NIH Stroke Scale (NIHSS)
A structured examination that scores stroke severity from 0 to 42, at presentation and over time.
By Claire White
- Clinician-administered by a trained and certified examiner; under 10 minutes.
- 11 categories, scored as 15 components, covering consciousness, gaze, visual fields, facial palsy, limb strength, ataxia, sensation, language, dysarthria, and extinction.
- Total 0 to 42; higher scores indicate greater severity.
- Public domain. The scale is published by NINDS and free to download.
Psychometrics
Per-item agreement is the mean kappa across 24 stroke patients in Brott et al. (1989). Total-score agreement is the interobserver ICC among 59 trained raters in Goldstein and Samsa (1997). Higher is better.
What does the NIHSS measure?
The NIHSS tests the major functions affected by stroke. NINDS sets out 11 categories: level of consciousness, horizontal gaze, visual fields, facial palsy, arm motor function, leg motor function, limb ataxia, sensation, language, dysarthria, and extinction or inattention. Consciousness is rated three ways, and the arms and legs are rated separately on each side, so those 11 categories produce 15 scored components. Each component is scored from 0, meaning normal, up to a ceiling of 2, 3, or 4 depending on what is being tested.
Because the scale is in the public domain, the full form is free to download from NINDS, and the item definitions and scoring rules can be read in advance.
What does an NIHSS score mean?
The total runs from 0 to 42, where higher means a more severe stroke. The convention in general clinical use groups the total into five bands.
| NIHSS total | Severity |
|---|---|
| 0 | No stroke symptoms |
| 1 to 4 | Minor stroke |
| 5 to 15 | Moderate stroke |
| 16 to 20 | Moderate to severe stroke |
| 21 to 42 | Severe stroke |
It is worth naming the scheme when a score is reported, because more than one is in use. The team that published the scale stratified the same totals differently, as mild for 1 to 5, mild to moderately severe for 5 to 14, severe for 15 to 24, and very severe above 25. A score of 12 is moderate on one reading and mild to moderately severe on the other.
A higher score predicts a poorer outcome. In the TOAST trial, each additional baseline point cut the odds of an excellent outcome at three months by 17 percent. The total is reference information used to gauge severity and track change, read alongside imaging and the rest of the examination.
How the NIHSS is administered
- Administered by
- A trained clinician. Neurologists, other physicians, and nurses score it comparably.
- Time to administer
- Under 10 minutes for a trained examiner
- Equipment
- The NIH Stroke Scale form, downloadable from NINDS
- Cost
- Free to take here.
- Published by
- NINDS
- Training required
- Certification is expected in routine stroke care. Apex Innovations and NIHStrokeScale.org (BlueCloud) both offer free, CME-accredited online courses. NINDS also links to an American Heart Association course.
- Recertification
- Commonly required annually, though the evidence supports a longer interval
- Age range
- Adults
Which conditions is the NIHSS used for?
Related instruments
- Modified Rankin Scale (mRS)One grade for disability, months after the stroke
- Barthel IndexIndependence across ten everyday activities
- Fugl-Meyer Assessment, Upper Extremity (FMA-UE)How much arm and hand movement has returned
- Frenchay Dysarthria Assessment (FDA-2)Dysarthria in depth, which the NIHSS covers in one item
NIHSS questions, answered
What does an NIHSS score mean?
The total runs from 0 to 42, and higher means a more severe stroke. A common convention reads 1 to 4 as minor, 5 to 15 as moderate, 16 to 20 as moderate to severe, and 21 and above as severe. Different schemes exist, and a higher baseline score predicts a poorer outcome at three months. The number is reference information. The stroke team reads it alongside imaging and the exam.
How do I get NIHSS certified, and is it free?
Free certification exists. Two independent online programs, Apex Innovations and NIHStrokeScale.org run by BlueCloud, both offer NIHSS training and certification at no cost, and both carry continuing education accreditation. Each is built around scoring recorded patient examinations. The Apex course runs about three hours and asks for 93 percent on the test to pass. The National Institute of Neurological Disorders and Stroke also links to an American Heart Association course. The scale itself is in the public domain and free to download from ninds.nih.gov. Recertification is commonly required every year, although published data covering more than a million test administrations found scoring accuracy holds steady over time, and the authors argue the interval should be longer. Standardized video training measurably improves scoring reliability.
How many items does the NIHSS have?
It depends on how you count, which is why you will see both 11 and 15. NINDS describes 11 categories. Four of them are recorded twice or three times over, because consciousness is rated three ways and the arms and legs are rated separately on each side, so the form has 15 scored rows. The original 1989 paper described it as a 15-item scale. Either way the maximum total is 42.
Is there an online NIHSS calculator?
The assessment here walks through all 15 scored components and totals them for you. The total is reference information, not a diagnosis.
How long does the NIHSS take?
Under ten minutes for a trained examiner.
Evidence, psychometrics and provenance
Designed by Thomas Brott and colleagues at the University of Cincinnati for use in acute stroke therapy trials, and published in 1989. Standardized video training and certification followed in 1994, developed for the NINDS trial of tissue plasminogen activator, and remain how examiners are prepared today.
- Intra-rater reliability
- ICC 0.93 on re-rating after three months (Goldstein and Samsa, 1997)
- Test-retest agreement
- Mean kappa 0.66 to 0.77 (Brott et al., 1989)
- Construct validity
- Correlates with infarct size on CT at one week (r 0.68) and with 3-month outcome (r 0.79) (Brott et al., 1989)
- Predictive validity
- In TOAST, each extra baseline point cut the odds of an excellent outcome at 3 months by 17 percent (Adams et al., 1999)
References
- 1.Brott T, Adams HP Jr, Olinger CP, et al. Measurements of acute cerebral infarction: a clinical examination scale. Stroke. 1989;20(7):864-870.
- 2.Goldstein LB, Bertels C, Davis JN. Interrater reliability of the NIH stroke scale. Arch Neurol. 1989;46(6):660-662.
- 3.Lyden P, Brott T, Tilley B, et al. Improved reliability of the NIH Stroke Scale using video training. NINDS TPA Stroke Study Group. Stroke. 1994;25(11):2220-2226.
- 4.Goldstein LB, Samsa GP. Reliability of the National Institutes of Health Stroke Scale. Extension to non-neurologists in the context of a clinical trial. Stroke. 1997;28(2):307-310.
- 5.Adams HP Jr, Davis PH, Leira EC, et al. Baseline NIH Stroke Scale score strongly predicts outcome after stroke: a report of the Trial of Org 10172 in Acute Stroke Treatment (TOAST). Neurology. 1999;53(1):126-131.
- 6.Anderson A, Klein J, White B, et al. Training and certifying users of the National Institutes of Health Stroke Scale. Stroke. 2020;51(3):990-993.
- 7.National Institute of Neurological Disorders and Stroke (NINDS). NIH Stroke Scale. ninds.nih.gov.
- 8.Shirley Ryan AbilityLab. Rehabilitation Measures Database: National Institutes of Health Stroke Scale.
This assessment uses a validated instrument and is reference information, not a diagnosis.