Things I got wrong
If I publish my own medical data and ask people to use it, I have to publish the mistakes too. Every correction anyone has flagged is listed here, with what changed and who spotted it. Newest first.
Haemoglobin and red cell count were displaying a HIGH flag when both results are below the reference range. The underlying data was correct; one page template printed “HIGH” for any out-of-range value regardless of direction.
What changed: Flags now render HIGH or LOW correctly on every page, the card and the plain-text version. Reference ranges are shown beside each value so any flag can be checked at a glance. A validator now recomputes every flag from its own reference range at build time and stops the build on a mismatch, so this cannot recur silently.
Flagged by my integrative oncologist during a review of this site.
Archived dosing tables and laboratory assay results sat in the same document as current information, with no warning directly above them.
What changed: Every archived dosing table now carries a historical warning immediately above it, and the ex-vivo assay lists carry a separate warning that those agents were tested in a dish and are not medicines I take. The machine-readable data feed now has an explicit reading guide and a current_medicines block, so an AI agent reading it cannot mistake an archived or assay drug for something I am taking.
Flagged by my integrative oncologist during a review of this site.