ACCESS IS NOT ENOUGH: DIGITAL COMMUNICATION AND THE SOCIOECONOMIC GRADIENT IN OLDER ADULTS' HEALTH ACROSS SIX HARMONISED COHORTS
Author(s)
Linwei Liu, MSc1, Qiu Zhang, PhD2.
1School of Pharmacy, Guangdong Pharmaceutical University, Guangzhou, China, China, 2Guangdong Research Center for Health Economics and Health Promotion, Guangdong Pharmaceutical University, Guangzhou, China, China.
1School of Pharmacy, Guangdong Pharmaceutical University, Guangzhou, China, China, 2Guangdong Research Center for Health Economics and Health Promotion, Guangdong Pharmaceutical University, Guangzhou, China, China.
OBJECTIVES: Digital communication is widely promoted to support healthy ageing, yet whether it narrows or widens the socioeconomic status (SES)-health gradient in later life—and how this varies across health systems—remains unclear, leaving a gap for digital-inclusion policy. We quantified the share of the SES-health gradient attributable to digital communication across six harmonised ageing cohorts, including 28 European countries, and tested its robustness.
METHODS: We analysed 348,458 person-waves from HRS (US), ELSA (England), SHARE (28 European countries), CHARLS (China), MHAS (Mexico), and KLoSA (Korea), harmonised via the Gateway to Global Aging Data. Composite SES and four-domain composite health indices were standardised within cohort. For the five cohorts with comparable internet-use measures, a linear mediation decomposition (1,000-replicate person-clustered bootstrap CIs) estimated the gradient share transmitted through digital communication. Robustness used Cinelli-Hazlett sensitivity analysis, E-values, counterfactual mediation, lagged-dependent-variable models, and a causal forest.
RESULTS: The per-SD SES-health gradient ranged from 0.19 (Korea) to 0.44 (US), roughly twice as steep in English-speaking cohorts. Digital communication accounted for 11-25% of the gradient across internet-comparable cohorts (HRS 18%, ELSA 11%, SHARE 25%, CHARLS 15%) and 6% for the bundled MHAS measure; E-values (1.7-2.2) indicated robustness to unmeasured confounding. Cross-national differences were driven primarily by the SES-to-adoption gradient (who gets online) rather than the health return to use, and pooled returns did not differ systematically by welfare regime. Lagged returns concentrated among higher-educated adults in HRS, ELSA, and MHAS.
CONCLUSIONS: Digital communication is a measurable, robust contributor to later-life health inequalities, but it equalises health chiefly at the access margin, and returns favour higher-educated adopters. For European health systems advancing the EU Digital Decade and healthy-ageing agendas, expanding connectivity alone is unlikely to narrow the gradient; equity gains require targeted digital-inclusion programmes that pair access with skills support for older, lower-SES adults.
METHODS: We analysed 348,458 person-waves from HRS (US), ELSA (England), SHARE (28 European countries), CHARLS (China), MHAS (Mexico), and KLoSA (Korea), harmonised via the Gateway to Global Aging Data. Composite SES and four-domain composite health indices were standardised within cohort. For the five cohorts with comparable internet-use measures, a linear mediation decomposition (1,000-replicate person-clustered bootstrap CIs) estimated the gradient share transmitted through digital communication. Robustness used Cinelli-Hazlett sensitivity analysis, E-values, counterfactual mediation, lagged-dependent-variable models, and a causal forest.
RESULTS: The per-SD SES-health gradient ranged from 0.19 (Korea) to 0.44 (US), roughly twice as steep in English-speaking cohorts. Digital communication accounted for 11-25% of the gradient across internet-comparable cohorts (HRS 18%, ELSA 11%, SHARE 25%, CHARLS 15%) and 6% for the bundled MHAS measure; E-values (1.7-2.2) indicated robustness to unmeasured confounding. Cross-national differences were driven primarily by the SES-to-adoption gradient (who gets online) rather than the health return to use, and pooled returns did not differ systematically by welfare regime. Lagged returns concentrated among higher-educated adults in HRS, ELSA, and MHAS.
CONCLUSIONS: Digital communication is a measurable, robust contributor to later-life health inequalities, but it equalises health chiefly at the access margin, and returns favour higher-educated adopters. For European health systems advancing the EU Digital Decade and healthy-ageing agendas, expanding connectivity alone is unlikely to narrow the gradient; equity gains require targeted digital-inclusion programmes that pair access with skills support for older, lower-SES adults.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
HPR58
Topic
Health Policy & Regulatory
Disease
Geriatrics