Digital climate-health platforms are web, mobile or dashboard-based architectures that link climate, environmental, health and socio-economic data to make climate-related health risks visible and actionable. This study maps stakeholder collaboration patterns in international platforms, focusing on who contributes, how collaboration is organized, and how platform-specific features shape inclusion, decision-making and use across varied institutional contexts.

We employed a sequential mixed-method design combining internet-based scoping, survey responses and 22 semi-structured interviews across 20 projects. Survey data were analyzed descriptively, and interview material was coded inductively to construct ideal-type collaboration patterns. Findings were triangulated with project documents and platform information to identify stakeholder roles, collaboration mechanisms and case-level pattern assignments.

Digital climate-health platforms involved developers, governments, international agencies, researchers, industry partners, community organizations and end-users. Five ideal-types were identified: central coordination, hierarchical distribution, network-based distribution, iterative co-evolution, and co-leadership. These patterns were not mutually exclusive; each case was assigned to a dominant pattern while secondary features were recorded. Collaboration was shaped by authority over data access, indicators, models, interfaces, feedback interpretation, implementation and post-launch maintenance.

The findings show that collaboration in digital climate-health platforms is both a platform design and a governance issue. Centralized and hierarchical arrangements provide accountability, continuity and authoritative communication but limit end-user influence. Network-based, iterative, and co-leadership arrangements can widen participation and responsiveness, yet they do not automatically constitute co-production. The findings point to the importance of clarifying decision rights, embedding co-design, and evaluating who can influence data, models, interfaces and use.