**Background:** As populations age globally, understanding factors that influence functional disability in older adults is critical. Perceived neighborhood social cohesion—defined as trust, safety, and participation within a community—may protect against functional decline. However, evidence from low- and middle-income countries like Ghana is limited. This study examined the association between perceived neighborhood social cohesion and functional disability among older Ghanaians, and whether sex, physical activity, and multimorbidity moderate this relationship.
**Methods:** Data came from the WHO Study on global AGEing and adult health (SAGE) Ghana Wave 2 (2014–2015), a nationally representative cross-sectional survey. The analytic sample included 4,446 adults aged 50 years and older. Functional disability was measured using the 12-item WHO Disability Assessment Schedule 2.0 (WHODAS 2.0), scored 0–100; those scoring ≥90.18% were classified as having disability. Perceived neighborhood social cohesion was assessed via three domains: community-level participation (9 items, range 9–45, α=0.87), perceived trust (3 items, range 3–15, α=0.889), and perceived safety (2 items, range 2–10, α=0.847). These were combined into an overall cohesion score (range 14–70, α=0.839). Physical activity was categorized as engagement in any of walking, moderate, or vigorous activity (yes/no). Multimorbidity was defined as 0, 1, or ≥2 chronic conditions (stroke, hypertension, depression, diabetes, angina, arthritis, chronic lung disease, asthma, cataract, oral health). Covariates included age, sex, marital status, education, rural/urban residence, and self-rated health. Generalized logistic regression models estimated odds ratios (OR) and 95% confidence intervals (CI). Interaction terms tested moderation by sex, physical activity, and multimorbidity.
**Key Results:** Participants with functional disability had a mean age of approximately 74 years, with higher prevalence among females (64.5% vs. 35.5%), widowed (46.5%), rural dwellers (61.1%), those with bad self-rated health (68.2%), and those with ≥2 chronic conditions (45.0%). Older people with no functional disability had a higher mean community-level participation score (24.8 vs. 18.3, P<0.001). In unadjusted analysis, overall perceived neighborhood social cohesion was significantly associated with lower odds of functional disability (OR = 0.94, 95%CI: 0.93, 0.94; P<0.001). After adjusting for confounders, the association remained significant (aOR = 0.94, 95%CI: 0.93, 0.95; P<0.001). Among individual domains, community-level participation (aOR = 0.94, 95%CI: 0.94, 0.95; P<0.001) and perceived trust (aOR = 1.00, 95%CI: 0.99, 1.00; P<0.001) were significantly associated with lower functional disability, while perceived safety was not significant after adjustment. Moderation analyses showed that sex and multimorbidity did not significantly moderate the overall cohesion–disability association. However, physical activity moderated the association between community-level participation and functional disability: among physically active older adults, higher community participation was associated with 2% lower odds of disability (OR = 0.98, 95%CI: 0.98, 0.99; P<0.01).
**Clinical Implications:** This study provides evidence from a low- and middle-income country that perceived neighborhood social cohesion is associated with reduced functional disability in older adults, with physical activity enhancing the protective effect of community participation. The findings suggest that interventions to improve social cohesion—such as fostering trust, safety, and community engagement—may help maintain functional ability in aging populations. Promoting physical activity alongside community participation could further reduce disability risk. Policymakers should consider community-based programs that create supportive social environments and encourage active lifestyles. Healthcare providers should recognize the role of social factors in functional health and advocate for cohesive neighborhoods. Limitations include the cross-sectional design (causality cannot be inferred), potential unmeasured confounders, and data from 2014–2015, though relevance remains high given current aging trends.