Dictate anywhere on your screen, capture telehealth audio straight from the call,
and skip the second login.
The National Institutes of Health Stroke Scale (NIHSS) is a validated 15-item tool that quantifies stroke severity and tracks neurological change over time. It scores core neurological functions, including consciousness, vision, motor strength, sensation, language, and speech. Originally built as a research tool for alteplase, it became the standard for stroke assessment and is now used in both clinical trials and routine practice. Calculate stroke scores below:
As you complete the assessment, the overall NIHSS score updates, with itemized criteria shown so you can independently verify each grade. From there, Heidi Evidence returns citation-backed answers to the clinical questions that follow. The next call stays yours.
Warning: Clinical review pending: not for clinical use
Level of Consciousness
Overall responsiveness
LOC Questions
Ask the patient's month and age
LOC Commands
Open/close eyes, grip/release non-paretic hand
Best Gaze
Horizontal eye movements only
Visual Fields
Confrontation testing
Facial Palsy
Ask patient to show teeth, raise eyebrows, close eyes
Motor Arm — Left
Extend arm 90°/45° for 10 seconds
Motor Arm — Right
Extend arm 90°/45° for 10 seconds
Motor Leg — Left
Hold leg at 30° for 5 seconds
Motor Leg — Right
Hold leg at 30° for 5 seconds
Limb Ataxia
Finger-nose-finger, heel-shin
Sensory
Pinprick to face, arm, trunk, leg
Best Language
Naming and reading tasks
Dysarthria
Rate clarity of speech
Extinction and Inattention
Double simultaneous stimulation
For clinical decision support only. Always integrate with full clinical assessment, local guidelines, and patient preferences.
Every summary carries citations with verbatim excerpts from the source, so you can open it, read it in context and confirm it says what the answer claims. What surfaces are decided by clinical quality, the governed alternative to pasting a clinical question into a general AI chatbot and hoping the answer holds.
Heidi Evidence draws on sources clinicians already trust, from Elsevier and DynaMed to MIMS Australia, and your organization can layer in its own regional pathways, policies, and formularies.
Heidi is the first in its field to hold ISO 42001 certification for AI management. Since Evidence is designed as a reference tool, the clinical call (and the responsibility for it) stays with you.
Evidence draws on the patient visit and its context, so the specifics travel with the question instead of being left behind. When a calculation is involved, the logic sits alongside the result, so you can see how it was reached and confirm it yourself rather than trust a black box.
Making that connection between the patient and the GP, like a puzzle fitting together.
Deborah Hawthorne
Clinicians at Connect2Care
Medical knowledge only. Not for autonomous decision making. Check sources and use your clinical judgement.
The scale scores 15 items across 11 neurological domains, with totals running from 0 to 42. Scoring follows the first response, not the best one, and Item 9 (Best Language) is the only exception. The design prioritizes reproducibility between raters over diagnostic precision, which is why the examiner scores observed performance rather than inferred capability.
Do not coach the patient through any item.Untestable items are handled as follows: physical limitations such as amputation or joint fusion affecting a specific motor or ataxia item are marked UN rather than scored as normal.
Items that cannot be fully examined for other reasons, including intubation or a language barrier and must still receive a score without leaving them left blank. Item 11 (Extinction/Inattention) can never be designated as untestable.Under Joint Commission Primary Stroke Center certification, an emergency department physician must assess a suspected stroke patient within 15 minutes of ED arrival, with the NIHSS used for the initial stroke assessment.