All-cause and cause-specific mortality among older migrant and non-migrant adults in Finland: a register study on all deaths, 2002–2020 Authors: Laura Kemppainen, Teemu Kemppainen, Jani Raitanen, Mari Aaltonen, Leena Forma, Anne Kouvonen, Jutta Pulkki Published: 2024 ABSTRACT AbstractMigrant mortality advantage is established in various studies, but there is a lack of evidence on migrant mortality trends in old age. Previous studies have primarily concentrated on all-cause mortality, and few include older age groups. Discussions about the migrant mortality advantage continue due to concerns about data availability and accuracy. Additionally, the mechanisms explaining the migrant mortality advantage remain unclear. This study examines all-cause and cause-specific mortality among older migrant and nonmigrant adults aged 70 and above using Finnish Cause of Death register data (2002–20) and the corresponding risk population. We investigate differences in overall and cause-specific mortality between migrant and Finnish-born population and by geographical region of origin. We calculated direct age-standardized mortality rates and age group-specific death rates followed by Poisson regression to study relative mortality differences. Age at death, sex, income, region of residence, and year of death were controlled for in the regression analysis. We found evidence of migrant mortality advantage across various causes of death, but there was variation by regions of origin groups and sex. Notably, women exhibit the strongest advantage in respiratory and digestive system diseases, while men demonstrated pronounced advantages in external causes and respiratory diseases. Our study challenges the notion of a general diminishing healthy migrant effect in old age. Our findings emphasize the need for nuanced investigations into socioeconomic factors and tailored interventions for older migrants. Introduction There is a large body of public health literature concerning differences in mortality rates between migrants and the majority populations [1, 2]. A recent meta-analysis [1] found evidence of migrants’ mortality advantage across most causes of death in high-income destination countries for those who migrated for work, education, or family reunification. Another recent review [2] identified a moderate mortality advantage for working-age migrants who moved to the Global North from specific regions (e.g. Northern Africa, Asia). Also, the mechanisms explaining migrant mortality advantage remain unclear [3]. Recent research has suggested that the observed migrant mortality advantage could be a data artifact, explained by over-coverage in data when people are recorded as residents even though they have left the country [4, 5]. However, Wallace and Wilson [6] have demonstrated with Swedish data that over-coverage adjustments do not radically change the results, especially in older age categories. Recent research has identified a U-shaped trend in migrant mortality advantage, observing the greatest mortality advantage among younger age groups [2, 4, 6]. It is argued, that as migrants age, their mortality advantage tends to converge with or even shift to a disadvantage compared to the local-born populations [4, 6]. This shift in mortality patterns may be attributed to a diminishing impact of the healthy migrant effect when health-based selection effects diminish over time due to the challenging living conditions in the destination country [7]. Migrants’ adaptation to local behavioural norms and lifestyles could also drive this convergence [4, 8]. However, this trend is not uniform across all migrant groups [4, 6]. Some populations continue to experience mortality advantage into old age, while others face a mortality disadvantage in specific older age cohorts [6]. To clarify and explain these variations, we need more detailed studies on mortality in older age groups [4]. Notably, previous studies on old-age migrant mortality have mainly examined all-cause mortality without scrutinizing specific causes of death. Cause-specific mortality reveals the diseases or conditions leading to deaths, helping us understand the unique health challenges and mortality risks of different migrant groups in each country. The only study concentrating solely on older age groups and specific causes of death is by Reus-Pons et al. [9] who conducted a mortality follow-up to year 2009 among migrants and non-migrants aged 50 years or more at the 2001 census living in Flanders and the Brussels-Capital Region. The study revealed an all-cause mortality advantage for migrants but also underscored specific mortality disadvantages among distinct origin groups. Our research is guided by three lines of thought stemming from the recent mortality research. First, it is crucial to study different kinds of population groups in detail to chart the existence and characteristics of mortality advantage. Given the rapid growth of the older migrant populations in Europe [10], it is important to examine the question of mortality advantage in older age groups [6]. Older migrants are often particularly susceptible to health inequalities due to different intersecting sources of disadvantage [11, 12]. Studying ethnic differences in causes of death helps identify public health issues, guide healthcare strategies for minorities and migrants, and tailor policies for ageing, ethnically diverse societies [13, 14]. Second, it is beneficial to differentiate between various causes of death since the extent of the mortality advantage is likely to vary across causes, and in certain instances, so too is the direction [9]. More generally, cause-specific results may help to uncover the mechanisms behind inequalities in mortality [15, 16]. Cause-specific studies including all age groups have not found evidence of mortality advantage in the cases of infectious diseases and external causes of death [1]. A study on migrants aged below 70 years in six European countries found that migrants typically had mortality disadvantage in infectious diseases and homicide, but mortality advantage in cancer and suicides [17]. Reus-Pons et al. [9] identified specific origin groups of older migrants with mortality disadvantage in infectious diseases, diabetes-related causes, respiratory diseases, cardiovascular diseases, and lung cancer. Third, we need to pay attention to the societal setting in its complexity. Theorizing the societal conditions of migrant mortality advantage, Roelfs and Shor [2, 18] followed van Tubergen and Kalmijn [19] and distinguished effects attributable to the country of origin, destination country and the relationship between the two, referred as the ‘setting’ [6]. Our case will contribute to these discussions by comparing older migrants from diverse geographical regions (Global North and West, Global South and East, Former USSR, Sweden, as well as Eastern Europe and Former Yugoslavia) with the corresponding majority population in Finland (Supplementary Fig. 1). Building on the research initiated by Reus-Pons et al. [9] on older migrants, our study contributes to the existing literature by utilizing more recent nationwide data spanning a longer timeframe (2002–20). Furthermore, we focus specifically on individuals aged 70 or older within a distinct societal context. Finland as a destination country can be described as a Nordic welfare state with universalist welfare benefits and services, a fairly low level of migrant population, relatively large socio-economic health inequalities [20] and a somewhat lower life expectancy [21] compared to other Nordic countries. The implications of different geographical regions of origin (‘setting’) range from cultural and social proximity (Sweden) to a history of geopolitical tensions (Russia) and substantial differences in multiple aspects (most countries of the Global South). Through these three aspects—focus on older age, cause-specific analysis, and attention to the societal setting—our article aims to further our understanding of migrant mortality advantage. In practise, we examine both all-cause and cause-specific mortality in old age comparing migrants from different regions of origin and Finnish-born population in Finland in 2002–20. Using register data covering all deaths of people aged 70 years and older during this period, we investigate: differences in overall and cause-specific mortality (ICD-10) between migrant and Finnish-born population in general and by age group, and differences in overall and cause-specific mortality (ICD-10) between migrant and Finnish-born population in general and by age group, and differences in overall and cause-specific mortality by geographical region of origin. differences in overall and cause-specific mortality by geographical region of origin. Following the typical approach in the literature [1, 9, 22], we stratify the analyses by sex. Given the context of a high-income destination country [1], it is reasonable to expect that there is some form of mortality advantage between migrant and the majority population in most causes of death [1], but it is not clear whether this applies to the older age groups or to all regions of origin [18]. At the end of the year 2022, there were nearly half a million people whose native language was not Finnish, Swedish, or Sami (the official languages) in Finland. Their number has doubled in 11 years. The proportion of these foreign-language speakers in the total population was 9% [23]. The largest language groups consisted of individuals speaking Russian, Estonian, and Arabic. About 3% of these foreign language speakers were 70 years old or older [24]. Compared to the 18% of Finnish speakers the number is still low but the number of older migrants is growing fast [10]. The highest number of individuals over 70 years old are among speakers of Russian, Estonian, and English [24]. The most common countries of birth for this demographic are the Former USSR, Sweden, Germany, and the United Kingdom [25]. Methods The study is based on register data including all deaths of individuals who died at age 70 or older in Finland in 2002–20 (N = 718 717) drawn from the Causes of Death Register and the corresponding risk population (N = 13 241,620 person years) from the Statistics Finland Population Structure. Of the recorded deaths, 1.1%, and of the risk population 1.3% were migrants. Migrant was defined as having been born outside of Finland. The data included 1,338 observations with missing country of origin, which were dropped from the regression analysis resulting in N = 717 379. Causes of death were obtained from the Finnish Causes of Death Register (Statistics Finland). For our study, we distinguished the following cause of death categories (ICD-10 codes): Categories with limited occurrences were excluded from our analysis. A00–B99: Certain infectious and parasitic diseases A00–B99: Certain infectious and parasitic diseases C00–D48: Neoplasms C00–D48: Neoplasms E00–E90: Endocrine, nutritional, and metabolic diseases E00–E90: Endocrine, nutritional, and metabolic diseases F00–F99: Mental and behavioural disorders F00–F99: Mental and behavioural disorders G00–G99: Diseases of the nervous system G00–G99: Diseases of the nervous system I00–I99: Diseases of the circulatory system I00–I99: Diseases of the circulatory system J00–J99: Diseases of the respiratory system J00–J99: Diseases of the respiratory system K00–K93: Diseases of the digestive system K00–K93: Diseases of the digestive system V01–Y98: External causes of morbidity and mortality V01–Y98: External causes of morbidity and mortality R00–R99: Symptoms, signs, and abnormal clinical and laboratory findings, not elsewhere classified, includes no death certificate (R999x) R00–R99: Symptoms, signs, and abnormal clinical and laboratory findings, not elsewhere classified, includes no death certificate (R999x) We also obtained the following register information from Statistics Finland: age at death (70–78, 79–84, 85–89, ≥90 years) (the age group categories correspond to the age quartiles of the migrant population in the causes of death register), age at death (70–78, 79–84, 85–89, ≥90 years) (the age group categories correspond to the age quartiles of the migrant population in the causes of death register), sex (female, male), sex (female, male), disposable personal yearly income (≤11 999, 11 200–14 399, 14 400–18 399, ≥18 400 EUR) [missing information on income (n = 767) was imputed using regression predictions. Our imputation model used the following predicting variables: interaction of age (10 categories) and sex, subnational region of residence, country of birth and the year of death), disposable personal yearly income (≤11 999, 11 200–14 399, 14 400–18 399, ≥18 400 EUR) [missing information on income (n = 767) was imputed using regression predictions. Our imputation model used the following predicting variables: interaction of age (10 categories) and sex, subnational region of residence, country of birth and the year of death), region of residence in Finland (NUTS codes: FI1B Uusimaa, FI1C Southern Finland, FI1D Northern and Eastern Finland, FI19 Western Finland, FI20 Åland), region of residence in Finland (NUTS codes: FI1B Uusimaa, FI1C Southern Finland, FI1D Northern and Eastern Finland, FI19 Western Finland, FI20 Åland), year of death (2002–2020), and year of death (2002–2020), and geographical region of origin (Finland, Former USSR, Sweden, Eastern Europe and Former Yugoslavia, Global North and West, Global South and East). geographical region of origin (Finland, Former USSR, Sweden, Eastern Europe and Former Yugoslavia, Global North and West, Global South and East). To calculate mortality rates, we obtained the corresponding risk population sizes for all combinations of the mentioned variables from the Statistics Finland (data available by request). As the risk population reflected the situation in the last day of the year, to obtain an estimate of the true risk population we added the number of all deaths per year to the end-of-year risk population. Mortality differences were first analyzed by direct age-standardized mortality rates (ASMR) and crude death rates (CDR) for each age category. This was followed by Poisson regression to study relative mortality differences. Poisson regression has been found a reliable choice even in the case of over-dispersion [26, 27]. Robust standard errors (Huber-White/sandwich) were estimated. Two models were specified: Model 1 includes only age and year of death as explanatory variables, while Model 2 adds also the region of residence and disposable income. Analyses were performed with Stata version 17 (StataCorp, College Station, Texas). Results We begin our exploration with the study population. The proportion of women varies considerably by geographical region of origin, ranging from 42% (Global North and West) to 71% (Former USSR). The youngest age group is rather pronounced among those from Sweden (73%) and those from the Global South and East (73%). Socio-economic differences are revealed by income data: while those with a Swedish background are relatively well-off (48% in the highest income quartile), the case was the opposite for those from the Former USSR (57% in the lowest income quartile). Geographically, the migrant populations are somewhat concentrated in the region Helsinki-Uusimaa, especially those from the Global South and East (62%). Descriptive statistics for the study population are found in Supplementary Table 1. Diseases of the circulatory system, neoplasms and diseases of the nervous system were the three most common causes of death categories in all country groups (Supplementary Fig. 2). The rate of neoplasms was somewhat more pronounced among those from the Global South and East and those from Sweden. A relatively high rate of endocrine, nutritional, and metabolic diseases and a low rate of mental and behavioural disorders further characterizes the classification of deaths among those from the Global South and East. The share of non-classified cases is relatively high among those from former Yugoslavia and Eastern Europe. Compared to the Finnish-born population, migrants generally exhibited lower mortality rates (Table 1). Taking the Finnish-born as the standard population, ASMR for Finnish-born men was 56.2 per 1000 (95% CI: 56.0–56.4), while the rate was clearly lower for the male migrants, 48.9 (47.2–50.5). Among women, the difference between the Finnish-born and migrants was clear, although less pronounced: ASMR for the Finnish-born women was 48.5 per 1000 (48.3–48.6) and 44.0 (42.7–45.2) for migrant women. Decomposing migrant populations by country of origin showed a notable variation in ASMR. Among both men and women, the lowest ASMR was found for migrants from the Global South and East, while the highest rates were among those from the former USSR and Sweden for men, and Former Yugoslavia and Eastern Europe for women, which come close to those of the majority population. Table 1.ASMR using the Finnish-born population as the standard, number of deaths, person-years at riskMenWomenOriginASMR per 1000 [95% CI]DeathsPerson-years at riskASMR per 1000 [95% CI]DeathsPerson-years at riskFinnish-born56.2 [56.0–56.4]308898549903748.5 [48.3–48.6]4008168268360Migrant (all)48.9 [47.2–50.5]31386898244.0 [42.7–45.2]4527106644Former USSR54.5 [52.0–57.0]16443046645.3 [43.9–46.8]327374318Sweden55.5 [48.8–62.1]273623544.5 [38.9–50.1]2336220Former Yugoslavia and Eastern Europe44.8 [37.9–51.7]161423353.6 [45.7–61.4]1734384Global North and West44.3 [41.3–47.3]7811834040.8 [37.7–43.9]66813159Global South and East34.6 [30.4–38.9]279970826.9 [22.8–31.0]1808563 ASMR using the Finnish-born population as the standard, number of deaths, person-years at risk Next, we describe how crude death rates (CDR) differ by age categories (Fig. 1) and regions of origin (for full results with confidence intervals, see Supplementary Table 2). As above, we stratify the analysis by sex. In men, the differences in CDR by region of origin are clearly larger in the oldest age group, and especially the Global South and East, and somewhat less strikingly, Eastern Europe and Yugoslavia, stand apart with lower death rates. In women, only the Global South and East shows such a diverging result. Crude death rates (CDR) by age group for men and women. Regarding relative differences in all-cause mortality (Table 2), we found an overall mortality advantage for migrants both in the simple model (Model 1: age and year of death) and in the model with full adjustments (Model 2: age, year of death, region of residence, disposable income) using Finnish-born individuals as the reference group. Figure 2 depicts the results from the fully adjusted models and shows that the advantage was somewhat stronger for men with an incidence rate ratio (IRR) of 0.81 (95% CI: 0.78–0.84) in the full model, compared to an IRR of 0.90 (0.87–0.93) for women. Mortality among migrant vs. Finnish-born population for women and men. IRRs and 95% CI from Poisson regression models. Nonclassified cases not shown. Table 2.Incident Rate Ratios (IRR) from Poisson regression model (robust stand. err.).All-cause mortalityMaleFemalen=312,036n=405,343MODEL 1MODEL 2MODEL 1MODEL 2Finnish-born populationref....Migrant population0.88 [0.84–0.91]0.81 [0.78–0.84]0.91 [0.88–0.94]0.90 [0.87–0.93]Former USSR countries0.97 [0.93–1.02]0.86 [0.82–0.91]0.94 [0.91–0.97]0.92 [0.89–0.96]Sweden0.95 [0.84–1.06]0.99 [0.89–1.11]0.93 [0.83–1.06]0.95 [0.84–1.07]Eastern Europe and Former Yugoslavia0.81 [0.70–0.93]0.75 [0.64–0.86]1.11 [0.96–1.27]1.09 [0.95–1.25]Global North and West0.79 [0.74–0.85]0.77 [0.71–0.83]0.85 [0.79–0.92]0.85 [0.79–0.92]Global South and East0.66 [0.59–0.75]0.58 [0.52–0.66]0.61 [0.53–0.70]0.60 [0.52–0.69] Certain infectious and parasitic diseases A00–B99 Male Female n=1,980 n=2,757 MODEL 1 MODEL 2 MODEL 1 MODEL 2 Finnish-born populationref....Migrant population1.24 [0.85–1.81]1.16 [0.79–1.71]0.96 [0.68–1.36]0.95 [0.67–1.35]Former USSR countries1.04 [0.58–1.87]0.95 [0.52–1.73]0.77 [0.49–1.21]0.76 [0.48–1.19]Sweden0.59 [0.08–4.22]0.59 [0.08–4.26]1.78 [0.56–5.51]1.82 [0.59–5.66]Eastern Europe and Former Yugoslavia1.72 [0.43–6.86]1.61 [0.40–6.43]0.95 [0.13–6.73]0.95 [0.13–6.74]Global North and West1.65 [0.89–3.06]1.59 [0.86–2.95]1.50 [0.75–3.00]1.52 [0.76–3.05]Global South and East1.32 [0.42–4.10]1.19 [0.38–3.73]1.07 [0.27–4.28]1.06 [0.26–4.24] Neoplasms C00–D48MaleFemalen=75,704n=72,295 MODEL 1MODEL 2MODEL 1MODEL 2Finnish-born populationref....Migrant population0.95 [0.89–1.02]0.83 [0.78–0.89]0.98 [0.91–1.04]0.89 [0.83–0.95]Former USSR countries1.05 [0.96–1.16]0.89 [0.81–0.97]1.00 [0.93–1.08]0.90 [0.84–0.98]Sweden1.16 [0.93–1.44]1.14 [0.92–1.42]1.22 [0.96–1.56]1.20 [0.94–1.53]Eastern Europe and Former Yugoslavia0.79 [0.59–1.05]0.69 [0.52–0.93]1.15 [0.84–1.56]1.04 [0.77–1.42]Global North and West0.85 [0.74–0.97]0.80 [0.69–0.92]0.85 [0.70–1.03]0.82 [0.68–1.00]Global South and East0.73 [0.59–0.91]0.61 [0.49–0.75]0.68 [0.51–0.92]0.60 [0.44–0.81] Endocrine, nutritional and metabolic diseases E00–E90MaleFemale n=3,600n=4,509MODEL 1MODEL 2MODEL 1MODEL 2Finnish-born populationref....Migrant population0.84 [0.60–1.18]0.76 [0.54–1.06]1.12 [0.86–1.45]1.06 [0.82–1.38]Former USSR countries0.66 [0.39–1.14]0.57 [0.33–0.99]1.00 [0.72–1.39]0.93 [0.67–1.30]Sweden0.56 [0.14–2.24]0.57 [0.14–2.28]2.10 [0.95–4.64]2.18 [0.98–4.86]Eastern Europe and Former Yugoslavia1.62 [0.61–4.32]1.45 [0.54–3.88]1.58 [0.51–4.92]1.49 [0.48–4.61]Global North and West0.61 [0.26–1.40]0.58 [0.25–1.34]0.99 [0.50–1.97]1.01 [0.51–2.01]Global South and East 1.83 [1.01–3.32] 1.56 [0.86–2.83]1.66 [0.75–3.67]1.54 [0.69–3.43] Mental and behavioural disorders F00–F99MaleFemale n=13,160n=30,309MODEL 1MODEL 2MODEL 1MODEL 2Finnish-born populationref....Migrant population1.03 [0.88–1.21]0.92 [0.78–1.08]0.99 [0.89–1.09]0.93 [0.84–1.03]Former USSR countries1.30 [1.06–1.60]1.11 [0.90–1.37]0.98 [0.87–1.11]0.91 [0.81–1.03]Sweden0.53 [0.24–1.18]0.58 [0.26–1.28]0.91 [0.57–1.45]0.92 [0.58–1.46]Eastern Europe and Former Yugoslavia1.33 [0.72–2.47]1.18 [0.64–2.19]1.02 [0.56–1.86]0.95 [0.52–1.74]Global North and West0.82 [0.59–1.14]0.77 [0.55–1.07]1.19 [0.95–1.48]1.17 [0.94–1.45]Global South and East0.56 [0.28–1.10]0.45 [0.23–0.90]0.29 [0.12–0.69]0.26 [0.11–0.62] Diseases of the nervous system G00–G99MaleFemale n=37,154n=68,172MODEL 1MODEL 2MODEL 1MODEL 2Finnish-born populationref....Migrant population0.69 [0.62–0.79]0.83 [0.74–0.93]0.77 [0.71–0.84]0.91 [0.83–0.98]Former USSR countries0.82 [0.71–0.96]1.03 [0.88–1.20]0.82 [0.74–0.90]0.99 [0.90–1.09]Sweden1.12 [0.80–1.58]1.21 [0.87–1.70]0.82 [0.59–1.14]0.82 [0.59–1.13]Eastern Europe and Former Yugoslavia0.51 [0.29–0.89]0.60 [0.34–1.05]0.67 [0.43–1.05]0.79 [0.51–1.24]Global North and West0.52 [0.40–0.67]0.56 [0.44–0.72]0.78 [0.65–0.98]0.81 [0.66–0.99]Global South and East0.44 [0.29–0.68]0.57 [0.37–0.88]0.28 [0.17–0.45]0.35 [0.21–0.57] Diseases of the circulatory system I00–I99MaleFemale n=132,117n=178,769MODEL 1MODEL 2MODEL 1MODEL 2Finnish-born populationref....Migrant population0.82 [0.77–0.87]0.75 [0.71–0.80]0.90 [0.86–0.94]0.89 [0.85–0.93]Former USSR countries0.92 [0.86–1.00]0.80 [0.74–0.87]0.96 [0.91–1.01]0.93 [0.88–0.98]Sweden0.85 [0.70–1.03]0.91 [0.75–1.11]0.81 [0.66–1.00]0.86 [0.70–1.05]Eastern Europe and Former Yugoslavia0.62 [0.48–0.81]0.57 [0.44–0.74]1.03 [0.81–1.30]1.01 [0.79–1.28]Global North and West0.77 [0.69–0.86]0.75 [0.67–0.84]0.76 [0.68–0.86]0.78 [0.69–0.87]Global South and East0.59 [0.48–0.72]0.51 [0.42–0.63]0.66 [0.53–0.82]0.64 [0.52–0.80] Diseases of the respiratory system J00–J99MaleFemale n=20,284n=15,150MODEL 1MODEL 2MODEL 1MODEL 2Finnish-born populationref....Foreign-born population0.82 [0.71–0.95]0.71 [0.61–0.83]0.77 [0.64–0.91]0.72 [0.60–0.86]Former USSR countries0.89 [0.73–1.10]0.75 [0.60–0.92]0.73 [0.59–0.91]0.68 [0.55–0.85]Sweden1.01 [0.64–1.59]1.04 [0.66–1.64]0.86 [0.43–1.71]0.79 [0.40–1.58]Eastern Europe and Former Yugoslavia1.07 [0.63–1.83]0.95 [0.55–1.62]1.74 [0.94–3.21]1.62 [0.88–3.00]Global North and West0.72 [0.53–0.96]0.68 [0.51–0.92]0.79 [0.53–1.18]0.76 [0.51–1.13]Global South and East0.49 [0.27–0.89]0.39 [0.21–0.71]0.58 [0.26–1.27]0.51 [0.23–1.13] Diseases of the digestive system K00–K93MaleFemale n=8,881n=12,281MODEL 1MODEL 2MODEL 1MODEL 2Finnish-born populationref....Foreign-born population0.93 [0.76–1.14]0.81 [0.66–0.99]0.76 [0.63–0.92]0.74 [0.62–0.89]Former USSR countries0.81 [0.59–1.11]0.66 [0.48–0.91]0.78 [0.62–0.96]0.75 [0.60–0.93]Sweden0.34 [0.11–1.08]0.37 [0.12–1.16]1.30 [0.70–2.40]1.35 [0.73–2.50]Eastern Europe and Former Yugoslavia1.34 [0.67–2.68]1.16 [0.58–2.31]0.61 [0.20–1.89]0.59 [0.19–1.82]Global North and West1.34 [0.98–1.84]1.26 [0.92–1.72]0.47 [0.26–0.85]0.48 [0.27–0.86]Global South and East0.70 [0.36–1.33]0.55 [0.29–1.06]0.97 [0.51–1.84]0.93 [0.49–1.76] External causes of morbidity and mortality V01–Y98MaleFemale n=13,999n=12,412MODEL 1MODEL 2MODEL 1MODEL 2Finnish-born populationref....Foreign-born population0.71 [0.59–0.85]0.66 [0.55–0.79]0.85 [0.72–1.02]0.83 [0.70–0.99]Former USSR countries0.74 [0.57–0.95]0.65 [0.50–0.84]0.83 [0.67–1.02]0.80 [0.65–0.99]Sweden0.97 [0.57–1.64]1.05 [0.62–1.78]0.78 [0.35–1.73]0.80 [0.36–1.77]Eastern Europe and Former Yugoslavia0.54 [0.23–1.30]0.50 [0.21–1.20]1.03 [0.43–2.47]1.00 [0.42–2.40]Global North and West0.74 [0.53–1.03]0.72 [0.51–1.00]1.22 [0.84–1.76]1.21 [0.84–1.75]Global South and East0.45 [0.24–0.86]0.40 [0.21–0.75]0.22 [0.55–0.88]0.21 [0.05–0.84] Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified R00–R99 (includes no death certificate (R999x))MaleFemale n=908n=1054MODEL 1MODEL 2MODEL 1MODEL 2Finnish-born populationref....Foreign-born population9.50 [7.64–11.8]6.34 [4.99–8.05]7.70 [6.20–9.58]6.28 [5.02–7.85]Former USSR countries12.7 [9.69–16.6]7.93 [5.90–10.7]7.31 [5.60–9.53]5.99 [4.56–7.85]Sweden1.12 [0.16–8.03]0.78 [0.11–5.79]1.53 [0.22–10.9]1.13 [0.16–8.12]Eastern Europe and Former Yugoslavia19.6 [11.1–34.6]13.7 [7.7–24.4]32.5 [19.2–54.8]26.7 [15.7–45.5]Global North and West4.78 [2.77–8.27]3.78 [2.17–6.60]5.61 [3.00–10.5]4.91 [2.62–9.20]Global South and East9.33 [5.22–16.7]5.95 [3.30–10.7]8.38 [4.00–17.5]6.29 [3.00–13.2]*Model 1 adjusted for age and year of death. Model 2 adjusted for age, year of death, region of residence and disposable income. Incident Rate Ratios (IRR) from Poisson regression model (robust stand. err.). Model 1 adjusted for age and year of death. Model 2 adjusted for age, year of death, region of residence and disposable income. There were clear differences in the all-cause mortality advantage by the region of origin (Table 2). Among women, the fully adjusted IRR ranged from 0.60 (0.52–0.69) for those from the Global South and East to 0.92 (0.89–0.96) for those from the Former USSR. Among men, the corresponding IRR range extended from 0.58 (0.52–0.66) for those from the Global South and East to 0.86 (0.82–0.91) among people from the former USSR. The all-cause mortality did not show a significant difference from the Finnish-born population for men and women from Sweden and women from Eastern Europe and former Yugoslavia. Cause-specific analysis showed that the strength of migrants’ mortality advantage varied by the cause of death (Fig. 2; Table 2). Among women, the strongest advantage was found in diseases of the respiratory system and diseases of the digestive system with the fully adjusted IRRs of 0.72 (0.60–0.86) and 0.74 (0.62–0.89), respectively. In men, the most pronounced mortality advantage was in external causes (IRR = 0.66; 0.55–0.79) and in respiratory diseases (0.71; 0.61–0.83). Regarding the properly classifiable causes of death (i.e. other than R00–R99: not elsewhere classified), almost all results pointed to the direction of mortality advantage for migrants. The only exception was observed in mental and behavioural disorders where men from the former USSR had a mortality disadvantage (1.30; 1.06–1.60) relative to their Finnish-born counterparts. However, this difference did not remain statistically significant in the full model. The migrants from the Global South and East tended to have the lowest cause-specific mortality rates in most causes of death, except for endocrine, nutritional, and metabolic diseases, where men from the USSR exhibited the greatest advantage (0.57; 0.33–0.99) and diseases of the digestive system where women from the Global North and West showed the greatest advantage (0.48; 0.27–0.86). In the case of R00-R99 (not elsewhere classified), the situation was reversed, as migrants were over-represented in this category. The fully adjusted IRRs ranged from non-significance (Sweden) to 26.7 (15.7–45.5) for women from Eastern Europe and Yugoslavia. Regarding our full regression adjustments for income and region of residence (Model 2), all-cause mortality advantage becomes stronger in men and remains unchanged in women. The strengthening of mortality advantage is most notable in cases of neoplasms, respiratory and digestive diseases, and external causes, particularly for men. Only in nervous system diseases does the trend reverse, with adjustments attenuating the advantage. Discussion In this article, we studied mortality among migrants from different origin regions focusing on older age, all-cause, and cause-specific analysis in the context of Finland, a Nordic welfare state. Our study uncovers an all-cause mortality advantage among migrants in old age, challenging the assertion of a diminishing healthy migrant effect with advancing age [4, 6]. Similar to previous findings from the UK, the USA, France, and Sweden [4, 6], our age group-specific examination showed that migrants from the Global South stood out from other groups with lower death rates at the oldest age groups. Moreover, our cause-specific analysis displayed variations across different causes of death and region of origin groups. Nevertheless, when compared to older Finnish-born individuals, older migrants consistently demonstrate a mortality advantage across many categories, except for mental and behavioural disorders (disadvantage for men from the Former USSR). Future studies should analyze cause-specific mortality across different age categories to provide more detailed results. Lower socioeconomic position has been associated with higher mortality also in old age [28]. Because in our data older migrants had mainly lower income than their Finnish-born counterparts, income differences could not explain the mortality advantage. In fact, controlling for income and region of residence typically strengthened the estimated advantage. This illustrates a clear example of a suppression effect wherein the covariates were concealing the actual relationship between migrancy and mortality, and their inclusion revealed a more accurate and pronounced mortality advantage. In the mentioned case of mortality disadvantage we could attribute the result to migrants’ lower income level. It has been argued that the socioeconomic gradient of health operates differently for migrants; however, the processes underlying this observation remain largely unknown [29]. Future studies should delve deeper to the complex associations of socio-economic factors with migrant mortality in old age. Our finding on the male mortality advantage in external causes contrasts the general finding of Aldridge et al. [1], which found migrant mortality disadvantage in external causes. Notably, external causes include alcohol-related deaths. Given that alcohol use is a well-known risk factor associated with increased mortality among Finnish-born people, particularly in men, the cultural setting may accentuate the migrant advantage [30]. Thus, migrant mortality advantage may be attributed to a healthier lifestyle in specific aspects compared to the Finnish-born population, particularly evident in lower mortality rates for external causes, cancers, and respiratory diseases (e.g. tobacco smoking). This example emphasizes the importance of careful consideration of the societal setting, when attempting to understand mortality patterns [3]. Furthermore, older migrants demonstrate a mortality advantage in the category of diseases of the nervous system encompassing dementias, a noteworthy finding considering Finland’s high dementia rates in international comparisons [31]. This could point to a genetic explanation, but we should also note that the risk of dementia is a complex interplay between genes, lifestyle, and environmental factors [32, 33]. Moreover, recent studies have documented underdiagnosing of dementia and barriers to timely diagnosis and care among ethnic minorities [34]. All these complex dynamics merit more detailed future research. Additionally, the prevalence of non-classified causes of death is higher among migrants except those from Sweden, signifying a potential ‘salmon bias’ effect, which refers to migrants who return to their origin destinations when ill or shortly before death [1, 35]. Our data reflect the fact that when a person dies in Sweden, the death certificate will be automatically sent to the Finnish authorities, but in other countries, it will be up to the relatives to inform the authorities [36]. Another Finnish mortality study [30] found that ill-defined cases and missing death certificates were from cases where an individual had died abroad and the body had not been brought back to Finland. Recently, concerns have been voiced on whether migrant mortality advantage is merely a data artifact stemming from over-coverage [5]. We had only aggregated data in at our disposal not allowing us to deal with over-coverage at the individual level. However, as Wallace and Wilson [6] demonstrated over-coverage plays only a small part in explaining the migrant mortality advantage, especially in the older age groups. A better understanding of over-coverage and the salmon bias effect would increase accuracy of mortality studies and help to explain mechanisms behind migrant mortality advantage. Moreover, the reliability of cause-of-death classification might decrease with old age due to an increase in competing causes, resulting in greater difficulty in establishing the root cause of death [37]. The comparability of causes of death across different countries can be problematic due to differences in diagnosis tools and scoring methods. For example, there is an established need to harmonize the assessment of dementia across different European countries [38]. Improving the accuracy of death reporting and cross-border coordination would enhance mortality data, leading to a better understanding of migrants' health outcomes. Strengths and limitations Our study is the first one to examine the causes of death among older migrants in Finland. It contributes to the limited international evidence on the migrant mortality advantage in older age. Compared to earlier migrant mortality studies, our study furthers understanding of the issue by including a longer period (2002–20), which enables the cause-specific investigation across different origin groups. However, as the numbers of migrants over 70 years old are still low in Finland, in some categories our sample is limited and some ICD categories were not included in the analysis [5]. Conclusion Our study challenges the notion of a diminishing healthy migrant effect in old age, revealing a rather consistent mortality advantage among older migrants across various causes of death. Our findings emphasize the need for nuanced investigations into socioeconomic factors. Improved death reporting and cross-border coordination are important for understanding of migrants’ mortality advantage, enhancing the accuracy of mortality data. Supplementary Material