Context at the source and the European Health Data Space

Extended companion to the paper: Freriks G, Rossi Mori A. "Isolated Data Do Not Become Information: Preserving Context at the Source for the European Health Data Space". Medical Informatics Europe (MIE) 2027, submitted. (reference [13] of the paper). This document (32 pages) develops the paper chapter by chapter: information arises from ensembles of data, in the human mind and in the concept networks of AI; the three levels of context (micro, meso, macro); the four layers of context to be preserved at the source (problem-oriented profiles, the succession of care episodes, the chains of metamorphosis, the annotations of all actors); linking chains and episodes, with five link types (derived from, in reply to, read for, parallel to, and the proposed "suspended"); tolerance of imprecision and the role of AI; the principle of no extra burden for professionals, patients and receivers; and the implications for primary and secondary use in the EHDS, including the data quality and utility label. Related records: the compact supplementary material (doi:10.5281/zenodo.23926086, reference [6] of the paper) and the complete series with its annotated bibliography (doi:10.5281/zenodo.23926090, reference [14]). The documents are working texts meant to open a discussion, not to settle it. They were written by the authors with the support of Claude (Anthropic) through an approach of augmented reasoning: the sources used are made explicit, unverified statements are flagged, and the most significant statements carry a reliability marker ({T} full convergence of independent sources, {P} partial convergence, {L} insufficient or non-independent sources, {I} inference from general knowledge). The persons in the cases are constructed teaching cases, not real patients. Each document is provided as PDF and as editable DOCX.