Recovery after stroke varies widely. The concepts of frailty and physical resilience offer complementary frameworks: frailty denotes reduced physiological reserve and vulnerability, while physical resilience denotes capacity to maintain or regain function after a health stressor. The relationship between pre-stroke social resources and post-stroke adaptive capacity is unclear. Social participation—engagement in interpersonal and community activities—may influence recovery via behavioral, emotional, and practical pathways, but measurement after stroke risks conflating pre-stroke context with stroke-related limitations. This study evaluated whether retrospectively reported pre-stroke social participation was associated with current physical resilience across frailty levels in middle-aged and older stroke survivors and tested frailty category as a potential effect modifier.
Design and setting: This was a single-center cross-sectional analysis using enrollment data from a longitudinal project at The Central Hospital of Wuhan, China. Participants were consecutively recruited between February 1 and April 30, 2026, with ethics committee approval and written informed consent.
Participants: Inclusion criteria were imaging-confirmed stroke, age ≥45 years, clinical stability at least three months after stroke onset, and Mini-Mental State Examination score ≥24. Exclusions included severe aphasia preventing communication, Charlson Comorbidity Index ≥3, or severe pre-existing psychiatric disorder.
Measures:
Covariates prespecified for adjustment were age (mean-centered), sex, monthly income (categorical: low/medium/high), hypertension, and diabetes mellitus. Data on pre-stroke physical activity and functioning, stroke severity, acute treatment, rehabilitation dose, and community services were not comprehensively available and therefore were not included.
Statistical analysis: Descriptive statistics by frailty category used ANOVA and chi-square or exact tests. Pearson correlations quantified unadjusted associations. The primary analysis used multivariable linear regression with HC3 robust standard errors; PRIFOR score was the dependent variable and social participation the primary explanatory variable, adjusting for the prespecified covariates and frailty category. Interaction terms tested effect modification by frailty. Sensitivity analyses were restricted to participants aged ≥60 years, ≥65 years, and to a combined pre-frail/frail subgroup.
Sample characteristics: The analysis included 314 participants: 140 non-frail (44.6%), 149 pre-frail (47.5%), and 25 frail (8.0%). Mean age was 68.4 years (SD 10.0); 52.5% were male. Mean social participation score was 5.2 (SD 3.8) and mean PRIFOR score 64.3 (SD 10.9). Both scores declined across increasing frailty categories.
Correlation and adjusted association: In the full sample, pre-stroke social participation and current physical resilience were weakly positively correlated (r = 0.295, 95% CI 0.190 to 0.392; p < 0.001). Category-specific correlations were r = 0.224 (non-frail), r = 0.267 (pre-frail), and r = 0.245 (frail), with the frail estimate imprecise (95% CI −0.166 to 0.584).
In the multivariable model adjusted for age, sex, monthly income, hypertension, diabetes, and frailty category, each one-point higher social participation score was associated with a 0.725-point higher PRIFOR score (95% CI 0.405 to 1.044; p < 0.001). The model explained 21.2% of outcome variance (adjusted R2 = 0.188); maximum variance inflation factor was 1.36.
Effect modification and sensitivity: Interaction terms for social participation by frailty category were not statistically significant (joint p = 0.314), indicating no evidence that the association differed by frailty category; however, the frail subgroup was small (n = 25) and stratum estimates were imprecise. Age-restricted sensitivity analyses among participants aged ≥60 and ≥65 years produced consistent positive adjusted coefficients. The combined pre-frail/frail subgroup analysis also showed a positive adjusted association.
In this cross-sectional cohort of middle-aged and older stroke survivors, retrospectively reported pre-stroke social participation had a weak positive association with current physical resilience after adjustment for measured covariates and frailty category. The absence of a significant interaction suggests the association did not statistically differ across non-frail, pre-frail, and frail categories in this sample, but limited numbers in the frail subgroup reduce precision and limit inference.
Mechanisms that could underlie the observed association include greater pre-stroke mobility, cognitive engagement, health behaviors, or social support among those more socially active; conversely, better pre-stroke function could enable participation, producing confounding. The adjusted model explained a modest portion of variance in resilience, and the correlation magnitude was categorized as weak.
Key limitations reported by the authors include the cross-sectional design that precludes causal inference, retrospective recall of pre-stroke participation subject to bias, concurrent measurement of frailty and resilience with possible conceptual overlap, and unmeasured confounders such as pre-stroke function, stroke severity, treatments, rehabilitation intensity, depression, and community services. The single-center sample and exclusion criteria restrict generalizability, and the frail subgroup size (n = 25) limited precision for interaction analyses.
Retrospectively reported pre-stroke social participation was weakly and positively associated with current physical resilience among middle-aged and older stroke survivors after adjustment for measured covariates and frailty category. There was no statistical evidence of effect modification by frailty level, but the small frail subgroup yields imprecise estimates. The authors recommend prospective, multicenter studies incorporating comprehensive pre-stroke function, stroke severity, treatment, rehabilitation exposure, and community support before considering prognostic or interventional implications.