Hypertension Hot Spots Shift Across China as Pollution and Clinic Access Emerge as Key Drivers

Hypertension among middle-aged and older adults in mainland China is not distributed evenly across the country, and its geographic pattern has changed markedly over the past decade. A new spatial analysis drawing on five waves of the China Health and Retirement Longitudinal Study (CHARLS) between 2011 and 2020 maps hypertension prevalence across 27 provinces and identifies the factors that best explain where the condition clusters. The study, published in BMC Public Health, finds that high-prevalence hot spots that once concentrated along the southeastern coast had dissipated by 2015, drifted westward, and largely vanished by 2020, leaving a more spatially homogeneous national picture.

The research team, led by Dawei Wei and colleagues, classified the 27 provinces into eastern, central, and western regions and calculated province-level hypertension prevalence from CHARLS data using adjusted sampling weights. To explain the observed geography, the researchers assembled eleven candidate factors spanning demographic, climatic, economic, and health-resource dimensions, drawn from the China Statistical Yearbook and the China Meteorological Yearbook. The analytical core of the study combines several spatial statistics techniques.

Global Moran’s I was used to test whether hypertension prevalence was clustered more than chance would allow, while Local Indicators of Spatial Association (LISA) cluster maps pinpointed where neighboring provinces shared similarly high or low rates. The Getis-Ord Gi* index, computed in ArcMap, flagged statistically significant hot spots and cold spots. Together these tools traced the movement of high-burden areas over the five survey waves, revealing a clear trajectory: concentration in the southeast during 2011 to 2013, complete dissipation of those hot spots by 2015, a westward shift in the interim, and near-uniform distribution across provinces by 2020.

To move beyond description and quantify why prevalence varies from province to province, the team applied GeoDetector, a method designed to measure the explanatory power of categorical or stratified variables over a spatially distributed outcome. The single-factor detector returned a statistic, q, expressing the share of variance in hypertension prevalence that each factor explains. Two variables dominated: industrial sulfur dioxide emissions, with q = 0.277, and the number of primary healthcare institutions, with q = 0.191.

In plain terms, provinces with heavier industrial air pollution and, separately, those with different densities of grassroots clinics showed systematically different hypertension burdens, and these two factors outperformed the nine other candidates. The single-factor results were not stable across time or space. The analysis revealed significant region-year heterogeneity, meaning that the factors explaining hypertension patterns in the eastern region in one survey wave did not necessarily carry the same weight in the central or western regions, or in later years.

This heterogeneity matters for policy: a national one-size-fits-all interpretation of hypertension risk factors would obscure the regional differences the data reveal. Perhaps the most striking finding came from the interaction detector. When pairs of factors were evaluated jointly, their combined explanatory power frequently exceeded the sum of their individual contributions, a signature of nonlinear synergistic effects.

The interaction between primary healthcare institutions and industrial SO2 emissions was particularly strong, reaching an interaction q of 0.51. This suggests that the geography of hypertension in China’s aging population is shaped not by isolated drivers but by the conjunction of environmental pollution exposure and the local capacity of the primary care system to detect and manage elevated blood pressure. Because GeoDetector results can depend on how continuous variables are discretized into strata, the authors ran a sensitivity analysis comparing an enhanced variant, the Optimal Parameters-based Geographical Detector (OPGD), against a fixed 5-quantile benchmark. The core factor rankings held up under both approaches: industrial SO2 emissions and primary healthcare institutions consistently occupied the top two positions across specifications.

This robustness check strengthens confidence that the identified drivers are not artifacts of an arbitrary classification scheme. The study’s conclusions carry direct policy weight. The authors argue that targeted interventions are urgently needed for economically underdeveloped and environmentally polluted provinces, where regional disparities in hypertension detection and management appear most acute.

Several caveats frame the findings. The analysis is ecological, operating at the province level rather than the individual level, so the identified associations describe places rather than proven causal pathways for any single patient. The CHARLS waves analyzed run from 2011 to 2020, and the published version notes it was shared early as accepted peer-reviewed research subject to final editorial edits.

Subject of Research: Provincial spatial distribution and driving factors of hypertension among adults aged 45 and older in mainland China

Article Title: Provincial distribution of hypertension among adults aged 45 and older in mainland China: a GeoDetector analysis based on CHARLS 2011–2020

Article References: Wei, D., Hu, Y., Li, M., Sun, Y., & Wu, J. (2026). Provincial distribution of hypertension among adults aged 45 and older in mainland China: a GeoDetector analysis based on CHARLS 2011–2020. BMC Public Health. https://doi.org/10.1186/s12889-026-29982-7

Image Credits: AI Generated

DOI: 10.1186/s12889-026-29982-7

Keywords: hypertension, China, CHARLS, GeoDetector, spatial analysis, air pollution, primary healthcare, SO2 emissions, health disparities, aging population, BMC Public Health, provincial distribution