Abstract
OBJECTIVE: To examine the association between community-based elderly care services (CECS) and outpatient service utilization among older adults in urban China. STUDY SETTING AND DESIGN: This observational cross-sectional study focused on CECS provided in City X, China, during 2022. The analysis examined how participation in CECS activities—health education, physical activities, cultural engagement, health screenings, and companionship—affected the probability and frequency of outpatient visits among adults aged 60 and above. DATA SOURCES AND ANALYTIC SAMPLE: Secondary data were obtained from City X’s Smart Elderly Care Information Service Platform, comprising health service utilization records linked to medical insurance reimbursement data (January 2019–April 2023). The final sample included 58,445 urban-registered, non-disabled older adults aged 60 and above with complete records. Instrumental variable methods and two-part models were used to address endogeneity. PRINCIPAL FINDINGS: Participation in CECS was associated with lower probability of any outpatient visit by 18.9% points (SE = 0.080, p = 0.018) and lower annual visits by 11.0 (p < 0.001). All five activity types were associated with lower outpatient probability, with the strongest association observed for companionship activities (β = -0.550, p = 0.018), followed by cultural/artistic (β = -0.406, p = 0.017), physical exercise (β = -0.383, p = 0.017), health screening (β = -0.237, p = 0.018), and health education (β = -0.224, p = 0.018). For visit frequency, publicity/education, health screening, physical exercise, and cultural activities showed significant reductions, whereas companionship showed no statistically significant association (β = -0.505, p = 0.645). Associations were stronger among younger-old (60–79), literate, and co-residing individuals, but the oldest-old (≥ 80) showed increased outpatient probability. CONCLUSIONS: Participation in CECS is associated with lower outpatient care utilization among urban older adults and appears consistent with patterns of healthy aging. However, heterogeneous associations suggest that service models may need further adaptation for the oldest-old, illiterate individuals, and those living alone. Future research should incorporate comprehensive health outcomes and longitudinal designs to explore underlying pathways and longer-term implications. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s13561-025-00712-w.