Both table versions enable direct verification
The notice reproduces the complete corrected and old Table 3, allowing readers to inspect the affected cells rather than relying solely on a correction statement.
↳ Corrected and old Table 3
Limitation: This content is identified as Editorial / Commentary. Evaluation of this content might be partial, and should be treated as indicative.
Abstract not available — view at DOI
Key Claims
This evaluation did not establish enough distinct central claims to display a claim list.
Strengths · Limitations
Strengths
Affected column is identified without ambiguity
The opening paragraph names the event/risk-count column and describes its month-like rendering, giving readers a specific target for comparison.
↳ Opening paragraph
Numerical repair restores usable event counts
Replacing entries such as "29-Sep" with "9/29" makes the original event and at-risk counts interpretable.
↳ Corrected and old Table 3, very mild to mild AD dementia, CKD row
Limitations
Correction process and cause remain unspecified
The notice does not explain how the display error arose or how the replacement counts were validated.
↳ Opening paragraph; corrected Table 3
Unchanged estimates require manual table comparison
Although the paired tables retain the same hazard ratios, confidence intervals, and p-values, the prose does not explicitly state that these entries are unchanged.
↳ Opening paragraph; corrected and old Table 3
Interpretive consequences receive no explicit discussion
The correction does not discuss whether the erroneous event/risk displays affected prior interpretation of the prognostic table.
↳ Correction text surrounding Table 3
Nabu Commentary
The paired versions of Table 3 make the numerical repair directly inspectable and support a higher assessment of corrective execution than of scientific contribution. Restoring event/risk fractions improves usability but adds no independent prognostic evidence. The concise opening identifies the affected column, while dense table rendering and the absence of an explicit unchanged-estimates statement leave some verification work to the reader. The notice links the repair to the original article without explaining the error's cause, validation process, or consequences for earlier interpretation.
Check another paper
Author Note / Field Feedback
Nabu’s assessment, alongside the field’s view.
Are you an author of this paper?
Dimension breakdown
Study Quality
Sound3.5
Confidence mediumContribution
2.7
Restoring event/risk fractions makes the original table interpretable and constitutes an assessable corrective contribution. The score includes a 0.4-point deduction for providing no independent scientific advance.
Opening paragraph; corrected and old Table 3
Methodological Rigour
4.0
Complete corrected and old tables permit direct checking of the repair, while the retained footnote preserves the model description. The notice does not explain how the error arose or how the correction was validated.
Corrected and old Table 3, including model footnotes
Reporting
4.2
The notice identifies the affected column and presents a direct problem-to-remedy sequence. Dense table rendering and the absence of an explicit statement about unchanged estimates require readers to perform some comparison themselves.
Opening paragraph; corrected and old Table 3
"some numbers under column \"No. events/No. people at risk\" in Table 3 are shown as months"
Positioning
2.9
The notice establishes its relationship to the original article but does not explain consequences for prior interpretation of the event/risk counts. Broader literature positioning is not required for this corrective purpose and is not treated as a serious omission.
Opening paragraph; final correction statement
Lower confidence on Positioning — domain match limited.
Reliability
Caveats4 of 4 checks
Internal CoherenceNoted
The old Table 3 contains month-like strings where event/risk counts belong. The replacement table resolves that display defect, and the paired versions show unchanged hazard ratios, confidence intervals, and p-values.
Research ConductClean
Unanimous agreement supports no visible research-conduct concern in this correction notice. The item reports a table correction rather than a new human-participant study.
Flags: 0 declared / 5 total
Reference IntegrityClean
1 reference in manuscript 1 of 1 checkable references found in an index
Post-Publication RecordClean
No retraction notice found in Retraction Watch.
Last checked: 16 Oct 2026
Sources: Retraction Watch ✓
Readiness
Where this paper’s evidence sits on the path from initial observation to real-world use.
Not assessed
Not assessed — the paper does not report enough for a readiness position.
Readiness is not applicable because the notice supplies no new clinical evidence or implementation guidance. Its practical function is to restore readability of the original table.
Opening paragraph; corrected Table 3
AI-generated, human-governed. Something look off? Contact us to request a review.