CROSS-NATIONAL CONSISTENCY OF PHYSICAL ACTIVITY AND SLEEP ASSOCIATIONS WITH EPISODIC MEMORY: LONGITUDINAL EVIDENCE FROM FOUR AGING COHORTS
Author(s)
Linwei Liu, MSc1, Qiu Zhang, PhD2.
1School of Pharmacy, Guangdong Pharmaceutical University, Guangzhou, China, 2Guangdong Research Center for Health Economics and Health Promotion, Guangdong Pharmaceutical University, Guangzhou, China, China.
1School of Pharmacy, Guangdong Pharmaceutical University, Guangzhou, China, 2Guangdong Research Center for Health Economics and Health Promotion, Guangdong Pharmaceutical University, Guangzhou, China, China.
OBJECTIVES: To examine whether physical activity and sleep health show different cross-national consistency in their within-person associations with episodic memory among older adults, using harmonized longitudinal aging cohorts spanning diverse health-system contexts.
METHODS: We conducted a coordinated secondary analysis of CHARLS (China), ELSA (England), HRS (United States), and SHARE (27 European countries), including 219,251 adults aged ≥50 years and 727,035 person-wave observations from 2000-2021. Episodic memory was cohort-standardized verbal recall. Physical activity was harmonized as a frequency-intensity index. Sleep was assessed using the common sleep satisfaction item and cohort-specific multidimensional composites where available. Random-effects within-between models estimated within-person associations, with pooled estimates from meta-analysis. Robustness was evaluated using one-stage individual-participant models, Hartung-Knapp-Sidik-Jonkman pooling, item-response-theory sleep scoring, negative controls, specification curves, continuous-time cross-lagged models, and an extension including MHAS (Mexico).
RESULTS: Physical activity was positively associated with episodic memory in all four cohorts (pooled β=0.109; 95% CI, 0.056-0.163), remained robust across pooling approaches, and was directionally positive after adding MHAS (HKSJ β=0.033; p=0.021). Sleep-memory associations were less consistent: estimates were positive in three cohorts but negative in HRS (I²=84%), were borderline using the common item (HKSJ β=0.034), became null after adjustment for physical activity, and reversed direction across sleep measurement models. The exploratory healthcare-expenditure gradient was not robust to cohort inclusion.
CONCLUSIONS: In harmonized multinational aging cohorts, physical activity showed a more consistent within-person association with episodic memory than sleep. These findings support prioritizing physical activity as a candidate behavioral target for cognitive-aging strategies, while underscoring the need for harmonized multidimensional sleep measures before drawing cross-national conclusions about sleep and memory.
METHODS: We conducted a coordinated secondary analysis of CHARLS (China), ELSA (England), HRS (United States), and SHARE (27 European countries), including 219,251 adults aged ≥50 years and 727,035 person-wave observations from 2000-2021. Episodic memory was cohort-standardized verbal recall. Physical activity was harmonized as a frequency-intensity index. Sleep was assessed using the common sleep satisfaction item and cohort-specific multidimensional composites where available. Random-effects within-between models estimated within-person associations, with pooled estimates from meta-analysis. Robustness was evaluated using one-stage individual-participant models, Hartung-Knapp-Sidik-Jonkman pooling, item-response-theory sleep scoring, negative controls, specification curves, continuous-time cross-lagged models, and an extension including MHAS (Mexico).
RESULTS: Physical activity was positively associated with episodic memory in all four cohorts (pooled β=0.109; 95% CI, 0.056-0.163), remained robust across pooling approaches, and was directionally positive after adding MHAS (HKSJ β=0.033; p=0.021). Sleep-memory associations were less consistent: estimates were positive in three cohorts but negative in HRS (I²=84%), were borderline using the common item (HKSJ β=0.034), became null after adjustment for physical activity, and reversed direction across sleep measurement models. The exploratory healthcare-expenditure gradient was not robust to cohort inclusion.
CONCLUSIONS: In harmonized multinational aging cohorts, physical activity showed a more consistent within-person association with episodic memory than sleep. These findings support prioritizing physical activity as a candidate behavioral target for cognitive-aging strategies, while underscoring the need for harmonized multidimensional sleep measures before drawing cross-national conclusions about sleep and memory.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
EPH64
Topic
Epidemiology & Public Health
Topic Subcategory
Public Health
Disease
Cardiovascular Disorders (including MI, Stroke, Circulatory), Generics, Geriatrics, Mental Health (including addiction), Nutrition