AI Science Tracker
EN

Claims-based models estimate later inflammatory-bowel-disease care events

University of California, Los Angeles; Leiden University Medical Center; Cedars-Sinai Medical Center · 27 Apr 2022 · medicine

Discovery

Claims-based models estimate later inflammatory-bowel-disease care events

Source check: Source checked. Not independent scientific confirmation.

Finding: Random forests and LASSO models used one year of insurance claims to estimate subsequent-year hospitalizations and surgery in patients with inflammatory bowel disease. In a temporally shifted cohort of 69,165 patients, reported AUCs were 0.73 for hospitalization and 0.71 for surgery. The training cohort contained 72,178 patients. This retrospective utilization prediction does not establish prevention benefit; possible patient overlap between annual cohorts, claims-based outcomes and validation-set threshold selection limit interpretation.

Technology: Random forest; LASSO logistic regression

Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.

AI contribution
Forests and penalized logistic models learn associations between prior-year utilization features and next-year claims-defined care events, enabling measured retrospective outcome-prediction comparisons.
Test and result
2015 baseline claims/2016 events supplied training data, while 2016 baseline/2017 events supplied temporal evaluation. Outcomes were claims-defined IBD hospitalization and surgery; the model compared fitted predictions with later recorded events.
Limitation / still unknown
Independent-patient generalization, biological disease mechanisms and improved treatment decisions remain unestablished. The source does not explicitly establish nonoverlapping patients across annual cohorts. Thresholds were selected on validation predictions; calibration was imperfect. A linked editorial was authenticated as an Editorial/Comment from primary metadata, but its subscription-only body was not reviewed.
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/PMC9515047/

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 relevant cohort/timing, outcome definitions, model/threshold Methods, Results and limitations. Current linked Editorial/Comment metadata and publisher/author-repository access checked; editorial body and supplementary feature tables not read. 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.

Open in the full catalog for filters, charts and related discoveries. Browse medicine discoveries.

Related catalog entries

Links use documented relationships where available, plus shared catalog context. A shared topic or model family does not establish a scientific dependency.