Real-world implementation of a clinical decision support system supporting community health workers in hypertension and diabetes care in vulnerable Brazilian municipalities

Background Community health workers (CHWs) play a central role in the longitudinal management of patients with hypertension and diabetes, particularly in socioeconomically disadvantaged settings. Clinical decision support systems (CDSS) may enhance CHWs’ ability to detect early signs of clinical deterioration and facilitate timely referrals within primary healthcare networks. Purpose To describe the development, implementation, and real-world use of a CDSS embedded in a multifaceted digital health intervention designed to support CHWs during home visits to patients with hypertension and diabetes. Methods This mixed-methods implementation study was conducted as part of the Control of Hypertension and Diabetes in Minas Gerais (CHARMING) Project in 18 primary care centers across five socioeconomically vulnerable municipalities in Minas Gerais, Brazil, guided by the RE-AIM framework. During home visits, CHWs used a tablet-based CDSS integrating automated blood pressure measurement and urine glucose testing via reagent strips, generating alerts and care recommendations based on predefined thresholds. The intervention also included structured training and ongoing tele-tutoring to support clinical assessment, lifestyle counselling, and tool utilization. Quantitative data were analyzed descriptively, and qualitative data were collected through focus groups to explore adoption, usability, and implementation determinants. Results Between March 2024 and January 2025, 984 CDSS-supported assessments were conducted among 680 patients. Among 648 patients with hypertension, 26.2% had inadequate blood pressure control. Among 242 patients with diabetes, 33.9% reported recent blood glucose levels above 250 mg/dL, 31.4% reported hyperglycemic symptoms, and 24.8% had positive urine glucose tests. The intervention demonstrated both opportunities and challenges in integrating digital decision support into community-based care. Qualitative findings indicated that digital tools, institutional support, and tele-tutoring facilitated work reorganization, professional empowerment, and improved communication with patients and primary care teams. Key barriers included limited internet connectivity in rural areas, misalignment between CDSS recommendations and local service capacity, and cultural and educational constraints affecting lifestyle change. Conclusion The multifaceted CDSS-supported intervention was feasible in vulnerable primary care settings and strengthened CHWs’ capacity for early risk identification. It enabled the detection of uncontrolled hypertension and diabetes among patients with unmet needs. Sustainable impact will depend on workflow integration, interoperability, and continuous professional support, underscoring the importance of sociotechnical approaches in underserved contexts.