Understanding the NIH Stroke Scale Group A Assessment
The NIH Stroke Scale is broken into item groups. Group A covers the first eight items — level of consciousness questions, gaze, visual fields, facial palsy, and motor function of the arms and legs. These are the items you use first when you walk into a room with a potentially stroking patient. Most EMS protocols and stroke center checklists start here because the early items separate awake from comatose and establish a baseline quickly. I am going to walk through each Group A item, the scoring, and where people typically mess up. The scale runs from 0 to 2 for most items, with a couple of 3-point items. The total Group A range is 0 to 19 before you even hit the rest of the scale. Item 1a is level of consciousness questions. You ask the patient their age and the current month. A score of 0 means they answer both correctly. Score of 1 means one is wrong. Score of 2 means neither is correct or the patient is intubated and cannot respond. The tricky part here is not the scoring itself. It is the border cases. I have had patients who are 82 years old and clearly confused but can still tell you the month if you phrase the question slowly. Some folks will automatically give a 2 just because the patient seems foggy. That is wrong. You only score based on whether the specific questions are answered correctly, not on general confusion. I once scored a patient a 2 on this item because they were under heavy sedation from a benzodiazepine administration prior to transfer. My neurology team pointed out that the sedation could be masking their true neurological status, so I re-rated them as 0 after the meds wore off two hours later. That changes the total score enough to affect tPA eligibility thresholds in some systems.
Item 1b is level of consciousness questions with a slight variation for intubated or language-barrier patients. You ask simple math like 1 dollar equals how many pennies. Correct answers score 0. One error scores 1. Two errors or no response scores 2. This item is almost always skipped in practice because most stroke patients are not being asked about currency. When you do use it, stay consistent. Do not switch to asking something easier mid-assessment just to get a 0. Item 1c is level of consciousness commands. You ask the patient to open and close their eyes, then grip and release your hand. Score 0 for both correct, 1 for one incorrect, and 2 for both incorrect. The common pitfall here is giving encouragement that helps the patient. If you say "just do it like this" and demonstrate the movement, some scorers count that as a 0. It is not. You present the command, wait, and score based on independent execution. I ran into a case where a right hemiparesis patient could only move their left arm due to the stroke. They opened their eyes but could not grip my hand with the affected side. I originally marked a 1 because they completed one of the two commands. That was correct. The key is separating motor deficit from comprehension deficit. If they clearly understood you and just could not execute due to weakness, that belongs on the motor items, not here. Item 2 is best gaze. You test the oculocephalic reflex or ask the patient to follow your finger. Horizontal gaze palsy scores 1. Partial or complete gaze palsy that cannot be overcome by reflex scores 2. Normal scores 0. The subtle point most people miss is that a forced gaze deviation toward a hemispheric lesion is actually a positive finding and should be scored as 1, not ignored because "everyone knows stroke patients look toward the lesion." Documentation matters. Write down that the gaze is deviated toward the right with a suspected left MCA stroke. That detail becomes critical if you need to justify a score change during a chart review.
Item 3 is visual fields. You test confrontation or formal perimetry if available. Complete hemianopia scores 2. Partial hemianopia scores 1. No visual loss scores 0. Blindness from another cause also scores 0. The practical issue here is distinguishing a true field cut from inattention. I had a patient who missed stimuli on the left during confrontation testing but was clearly not paying attention because they were extremely agitated. I waited ten minutes, rechecked, and the field was intact. Scoring an incomplete hemianopia on a frantic exam leads to inflated NIHSS scores. Retest if the patient is unstable or anxious. Item 4 is facial palsy. You ask the patient to show teeth or close their eyes tightly. Asymmetry scores 1. Complete paralysis scores 2. Normal scores 0. Bell's palsy complicates this item significantly. A patient with a known peripheral facial nerve palsy will score 2 on facial palsy regardless of stroke status. The workaround I use is to document the pre-existing condition clearly and note whether the stroke-related asymmetry is superimposed on it. Some stroke scales allow you to adjust for known peripheral deficits, but the standard NIHSS does not. Be explicit in your notes so the receiving team knows you accounted for it. Item 5 is left arm motor drift. You hold both arms extended at 90 degrees for 10 seconds or at 45 degrees for 5 seconds. No drift scores 0. Drift scores 1. Some effort against gravity but hits the bed scores 2. No effort at all scores 3. Not testable scores 4. The most common error I see is scorers using 45 degrees for everyone because it is faster. It is not equivalent. The 45-degree position is more demanding and can mask mild drift that would be obvious at 90 degrees. Stick to the standardized position unless the patient genuinely cannot maintain 90 degrees. If you modify the position, document it.
Get the Full Details
Item 6 is left leg motor drift. Same scoring structure as the arm. Test by holding the leg at 30 degrees for 5 seconds. I once assessed a patient with severe osteoarthritis in the right hip who could not abduct the leg properly. The arm scores were normal but the leg score was inflated because of the orthopedic limitation. The workaround was to test ankle dorsiflexion instead and document the reason. This is not ideal but it prevents an artificially high score from masquerading as a worse stroke. Item 7 is right arm motor drift. Same methodology as item 5. Item 8 is right leg motor drift. Same methodology as item 6.
Here is a practical tip for running through Group A quickly. Do not stop after each item to calculate. Run through all eight and jot down the scores in order. Then total. This takes about 3 to 5 minutes in a cooperative patient and closer to 8 to 12 minutes in someone who is agitated, intubated, or has severe dysarthria. The time difference matters when you are trying to meet door-to-needle targets. One thing the NIHSS does not do well is capture aphasia severity or neglect independently in Group A. Those come in later items. Group A only indirectly reflects cognitive and language function through the level of consciousness questions. If a patient scores 0 on items 1a through 1c but then scores high on item 9 (language) and item 11 (neglect), the Group A subtotal will not reflect the full clinical picture. This is a known limitation. Some institutions use the Full Outcomes Scale or the Canadian Neurological Scale as supplementary tools when the NIHSS feels incomplete for a particular patient phenotype. The total Group A score can range from 0 to 19. In practice, most ischemic stroke patients fall between 2 and 12 in this section. Scores above 15 in Group A alone usually indicate a large vessel occlusion or severe global dysfunction. I have seen scores of 17 in Group A in patients with bilateral anterior circulation involvement or brainstem strokes. Do not treat the number in isolation. Correlate with imaging and clinical evolution.
If you want a printable quick-reference sheet, most hospital pharmacies and stroke networks distribute laminated NIHSS cards. The American Stroke Association maintains free downloads on their website. I keep one in the trauma bay pocket and another on the coding station monitor. The format varies slightly between agencies but the scoring criteria are standardized. The main takeaway is that Group A items are straightforward in theory but full of small scoring traps in practice. The biggest one is conflating comprehension problems with motor problems on the consciousness items. The second is skipping retests on borderline visual field exams. The third is inflating facial palsy scores in patients with pre-existing facial nerve issues. Get those three right and your Group A scoring will be reliable enough to stand up to audit and peer review.
