Surgeons’ risk judgments diverge from forest-model estimates
Source check: Source checked. Not independent scientific confirmation.
Finding: Thirty-nine spine surgeons estimated complication and reoperation risks for nine adult-spinal-deformity case vignettes. Their judgments varied widely and had low agreement with estimates from an existing random-survival-forest calculator. This measured disagreement does not show that the calculator was more accurate for these cases: actual patient outcomes were not the comparator. A 39% response rate, selected vignettes and inconsistent directional wording in the paper limit interpretation.
Technology: Random survival forest surgical-risk calculator
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- A previously trained random survival forest generates case-specific risk estimates that provide the learned comparator for measuring agreement with human judgments.
- Test and result
- Comparison of 39 surgeons’ elicited risk estimates for nine fixed case vignettes against an existing random-survival-forest calculator. Mean indices of agreement with the model ranged .122–.416 across complication/reintervention endpoints and follow-up horizons.
- Limitation / still unknown
- Accuracy versus actual outcomes, improved counseling and patient benefit were not established by this comparison. Group-average differences were not statistically significant; low individual agreement should not be recast as proven machine superiority. One Results sentence reverses the temporal direction shown in tables and discussion, so directional over/underestimation claims are omitted.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://dukespace.lib.duke.edu/server/api/core/bitstreams/bbceca7d-5fa4-4eea-9643-1a5e191cb3fa/content
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: Primary Duke author-deposited journal PDF front matter, relevant calculator/survey Methods, Results/Tables1–3 and discussion/limitations. Publisher canonical identity and journal online date corroborated. Exact primary identity/date and fresh primary PubMed correction/retraction relationships checked..
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.