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 (frontmatter_schema, 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 08:40:30 +00:00
parent d33c4df609
commit 0e4da00ca2
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 ```markdown
## The Claim (current version)
## Challenging Evidence ## Challenging Evidence
**Source:** Burden of Proof study group, PMC12726400, 41 studies from 1,225 screened **Source:** Burden of Proof study group, PMC12726400, 41 studies from 1,225 screened
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). This contrasts with standard meta-analyses, which have reported hazard ratios for social isolation and dementia ranging from 1.189 (after depression adjustment) to 1.31. The BoP methodology's greater conservatism (specifically correcting for biases that inflate observational estimates) explains the gap between methods, suggesting that the overall association between social isolation and dementia may be weaker or more uncertain than previously estimated by standard meta-analyses.
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). This contrasts with standard meta-analyses, which have reported hazard ratios for social isolation and dementia ranging from 1.189 (after depression adjustment) to 1.31. The BoP methodology's greater conservatism (specifically correcting for biases that inflate observational estimates) explains the gap between methods.
``` ```