Bayesian modeling framework for optimizing pre-hospital stroke triage decisions
Bibliographic record
Abstract
Ischemic stroke is responsible for significant morbidity and mortality in the United States and worldwide. Stroke treatment optimization requires emergency medical personnel to make rapid triage decisions concerning destination hospitals that may differ in their ability to provide highly time-sensitive pharmaceutical and surgical interventions. These decisions are particularly crucial in rural areas, where transport decisions can have a large impact on treatment times - often involving a trade-off between delay in pharmaceutical therapy or a delay in endovascular thrombectomy. In this work, we explore a Bayesian modeling framework to address this decision-making process, showing how these techniques may be used to fully account for diagnostic and therapeutic uncertainty. We demonstrate how these techniques can contextualize triage decision at a fine-grained spatial scale. We further show the application of this modeling approach in the US State of Iowa, using data from the Virtual International Stroke Trials Archive (VISTA), and describe potential next steps for improved triage. ABBREVIATION LVO: large vessel occlusion; non-LVO, non-large vessel occlusion; IVT: intravenous tissue plasminogen activator; EVT: endovascular thrombectomy; CSC: comprehensive stroke centers; PSC: primary stroke centers; DS: drip and ship; MS, mothership; EMS: Emergency Medical Service; BGLM: Bayesian Generalized Linear Model; BGAM: Bayesian Generalized Additive Model; BART: Bayesian Additive Regression Trees; VISTA: Virtual International Stroke Trials Archive; NIHSS: National Institute of Health Stroke Severity Scale; ASPECTS: Alberta Stroke Programme Early CT Score; mRS, modified Rankin score; ROCAUC: Area under the receiver operating characteristic curve; ELPD: Expected Log pointwise Predictive Density; SE: Standard Error; ICA: Internal Carotid Artery; M1: Middle Cerebral Artery segment 1; M2: Middle Cerebral Artery segment 2; TIA: Transient Ischemic Attack; Cr-I: Credible Intervals; LKW: Last Known Well.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.008 | 0.020 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.003 | 0.002 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.003 | 0.002 |
| Open science | 0.003 | 0.002 |
| Research integrity | 0.003 | 0.003 |
| Insufficient payload (model declined to judge) | 0.005 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".