Clinical research requires reliable data, which remains out of reach despite decades of efforts. The proliferation of technical and terminological standards has fragmented healthcare data, improving interoperability at a structural level but often diluting clinical meaning and context by prioritizing storage and exchange over actual clinical meaning. This paper proposes a different approach, linking raw data to explicit semantic expressions built from existing terminologies using a formal descriptive grammar. Implemented through knowledge graphs, this approach preserves the original clinical context of the data. By treating data as an expression of reality rather than an object to be normalized, this method enables precise queries and supports exploratory research, without replacing current standards but combining them more intelligently.