Learned symptom clusters track heart-failure emergency outcomes
Source check: Source checked. Not independent scientific confirmation.
Finding: Fitted k-means partitioned nine recorded symptoms in438 emergency patients with prior heart failure into three groups. The indigestion group had adjusted odds ratios of1.8 for both acute heart-failure exacerbation and30-day adverse cardiac events compared with the group lacking indigestion and dyspnea. These are exploratory same-cohort associations, not a validated triage intervention.
Technology: K-means clustering
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Fitted k-means learns symptom-based patient partitions; the resulting assignments enable clinical outcome comparisons. PCA performed poorly and was not the retained partition.
- Test and result
- A secondary analysis of prospectively collected emergency presentations used438 patients and nine binary symptoms. Silhouette analysis selected three clusters of178,100 and160 patients. Two reviewers adjudicated clinical outcomes; regression adjusted for selected baseline variables.
- Limitation / still unknown
- The same selected cohort was used to discover and test clusters, with univariate-informed adjustment. Confidence intervals for the reported odds ratios round to1.0; no independent replication or causal interpretation is established. Outcomes include a heterogeneous composite; clustering is not proof of distinct disease mechanisms or improved emergency decisions.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC10526960/ · https://doi.org/10.1016/j.hrtlng.2023.05.012
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 21 Nov 2026. Review scope: Original primary clinical/computational Methods, reference definitions, model evaluation and Results independently read; no operational laboratory sections.
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.