Large-scale macroeconomic changes and their impact on inequalities in mortality: a register-based study of mortality in the countries of the Baltic Sea region 2000-2011.
Enormous macroeconomic changes occurred in the Baltic countries in the 2000s. In this project we will study how unprecedented economic growth from the mid-2000s and the equally large economic downturn after the 2008 global financial crisis impacted on the mortality rates of people with different socioeconomic backgrounds in Estonia, Latvia and Lithuania. To better contextualize the findings from these countries they will be compared with those from Sweden and Finland to explore the potential buffering effect of more ambitious welfare policies. Information about demographic and socioeconomic status will be obtained from population censuses and will be linked to national mortality registries to obtain cause of death data. Three distinct periods will be examined: 2000-2003, 2004-2007 and 2008-2011. The magnitude of the changes in overall and cause-specific mortality across these years will be compared between and within the countries. We focus on alcohol-related causes of death, on suicide and causes amenable to medical care. The observed variations will be interpreted in relation to health and social policies in the respective countries. By using this unique natural experiment to explore the effects of economic cycles on human health in diverse social groups that have been comparatively little researched to date, the project will add to the theoretical understanding of how socioeconomic change impacts on health as well as elucidate how its effects can differ between contexts.
Final report
1. PURPOSE OF THE PROJECT AND HOW IT WAS DEVELOPED DURING THE PROJECT PERIOD
The general purpose of this project was to further enhance understanding of the relationship between macro-social processes and population health, which still remains very little researched in Eastern Europe. More specifically, we wanted to study how large-scale macroeconomic changes in the 2000s affected the mortality rates of people with different socioeconomic backgrounds in Estonia, Latvia and Lithuania in comparison to the neighboring countries of Sweden and Finland with more extensive welfare policies. Although we did not change the general purpose or the main research questions of the project, a few adjustments were made to the research plan. First, we extended the coverage of the study period, by adding a fourth sub-period, 2012–2015, that was based on the 2011 census follow up. After an initial data inquiry the list of social variables and causes of death was extended to cover marital status, place of residence and 52 causes of death in total. At the same time, we were not able to acquire income data due to methodological considerations and we had to exclude Sweden as data were not available during the project time. As mental health has been found to be particularly sensitive to economic recession, we conducted an additional study using survey data from Estonia to explore the effect of economic fluctuations on long-term trends in depression, a main single determinant of suicide, to support our findings on inequalities in suicide mortality.
2. A SHORT DESCRIPTION OF HOW THE PROJECT WAS IMPLEMENTED
To study the impact of large-scale macroeconomic processes on mortality in different socioeconomic groups, a series of longitudinal register-based mortality studies were conducted. As census-based mortality follow-up studies had not been conducted before in either Estonia or Latvia, it presented a challenging and complex task not only methodologically but also because of the legal aspects of personal data processing, especially in light of tightened EU regulations. After agreement with National Statistical Offices for the data linkage of two population censuses (2000 and 2011) and mortality registers, a thorough data inquiry was conducted regarding the coverage of census/registry records to ensure data comparability across countries and between the two censuses. The collaboration between the project partners was arranged via electronic communication, project meetings and site visits for practical guidance. All data linkage was performed by Statistical Offices. Data were anonymized and aggregated into multidimensional frequency tables combining deaths and population exposures split by study periods and sociodemographic variables before they were delivered for research purposes. Additional data collection for methodological purposes was later performed in Estonia and Latvia.
3. THE PROJECT’S THREE MOST IMPORTANT RESULTS
Overall mortality changes. A procyclical mortality pattern suggests that overall mortality rises when the economy is expanding and falls during recessions. Our results on all-cause mortality were in line with previous research showing that overall mortality declined slightly, or it increased during the period of strong economic growth, whereas the mortality decline accelerated during the economic recession with a larger effect seen in the Baltic countries. Changes in alcohol-related mortality, i.e. an increase during economic expansion and decrease during recession indicate that macroeconomic processes are likely to underpin the observed mortality pattern via their link to living standards and alcohol affordability. Economic contraction also accelerated the road-transport accident (RTA) mortality decline, possibly via reduced traffic density caused by increased fuel prices and loss of, or reduction in income with a larger effect seen in the Baltic countries. During the period of economic growth, mortality decline from causes amenable to health care was larger in Finland; during the recession the decline accelerated more in the Baltic countries. The results on suicide mortality among working-age men were in sharp contrast to all-cause mortality: overall suicide mortality declined substantially in all countries during the economic expansion, but the decline slowed during the recession. Using survey data from Estonia we found a similar pattern: the prevalence of depression decreased during the period of economic growth but increased during the recession. However, the impact of recession on the prevalence of depression was short-term.
Inequalities in cause-specific mortality. Educational inequalities in alcohol-related mortality increased across the study period with a much larger increase in the Baltic countries. During economic expansion mortality increased faster among the low educated while in the recession mortality decreased more among the high educated. For RTA mortality educational inequalities increased substantially among men (more in the Baltic countries) but not among women (excepting Lithuania). Lower educated men became increasingly disadvantaged in terms of RTA mortality, possibly because of reduced access to safer cars and because of a more harmful drinking and driving culture; higher educated women did not have a larger RTA mortality decline during economic growth as they possibly became more exposed to driving compared to low educated women. Educational inequalities in amenable mortality were larger in the Baltic countries where they also increased more across the study period: mortality decline was larger among the high educated during periods of both economic growth and recession in all countries. Educational inequalities in suicide mortality among working-age men were particularly large in Latvia and Lithuania. In the Baltic countries, lower educated men had a smaller decline in suicide mortality than higher educated men during a period of rapid economic expansion, however, they were not more disadvantaged during the recession, possibly because of being less exposed to financial loss. Inequalities in suicide mortality may increase during economic booms and decrease during recessions. Similarly, the prevalence of depression did not increase more among disadvantaged groups during the recession.
Methodological contributions. The new longitudinal database enabled us to extend prior scarce research evidence about the numerator-denominator bias in unlinked data on mortality differentials by comparing linked and unlinked mortality estimates using data from Estonia. The results showed that not only bias can change across time but also, that the misreporting of educational level on death records seems to be country-specific, indicating that numerator-denominator bias can take different forms in different contexts.
The censuses in the Baltic countries combined traditional survey-based enumeration and register-based enumeration, the latter covering only a limited set of socio-demographic indicators. The share of the population enumerated using only survey-based enumeration varied from 91% in Latvia to 98% in Estonia. A sensitivity analysis performed for Latvia showed that by excluding register-based records we slightly underestimated mortality but the effect on changes between the periods was minimal.
4. NEW RESEARCH QUESTIONS GENERATED THROUGH THE PROJECT
Already during the time of the project, the new longitudinal database covering all three Baltic countries and Finland over a 15-year period was recognized as a unique, high-quality data resource for comparative studies on socioeconomic inequalities in mortality in the Baltic Sea region. The intention is to now extend the research scope to causes of death that are less sensitive to short-term macroeconomic changes and thus, were not covered in the current project. One possible research area would cover long-term trends in inequalities in site-specific cancer mortality. Another possible research topic is to examine the long-term mortality trends of Russian minorities in the Baltic countries compared to the Russian population in neighboring Russian regions with a specific focus on alcohol-related mortality.
5. THE PROJECT’S INTERNATIONAL DIMENSIONS, SUCH AS CONTACTS AND MATERIALS
The project was originally designed as a collaborative undertaking between research/academic institutions and public authorities in five European countries, thus being international in nature. As many of the project participants were affiliated with several research/academic institutions, often in different countries, it further extended the project’s international dimensions. The methodological skills developed during the project concerning longitudinal register-based mortality studies in the Baltic countries will further enhance the opportunities for participation in European collaborations. Due to data protection regulations we are not allowed to make data available to third parties and thus potential research collaboration will need to be enquired about.
6. HOW THE PROJECT TEAM HAS DISSEMINATED THE RESULTS TO OTHER RESEARCHERS
The project results will be disseminated mostly through research articles published in international peer-reviewed journals. We aim to publish all articles through open access. The project results have been introduced to other researchers at the international seminar “Persisting burden of alcohol in Central and Eastern Europe: recent evidence and measurement issues” in Vilnius, Lithuania, 2017, and at the 21st Nordic Demographic Symposium in Reykjavik, Iceland, 2019. The project was also introduced at the Research Day of the School of Social Sciences, Södertörn University. The research on health inequalities has been integrated to the teaching at the Sociology Master’s Degree Course “Health and Social Change” at Södertörn University.
The general purpose of this project was to further enhance understanding of the relationship between macro-social processes and population health, which still remains very little researched in Eastern Europe. More specifically, we wanted to study how large-scale macroeconomic changes in the 2000s affected the mortality rates of people with different socioeconomic backgrounds in Estonia, Latvia and Lithuania in comparison to the neighboring countries of Sweden and Finland with more extensive welfare policies. Although we did not change the general purpose or the main research questions of the project, a few adjustments were made to the research plan. First, we extended the coverage of the study period, by adding a fourth sub-period, 2012–2015, that was based on the 2011 census follow up. After an initial data inquiry the list of social variables and causes of death was extended to cover marital status, place of residence and 52 causes of death in total. At the same time, we were not able to acquire income data due to methodological considerations and we had to exclude Sweden as data were not available during the project time. As mental health has been found to be particularly sensitive to economic recession, we conducted an additional study using survey data from Estonia to explore the effect of economic fluctuations on long-term trends in depression, a main single determinant of suicide, to support our findings on inequalities in suicide mortality.
2. A SHORT DESCRIPTION OF HOW THE PROJECT WAS IMPLEMENTED
To study the impact of large-scale macroeconomic processes on mortality in different socioeconomic groups, a series of longitudinal register-based mortality studies were conducted. As census-based mortality follow-up studies had not been conducted before in either Estonia or Latvia, it presented a challenging and complex task not only methodologically but also because of the legal aspects of personal data processing, especially in light of tightened EU regulations. After agreement with National Statistical Offices for the data linkage of two population censuses (2000 and 2011) and mortality registers, a thorough data inquiry was conducted regarding the coverage of census/registry records to ensure data comparability across countries and between the two censuses. The collaboration between the project partners was arranged via electronic communication, project meetings and site visits for practical guidance. All data linkage was performed by Statistical Offices. Data were anonymized and aggregated into multidimensional frequency tables combining deaths and population exposures split by study periods and sociodemographic variables before they were delivered for research purposes. Additional data collection for methodological purposes was later performed in Estonia and Latvia.
3. THE PROJECT’S THREE MOST IMPORTANT RESULTS
Overall mortality changes. A procyclical mortality pattern suggests that overall mortality rises when the economy is expanding and falls during recessions. Our results on all-cause mortality were in line with previous research showing that overall mortality declined slightly, or it increased during the period of strong economic growth, whereas the mortality decline accelerated during the economic recession with a larger effect seen in the Baltic countries. Changes in alcohol-related mortality, i.e. an increase during economic expansion and decrease during recession indicate that macroeconomic processes are likely to underpin the observed mortality pattern via their link to living standards and alcohol affordability. Economic contraction also accelerated the road-transport accident (RTA) mortality decline, possibly via reduced traffic density caused by increased fuel prices and loss of, or reduction in income with a larger effect seen in the Baltic countries. During the period of economic growth, mortality decline from causes amenable to health care was larger in Finland; during the recession the decline accelerated more in the Baltic countries. The results on suicide mortality among working-age men were in sharp contrast to all-cause mortality: overall suicide mortality declined substantially in all countries during the economic expansion, but the decline slowed during the recession. Using survey data from Estonia we found a similar pattern: the prevalence of depression decreased during the period of economic growth but increased during the recession. However, the impact of recession on the prevalence of depression was short-term.
Inequalities in cause-specific mortality. Educational inequalities in alcohol-related mortality increased across the study period with a much larger increase in the Baltic countries. During economic expansion mortality increased faster among the low educated while in the recession mortality decreased more among the high educated. For RTA mortality educational inequalities increased substantially among men (more in the Baltic countries) but not among women (excepting Lithuania). Lower educated men became increasingly disadvantaged in terms of RTA mortality, possibly because of reduced access to safer cars and because of a more harmful drinking and driving culture; higher educated women did not have a larger RTA mortality decline during economic growth as they possibly became more exposed to driving compared to low educated women. Educational inequalities in amenable mortality were larger in the Baltic countries where they also increased more across the study period: mortality decline was larger among the high educated during periods of both economic growth and recession in all countries. Educational inequalities in suicide mortality among working-age men were particularly large in Latvia and Lithuania. In the Baltic countries, lower educated men had a smaller decline in suicide mortality than higher educated men during a period of rapid economic expansion, however, they were not more disadvantaged during the recession, possibly because of being less exposed to financial loss. Inequalities in suicide mortality may increase during economic booms and decrease during recessions. Similarly, the prevalence of depression did not increase more among disadvantaged groups during the recession.
Methodological contributions. The new longitudinal database enabled us to extend prior scarce research evidence about the numerator-denominator bias in unlinked data on mortality differentials by comparing linked and unlinked mortality estimates using data from Estonia. The results showed that not only bias can change across time but also, that the misreporting of educational level on death records seems to be country-specific, indicating that numerator-denominator bias can take different forms in different contexts.
The censuses in the Baltic countries combined traditional survey-based enumeration and register-based enumeration, the latter covering only a limited set of socio-demographic indicators. The share of the population enumerated using only survey-based enumeration varied from 91% in Latvia to 98% in Estonia. A sensitivity analysis performed for Latvia showed that by excluding register-based records we slightly underestimated mortality but the effect on changes between the periods was minimal.
4. NEW RESEARCH QUESTIONS GENERATED THROUGH THE PROJECT
Already during the time of the project, the new longitudinal database covering all three Baltic countries and Finland over a 15-year period was recognized as a unique, high-quality data resource for comparative studies on socioeconomic inequalities in mortality in the Baltic Sea region. The intention is to now extend the research scope to causes of death that are less sensitive to short-term macroeconomic changes and thus, were not covered in the current project. One possible research area would cover long-term trends in inequalities in site-specific cancer mortality. Another possible research topic is to examine the long-term mortality trends of Russian minorities in the Baltic countries compared to the Russian population in neighboring Russian regions with a specific focus on alcohol-related mortality.
5. THE PROJECT’S INTERNATIONAL DIMENSIONS, SUCH AS CONTACTS AND MATERIALS
The project was originally designed as a collaborative undertaking between research/academic institutions and public authorities in five European countries, thus being international in nature. As many of the project participants were affiliated with several research/academic institutions, often in different countries, it further extended the project’s international dimensions. The methodological skills developed during the project concerning longitudinal register-based mortality studies in the Baltic countries will further enhance the opportunities for participation in European collaborations. Due to data protection regulations we are not allowed to make data available to third parties and thus potential research collaboration will need to be enquired about.
6. HOW THE PROJECT TEAM HAS DISSEMINATED THE RESULTS TO OTHER RESEARCHERS
The project results will be disseminated mostly through research articles published in international peer-reviewed journals. We aim to publish all articles through open access. The project results have been introduced to other researchers at the international seminar “Persisting burden of alcohol in Central and Eastern Europe: recent evidence and measurement issues” in Vilnius, Lithuania, 2017, and at the 21st Nordic Demographic Symposium in Reykjavik, Iceland, 2019. The project was also introduced at the Research Day of the School of Social Sciences, Södertörn University. The research on health inequalities has been integrated to the teaching at the Sociology Master’s Degree Course “Health and Social Change” at Södertörn University.