Where Ageing Meets Fragile Landscapes: Sustainability’s Shifting Grip on Healthy Life Expectancy

In the high, thin air of the Qinghai-Tibet Plateau, on the eroded slopes of the Loess Plateau, and across the erosion-prone hills of northeastern China, a quiet demographic experiment has been unfolding for decades. These are regions scarred by Kashin-Beck disease, a disabling osteoarthropathy once endemic across China’s ecologically fragile zones, and they are now ageing faster than much of the country. By 2020, people aged 60 and above made up 13.8 percent of the population on the Qinghai-Tibet Plateau, 20.6 percent on the Loess Plateau, and a striking 27.1 percent in the northeastern hilly areas, figures that may even understate reality because so many younger people have migrated to cities. A new study published in Environmental and Sustainability Indicators asks a question with profound consequences for hundreds of millions of older people worldwide: which dimensions of sustainability actually determine whether those extra years of life are lived in good health?

The research team, led by Lantian Zhang of the Institute of Geographic Sciences and Natural Resources Research with collaborators including Mark W. Rosenberg of Queen’s University, assembled an unusually ambitious dataset covering 1,376 counties across China, of which 379 are classified as Kashin-Beck disease endemic. Rather than treating the disease itself as the focus, the researchers used these endemic counties as a natural laboratory for studying healthy ageing under conditions of ecological fragility and long-term socioeconomic disadvantage. Their outcome measure was old-age healthy life expectancy at age 60, a metric endorsed by the World Health Organization that combines mortality data with the prevalence of good health to estimate how many years a 60-year-old can expect to live in reasonable health.

Calculating this measure for more than a thousand counties across fifteen years was itself a technical feat. The team drew on census microdata from 2005, 2010, and 2020 provided through the Peking University Research Data Center, applying the Sullivan method to construct life tables from age 60 onward. Self-rated health, in which individuals assess their own overall condition, served as the health indicator, with people classified as healthy or basically healthy counted among the healthy states. Because no nationally representative mortality survey was conducted in 2005, the researchers estimated that year’s mortality using the 2000 national survey, refined through small-area estimation models that incorporated county-level covariates such as education rates, ethnic composition, income, and population density, then scaled the results to match provincial statistics.

On the explanatory side, the researchers built a multidimensional sustainability indicator system aligned with the United Nations Sustainable Development Goals, spanning the classic three pillars of sustainable development. The ecological dimension included forest coverage, vegetation indices measured by satellite, soil conservation, air pollutants such as fine particulate matter and ozone, and carbon emissions. The economic dimension captured income, employment, industrial structure, and financial resources. The social dimension encompassed safe drinking water, sanitation, housing, road density, education, and critically, the density of health facilities, health technicians, and hospital beds. Each indicator was standardized, and dimension-specific indices were constructed using a hybrid weighting scheme that combined expert-driven analytic hierarchy process weights with data-driven entropy weights in a 70:30 ratio, a balance the team tested for robustness against alternative weightings.

The analytical engine of the study was a hierarchical Bayesian framework implemented through Integrated Nested Laplace Approximation, a computational approach that has revolutionized spatial statistics by making complex models tractable. The researchers started with a baseline model decomposing healthy life expectancy into structured and unstructured spatial effects, a common temporal trend, and region-specific deviations from that trend. They then layered on increasingly sophisticated specifications, culminating in a spatiotemporal varying coefficient model in which the regression coefficients themselves shift across both space and time, and a further extension allowing county-by-time interactions. Model selection relied on the deviance information criterion, the Watanabe-Akaike information criterion, and cross-validation, while a spatiotemporal variance partitioning index quantified how much of the variation in healthy life expectancy each sustainability dimension explained.

The headline finding is stark: social sustainability matters most, and it matters most precisely where societies are weakest. Across all counties, a one-unit increase in the standardized social sustainability index was associated with an 11.66-year difference in healthy life expectancy at 60, with a 95 percent credible interval of 10.85 to 12.48 years. Ecological and economic sustainability showed positive but far smaller associations of roughly 2.3 and 2.2 years respectively. Yet the averages conceal dramatic heterogeneity. In Kashin-Beck endemic counties, the social sustainability coefficient rose to 13.56 years, while economic sustainability showed no detectable association at all. In non-endemic counties, the pattern reversed: economic sustainability was positively linked to healthy ageing, and social effects were somewhat weaker. On the Qinghai-Tibet Plateau, endemic counties showed the strongest ecological associations anywhere in the study, at 6.15 years, suggesting that in harsh high-altitude environments the natural surroundings themselves become a meaningful determinant of healthy old age.

Perhaps the most striking result is temporal. When the researchers decomposed the variation in healthy life expectancy, temporal dynamics of the sustainability indicators explained more than 90 percent of the variance in both endemic and non-endemic areas, dwarfing spatial effects, with model residuals accounting for less than 5 percent. This dominance of change over time reflects the sweeping nationwide improvements in living conditions and healthcare that China experienced between 2005 and 2020, including the 2009 New Healthcare Reform that expanded insurance coverage and strengthened primary care. In the Kashin-Beck affected areas specifically, decades of government intervention, including selenium supplementation, population relocation, and ecological compensation, appear to have paid dividends as the cohorts affected by earlier epidemics reached old age. The study also documented a narrowing of regional health inequalities, with non-endemic counties averaging 15.96 healthy years at age 60 compared with 15.15 years in endemic counties, a significant but not overwhelming gap.

The geography of these associations shifted in revealing ways. In 2005, the strongest links between social sustainability and healthy ageing were concentrated on the Qinghai-Tibet Plateau, where geographic inaccessibility and weak infrastructure made clinics, sanitation, and schools the binding constraints on wellbeing. By 2020, those strongest associations had migrated toward more economically developed regions, a redistribution the authors interpret through the lens of evolving policy: as basic healthcare shortages were alleviated in poorer areas, the marginal importance of social conditions migrated to wealthier regions grappling with chronic disease burdens that demand long-term, integrated health and social care. Meanwhile, the influence of ecological sustainability grew steadily over the fifteen-year window, hinting that as other determinants improve, environmental quality becomes an increasingly relevant pillar of healthy ageing.

For policymakers, the message is that one size does not fit all. In regions with fragile socioeconomic foundations, investments in social infrastructure, healthcare accessibility, sanitation, and education offer the greatest returns for healthy ageing, while sustained economic development and environmental governance matter over the longer horizon. The authors are candid about limitations: self-rated health captures a subjective dimension of wellbeing, the Sullivan method relies on cross-sectional data and cannot disentangle birth cohort effects, and as an ecological analysis the study describes regional patterns rather than individual-level mechanisms. Still, by demonstrating that the sustainability-health relationship is neither fixed in place nor uniform across the landscape, the study delivers a methodological lesson as important as its substantive one. Static, aspatial analyses of healthy ageing, it suggests, are not merely incomplete; they can obscure the very dynamics that determine whether the world’s fastest-ageing regions, many of them its most environmentally fragile, can convert rising life expectancy into genuinely healthy longevity.

Subject of Research: Spatiotemporal associations between ecological, economic, and social sustainability and old-age healthy life expectancy in ecologically fragile regions of China

Article Title: Spatiotemporal effects of sustainability dimensions on old-age healthy life expectancy in ecologically fragile regions

Article References: Zhang, L., Gu, L., Yang, L., Li, H., Wang, L., & Rosenberg, M. W. (2026). Spatiotemporal effects of sustainability dimensions on old-age healthy life expectancy in ecologically fragile regions. Environmental and Sustainability Indicators, 32, Article 101546. https://doi.org/10.1016/j.indic.2026.101546

Image Credits: AI Generated

DOI: 10.1016/j.indic.2026.101546

Keywords: healthy life expectancy, population ageing, sustainability indicators, Kashin-Beck disease, Bayesian spatiotemporal modeling, Qinghai-Tibet Plateau, Loess Plateau, social sustainability, Sullivan method, spatial heterogeneity, Sustainable Development Goals, China