The Stroke Triage Gap: Solving the Discordance in Emergency Care

One Disease, Two Severity Scores
Stroke triage in the United States operates at scale, and the core challenge is balancing volume with precision. In 2022, emergency departments recorded about 686,000 visits with stroke as the primary diagnosis (2022 annual count, primary-diagnosis visit basis). Even small mismatches between ambulance and hospital severity scores can therefore affect a large treatment queue. By 2024, the national stroke death rate was 49.1 per 100,000 (2024 annual mortality rate basis), underscoring that score discordance is a system-level risk, not a clerical issue. In practice, the first assessment can shape destination decisions, treatment timing, and survival probability.
That risk is concentrated at the front end of care delivery. Data spanning more than 500,000 patients across nearly 700 hospitals shows that 60% of stroke cases arrive via emergency medical services (EMS) (multi-center patient cohort basis). Prehospital judgment is therefore the entry point for most stroke pathways. In response, regions are adopting real-time dashboards to track gaps between field assessments and final emergency department (ED) diagnoses. The emphasis shifts from individual performance to network integrity, treating triage discordance as an operational risk that must be systematically reduced.
The Environmental Constraint: Motion vs. Instrument
Differences between ambulance and ED scores are often a predictable product of different working environments. Most patients are first assessed in motion under prehospital constraints, then reassessed in a controlled hospital setting with broader diagnostic tools. Divergence is thus driven largely by information conditions and destination rules, not clinician intent. When accountability is assigned at the network level, policy can move from ad hoc fixes to standards designed for emergency transport conditions.
This divide sustains an active policy debate over speed versus data precision. One view prioritizes rapid transport to reduce protocol friction and preserve patient choice; another argues that standardized prehospital data fields are a baseline public-safety requirement. A workable middle path is to standardize the core fields collected by paramedics and reconcile them immediately at arrival. Once discordance is measured as a recurring signal, systems can convert recurring disputes into quarterly protocol updates.
The Large-Vessel Precision Test
The clearest test is whether field scoring can identify large-vessel occlusion (LVO), where rapid procedural treatment is critical. Reported performance metrics indicate that structured EMS neurological severity scores reached an AUC of 0.727, compared with 0.769 for ED physician use (same diagnostic performance framework basis). The proximity suggests EMS scores can function as primary routing inputs rather than preliminary notes. A 2026 policy objective is to align destination logic with these scores so high-severity patients reach intervention-capable centers earlier.
National frameworks already recommend destination planning by rural, suburban, and urban geography. Geography-specific routing is now treated as mainstream policy design rather than a regulatory edge case. The larger barrier is governance: how scores are audited, interpreted, and tied to outcomes. A 12-month competency audit cycle linked to destination adherence and outcomes offers states a mechanism to supervise implementation without imposing rigid one-size-fits-all rules.
Measuring the Access Dividend
When destination rules and severity scoring are synchronized, treatment access can shift quickly. In large metropolitan settings that adopted severity-based destination policies, endovascular therapy (EVT) use increased from 4.8% to 13.6% (pre/post policy comparison basis). This change indicates that routing design is a high-leverage policy tool for expanding access to catheter-based stroke intervention. Reform benchmarks are therefore broadening from transport speed alone to therapy-conversion performance.
A common concern is that aggressive routing could delay thrombolysis, the use of clot-dissolving medication. Regional analyses, however, indicate that higher EVT access and stable thrombolysis timing can coexist when protocols are coherently designed. This reduces the need to frame policy as a zero-sum tradeoff between medication timing and procedural access. A practical rollout model includes a 90-day feasibility review to test institutional comparability and execution capacity before statewide expansion.
The Implementation Paradox and Judgment Standards
Even with stronger policy, full agreement between field and hospital scores remains unlikely. Current reliability estimates show moderate agreement, with a weighted kappa of 0.572 (inter-rater agreement statistic basis). Some discordance is therefore a persistent operating condition, especially in early presentations with subjective features. Effective systems do not assume perfect alignment; they build monthly discrepancy reviews and rapid-correction pathways into routine operations. The target is controlled disagreement followed by immediate feedback.
Policy Choice: Success and Failure Summary
Severity-based routing represents a structural choice between local hospital autonomy and centralized network efficiency. Success is defined by higher advanced-therapy use while maintaining standard door-to-needle performance. Failure appears as meaningful delays in initial treatment or congestion at specialized centers caused by weak field discrimination. Execution depends on three conditions:
- Regional Topology Mapping: Protocols should vary by urban density and transport interval to avoid risk misallocation in rural settings.
- Feedback Loop Integration: Monthly reconciliation between EMS scores and hospital outcomes should be required to improve field accuracy.
- Outcome-Linked Review: Destination adherence should be tracked with recovery indicators to confirm clinical benefit.
10-Day Judgment Criteria
To assess near-term routing impact, systems can monitor the following on a rolling basis:
- Therapy Conversion Rate: EVT utilization above 10% in the target population, triggered by 90% adherence to regional routing protocols.
- Thrombolysis Time Stability: Door-to-needle times maintained within 10% of historical baseline, triggered by transport intervals staying below a 15-minute delay threshold.
- Systemic Accuracy: Field-to-hospital severity agreement kappa above 0.55, triggered by standardized pre-notification checklist implementation.
In 2026, stroke-care performance increasingly depends on whether networks function as closed loops rather than linear handoffs. When EMS scoring, destination decisions, and treatment timelines are reconciled within one feedback cycle, the triage gap can become an instrument for continuous system improvement rather than a fixed constraint.
Sources & References
Cerebrovascular Disease or Stroke (FastStats)
CDC National Center for Health Statistics • Accessed 2026-04-18
U.S. stroke burden remains high in emergency and mortality data. The page reports national counts for ED visits, deaths, and prevalence.
View OriginalEmergency Medical Services Utilization for Acute Stroke Care: Analysis of the Paul Coverdell National Acute Stroke Program, 2014–2019
CDC / Paul Coverdell National Acute Stroke Program • Accessed 2026-04-18
Large multi-hospital CDC dataset shows EMS use patterns and pre-notification trends for stroke/TIA patients.
View OriginalStroke Systems of Care: Policy Evidence Assessment Reports
CDC Paul Coverdell National Acute Stroke Program • Accessed 2026-04-18
CDC classifies prehospital stroke policy interventions by evidence level and highlights EMS pre-notification and triage-routing as high-impact policy levers.
View OriginalPrehospital Comprehensive Stroke Center vs Primary Stroke Center Triage in Patients With Suspected Large Vessel Occlusion Stroke
JAMA Neurology (University of Chicago/Chicago EMS multi-center study) • Accessed 2026-04-18
Regional EMS routing policy for suspected LVO was associated with higher endovascular therapy use without harming thrombolysis timing.
View OriginalCan Prehospital Personnel Accurately Triage Patients for Large Vessel Occlusion Strokes?
Journal of Emergency Medicine • Accessed 2026-04-18
Field FAST-ED scoring by EMS showed comparable predictive performance to emergency physician FAST-ED scoring, directly informing EMS-vs-ED triage consistency.
View OriginalInter-rater reliability of the Korean Triage and Acuity Scale performed between the prehospital and hospital stages
Journal of The Korean Society of Emergency Medicine • Accessed 2026-04-18
Direct prehospital-vs-hospital KTAS agreement study found moderate reliability, with better agreement when symptoms were objective or matched.
View OriginalRecommendations for Regional Stroke Destination Plans in Rural, Suburban, and Urban Communities
American Heart Association/American Stroke Association consensus statement (Stroke journal) • Accessed 2026-04-18
Multisociety U.S. consensus provides formal destination-planning recommendations for EMS stroke routing across geographic settings.
View OriginalU.S. ED visits with stroke as primary diagnosis: 686,000
CDC NCHS (NHAMCS 2022 table) • Accessed 2026-04-18
U.S. ED visits with stroke as primary diagnosis recorded at 686,000 (2022)
View OriginalU.S. stroke deaths: 166,852 (49.1 per 100,000)
CDC NCHS / NVSS via CDC WONDER • Accessed 2026-04-18
U.S. stroke deaths recorded at 166,852 (49.1 per 100,000) (2024)
View OriginalStroke/TIA patients analyzed in CDC Coverdell EMS study: 500,829 patients from 682 hospitals
CDC Coverdell study • Accessed 2026-04-18
Stroke/TIA patients analyzed in CDC Coverdell EMS study recorded at 500,829 patients from 682 hospitals (2014-2019)
View OriginalEMS arrival share in acute stroke/TIA cohort: 60%
CDC Coverdell study • Accessed 2026-04-18
EMS arrival share in acute stroke/TIA cohort recorded at 60% (2014-2019)
View OriginalEVT rate among EMS-transported AIS (<=6h) after regional triage policy: 4.8% pre vs 13.6% post
JAMA Neurology multicenter Chicago study • Accessed 2026-04-18
EVT rate among EMS-transported AIS (<=6h) after regional triage policy recorded at 4.8% pre vs 13.6% post (2021)
View OriginalFAST-ED discrimination for LVO in field vs ED: AUC 0.727 (EMS) vs 0.769 (ED physician)
Journal of Emergency Medicine • Accessed 2026-04-18
FAST-ED discrimination for LVO in field vs ED recorded at AUC 0.727 (EMS) vs 0.769 (ED physician) (2020)
View OriginalPre-KTAS vs KTAS agreement: Weighted kappa 0.572 (95% CI 0.513-0.632)
Journal of The Korean Society of Emergency Medicine • Accessed 2026-04-18
Pre-KTAS vs KTAS agreement recorded at Weighted kappa 0.572 (95% CI 0.513-0.632) (2025)
View OriginalShyam Prabhakaran, Professor and Chair of Neurology
University of Chicago Medicine • Accessed 2026-04-18
Regional policies for stroke care are still in their nascence compared to cardiac or trauma care.
View OriginalMaria Roznik, RN, BSN, ED Clinical Instructor for Nursing Education
Cleveland Clinic Hillcrest ED • Accessed 2026-04-18
Using AHRQ's evidence-based ESI tool allowed us to manage patients in a quick and efficient way. [URL unavailable]
Tareq Kass-Hout, Assistant Professor of Neurology
University of Chicago Medicine • Accessed 2026-04-18
Implementation of a prehospital transport policy for comprehensive stroke center triage ... [was] significant, rapid, and sustained. [URL unavailable]
Study finds recent change in EMS transport policy could improve stroke outcomes
EurekAlert • Accessed 2021-08-09
University of Chicago release contextualizes the Chicago regional triage policy and its association with higher EVT use.
View OriginalRegional Stroke Triage by EMS Ups Endovascular Therapy Rate
HealthDay • Accessed 2021-08-20
Summarizes JAMA Neurology findings for clinicians, emphasizing rapid post-policy EVT increase.
View OriginalStudy: Improved treatment for stroke patients after EMS transport policy change
EMS1 • Accessed 2021-08-10
EMS-focused coverage highlighting operational implications of destination triage to comprehensive stroke centers.
View OriginalWhat do you think of this article?