A consensus-based evaluation framework for telehealth in resource-constrained settings: evidence from a community Delphi study in Colombia.

Introduction: Telehealth has expanded rapidly as a strategy to improve access to health care for underserved populations, yet robust, equity-focused evaluation frameworks remain limited in low-resource and displacement-affected settings. In Colombia, rural and peri-urban communities such as Granizal face persistent geographic, financial, and infrastructural barriers to care despite formal insurance coverage. Community-informed approaches are needed to ensure that telehealth programs are responsive, effective, and sustainable. Methods: This study employed a three-round Delphi methodology to develop a consensus-based evaluation framework for a community telehealth program in Granizal, Antioquia. Fifteen panelists representing academic researchers, community health workers (promotoras de salud), and municipal decision-makers participated across all rounds. The panelists were selected and included according to their experience and expertise with telehealth services and also based on their community involvement specifically in Granizal. Phase 1 used open-ended questionnaires to elicit relevant data elements, barriers, engagement strategies, and access solutions. Responses were analyzed using thematic analysis. In Phase 2, panelists rated identified items using a five-point Likert scale. Phase 3 prioritized items meeting predefined consensus thresholds. Ethical approval was obtained from institutional review boards in Colombia and the United States. Results: Consensus was achieved across seven evaluation domains: pre-visit data, post-visit data, follow-up data, health post resources, barriers, community engagement strategies, and access solutions. Panelists prioritized comprehensive clinical and sociodemographic data, continuity of care indicators, patient experience measures, and follow-up outcomes related to adherence and quality of life. Key barriers included limited connectivity, insufficient infrastructure, fragmented insurance coverage, workforce instability, and lack of political commitment. CHWs were identified as central to patient engagement, digital literacy support, and program sustainability. The resulting framework integrates clinical, social, infrastructural, and behavioral indicators aligned with local realities. Conclusions: This study demonstrates the value of participatory, consensus-based approaches for designing telehealth evaluation frameworks in underserved settings. By embedding community health workers and local stakeholders in the evaluation process, the framework advances equity, contextual relevance, and implementation feasibility. The Granizal model offers a replicable approach for strengthening telehealth evaluation and decision-making in similar low-resource and displacement-affected communities.