Preprint overview
This demonstration paper evaluates a simple interpretable triage model trained on synthetic rural clinic operations data. It highlights model cards, fairness checks, negative results, and reviewer-visible limitations for operational health research.
Preprint: Public immediately; not yet formally reviewed.
DOI: 10.5555/scienceorbit.demo.002
AI summary
Problem
Clinics need simple ways to predict wait times without hiding decisions in opaque models.
What was discovered
The sample model helps in some settings but performs poorly during staffing shortages.
Key findings
Summaries are generated to improve accessibility. Authors review and approve them; always consult the full paper for technical detail.
Research assistant
Answers are grounded in this paper's abstract and AI summary. Always verify against the full text.
Automatic checks for linked repositories, datasets, and notebooks.
Authors stake testable claims. Later, outcomes are scored for accountability.
No predictions registered yet.
26/100
Reproducibility not yet established
Score rewards linked code/data, verification signals, and independent replication — not popularity.
Pre-publish checks for statistical anomalies, template patterns, citation padding, and duplication signals.
Risk score: 4/100 · Passed
Synthetic-data disclosure present
Full text
This demonstration paper evaluates a simple interpretable triage model trained on synthetic rural clinic operations data. It highlights model cards, fairness checks, negative results, and reviewer-visible limitations for operational health research.
Cited
Structured, public peer review.
The synthetic-data disclosure is clear, and the failure mode under staffing shocks is exactly the kind of result worth publishing. A sensitivity table would make the limitations easier to scan.
Connected by shared topics, tags, and fields.
We combine open Sentinel imagery, field plots, and transparent uncertainty estimates to identify coastal wetlands with high carbon recovery potential. The sample workflow demonstrates how authors can attach datasets, code repositories, AI summaries, public review, and replication signals to a preprint.
Ada Mitchell, Ben Carver