Health inequities across the life course among people with Down syndrome in Brazil: a data-driven analysis using clustering and principal component methods

People with Down syndrome (PwDS) experience significant health inequities driven by social determinants and unequal access to care. However, data-driven approaches to identify multidimensional health profiles across the life course–particularly in middle-income countries–remain limited. This observational study surveyed 366 PwDS (5 months–59 years) in Brazil using a structured questionnaire. Given the mixed variable types, primary clustering used Gower’s dissimilarity coefficient with Partitioning Around Medoids (PAM), while K-means clustering on standardized variables was used as a sensitivity analysis and evaluated using the Elbow, Silhouette, Calinski–Harabasz, and Davies–Bouldin indices. Principal Component Analysis (PCA) identified latent dimensions. Associations were examined using chi-square tests and Cramér’s V, and Body Mass Index (BMI) was analyzed by developmental stage, using World Health Organization (WHO) BMI-for-age z-scores in participants under 19 years and a gamma-family generalized linear model of raw BMI in adults. A system-level structural indicator, predominant reliance on the public (Unified Health System, SUS) versus private health system, was also incorporated into the clustering model. Profiles reflect biological and developmental differences together with structural and social inequities observed across age groups. Educational attainment was low despite high caregiver education, and healthcare access relied heavily on private services. Overweight and obesity increased with age, affecting 57.2% of females aged 11–59 years. Among participants under 19 years, mean BMI-for-age z-score was + 0.83 (SD = 1.64), significantly above the WHO reference, with approximately 44% overweight or obese. Thyroid disorders (45.4%), visual impairment (43.3%), and congenital heart disease (40.3%) were prevalent. Only 29.8% of caregivers rated public healthcare as adequate. Clustering identified two age-related profiles: a younger, lower-burden group and an older group with higher multimorbidity and healthcare dissatisfaction. PCA revealed four dimensions; in adjusted Poisson regression, household income was not associated with comorbidity burden. Incorporating health-system reliance revealed a distinct adolescent subgroup with predominant reliance on the public system and reported the lowest household income, underscoring the structural patterning of health profiles. Health inequities among PwDS in Brazil reflect age-related differences together with structural inequities in healthcare access observed across age groups. Data-driven approaches can identify vulnerability profiles and support equity-oriented public health strategies.