Several validated stroke severity tools help EMS providers identify strokes and gauge their seriousness in the field. The most widely recognized is the Cincinnati Prehospital Stroke Scale (CPSS), but it is far from the only one. Over a dozen prehospital stroke scales exist, and newer tools like RACE, FAST-ED, LAMS, and VAN go beyond simple stroke detection to estimate whether a patient has a large vessel occlusion (LVO), the most dangerous and time-sensitive type of stroke. Understanding what each tool measures and when to use it can directly affect where a patient is transported and how quickly they receive treatment.
CPSS and FAST: The Foundation
The Cincinnati Prehospital Stroke Scale and the Face-Arm-Speech-Time (FAST) test are the most commonly taught stroke detection tools in EMS education. They check the same three physical signs: facial droop, arm weakness, and speech disturbance. A positive finding on any one of these three items suggests a stroke is occurring. Because they use identical clinical items (CPSS simply adds a specific sentence for the patient to repeat), researchers often group them together in studies.
These tools are fast and easy to perform, but they have an important limitation. They tell you whether a stroke is likely happening. They do not tell you how severe the stroke is or whether a large artery in the brain is blocked. That distinction matters because LVO strokes require specialized treatment at a Comprehensive Stroke Center, not just the nearest Primary Stroke Center. This gap led to the development of severity scales.
Cincinnati Prehospital Stroke Severity Scale
The CPSSS is a separate tool from the basic CPSS. It scores three specific findings on a 0 to 4 point scale: conjugate gaze deviation (eyes forced to one side) earns 2 points, arm weakness earns 1 point, and abnormal responses to basic commands and questions earns 1 point. Higher scores suggest a more severe stroke and a greater likelihood of a large vessel occlusion. This scale was validated using items drawn from the National Institutes of Health Stroke Scale (NIHSS), the gold-standard severity measure used in hospitals.
RACE Scale
The Rapid Arterial oCclusion Evaluation scale was designed specifically to help EMS identify LVO strokes in the field. It checks five things: facial palsy (0 to 2 points), arm motor function (0 to 2), leg motor function (0 to 2), gaze deviation (0 to 1), and aphasia or agnosia (0 to 2). The total score ranges from 0 to 9.
A score of 5 or higher was originally identified as the optimal cutoff for flagging an LVO, but more recent meta-analyses suggest using a cutoff of 3 or 4 to catch more cases. At a cutoff of 3 or above, the RACE scale picks up about 86% of LVO strokes, though it also flags some non-LVO patients. At a cutoff of 4, sensitivity drops to around 78% while specificity rises to 68%. In a large network meta-analysis across 22 studies and over 25,000 patients, RACE showed a pooled sensitivity of 71% and specificity of 73% for detecting LVO, making it one of the better-performing field tools.
FAST-ED
The Field Assessment Stroke Triage for Emergency Destination scale also draws from NIHSS items and scores five areas: speech changes (0 to 2), eye deviation (0 to 2), facial palsy (0 to 2), arm weakness (0 to 2), and denial or neglect of one side of the body (0 to 2). The total ranges from 0 to 10, and the score maps to a probability of LVO. A score of 1 or below corresponds to less than a 15% chance of LVO. A score of 2 to 3 suggests roughly a 30% chance. A score of 4 or higher puts the probability at 60% to 85%.
Across 18 studies involving over 21,000 patients, FAST-ED showed a pooled sensitivity of 63% and the highest specificity of the major severity scales at 80%. That means it is somewhat more conservative: it misses a few more LVO cases but produces fewer false alarms.
Los Angeles Motor Scale (LAMS)
LAMS is one of the simplest severity tools. It scores just three motor findings: facial droop (0 or 1), arm strength (0, 1, or 2 depending on whether the arm drifts, falls rapidly, or stays up), and grip strength (0, 1, or 2). The total ranges from 0 to 5. Higher scores point to more severe strokes. Pooled data from 13 studies show a sensitivity of 63% and specificity of 75% for LVO detection. Its simplicity makes it quick to perform, though it sacrifices some diagnostic power by skipping gaze deviation and language assessment, two findings that strongly correlate with large vessel blockages.
VAN Assessment
The Vision, Aphasia, Neglect (VAN) screen takes a two-step approach. First, the provider checks for motor weakness by asking the patient to raise both arms and hold them for 10 seconds. If there is no drift, weakness, or paralysis, the patient is VAN negative and the assessment stops. If weakness is present, the provider then checks for any of three additional findings: visual field problems, language difficulties (aphasia), or neglect of one side of the body. A patient who has weakness plus at least one of those three findings is considered VAN positive, suggesting a possible large vessel occlusion.
VAN was originally developed as an emergency department tool but has been piloted for prehospital use. Its two-step design means many patients can be screened out quickly at the first step, saving time on scene.
Other Detection Scales
Beyond these severity-focused tools, several other scales exist primarily for stroke detection rather than severity grading. The Los Angeles Prehospital Stroke Screen (LAPSS) checks facial droop, arm strength, and grip strength but also incorporates screening criteria like blood sugar level and medical history. The Melbourne Ambulance Stroke Screen (MASS) adds grip strength to the standard facial, arm, and speech checks. The Ontario Prehospital Stroke Screen (OPSS) includes leg motor function. The MedPACS scale adds gaze deviation and leg assessment. The shortened NIHSS for EMS (sNIHSS-EMS) is the most comprehensive field adaptation, scoring nine areas including consciousness, sensory deficits, and gaze on a multi-point scale.
How These Tools Affect Transport Decisions
The practical purpose of severity scales is triage. A basic detection tool like CPSS or FAST tells paramedics that a stroke is occurring and that rapid transport and hospital prenotification are needed. A severity tool like RACE, FAST-ED, or LAMS goes further by estimating whether the stroke involves a large vessel occlusion. When severity scores cross certain thresholds, EMS protocols in many regions direct the crew to bypass the nearest Primary Stroke Center and transport directly to a Comprehensive Stroke Center, where clot-retrieval procedures are available.
These thresholds vary by region. Some systems use a RACE score of 5 or above, others use FAST-ED of 4 or above, and some use LAMS of 4 or above. Local EMS protocols determine which scale is adopted and what cutoff triggers a longer transport. The American Heart Association’s severity-based stroke triage algorithm recommends that EMS systems incorporate a validated severity tool, though it does not mandate one specific scale over another.
Time on Scene
Every minute matters during a stroke. Median total prehospital time for ischemic stroke patients is about 39 minutes, with roughly 15 of those minutes spent on scene. That on-scene window, about 38.5% of total prehospital time, is where stroke screening happens alongside other assessments and stabilization. Severity scales are designed to fit within that window without adding significant delay. Most can be completed in under two minutes by a trained provider. Despite this, data show that the basic CPSS is documented in only about 65% of stroke patients, and hospital prenotification happens only about 52% of the time, highlighting gaps in consistent use.
Comparing Accuracy Across Scales
No single prehospital scale perfectly identifies every LVO. A 2024 network meta-analysis in the journal Neurology compared the major tools head to head. RACE had the highest sensitivity at 71%, meaning it caught the most LVO cases. FAST-ED had the highest specificity at 80%, meaning it was best at ruling out patients who did not have an LVO. LAMS fell slightly behind both, with 63% sensitivity and 75% specificity. All three performed in a similar overall range, and none reached the accuracy levels of hospital-based imaging.
The tradeoff between sensitivity and specificity is real. A more sensitive tool will flag more true LVO cases but also send more non-LVO patients on longer transports to Comprehensive Stroke Centers. A more specific tool will reduce unnecessary bypasses but risk missing some patients who needed the advanced center. Most EMS medical directors choose the scale that best fits their regional geography, transport times, and available stroke centers.