SciK-Health: an open-data dashboard for the multidimensional evaluation of Italian academic health science centers
摘要
Academic Health Science Centers (AHSCs) integrate research, education, and clinical care within complex organizational settings, making performance assessment inherently multidimensional. Traditional evaluation approaches based on single-source bibliometric data provide only a partial representation of this complexity, failing to capture the full translational spectrum from basic research to clinical application and societal impact. This paper presents SciK-Health, an integrated knowledge management platform designed to transform heterogeneous research data into actionable insights for multiple stakeholders. The platform addresses three key challenges: fragmentation across data sources, institutional heterogeneity requiring standardized metrics, and differentiated user needs. Methodologically, it implements a transparent three-stage workflow integrating OpenAlex (publications), Dimensions (clinical trials, patents, grants, datasets), and Altmetric (social impact), using Research Organization Registry identifiers and DOI-based cross-linking. The system operationalizes over 140 standardized indicators across five dimensions: bibliometric impact, social engagement, industrial innovation, clinical research activity, and competitive funding success. Applied to 49 Italian public AHSCs, SciK-Health adopts a user-centered design serving distinct audiences: citizens accessing institutional expertise, clinical managers supporting strategic monitoring, and policy makers conducting portfolio-level analyses. Empirical results highlight substantial institutional heterogeneity and show that single-source bibliometric approaches provide a systematically incomplete assessment of performance. Relative to existing platforms (SciVal, InCites, CWTS Leiden Open Edition), SciK-Health contributes a fully open-data infrastructure, multi-output integration, and a citizen-oriented LLM-assisted search interface. While the methodological framework is transferable, cross-national application requires substantial adaptation of data sources, institutional taxonomies, and indicator systems.