Association between knowledge, attitude, and practice regarding family doctor contract services and perceived healthcare barriers among contracted residents: a cross-sectional study in Shanghai, China

This study aimed to assess the level of knowledge, attitude, and practice (KAP) regarding family doctor services among contracted residents in Shanghai, examine its correlation with perceived healthcare barriers, and identify factors influencing KAP. This cross-sectional study surveyed 304 contracted residents of family doctors in Minhang District, Shanghai, between January and April 2026. The questionnaire, designed based on the Andersen Behavioral Model of Health Services Use, assessed KAP and six dimensions of perceived healthcare barriers: imperceptibility, unavailability, inaccessibility, inadaptability, unaffordability, and unacceptability. We used one-way ANOVA, Spearman’s correlation, and multivariate linear regression to identify factors associated with KAP. Propensity score matching (1:1 nearest-neighbor) was applied to control for confounders and estimate the independent association between perceived healthcare barriers and KAP as well as service use behaviors. The mean total KAP score was 39.37 ( SD 7.99). Mean scores for knowledge, attitude, and practice were 13.34 ( SD 2.74), 13.05 ( SD 2.79), and 12.97 ( SD 3.00), respectively, indicating moderate-to-high levels. Univariate analysis showed significant differences in KAP by age, education, household registration, visit frequency, chronic disease status, source of medical expenditure, source of medical care, and level of perceived barriers (all P < 0.05). Spearman correlation showed that total perceived barriers and all subscales were positively correlated with total KAP and each KAP domain (all P < 0.05), with the strongest correlation for the practice domain ( r s =0.298, P < 0.001). Multivariate linear regression identified age ≥ 35 years, education above high school, urban resident basic medical insurance, and self-funded healthcare as factors correlated with higher KAP (all P < 0.05; adjusted R ²=0.249, F = 7.694, P < 0.001). Notably, multivariate linear regression failed to detect a significant association between total perceived healthcare barriers and overall KAP. After propensity score matching to balance baseline confounders, total perceived barriers and the inadaptability dimension were positively associated with KAP ( ATT = 0.281 and 0.174, both P < 0.05). Higher scores for total barriers, imperceptibility, unavailability, and inadaptability corresponded to higher practice scores (all P < 0.05). Contracted residents have moderate-to-good KAP regarding family doctor services, which is associated with age, education, and insurance type. Within this specific cohort of actively-utilizing contracted residents, perceived healthcare barriers demonstrated a statistically positive, non-causal association with KAP. Information, resource, and process-related barriers showed the strongest correlations with active service use, suggesting they may be potential leverage points for quality improvement, pending validation in longitudinal designs. The absence of association for traditional barriers (geographic, financial, psychological) should be interpreted cautiously, as it may reflect the unique primary care environment of urban Shanghai rather than universal trends. Notably, all associations are statistical, not causal; the cross-sectional design cannot determine causal direction, and further longitudinal studies are required for verification.