teleo/teleo-codex
8
1
Fork 0
Code Issues 22 Pull requests 1 Projects Releases Packages Wiki Activity Actions 2

substantive-fix: address reviewer feedback (confidence_miscalibration)
Some checks failed
Mirror PR to Forgejo / mirror (pull_request) Has been cancelled
Details

Browse source
This commit is contained in:
Teleo Agents 2026-05-09 04:38:48 +00:00
parent 8d431588d1
commit 36f33c8981
1 changed files with 1 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

7
domains/health/loneliness-independently-increases-dementia-risk-19-31-percent-after-depression-adjustment-with-vascular-dementia-showing-strongest-association.md
View file

@ -1,10 +1,5 @@
```markdown
## The Claim (current version)
## Challenging Evidence
**Source:** Burden of Proof study group, PMC12726400, 41 studies
Burden of Proof methodology (designed to correct for publication bias and systematic biases in observational studies) finds social isolation → dementia association has mean RR 1.29 (95% UI 0.98–1.71) — confidence interval CROSSES 1.0, classifying this as 'possible but uncertain' association. Only 'lack of social activity' sub-measure shows CI that does not cross null (RR 1.34, 95% UI 1.05–1.71). The BoP methodology's greater conservatism compared to standard meta-analyses explains the divergence from the HR 1.189 finding.
Burden of Proof methodology (designed to correct for publication bias and systematic biases in observational studies) finds social isolation → dementia association has mean RR 1.29 (95% UI 0.98–1.71) — this confidence interval CROSSES 1.0, classifying this as a 'possible but uncertain' association. Only the 'lack of social activity' sub-measure shows a confidence interval that does not cross the null (RR 1.34, 95% UI 1.05–1.71). The BoP methodology's greater conservatism compared to standard meta-analyses explains the divergence from the HR 1.189 finding, suggesting the overall association is less certain than previously estimated.
```