Böcker 2026 cites the Stavnichuk 2020 meta-analysis but never tests its pooled −0.8%/month rate — the two notes connect through an intermediate synthesis note, not a direct numeric check

space-medicineaerospace-physiologybone-densitybed-rest-analogmeta-analysisbridge-investigationembedding

A cosine-0.87 embedding pairing flagged claim-spaceflight-bmd-loss-lower-limbs-about-0-8-percent-per-month (the Stavnichuk et al. 2020 pooled meta-analysis) and claim-matched-pqct-comparison-finds-bone-loss-comparable-while-muscle-loss-under-runs (the Böcker et al. 2026 site- and instrument-matched pQCT study) as an unlinked pair. The connection is real, not an embedding false friend: Böcker's own reference list carries the Stavnichuk meta-analysis as entry 17, an exact match to the Stavnichuk note's source (author list, journal, volume, article number, and year all agree). The second paper's author team had read and cited the first.

The citation is load-light, however. It is attached to the general sentence "Although the applied musculoskeletal countermeasures have progressively improved over the past decades[15], musculoskeletal adaptations do still occur[16,17]" — Stavnichuk is cited as background support for "spaceflight bone loss still happens despite countermeasures," not as the source of any number Böcker's own analysis uses or tests against. Böcker's headline "bone loss was comparable" result comes from a within-study statistical test between its own 13 spaceflight subjects and 11 bed-rest subjects, scanned on cross-calibrated pQCT devices (a "two-sided, non-paired t-test … of the individual change scores"); Stavnichuk's pooled −0.8%/month figure never enters that test as a benchmark, prior, or sanity check.

So the two claim-notes do not connect by checking each other's numbers. They connect through the vault's own synthesis note claim-bed-rest-under-runs-spaceflight-bone-loss-depends-on-readout, which is the document that actually places Stavnichuk's spaceflight rate beside a bed-rest DXA rate (claim-bed-rest-bmd-loss-about-one-percent-per-month-hip-pelvis-heel) and flags the near-equality as a possible unmatched-population artifact — the artifact hypothesis that Böcker's matched study then tested with its own data. Unlike the vocabulary-only pairing in observation-hawks-jeffress-cosine-pairing-is-embedding-false-friend, this pair shares a real referent and a real citation, so the cosine is tracking a genuine shared topic. See also entity-mariya-stavnichuk and entity-jorn-rittweger.