Saved observations will appear here in the next integration step.
Temporal association only; not causation. Missing entries are not treated as absence.
The analysis design remains as specified: latency windows, repeated associations, exposure combinations, symptom clusters, contradictory observations and confounding/alternative explanations.