C-STAT Stroke Scale for Prehospital LVO Detection

The Cincinnati Stroke Triage Assessment Tool, commonly called C-STAT, is a bedside screening scale designed for paramedics and emergency medical technicians to identify patients who may be having a large vessel occlusion stroke, the kind most likely to benefit from advanced clot-retrieval procedures available only at specialized hospitals. Its accuracy is moderate, with studies reporting sensitivity in the range of roughly 63 to 71 percent and specificity around 67 to 77 percent for detecting these severe strokes. That performance places it in the middle of the pack among the dozen or so prehospital stroke scales in use today, but its real-world value lies less in raw numbers and more in how it reshapes ambulance routing decisions and the clock that ticks against a patient’s brain.

What C-STAT Actually Measures

C-STAT was built on top of the familiar Cincinnati Prehospital Stroke Scale, which checks three things: facial droop, arm drift, and speech abnormality. C-STAT adds assessments of gaze deviation (whether the patient’s eyes are stuck looking to one side), level of consciousness, and the severity of weakness in the arm and leg. Each item is scored, and a total at or above a set threshold flags the patient as “C-STAT positive,” suggesting a possible large vessel occlusion. The idea is that a paramedic can run through the assessment in the back of an ambulance without imaging equipment or lab work, and within minutes have a rough signal about whether the stroke is the type that needs a comprehensive stroke center rather than the nearest emergency room.

Large vessel occlusions account for a minority of all strokes but cause a disproportionate share of death and disability. They involve a clot blocking one of the brain’s major arteries, and the primary treatment is endovascular thrombectomy, a catheter-based procedure performed by neurointerventional specialists. Because that procedure is only available at comprehensive stroke centers, getting the right patient to the right hospital quickly is the central problem C-STAT tries to solve.

Accuracy in Detecting Large Vessel Occlusions

The numbers on C-STAT’s diagnostic accuracy vary depending on the study population and the threshold used. In one study from a regional stroke system, C-STAT’s sensitivity for large vessel occlusion was about 71 percent and its specificity about 67 percent, though the confidence intervals were wide because the number of confirmed occlusions was small.¹ A separate study comparing a numerical version of the Cincinnati Stroke Scale against several severity tools found C-STAT’s sensitivity at about 63 percent and specificity around 77 percent, with an area under the curve of 0.73.²

What those numbers mean practically: C-STAT catches roughly two-thirds to three-quarters of the patients who actually have a large vessel occlusion. That is better than a coin flip, but it means a meaningful fraction of occlusion patients will screen negative and initially go to a primary stroke center that cannot perform thrombectomy. At the same time, a fair number of patients who screen positive turn out not to have an occlusion at all. In one study, only about 14 percent of patients flagged as C-STAT positive actually had a confirmed large vessel occlusion, and nearly half of the positives turned out to not even be having a stroke.³

How C-STAT Compares to Other Prehospital Scales

There is no shortage of competing tools. A large prospective study in the Netherlands had trained paramedics score patients on eight different scales simultaneously, then compared each scale’s ability to detect large vessel occlusions confirmed by imaging. The RACE scale came out on top with an area under the curve of 0.83, followed by G-FAST and CG-FAST at 0.80 each, LAMS and CPSS at 0.79, PASS at 0.76, C-STAT at 0.75, and FAST-PLUS at 0.72.⁴ In other words, C-STAT fell in the lower-middle range, meaningfully behind the best-performing tool but not dramatically worse than the cluster in the middle.

The differences between these tools are not enormous once you get past the top two or three. Most of them are working with similar clinical signs, and no prehospital scale has managed to achieve the kind of accuracy that would make the triage decision obvious. The honest picture is that all of these scales are imperfect field-level approximations, and the choice among them often comes down to local factors: which one the regional EMS system has trained on, how many items it contains (more items can be harder to remember under stress), and how well it fits into existing dispatch and transport protocols.

A study from Eastern Taiwan found that most scales, C-STAT included, have particular trouble with strokes in the posterior circulation, meaning the arteries that supply the back of the brain. Sensitivity for those occlusions dropped below 30 percent for several common tools.⁵ That is a blind spot worth knowing about: the classic signs these scales check, like gaze deviation and arm weakness, are more typical of anterior circulation strokes. Posterior circulation strokes can present with dizziness, visual changes, or coordination problems that these tools were not designed to catch.

The Overtriage Problem

One of the biggest real-world headaches with C-STAT and similar scales is overtriage: sending patients to a comprehensive stroke center who do not actually need that level of care. A simulation study estimated that when using C-STAT with a ten-minute transport threshold, about 79 percent of patients directly routed to a comprehensive center were overtriaged.⁶ That is a striking number. It means for every patient who genuinely needed to be there, roughly four others were sent unnecessarily.

Overtriage is not just a statistical nuisance. Comprehensive stroke centers have finite capacity. Flooding them with patients who do not need thrombectomy can strain emergency departments, delay imaging for the patients who actually have occlusions, and create downstream bottleneck problems. Another analysis found that C-STAT’s positive predictive value was relatively low while its negative predictive value was high, meaning the tool is better at ruling out large vessel occlusions than confirming them.⁷ In practical terms, a negative C-STAT result gives you reasonable confidence the patient does not have a major occlusion. A positive result is much less conclusive.

This asymmetry is by design, at least partly. In stroke triage, the consequences of missing an occlusion (the patient arrives at a hospital that cannot perform thrombectomy, loses precious time being transferred) are generally worse than the consequences of sending a non-occlusion patient to a comprehensive center. So the scales are calibrated to cast a wide net, accepting many false alarms to avoid missing treatable cases. Whether that trade-off is appropriate depends on local geography, hospital capacity, and how long the bypass adds to transport time.

What Happens When a Region Actually Adopts C-STAT

The most concrete evidence of C-STAT’s impact comes from Oahu, Hawaii, where a bypass protocol was implemented island-wide in 2019. Under the protocol, patients who scored C-STAT positive were driven past the nearest local hospital and taken directly to a comprehensive stroke center. The results were clear: the median time from symptom onset to arrival at a comprehensive center dropped dramatically, from about 140 minutes before the protocol to 68 minutes after. The time from onset to the actual clot-retrieval procedure fell from roughly 233 minutes to 189 minutes. And this was achieved without significantly delaying the administration of tPA, the clot-dissolving drug that can be given at any stroke-capable hospital.⁸

Those time savings matter because brain tissue dies rapidly during a stroke. Every minute of delay can mean the difference between recovering independent function and permanent disability. The Oahu data also showed a reduction in interfacility transfers, which are the slow, resource-heavy process of stabilizing a patient at one hospital, arranging an ambulance or helicopter, and moving them to a second hospital. Bypassing that step entirely by getting the right patients to the right place from the start is the whole rationale behind prehospital triage tools.

One modeling study using the same simulation that flagged overtriage also estimated that with C-STAT and a ten-minute transport threshold, about 58 percent of large vessel occlusion patients were directly transported to a comprehensive center, 54 percent received thrombectomy within six hours, and roughly 37 percent achieved a good functional outcome afterward.⁹ Those numbers are not spectacular in isolation, but the relevant comparison is what happens without any severity-based triage: more patients going to the nearest hospital regardless, more transfers, and longer delays to definitive treatment.

How Easy Is C-STAT for Paramedics to Use

A prehospital tool is only useful if the people in the field can actually perform it under real-world conditions. A pilot prospective evaluation had EMS providers use C-STAT without any formalized training specific to the tool, and found that its test characteristics were comparable to other published scales even under those conditions.¹⁰ The fact that paramedics could pick it up without dedicated instruction is a point in its favor, since many EMS systems struggle with the time and resources needed for in-depth training programs on new clinical tools.

That said, C-STAT is more complex than the simplest scales. Tools like LAMS or the basic Cincinnati Prehospital Stroke Scale involve just a few quick checks. C-STAT adds items like gaze assessment and level of consciousness, which require a bit more clinical judgment. In the chaotic, time-pressured environment of an ambulance, every additional assessment step introduces a chance for error or delay. Whether those extra items improve the tool’s accuracy enough to justify the added complexity is something EMS medical directors weigh when choosing which scale to adopt.

The Stroke Mimic Challenge

One of the underappreciated challenges for any prehospital stroke scale is distinguishing actual strokes from conditions that look like strokes but are not. Seizures, severe migraines, low blood sugar, inner ear disorders, and even extreme anxiety can produce symptoms that mimic a stroke. In the study that examined C-STAT’s regional impact, nearly half of all C-STAT-positive patients turned out not to be having a stroke at all.¹¹ Some of those mimics involved severe neurological signs that would trick any field-level tool, not just C-STAT.

This is not a flaw unique to C-STAT. All prehospital stroke scales share this vulnerability because they rely on clinical signs that are not exclusive to strokes. The only definitive way to confirm or rule out a stroke is brain imaging, which is not available in an ambulance. Some researchers are exploring mobile stroke units, which are ambulances equipped with CT scanners, as a way to bring imaging into the field. Those units are expensive and rare, so for the foreseeable future, clinical scales like C-STAT remain the main decision-support tool for paramedics.

Telestroke and Broader Network Integration

C-STAT does not exist in a vacuum. It is increasingly being folded into telestroke networks, where a neurologist at a remote hospital evaluates a patient via video link and helps guide triage decisions. One study within a large academic telestroke network retrospectively calculated C-STAT and six other large vessel occlusion scales using data collected during telestroke consultations, comparing their performance to the full National Institutes of Health Stroke Scale.¹² The broader takeaway from that line of research is that simplified scales like C-STAT can hold their own against more comprehensive assessments for the narrow question of whether a clot is blocking a large artery. The full stroke scale captures more clinical detail, but much of that detail is not needed if the only question is “should this patient be sent to a comprehensive stroke center right now?”

The integration of these scales into digital platforms also makes them easier to deploy consistently. Mobile apps that walk a provider through the scoring items in a structured sequence reduce the chance of skipping a step or misremembering the threshold. Research on a mobile application for the full NIHSS, a more complex stroke assessment, found that scoring via an app produced results that closely matched paper-based scoring with strong agreement between the two methods.¹³ If that reliability holds for simpler scales like C-STAT, app-based administration could standardize scoring across EMS systems and reduce variability between individual paramedics.

When Geography Changes the Calculation

Whether C-STAT makes a meaningful difference for patient outcomes depends heavily on where you live. In a dense urban area where a comprehensive stroke center is ten minutes away and a primary stroke center is five minutes away, the bypass decision has low stakes: you are adding a short amount of time even if the patient does not have an occlusion. In a rural setting where the comprehensive center is an hour further than the nearest hospital, a false positive means a patient who might have a hemorrhagic stroke or a non-stroke condition is spending an extra hour in an ambulance without definitive care. The same tool, the same accuracy numbers, and very different risk profiles.

Modeling work on regional stroke networks has shown that adopting triage strategies that account for these geographic realities can meaningfully increase the number of patients eligible for advanced treatment. One analysis of the Montreal stroke network estimated that an optimized transfer protocol could increase the proportion of patients eligible for intervention by about 12.5 percent while also reducing emergency department crowding.¹⁴ The key insight is that the triage tool’s accuracy is only one variable. Transport distances, hospital capacity, time of day, and traffic all interact with the scale’s sensitivity and specificity to determine whether a particular patient ends up better off or worse off under a bypass protocol.

Some EMS systems handle this by setting a maximum bypass time. If the comprehensive stroke center is more than a set number of minutes further than the nearest hospital, the patient goes to the closer facility regardless of the C-STAT score, with the plan to transfer later if imaging confirms an occlusion. That kind of nuanced protocol design is where the real work of stroke systems happens, and it is far more consequential than which specific scale a region picks.

Posterior Circulation and Other Blind Spots

As mentioned in the comparison section, strokes affecting the back of the brain are a known weakness for C-STAT and most of its competitors. Posterior circulation occlusions can cause vertigo, double vision, difficulty swallowing, and loss of coordination rather than the classic face-arm-speech deficits that prehospital scales are built around. The Taiwan study found that sensitivity for posterior circulation occlusions dropped precipitously across nearly all scales, with some detecting as few as 20 percent of cases.¹⁵

This is a genuine gap in the current prehospital triage approach. Posterior circulation strokes account for roughly one in five ischemic strokes, and they can be just as devastating as anterior circulation occlusions. A patient with a basilar artery occlusion who presents with dizziness and vomiting may not trigger any alarm on a C-STAT assessment, potentially delaying recognition by hours. Some newer scales are attempting to incorporate items that capture posterior circulation signs, but none has yet been widely adopted by EMS systems. For now, paramedic education about atypical stroke presentations remains the main safeguard, alongside clinical suspicion that goes beyond whatever any scoring tool says.