lower degree of connectivity to the world of Eastern compared to many Western European countries.(9) In 2019 no Eastern European destination was included in the top 20 ranking of airports in Europe with the highest levels of direct connectivity, measured by numbers of destinations and frequency of flights to the same destination.(10)
Direct estimates of the prevalence of COVID-19 infection have been challenging throughout the pandemic. Several attempts have been made to estimate the prevalence of infection per capita in the early stages of the pandemic. Russell et al.(11) used mortality from COVID-19 to work backwards to estimate the prevalence of infection several weeks earlier making a series of assumptions about case-fatality and the period from infection to death. In mid-March 2020, at the time nearly all countries introduced their first range of public health measures/lockdowns, the estimated prevalence of COVID-19 infection was systematically lower in Eastern compared to Western countries, supporting the hypothesis that the later peaks of excess deaths that occurred in Eastern compared to Western Europe was due to lower prevalence at the start of the pandemic in Eastern Europe.
The underlying difference in transport connectivity between East and West can be seen as part of the longer-term legacy of the historical divisions across the European continent in the post-war period. This reflects differences in patterns and intensity of international business and commerce as well as favored tourist destinations. The development of the major airline hubs in the UK, France, Germany, Spain, and Italy has occurred over many decades. In Eastern Europe, however, entry and exit of people into and out of the countries of Eastern Europe was severely constrained up until the fall of the Berlin Wall and the collapse of the Soviet Union.
The summer of 2020 was accompanied by a relaxation in Europe of many of the most stringent restrictions on movement and contact between people, both within and between countries that drove the resurgence of infections in the early autumn.(12) In particular it would have resulted in the seeding of new infections in the Eastern European countries that they had avoided in the first wave of the pandemic which eventually drove the substantial peaks of excess mortality in these countries from October 2020 onwards.
In 2021, Eastern European countries lagged behind those of Western Europe in terms of pace and ultimate level of population coverage of COVID-19 vaccination (see Extended Data Figure 3). We have found that this is the most influential factor that reduces the size of the East-West difference in life-expectancy losses (Table 2). This will have been accounted for higher levels of vaccination hesitancy in countries of Eastern Europe particularly in Russia and Bulgaria.(13–16) Steinert et al.(17) found no adequate response to messages about the benefits of vaccination in several countries including Bulgaria and Poland, which was not the case in Germany and the UK.
There has been much scientific and political interest in quantifying the extent to which non-pharmacological interventions (NPIs) played a role in reducing the impact of the pandemic (18–20). Our analysis found that differences in NPIs as measured by mean levels of the stringency index in each country over 2021 did not explain any of the East-West differences in life expectancy losses in 2021. However, there are considerable challenges in making such assessments. Firstly, the relationship between the level of NPIs and SARS-CoV-2 infection was iterative and bidirectional: while the stringency of NPIs would depend on the level of SARS-CoV-2 in the community, the latter would also inform the subsequent level of NPI stringency. Secondly, the impact of NPIs on risk of transmission is related to compliance at personal and institutional levels. The stringency index we used summarises the government’s intent, rather than actual uptake and compliance. Thirdly, in retrospect it is clear the construction of the stringency index we used was not optimal as it was a simple sum of the NPI measures. Not all the NPIs have the same impact on the transmission of SARS-CoV-2. For example, a public information campaign may not have the same effect as a ban on public gathering (which also varied by duration and size of the public gathering). This might have resulted in exposure misclassification and a resultant failure to detect any contribution of the stringency index to statistically explaining East-West differences in life expectancy losses in 2021.
What is clear from the dynamics of the spread of COVID-19, as measured by excess mortality, is that the massive fall in cross-border movement that occurred quickly in Europe and elsewhere from March 2020 had a major impact on the spread of the virus.(18) This is the necessary corollary of our conclusion that it was the differences in the prevalence of SARS-CoV-2 infection in European countries that are key to understanding the later peaks of mortality in Eastern compared to Western countries in 2020.
In the 5-6 months from March 2020 there were very strictly enforced lockdowns in many Eastern countries which would have helped stop those few cases that were in the East generating an exponential increase in cases. Crucially, in most Eastern countries the lockdowns occurred in the absence of an epidemic spread of COVID-19. The public in these countries would have experienced major constraints on their work and family lives without hearing about people they knew or reports in their media about mounting numbers of deaths and cases as occurred in Western countries such as UK, Spain, and France. It is tempting therefore to speculate that this may have contributed to a level of fatigue and resistance to continuing to follow restrictions in these populations later on in the pandemic when cases and COVID-related cases did start rising steeply.
The extent to which the experience of communism in Eastern Europe in the 20th century has had a persistent negative effect on trust at many different levels of society has been extensively studied.(21) A large multi-regional study conducted in 2020 across 16 countries found that the four countries from Central and Eastern Europe showed the lowest level of trust in government.(22) In our analysis trust in government, and to a smaller degree, trust in science appears to explain a component of East-West differences in life expectancy losses in 2021. This could be through a moderating influence on the willingness of individuals to comply with recommended or required behaviour changes aimed at reducing person-to-person transmission. It could also operate through an association of trust with vaccine uptake.
Differences in the degree to which countries in Eastern and Western Europe tend to enforce laws and regulations in general may have an impact on differences in life expectancy losses.(6) This may operate both through things such as policing of rule-breaking social gatherings as well as institutions such as businesses being confident that breaches of regulations would not incur penalties. The level of trust in institutions as a factor that could influence the impact of the pandemic on populations has been explored in a number of studies.(22–24) A study looking at factors influencing adherence to NPIs, and its change over time as populations became fatigued, found that reductions in adherence to physical distancing occurred to a smaller degree in countries with high interpersonal trust.(25)
Pre-pandemic research on vaccine hesitancy has identified trust as an important factor.(26) This is consistent with the findings of studies of COVID-19 vaccine uptake(27), although another study specifically of Eastern Europe using a questionable measure of trust was inconclusive.(28) However, our results suggest that the combination of vaccine coverage together with trust in government reduces by half the observed East-West differences in life expectancy losses. We interpret this as indicating that trust in government may be important above and beyond its relationship to vaccine coverage, possibly through its influence on, among others, compliance with NPI regulations.
The pandemic resulted in a renewed divergence in life expectancy between Eastern and Western Europe that replaced the striking convergence that had been evident in the two decades up to 2019. What is particularly notable was the fact that none of the Eastern countries saw a rebound in life expectancy in 2021 which was observed in many Western countries. Whether this is to be explained by the pandemic taking hold 6 months or more later in the East compared to many Western countries remains to be seen and will require analyses extending into 2022 and 2023. However, it is notable that those Western countries that did have a late first peak in 2020 (Austria, Germany, and Switzerland) still managed to experience much smaller life expectancy loss in 2021 compared to the East.
Our analysis has a number of weaknesses. Firstly, the variables we used as proxies for the factors that might underlie East-West differences in losses in 2021 may only parially capture the underlying concept. For example, the extent to which different countries enforced the social distancing rules and other components of their NPI policies may not be adequately captured by the assessment of generic intent of a country to implement a broad range of laws and regulations, many unrelated to health. This would result in an underestimation of any real impact of this factor.
Secondly, we do not have proxy measures for some of the key pathways shown in the conceptual diagram, including availability and effectiveness of treatments for people with COVID-19. Moreover, we did not have measures of individual and institutional behaviour change that occurred in response to government advice/stipulation or simply individual responses to information available through the news and social media about the pandemic. However, we do have more upstream factors, such as enforcement and trust in government and science, which we have regarded as potentially important general influences on behaviours of individuals and institutions.
Finally, our conceptual model of life expectancy losses in 2020-2021 does not take account of the fact that there are very likely to have been differences between Eastern and Western European countries in the levels of immunity to SARS-CoV-2. Data on population-based sero-prevalence of antibodies in Europe at the start of 2021, or at any other point in the pandemic, is very limited.(29) However it is notable that in England, which had one of the earliest and most intense peaks in COVID excess deaths, prevalence of immunity in the population in mid-July 2020 were surprisingly low at around 6%.(30) Other time series data from the UK show that the proportion of the population with levels of immunity only started to increase above this level once mass vaccination began in early 2021.(31) However, it is notable that in the several Western European countries which had late first peaks in autumn 2020 (e.g. Austria, Switzerland, Germany, Denmark), like all of the Eastern countries, nevertheless ended up with far lower excess mortality and life expectancy losses in 2021. Overall we conclude that low rates of naturally acquired immunity at the start of 2021 would be unlikely to explain much of the East-West difference in life expectancy losses in 2021.
In conclusion, the main early impact of the pandemic in the spring of 2020 was felt in a number of Western European countries, with none of the former communist countries of Eastern Europe experiencing a major peak in excess mortality until the autumn of 2020. Despite this the overall level of losses up to the end of 2021 in life expectancy and excess death rates were far higher in the East compared to the West. We have argued that both of these distinctive features of the East-West difference can be ultimately related to the differences in the post-war history of the two blocs. Of particular importance are the lower degree of international transport connectivity of Eastern countries that is likely to explain a lower prevalence of infection in March 2020 when lockdowns were initiated around the world. Lower levels of trust in government and science and willingness to enforce laws and regulations in the East are likely to stand behind the far higher excess mortality rates seen in Eastern than Western countries in 2021. This appears to have been importantly mediated principally through lower levels of vaccine coverage as well as population adherence to NPIs.
Regardless of mechanism, the COVID-19 pandemic put into reverse the convergence in life expectancy between East and West that had been occurring for the past 15-20 years. When this positive trend will resume and overcome this set back remains to be seen.
Materials and Methods
We used annual mortality data for European countries with data available up to and including 2021 from the Human Mortality Database (HMD)(32) and weekly mortality data from the Short-Term Mortality Fluctuations (STMF) series.(33) We divided the resulting 28 populations into two groups: Eastern Europe (East) consisting of 11 former communist countries and Western Europe (West) consisting of the remaining 17 countries (Supplementary Table S1).
Excess crude death rates (ECDRs) for all ages and both sexes were calculated. For each country and every week of 2020-2021, ECDRs were estimated as differences between the observed and expected CDRs as detailed in the Supplementary Methods. We identified for each country the month of the initial major peak of ECDR and the total ECDR over 2020-21.
Simple arithmetic calculation of the longevity differences between two periods (eg 2019 and 2020-21) tends to underestimate any true losses.(34, 35) Instead life expectancies observed in 2020-21 should be compared not with life expectancies at some arbitrary time point in the past but rather with the expected values forecasted from past mortality trends that would be observed in 2020 and 2021 in the absence of the pandemic as illustrated in Supplementary Figure S3. Using this approach we calculated the life expectancy losses as the differences between the observed life expectancies in 2020 and 2021 and the corresponding values predicted by the Lee-Carter model using the period 2005-19 to estimate the baseline based on previous sensitivity analysis,(33, 36) as described in more detail in Supplementary Methods.
The same relative excess in age-specific death rates can generate a smaller or larger life expectancy loss depending on the absolute baseline level of death rates. As the baseline mortality is still substantially higher in the East than in the West, this difference could contribute to the East-West gap in the losses. To assess the contributions of the relative change (the pandemic per se) and the level (pre-existing population health), we used a novel approach to decompose the East-West difference in the life expectancy losses into these two components as described in more detail in Supplementary Methods.
The final part of our analysis focused on factors that could potentially explain the distinct East-West contrast in life expectancy losses that emerged in 2021. We regressed (using robust standard errors) male and female life expectancy losses on vaccination as of the 1st of December 2021, trust in government, trust in science, and the regulatory enforcement index. Finally, we examined the model of the East-West differences in male and female life expectancy losses (presented by the East-West dummy) and how this difference attenuates in response to adjustment for single explanatory variables and for two-variable combinations. To guide the specification and interpretation of these models we developed a conceptual diagram (Figure 6), based on first principles and previous work,(5, 6) which suggests potential pathways linking characteristics of Eastern and Western countries with life expectancy losses.
Life table calculations were carried out in statistical software R (version 4.2.2); the rest of the statistical analysis was conducted in Stata 17. We provide the scripts and corresponding input and output datasets (https://github.com/VMSdemo/East-West-contrast-in-life-expectancy-losses-in-2020-21).
Data Availability
The raw data originated from the open access sources listed in the References. We also provide figures and tables data in separate Excel files (see the link below). Some of the Excel files contain calculations (including decomposition of life expectancy losses) and final data manipulations.
https://github.com/VMSdemo/East-West-contrast-in-life-expectancy-losses-in-2020-21
Declarations
Author Contributions
V.M.S. and S.T. conceived the study. V.M.S., S.T., and D.J. collected, preprocessed and validated the data. V.M.S. carried out the statistical analyses with support from V.M.S. and D.J. V.M.S., S.T., and D.A.L. drafted the manuscript. V.M.S., S.T., D.J., N.I., and D.A.L. contributed to subsequent versions, and to the interpretation of the data and results. All authors reviewed and approved the final version of the manuscript.
Competing interests
The authors declare that they have no competing interests.
Funding
S.T. acknowledges support from the Australian Research Council (DP210100401). N.I. acknolwledges support from the UK National Institute for Health and Care Research (HDRUK2022.0313). The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.
Data sharing
The raw data originated from the open access sources listed in the References. We also provide figures and tables data in separate Excel files (https://github.com/VMSdemo/East-West-contrast-in-life-expectancy-losses-in-2020-21). Some of the Excel files contain calculations (including decomposition of life expectancy losses) and final data manipulations.
Acknowledgements
We would like to thank Professor John Edmunds for his valuable insights and suggestions on an earlier draft of the paper.
References
