Simulating the acute ischemic stroke code timeline — from ambulance doors to tPA bolus — and how parallelized workflow compresses time-critical decisions
Every minute of untreated large-vessel ischemic stroke costs an estimated 1.9 million neurons. The door-to-needle clock is not a paperwork metric — it is a direct proxy for salvageable brain tissue. The single highest-leverage intervention happens before the patient ever reaches the hospital: EMS pre-notification.
"Door-to-needle" (DTN) is defined precisely: the interval from ED arrival (registration timestamp or EMS handoff) to the start of IV alteplase (or tenecteplase) bolus administration. It deliberately excludes prehospital time, because hospitals can only control what happens after arrival — but prehospital coordination still shapes everything downstream.
EMS pre-notification is the force multiplier: paramedics use a validated stroke screen (Cincinnati Prehospital Stroke Scale, FAST-ED, or RACE) in the field and radio ahead with last-known-well time, suspected severity, and estimated arrival. This allows the receiving ED to:
• Activate the stroke team (neurology, radiology tech, pharmacy, nursing) before the patient arrives • Clear and reserve the CT scanner • Pre-mix or stage weight-based alteplase dosing calculations • Bypass triage entirely — the patient goes straight to a dedicated stroke bay or directly to the CT table
Hospitals with reliable EMS pre-notification report DTN times 15-20 minutes faster on average than those relying on walk-in or non-notified arrivals.
A "Code Stroke" or "Brain Attack" page fires a single overhead or pager alert that simultaneously activates every downstream resource: the on-call neurologist or tele-stroke service, a dedicated CT technologist, the ED charge nurse, pharmacy (for weight-based dosing), and often a point-of-care lab tech. The core design principle is that no single person becomes a bottleneck waiting to be individually paged — one activation, many parallel responders.
Modern EDs increasingly co-locate the stroke bay adjacent to (or literally next to) the CT suite, physically minimizing the door-to-CT distance to under 30 seconds of travel time.
Non-contrast head CT remains the fastest, most universally available tool to exclude intracranial hemorrhage — the single absolute contraindication to thrombolysis that must be ruled out before any drug is considered. The national benchmark is door-to-CT ≤25 minutes; leading centers routinely achieve under 15.
Rather than routing the patient through a standard ED bed, triage note, and physician order sequence, "direct-to-CT" protocols send the patient straight from the ambulance stretcher onto the CT table — often before a full history is taken. Vital signs, IV access, and a fingerstick glucose are obtained en route or on the table itself, not as a gating step beforehand.
A rapid neurological exam (abbreviated NIHSS or a "last known well" confirmation) occurs simultaneously while the patient is being positioned, not before transport begins. Radiology is notified in advance so the scanner is empty and the protocol is pre-loaded, eliminating queueing time entirely.
Non-contrast CT within the door-to-CT window must answer three yes/no questions fast: (1) Is there acute intracranial hemorrhage? (2) Is there an established, completed large infarct with hypodensity exceeding roughly one-third of the middle cerebral artery territory (a relative contraindication)? (3) Are there early ischemic signs (loss of gray-white differentiation, sulcal effacement) that inform — but do not by themselves exclude — thrombolysis eligibility?
CT angiography, when large-vessel occlusion is suspected (for mechanical thrombectomy triage), can often be acquired in the same table time without materially extending the door-to-CT interval, since it is appended immediately after the non-contrast series.
This is the single largest architectural lever in door-to-needle optimization. Legacy stroke workflows queued point-of-care glucose, coagulation labs, formal NIHSS scoring, and patient weight one after another, each waiting on the last. Modern protocols run every one of these tracks at the same time the patient is on the CT table.
Track 1 — Point-of-care glucose: obtained by fingerstick within the first minute; hypoglycemia and hyperglycemia can both mimic or worsen stroke symptoms and must be excluded/corrected immediately, but the result returns almost instantly and rarely blocks the timeline.
Track 2 — Point-of-care INR/coagulation: for patients on warfarin or with suspected coagulopathy, a bedside coagulometer returns an INR in minutes rather than waiting on a send-out lab result that can take 30-45 minutes through a central lab.
Track 3 — NIHSS scoring: the National Institutes of Health Stroke Scale is performed by a trained nurse or physician at the bedside or literally at the CT table, in parallel with imaging, not after it.
Track 4 — Weight estimation: alteplase dosing is weight-based (0.9 mg/kg, max 90 mg); rather than waiting for a formal scale weight, most protocols now use a rapid visual/algorithmic estimate or a bed-integrated scale so dosing can be pre-calculated before the CT even finishes.
If each of the four tracks above takes 5-15 minutes and they are run sequentially, they alone can consume 30-40 minutes — more than the entire target DTN. Run concurrently, the total added time collapses to whichever single track is slowest (typically the coagulation result, 5 minutes), not the sum of all four.
This is why "parallel workflow efficiency" is modeled as a continuous slider rather than a binary switch in this simulation: real EDs sit somewhere between fully sequential (legacy) and fully parallel (best-practice), and small process changes — moving the POC coagulometer to the CT suite, training nurses to score NIHSS en route — shift that percentage incrementally.
Modeled here: post-CT overhead falls from a 35-minute fully-sequential baseline toward a 10-minute floor as parallel workflow efficiency approaches 100%. That floor represents irreducible time — reading the scan, confirming eligibility, and physically drawing up and pushing the drug.
With imaging complete and labs/NIHSS/weight already in hand from the parallel tracks, the decision gate is the last checkpoint before drug administration. A radiologist or neurologist reviews the non-contrast CT to exclude hemorrhage and confirm eligibility against inclusion/exclusion criteria — ideally within a couple of minutes, not through a formal dictated report cycle.
Waiting for a formal, transcribed radiology report can add 15-30 minutes to the timeline. Time-critical stroke protocols instead rely on a "wet read": the neurologist or stroke-trained radiologist reviews images live at the PACS workstation — often while still standing at the CT console — and verbally confirms absence of hemorrhage directly to the treating team. The formal dictated report follows afterward for the medical record, but does not gate treatment.
Tele-stroke networks extend this capability to hospitals without in-house neurology coverage: a remote vascular neurologist reviews the CT and examines the patient via two-way video, rendering a decision within minutes, indistinguishable in speed from an in-person consult at high-performing centers.
By the time the scan is read, the team already has: last-known-well time (confirmed on arrival), NIHSS score (from the parallel track), glucose and INR (from point-of-care testing), and estimated weight (for dosing). The decision gate therefore only has to synthesize information that already exists rather than initiate new data collection — which is precisely why parallelizing Stage 3 pays off disproportionately at this checkpoint.
Common exclusion criteria checked at this gate: intracranial hemorrhage, recent major surgery or trauma, active internal bleeding, severe uncontrolled hypertension, or a large established infarct on imaging. If none apply and the patient is within the therapeutic time window, the gate opens.
The moment the weight-based alteplase (or tenecteplase) bolus begins infusing, the door-to-needle clock stops. Leading centers now routinely administer the bolus in the CT suite itself — the patient never leaves the table between the scan and the drug — collapsing what used to be a multi-department relay into a single continuous encounter.
Once the gate opens, pharmacy or nursing draws up the pre-calculated weight-based dose — ideally already prepared in anticipation during the parallel-workflow stage — and the bolus is pushed at the bedside. For alteplase, 10% of the total dose is given as an initial bolus over one minute, with the remainder infused over the following 60 minutes; tenecteplase, increasingly favored for its single-bolus simplicity, is given as one push with no follow-on infusion, which itself shaves administration complexity off the back end of the timeline.
The "door-to-needle" timestamp is the start of this bolus — not the order, not the pharmacy dispense, not the consent conversation. Every process improvement in Stages 1-4 exists purely to compress the interval leading up to this single moment.
The Get With The Guidelines-Stroke program set ≤60 minutes as the national DTN benchmark because outcomes data show a clear, near-linear relationship: faster treatment produces less disability at 90 days and lower in-hospital mortality, for every increment of time saved, not just past a single cutoff. The more aggressive ≤30-minute "stretch" target recognized by "Target: Stroke Elite Plus" designation represents what dedicated, protocol-driven centers can achieve when every stage in this timeline — pre-notification, direct-to-CT, parallel eligibility screening, wet-read decisions, and at-the-table bolus — is optimized simultaneously.
A commonly cited clinical heuristic (Saver, 2006) estimates that each minute of untreated large-vessel ischemic stroke costs approximately 1.9 million neurons, 14 billion synapses, and 12 km of myelinated fibers. The tissue-saved metric in this simulator applies that heuristic illustratively to the minutes reclaimed by parallel workflow — it is a teaching approximation, not a patient-specific clinical prediction.