Max Planck Institute for Demographic Research

Stillbirth rates have stalled or increased in some European countries during the last decade. We investigate to what extent time-trends and between-country differences in stillbirth rates are explained by the changing prevalence of advanced maternal age and teenage pregnancies or multiple births. We analysed data on stillbirths and live births by maternal age and multiplicity from 2010 to 2021 in 25 European countries using Kitagawa decomposition to separate rate differences into compositional and rate components. Rates significantly decreased in six countries, but increased in two. Changes in maternal age structure reduced national stillbirth rates by a maximum of 0.04 per 1000 in the Netherlands and increased rates by up to 0.85 in Cyprus. Changes in the prevalence of multiple births decreased rates by up to 0.19 in the Netherlands and increased rates by up to 0.01 across multiple countries. Maternal age differences explained between 0.11 of the below-European average stillbirth rate in Belgium and 0.13 of the above-average rate in Ireland. Excluding Cyprus, differences in multiple births explained between 0.05 of the below-average rate in Malta and 0.03 of the above-average rate in Ireland. For most countries, the increase in advanced-age pregnancies contributed to rising stillbirth rates over time, while reductions in multiples led to decreases in rates. However, large parts of the trends remain unexplained by those factors. By 2021, neither factor explained the differences between countries, due to increased compositional uniformity and declining stillbirth risk for advanced maternal age.

The association between parental separation and children’s education has been widely studied, but mainly at a single time point and for marital dissolution only. We examine whether the (generally negative) association has changed across cohorts for several educational outcomes and whether the association differs by parental union type (marriage, cohabitation) and socioeconomic family background (parental education).We use Finnish total population register data. We focus on child cohorts born between 1987 and 2003 (N = 967,242) and analyse grade point averages, secondary education and tertiary education using linear regression and linear probability models with standard errors clustered within families.The association between parental separation and educational achievement is negative and has remained similar across the birth cohorts. Differences according to parental union type and socioeconomic family background are rather small. The stability of the association over time suggests that the consequences of parental separation on children’s education have not changed over time, and they do not depend much on parental union type or family background.

Purpose The Swedish Families of the 1990s (SWIFT90) is a population-based national register cohort that follows everyone born between 1990 and 1999, their parents and siblings. The cohort was set up primarily to investigate factors associated with biological parents’ involvement with child welfare services and their outcomes following child(ren) placement in out-of-home care (OHC) under the research project ‘Drivers of inequalities of families involved in child welfare services (DRIVERS)’. Participants This cohort is defined as families consisting of parents and their children, of which at least one was born between 1990 and 1999 in Sweden, which totals 1 075 037 children. The children are linked to both (adoptive or biological) parents and their siblings, so the total number of individuals in the SWIFT90 yields a total sample of n=3 292 417. These families are followed through multiple national registers including information on income, education, inpatient care, mortality and criminal offences. SWIFT90 compiles administrative data spanning from 1960 to 2022, which reflects the most recent data available at the time of the data request. Findings to date SWIFT90 provides information on 61 982 parents (fathers=31 028; mothers=30 954) with children placed in OHC and their respective children (38,084). Several covariates could be assessed for the parental generation, therefore, providing a comprehensive picture of socioeconomic and health-related aspects of families with children born in the 1990s and placed in OHC in Sweden. Future plans SWIFT90 will be used to investigate the socioeconomic and health trajectories of families involved with the child welfare system. With this cohort study ia possible to examine whether the inclusion of certain covariates alters the association between parental characteristics and child welfare outcomes. Future studies could also guide initiatives to prevent vulnerable circumstances among biological parents from escalating to the point where child placement into OHC becomes necessary. Additionally, they could help enhance the conditions of biological families and support opportunities for reunification after placement has occurred.

Flexible working arrangements have become increasingly common and are considered a means to better reconcile paid and unpaid work. Therefore, the use of such measures can determine how couples divide their household and childcare tasks. While currently these tasks are dominantly female connotated, an increase in flexible work arrangements may contribute to a more gender-egalitarian distribution of unpaid work. However, empirical evidence on this association is mixed, and it remains unclear to what extent it differs by gender. Using a sample of 3244 individuals in the German Family Panel of 2018/2019 who were cohabiting with an opposite-sex partner and by applying linear regression models, we tested several hypotheses derived from economic, gender, and time-availability approaches. We separately addressed the division of housework and childcare tasks related to three flexible work measures, namely home-office, schedule flexibility, and working-time autonomy. Contrary to our hypotheses, no flexibility measure seemed to be related to a higher share of household tasks. Rather, any significant association we identified was fully explained through gender: Women took on a larger share of any household task, irrespective of their work flexibility. Only the share of childcare performed seemed to differ by the use of schedule flexibility, as well as by gender. Whereas mothers’ contributions to childcare were larger when they used flexibility, those of fathers were smaller. We conclude that flexible working arrangements do not contribute to a more gender-egalitarian division of unpaid work per se, but the (gendered) motivation to use such flexibility may be decisive.

Background The association of workload and performance with physical functioning is recognised among the ageing public sector workforce. The characteristics of working conditions and social- and health-related factors associated with physical functioning after statutory retirement are still unknown. Also, previous studies on changes in physical functioning have not used a person-oriented approach. We examined physical functioning trajectories over statutory retirement and how social- and health-related factors are associated with them. Our aim was to identify distinct developmental trajectories of physical functioning over statutory retirement and to examine how social- (age, gender, marital status, education) and health-related (physical workload, self-reported sleep problems, alcohol consumption, smoking, fruit and vegetable (F&V) consumption, leisure-time physical activity (LTPA), and body mass index (BMI)) factors before retirement were associated with the identified trajectories. Methods We used data from the Helsinki Health Study cohort. Participants consisted of 2736 employees of the City of Helsinki, Finland who retired during the follow-up. Growth mixture modelling was used to identify physical functioning trajectories and multinominal regression analyses to examine associations of social- and health-related factors with them. Results Three distinct developmental patterns in physical functioning before and after retirement were found among ageing and retired employees. Lower educational level, sleep problems, physical inactivity, and obesity were associated with the trajectory groups of ‘fast decreasing’ and ‘slowly increasing’, compared to the ‘stable high’ trajectory. Conclusion The results suggest that poor social- and health-related factors are key risk factors associated with declining and lower-level physical functioning over the retirement period. Supporting healthy lifestyles among older employees might maintaining good physical functioning until retirement and beyond.

A migração dos estudantes é um importante fenômeno que pode ter implicações para o resultado educacional. Poucos estudos avaliaram a influência da migração na escolaridade dos indivíduos, sobretudo no contexto dos países em desenvolvimento. Este trabalho tem como objetivo analisar a relação das migrações com o fluxo escolar da coorte de estudantes de Minas Gerais que estavam matriculados no 1º ano do ensino fundamental, em 2008. Utilizaram-se dados longitudinais preparados a partir dos Censos Escolares, de 2008 a 2019, para comparar a regularidade do fluxo escolar dos estudantes migrantes, por quantidade de migrações, com a dos não migrantes. Foram encontradas baixas diferenças na regularidade de fluxo escolar entre migrantes e não migrantes antes da migração. Por outro lado, o fluxo escolar de migrantes se torna mais irregular após o movimento, exceto para migrações realizadas em transições de níveis escolares. No caso de migrações na transição para o ensino médio, foram encontradas alta seletividade positiva antes da migração e baixa variação após. Isto indica que a migração pode ter como objetivo o investimento em educação. Os resultados evidenciam a necessidade de se considerar a migração em estudos sobre resultados educacionais e os ganhos de uma análise longitudinal para os estudos demográficos.

Background Low family socioeconomic position is a well-established determinant of poor health in youth. Much less is known about the social patterning of youth medication use, and the current evidence is mixed. Furthermore, previous studies have not assessed important confounders of the associations. We analyse differences in youth medication use by childhood family income and medication type. Methods Administrative register data on full Finnish cohorts born in 1979–2003 (n=1 490 666) and survival analysis were used to assess the risk of using common prescription medications between ages 16 and 20 according to mean household income in ages 11–15, accounting for several observed familial characteristics including parental health. We also compared siblings with discordant childhood income exposures to assess whether any differences are explained by unobserved familial confounding. Results For each 10% increase in childhood family income, there was a 0.6%–1.7% increase in the probability of using the most common prescription medications: antibiotics, painkillers, and allergy and asthma medications. In contrast, a 10% increase in childhood income was related to a 2.5% decrease in the probability of psychotropic medication use. In sibling comparisons, childhood income was not associated with any type of medication use. Conclusion Apart from psychotropics, the results may indicate medication underuse among youth from low-income families. The sibling comparisons suggest that moderate differences in childhood income are unlikely to cause differences in youth medication use and thus, in contexts of relative income equality, income differences in medication use are likely to reflect other, unobserved, family factors shared by siblings.

Objectives To test the effect of local, industry-specific labour demand on employment and recidivism following a person’s release from a prison in Finland. Methods This study was an observational study based on administrative data. The data included all working-aged men (N = 33,270) released from prison between 2007 and 2018 in Finland. We analysed the industry of the employers descriptively and studied the effect of labour demand on employment and recidivism rates post-release with a linear probability model. Labour demand was measured by counting new employments that began in the first 30 days post-release in the region the subject resided in at the time of release. Employment and recidivism were analysed with a 6- and 12-month follow-up. Results Of the study population, 11.1% were employed and 44.7% recidivated in the first 6 months. We found the most common employers to be the construction and manufacturing industries along with temporary work agencies. Our models indicated that the low-skill labour demand in construction increased the likelihood of employment (β = 0.0108, se = 0.0026) but did not reduce the risk of recidivism (β = 0.0030, se = 0.0045). Conclusions Higher labour demand may not lead to lower recidivism in all contexts. As the effect of labour demand may be driven by specific societal or labour market characteristics, further research should focus on identifying the characteristics that affect recidivism and post-prison employment. As this study may only be generalisable to the Finnish prison population, similar studies are needed from varied contexts.

This study examines the potential influence of selection on the association between re-partnering and single mothers’ mental health and life satisfaction in Germany and the United Kingdom. Drawing on extensive longitudinal panel data, we analyze the trajectories of 1694 separated single mothers in Germany (SOEP) and 1070 in the UK (BHPS/UKHLS). Employing fixed effects models, we examine the outcomes before and after entry into single motherhood and compare trajectories of stably single mothers and re-partnered single mothers. In both countries, the findings weakly indicate that prior to entering single motherhood, re-partnered mothers exhibit higher levels of life satisfaction, suggesting positive selection. Increasing differences in life satisfaction after the transition into single motherhood between mothers that re-partner and stably single mothers indicate a positive association of re-partnering and life satisfaction. No evidence of mental health selection into re-partnering was found in either country, but the trajectory of re-partnered mothers in Germany shows a stronger increase than that of mothers who remain single.

Background Seasonal fluctuations in mortality affect annual life expectancy at birth (e 0 ). Nevertheless, evidence on the impact of seasonal mortality on longevity is very limited and mainly restricted to assessing season-specific mortality levels due to shocks (e.g., heatwaves and influenza epidemics). We investigated the influence of seasonality in mortality on life expectancy levels and temporal trends across 20 European countries during 2000–2019. Data and methods We used harmonised weekly population-level mortality data from the Human Mortality Database. Seasonal contributions to life expectancy at birth and age 65, by sex, were estimated using the excess mortality approach and decomposition analysis. Time-series analysis was used to evaluate the impact on long-term mortality trends. Results Seasonal mortality had a substantial but stable impact on e 0 between 2000 and 2019. On average, we found an annual reduction in life expectancy due to seasonal excess mortality of 1.14 years for males and 0.80 years for females. Deaths in the elderly population (65+) were the main driver of this impact: around 70% and 90% of these reductions in life expectancy were attributable to older ages. Excess mortality in winter had the strongest impact on annual life expectancy, especially in Portugal and Bulgaria (around 0.8-year loss on e 0 ). Conclusions The study revealed significant cross-country variations in contributions of seasonal mortality. The most pronounced effects were observed in winter months and at older ages. These findings underscore the need for timely and targeted public health interventions to mitigate excess seasonal mortality.

This paper introduces a special issue of International Migration Review and overviews the promise and limitations of organic data for migration research. We explain different approaches to collecting and analyzing timely open-source organic data, and how it can be integrated with traditional administrative and survey data. After identifying the promise and limitations of specific organic data sources, we offer two empirical examples that illustrate how organic data may be used in research on migration and displacement. One example uses Ukrainian-language tweets to capture changing day-to-day dynamics during the initial stages of an emerging conflict. The second examines how a significant policy change in Colombia is associated with xenophobic conversation about Venezuelan migrants using Spanish-language tweets. Together, they illustrate how organic data may be used to examine the shifting dynamics of migration and forced migration.

STUDY QUESTION How are the changing maternal age structure and population growth expected to shape future twinning rates in low-income countries? SUMMARY ANSWER With maternal age at birth projected to shift toward older ages, twinning rates are also estimated to increase in most low-income countries by 2050 and even more by 2100. WHAT IS KNOWN ALREADY Many of the sub-Saharan African and South Asian countries are undergoing, and projected to further experience, the shift of maternal age at birth to older ages. Advanced maternal age is a well-established predictor of multiple births at the individual level, but currently, it is unknown how the changes in maternal age distribution are associated with the changes in twinning rates at the population level in low-income countries. STUDY DESIGN, SIZE, DURATION We first estimated age-specific twinning probability based on Demographic Health Surveys and World Fertility Surveys data. We then scaled up the age-specific twinning probability at the population level to estimate changes in the number of twin births in 2050 and 2100 attributable to the estimated shifts in maternal age toward older ages as projected by the UN World Population Prospects (WPP). PARTICIPANTS/MATERIALS, SETTING, METHODS We analyzed ∼3.19 million births that occurred within 10 years before the interview. Majority of the births in our data took place between 1980 and 2015 across 39 countries, where the uptake of medically assisted reproduction (MAR) is known to have been low during the observation period. We estimated country fixed-effects models to obtain country-specific twinning rates and age-specific twinning probability. We applied these estimates to the future number of births projected by the UN WPP, to estimate the number of twin births in 2050 and 2100. MAIN RESULTS AND THE ROLE OF CHANCE With maternal age at birth projected to shift toward older ages, twinning rates are also estimated to increase in most countries by 2050 compared to 2010 (increases from 0.3% to 63% depending on countries), and even more in all studied countries by 2100 (increases from 3.5% to 79%). Due to its large population size, India will continue to have among the largest share of twin births despite its estimated decline of twin births by 10.5% by 2100. Nigeria, due to its not only large and growing population size but also high twinning rate, is expected to contribute the second largest number of twin births. LIMITATIONS, REASONS FOR CAUTION Although the accuracy in maternal recall of multiple births tends to be high, our use of data based on recalled births from the past nonetheless imply a potential bias in our estimation of twinning rates. WIDER IMPLICATIONS OF THE FINDINGS The present study suggests that, even if the spread of MAR remains slow in many low-income countries, twinning rates and number of twin births are expected to grow as an increase in maternal age at birth and population growth continue. Our findings call for more public health attention and societal support to be paid to twins and their families in low-income countries, given that twins are at higher risk of developmental challenges and health disadvantages. STUDY FUNDING/COMPETING INTEREST(S) D.S.L. was supported by the European Union (ERC, BIOSFER, 101071773). K.J.B. was supported by a Pro Futura Scientia XIV Fellowship awarded by the Swedish Collegium for Advanced Study and Riksbankens Jubileumsfond. There are no competing interests to declare. TRIAL REGISTRATION NUMBER N/A.

This paper examines, using exogenous variation generated by a Finnish pension reform implemented in 2005, the interplay between health and financial incentives to postpone retirement. Based on detailed administrative data on individual health and retirement behavior, we focus on whether individual reactions to incentives vary according to health status and analyze whether individuals with ill health are also able to take advantage of the potential monetary benefits of delayed retirement created by the reform. We find that on average, individuals react to the financial incentives created by the reform as expected. This result holds for most of the health‐related subgroups we analyze. However, those with a long period of sickness absence are less likely to respond to changes in the financial incentives to postpone retirement.

Background Previous studies on socio-economic inequalities in mortality have documented a substantial contribution of alcohol-attributable mortality (AAM) to these inequalities. However, little is known about the extent to which AAM has contributed to time trends in socio-economic inequalities in mortality. Objective To study long-term trends in educational inequalities in AAM and assessed their impact on trends in educational inequalities in life expectancy in three European countries. Methods We analyzed cause-specific mortality data by educational group (low, middle, high) for individuals aged 30 and older in England and Wales, Finland, and Turin (Italy) over the 1972–2017 period. To estimate AAM, we used the multiple causes of death approach for England and Wales and Finland (1987–2017), and a recently introduced method for Turin (Italy). We used segmented regression analysis to study changes in absolute educational inequalities in AAM, measured by the Slope Index of Inequality (SII). We assessed the contribution of AAM to trends in educational differences in remaining life expectancy at age 30 (e30) using cause-deleted life tables. Results AAM increased more among the low-educated than the high-educated in England and Wales (1972–2017) and Finland (1987–2007). In contrast, in Finland (2007 onwards) and Turin (1972–2017), AAM decreased more among the low-educated than the high-educated. In England and Wales, AAM contributed 37% (males) and 24% (females) of the increase in educational inequalities in e30. In Finland in 1987–2007, AAM contributed 50% (males) and 34% (females) of the increase in educational inequalities in e30. AAM also contributed to recent trend breaks, such as to the onset of an increase in educational inequalities in e30 in England and Wales, and to the onset of a decline in educational inequalities in e30 in Finland after 2007. Discussion AAM mortality contributed substantially not only to levels of, but also to changes in educational inequalities in e30 in the studied populations. Reducing the impact of alcohol on mortality among low-educated groups may positively affect trends in educational inequalities in life expectancy.

Over the last 15 years, many European countries have experienced fertility declines. Existing research on this shift in fertility behavior points to economic aspects and increased levels of uncertainty as important drivers. However, in this debate little attention has been paid to how the relevance of individual- and contextual-level dimensions for understanding the new fertility patterns varies by level of urbanization. This is surprising given that urban and rural areas not only differ strongly in fertility timing and levels, but also in economic conditions. Our paper fills this important research gap by analyzing rich register data from Finland using multi-level event history models to study the transition to first birth among younger (under age 30) and older (ages 30 or older) women. We show that urban–rural differences in the transition to parenthood are particularly pronounced among younger women. In addition, the results indicate that economic circumstances and related uncertainties are more relevant for understanding first births probabilities for younger women than older women. Finally, among younger women, the relevance of economic circumstances and related uncertainties seems to be most relevant in the capital region of Helsinki and urban areas compared to semiurban and rural areas. Our findings underline that the urban–rural dimension should receive more attention in research on fertility in times of uncertainty.

Background Psychiatric disorder diagnoses are linked to long-term socioeconomic ‘shadows’ into adulthood, but little is known about how these diagnoses are associated with adulthood payment problems in Norway and whether these associations are moderated by parental and area characteristics. Methods We combine Norwegian registry data with payment problem records from the public debt enforcement agency for the 1995–1997 birth cohorts (n=173 750). We ascertain behavioural, neurotic, mood and substance use disorder diagnoses from specialised care at ages 13–17. Payment problems are defined as any records of mandatory deductions in wages or benefits up to age 23. Causal forests are used to explore the potential moderating role of sociodemographic variables. Results Some 6% of the sample have payment problems, but the share is higher among men and among people with low childhood socioeconomic status. After adjusting for background characteristics, behavioural problem diagnoses are linked to a 4.6 percentage point higher risk of payment problems (95% CI 4.1 to 5), mood disorder diagnoses to a 2.2 percentage point higher risk (95% CI 1.5 to 2.9), neurotic disorder diagnoses to a 2.5 percentage point higher risk (95% CI 2 to 3.1) and substance use disorder diagnoses to a 9.3 percentage point higher risk (95% CI 7.6 to 10.9). The associations between substance use and behavioural diagnoses and payment problems are stronger for people with low childhood socioeconomic status. Conclusion The associations between adolescent substance and behavioural diagnoses and adulthood payment problems are particularly strong among people with low childhood socioeconomic status. Reducing these associations is critical for preventing the accumulation of disadvantages.

Background In the winter of 2022/2023, excess death estimates for Germany indicated a 10% elevation, which has led to questions about the significance of this increase in mortality. Given the inherent errors in demographic forecasting, the reliability of estimating a 10% deviation is questionable. This research addresses this issue by analyzing the error distribution in forecasts of weekly deaths. By deriving empirical prediction intervals, we provide a more accurate probabilistic study of weekly expected and excess deaths compared to the use of conventional parametric intervals. Methods Using weekly death data from the Short-term Mortality Database (STMF) for 23 countries, we propose empirical prediction intervals based on the distribution of past out-of-sample forecasting errors for the study of weekly expected and excess deaths. Instead of relying on the suitability of parametric assumptions or the magnitude of errors over the fitting period, empirical prediction intervals reflect the intuitive notion that a forecast is only as precise as similar forecasts in the past turned out to be. We compare the probabilistic calibration of empirical skew-normal prediction intervals with conventional parametric prediction intervals from a negative-binomial GAM in an out-of-sample setting. Further, we use the empirical prediction intervals to quantify the probability of detecting 10% excess deaths in a given week, given pre-pandemic mortality trends. Results The cross-country analysis shows that the empirical skew-normal prediction intervals are overall better calibrated than the conventional parametric prediction intervals. Further, the choice of prediction interval significantly affects the severity of an excess death estimate. The empirical prediction intervals reveal that the likelihood of exceeding a 10% threshold of excess deaths varies by season. Across the 23 countries studied, finding at least 10% weekly excess deaths in a single week during summer or winter is not very unusual under non-pandemic conditions. These results contrast sharply with those derived using a standard negative-binomial GAM. Conclusion Our results highlight the importance of well-calibrated prediction intervals that account for the naturally occurring seasonal uncertainty in mortality forecasting. Empirical prediction intervals provide a better performing solution for estimating forecast uncertainty in the analyses of excess deaths compared to conventional parametric intervals.

Background Extreme temperatures are associated with negative health outcomes, in particular for older adults with pre-existing conditions. While climate change is expected to increase exposure to temperature levels that are detrimental for health, little is known about how dementia shapes vulnerability to extreme temperatures. Methods We leveraged repeated quarterly individual-level health claims from 2004 to 2019 on 250,000 individuals in Germany aged 50 years and above with information on key neurodegenerative diseases such as dementia. We linked data on the location of residence of these individuals with high resolution gridded meteorological data. In our empirical analysis, we applied an individual-level Fixed Effects model to estimate how temperature affects the single patient’s probability of hospitalization and death, adjusted for seasonality and comorbidities. Results Our findings reveal that heat and cold exposure increase the risk of death. Conversely, the association between extreme temperatures and hospital admissions is more nuanced showing an increase only with cold exposure. Stratifying the analysis by individuals affected by dementia, we observe heat to increase mortality only for individuals with dementia and cold to determine an 8 times larger impact on them and a larger increase in hospitalization. Also, we observe individuals aged above 80 and with dementia do be the most at risk of death with exposure to cold and in particular heat. Conclusion Our study contributes to the growing body of evidence on the health impacts of climate change and emphasizes the need for targeted strategies to protect vulnerable groups, particularly patients with dementia, from adverse temperature effects.