This study aimed to examine gender-specific risk factors and regional differences in suicidal ideation across Nordic and Visegrad countries, with loneliness as a key characteristic of social determinants in later life. Our findings indicate that suicidal ideation among older adults is strongly shaped by social and contextual factors rather than by age alone. Loneliness emerged as the strongest and most consistent predictor of suicidal thoughts for both men and women, while education and partnership status showed gender-specific effects. Regional context was also relevant, as men living in Visegrad countries exhibited higher odds of suicidal ideation than their counterparts in Nordic countries, whereas no significant regional differences were observed among women.

These results are broadly consistent with previous SHARE-based research on passive suicidal ideation. Using SHARE Wave 7, Lee (2021) reported substantial cross-national variation in suicidal thoughts and consistently higher prevalence among women and identified loneliness and poor health as key correlates of passive death wishes in later life. By using SHARE Wave 8 data and adopting a different analytical approach that emphasizes socio-geographical regions and gender interactions, the present study extends this work by showing that regional context operates differently for men and women and that social determinants remain central to understanding suicidal ideation in later life. Together, these findings reinforce the importance of treating suicidal ideation as a distinct outcome from suicide mortality and as a critical target for early intervention and prevention (Klonsky & May, 2015; Snowdon, 2018; Schouler-Ocak & Khan, 2024).

Previous research suggests that gender differences in loneliness are often small and inconsistent once confounding factors are considered, although women in later life may experience more intense loneliness due to higher rates of widowhood and longer life expectancy (Pagan & Malo, 2024). These patterns provide an important background for interpreting gendered experiences of loneliness in older populations.

The regression analyses in our study demonstrate a strong and consistent association between loneliness and suicidal thoughts for both men and women. As loneliness increases, the odds of reporting suicidal thoughts rise markedly in both subsamples. This finding is in line with a substantial body of literature identifying loneliness as a key risk factor for suicidal ideation across genders (Cacioppo et al., 2006). Importantly, the SHARE item used in this study primarily captures passive death wishes rather than active suicidal intent, as documented in previous SHARE-based research (Lee, 2021). Loneliness may therefore be particularly influential in shaping passive suicidal ideation, consistent with theoretical models of late-life suicidal risk that emphasise social disconnection, hopelessness, and perceived burdensomeness (Schouler-Ocak & Khan, 2024). In this sense, our findings also align with Durkheim’s classic argument that weakened social integration increases vulnerability to suicidal outcomes.

Education emerges as a significant protective factor against suicidal thoughts, with notable differences between genders. For men, both middle and high education levels significantly reduce the likelihood of suicidal thoughts, whereas for women, only a high education level shows a significant protective effect in our sample. These gender differences may reflect varying societal roles and expectations, where higher education potentially provides men with greater socioeconomic stability and resilience against stressors (Stack, 2000). For women, the benefits of higher education may also relate to increased social support networks and better access to mental health resources (Cutler & Lleras-Muney, 2010). The nuanced impact of education on suicidal thoughts highlights the need to consider gender- specific pathways when examining protective factors in later life.

Previous research indicates that living alone, especially for widows and widowers, increases the risk of loneliness, which in turn is associated with suicidal thoughts (Pagan & Malo, 2024). Our results show that living with a partner significantly decreases the likelihood of experiencing suicidal thoughts for women, but not for men. This gender-specific effect suggests that women may derive greater emotional and social support from cohabiting relationships, which buffers against suicidal ideation. In contrast, men may not experience the same protective effect, possibly due to traditional gender norms emphasizing self-reliance and emotional stoicism (Canetto & Sakinofsky, 2010). Some studies further indicate that risk factors for suicidal thoughts may operate differently across socio-demographic groups and social contexts (Fung & Chan, 2011). Additionally, the duration of the suicidal process tends to be shorter in men than in women, while the higher lethality of suicide methods among men may explain observed gender differences in completed suicides (Schrijvers et al., 2012; De Leo & Kolves, 2017), without contradicting patterns in suicidal thoughts.

Although our study focuses broadly on suicidal thoughts, evidence indicates that older adults may experience mainly passive ideation, while younger older adults may more often report active suicidal thoughts (Cabello et al., 2021; Wastler et al., 2022). This distinction supports including participants aged 50–64 in the “late-life” sample, as it allows capturing age-related differences in the prevalence and nature of suicidal ideation.

Contrary to expectations, our results show no significant differences in suicidal thoughts across age groups for either gender. This indicates that chronological age, within the studied range, does not independently influence the likelihood of suicidal thoughts. Instead, other factors such as loneliness, education, and partnership status appear to play more critical roles in shaping suicidal ideation in later life (Waern et al., 2003).

A notable pattern concerns the oldest female age group (81–94 years), whose association with suicidal ideation changes across model specifications. In the unadjusted model, women in this age group show different odds of suicidal ideation compared with those aged 50–64 years; however, once loneliness, partnership status, education, and regional context are included, this association becomes non-significant and shifts in direction. This pattern is consistent with confounding and suppression effects. Older women in this sample are more likely to possess characteristics that are protective against suicidal ideation, such as lower reported loneliness or higher likelihood of co-residing with a partner in some contexts, which may mask vulnerabilities associated with advanced age in crude models. After adjustment, the residual association of age may reflect age-specific risks such as frailty, bereavement, declining health, and reduced autonomy that are not fully captured by the included covariates (De Leo & Giannotti, 2021). Given the cross-sectional design, this pattern should be interpreted as a composition effect rather than evidence of a causal role of age itself.

Although chronological age did not independently predict suicidal thoughts in our sample, the same age may correspond to different health and functional profiles across countries. Such cross-national differences in life expectancy and functional health may influence the interpretation of age effects and regional patterns of suicidal ideation (see Appendix 2 for detailed indicators of life expectancy and Healthy Life Years at birth by country and sex).

These regional differences are consistent with prior evidence showing substantial cross-national variation in Healthy Life Years (HLY) and morbidity compression in Europe (Straka et al., 2024). Although we restrict the sample to respondents aged 50–94 years and adjust for age in the regression models, the same chronological age may correspond to different average health profiles across countries. Life expectancy and HLY differ substantially between the Nordic and Visegrad regions, implying that, for instance, a 75-year-old in Sweden may, on average, be healthier than a 75-year-old in Hungary. Such cross-national differences in health may confound the interpretation of regional effects if unobserved health status is correlated with both country context and suicidal ideation; therefore, we interpret regional contrasts cautiously and treat cross-country heterogeneity in health and functional aging as a limitation (Appendix 2).

As an additional robustness consideration, cross-country differences in population health could be proxied by country-level indicators such as Eurostat HLY or life expectancy. However, given the small number of countries (N = 7), we refrain from over-interpreting macro-level adjustments and instead report them, if included, as sensitivity checks.

Comparative research indicates that Nordic and Central–Eastern European countries can be understood as distinct but analytically meaningful regional groupings shaped by different historical and institutional trajectories. Hilmarsson (2021) conceptualizes these countries as small-state clusters within Europe, emphasizing how variations in political economy, welfare arrangements, and EU integration have produced divergent development paths. Research on post-socialist welfare transformations further shows that Central European countries, including the Visegrad group, have followed reform trajectories that differ substantially from those of Nordic welfare states (Szikra, 2014; Borevi, 2017).

Evidence from comparative health and welfare research supports the relevance of a welfare regime perspective for understanding population health outcomes. Welfare regime characteristics account for substantial national-level variation in self-perceived health and exposure to social risks across Europe, with Nordic welfare states generally associated with more favourable outcomes than Central and Eastern European regimes (Eikemo et al., 2008; Widding-Havneraas & Pedersen, 2020; Berkowitz et al., 2024). Nordic countries typically feature universalistic welfare systems with high public social spending, extensive primary care, and community-based services for older adults, reflecting long-standing commitments to prevention, early intervention, and social participation (Esping-Andersen, 1990; Vabø & Szebehely, 2012). In contrast, Visegrad countries combine public provision with a stronger role for family support and market mechanisms, exhibiting lower public expenditure on health and social services and more limited development of community-based mental health care (Fenger, 2007; Cerami & Vanhuysse, 2009; Szikra, 2014; WHO, 2025). Across Europe, mental health care integration into primary care varies, and unmet needs remain higher in countries with less developed community-based services (OECD, 2025; WHO, 2025). These structural differences provide a relevant framework for interpreting regional patterns of late-life suicidal ideation.

Nordic and Visegrad countries also differ in economic and cultural contexts that shape late-life mental health outcomes. Nordic countries generally display higher levels of economic security and GDP per capita, while Visegrad countries experienced rapid post-socialist socioeconomic transformations, resulting in greater heterogeneity in material conditions among older adults. Culturally, Nordic societies emphasize individual autonomy supported by public systems, whereas Visegrad countries place more emphasis on family networks and traditional social roles (Pagan & Malo, 2024). These factors may influence experiences of aging, social isolation, and mental health. Despite shared demographic trends, including population aging and rising demands on health and social care, regional institutional and socio-cultural differences are likely to shape suicidal ideation patterns.

Welfare state differences are reflected in social protection, eldercare, and mental health services. Nordic countries feature universal pensions with high replacement rates, broad social transfers, and low old-age poverty (Esping-Andersen, 1990; Eikemo et al., 2008). Eldercare relies primarily on municipally provided home-help and home-nursing services, with needs-based entitlement and low reliance on long-term institutional care (Szebehely & Meagher, 2018). Mental health care for older adults is integrated into primary and municipal services, emphasizing prevention and early intervention through mechanisms such as primary-care screening for depression and coordination between health and social services (OECD, 2025; WHO, 2025). By contrast, Visegrad countries show lower public spending, more stratified benefits, and higher old-age poverty (Fenger, 2007; Cerami & Vanhuysse, 2009). Eldercare relies more heavily on families, with fewer community-based services and patchy psychosocial support; mental health care often depends on hospitals or psychiatric institutions, resulting in higher unmet need (WHO, 2025). These differences in welfare arrangements and service provision provide a plausible explanation for cross-national variation in older adults’ mental health.

Maskileyson et al. (2021) highlight significant variations in depressive symptoms among older adults across different European countries. Older adults in Scandinavian countries, such as Denmark and Sweden, generally report fewer depressive symptoms compared to their counterparts in other European countries.

Our study finds that men living in Nordic countries have a significantly lower risk of experiencing suicidal thoughts compared to those in Visegrad countries. This regional disparity is not observed in women, indicating that regional effects on suicidal thoughts are more pronounced in men. The lower risk in Nordic countries for men might be attributed to comprehensive social welfare systems, better mental health services, and higher societal support prevalent in these countries (Hagerty et al., 2002). In contrast, we presume, the absence of significant regional differences for women suggests that other factors, possibly related to gender equality and social norms, may mitigate the regional effects observed in men.

Beyond welfare-state and service-related explanations, cultural factors may also shape regional differences in suicidal ideation (Snowdon, 2018). Culture influences how psychological distress is perceived and expressed, as well as norms surrounding help-seeking and disclosure. Previous research highlights that variations in suicide patterns across gender and age groups are partly rooted in sociocultural contexts (Schouler-Ocak & Khan, 2024).

Taken together, our findings suggest that regional differences in suicidal thoughts emerge from a complex interplay of institutional, social, and cultural mechanisms, with gender moderating these effects. It is important to note that suicidal ideation represents only one stage in the continuum of suicidal behaviour, and regional differences in ideation may not directly translate into differences in suicide attempts or completed suicides (Snowdon, 2018; Schouler-Ocak & Khan, 2024).

Several limitations should be considered when interpreting our findings. First, this study focuses on suicidal ideation rather than suicide attempts or completed suicides. As ideation represents an earlier stage in the suicidal continuum, the results should be interpreted with caution and may not fully reflect risk for more severe suicidal outcomes.

Second, potential biases inherent in the SHARE data, such as self-reporting and regional specificity, may limit the generalizability of our findings. Additionally, the cross-sectional design further restricts our ability to infer causality and restricts conclusions to associations observed at a single point in time.

Third, the single-item measure of suicidal thoughts used in this study does not distinguish between transient thoughts, passive death wishes, and active suicidal planning.

A further limitation concerns cultural differences in the reporting of suicidal ideation. Attitudes toward suicide and mental illness vary across societies, and stigma or social norms may influence individuals’ willingness to disclose suicidal thoughts in surveys. Consequently, cross-national differences may partly reflect variation in reporting rather than true differences in prevalence (Snowdon, 2018; Schouler-Ocak & Khan, 2024).

Questions related to suicidality may also be subject to stigma-related non-response. We therefore report the number of ‘Refusal’ and ‘Don’t know’ responses to the dependent-variable item mh004_ in the full SHARE Wave 8 sample (Appendix 3). Refusals were rare and showed no pronounced pattern by age, gender, education, or country, suggesting that non-response to the dependent variable is unlikely to be strongly systematic. Nevertheless, this limitation should be considered when interpreting prevalence estimates.

In addition, country-specific sample size in the analytic sample do not reflect national population sizes due to SHARE’s sampling design, country-specific fieldwork constraints, and the complete-case requirements of this study. The gender composition also varies across countries. Therefore, unweighted descriptive estimates should be interpreted as reflecting the analytic sample rather than nationally representative population proportions.

Finally, the scope of the current research restricts a more in-depth exploration of the underlying factors contributing to the findings. Due to constraints, our analysis could not examine potential causal mechanisms or broader contextual influences. Although we compared Nordic and Visegrad countries as regional groups, preliminary country-level analyses suggest heterogeneity in effect sizes, particularly for education and cohabitation. Some country-by-gender strata included few events, limiting precision and precluding definitive country-level comparisons. Despite these limitations, the results provide a solid empirical basis for future research employing longitudinal designs, richer measurement instruments, and more granular contextual data.