Door-to-needle (DTN) time — the interval from hospital arrival to the start of intravenous thrombolysis — is the part of the stroke chain you control completely. Onset-to-arrival depends on public awareness and EMS; what happens after the doors open is pure systems design. Guidelines have long set a DTN target of ≤60 minutes, yet the best centers now routinely treat in under 30, and the very fastest have published median times near 20 minutes.4 The difference is almost never a faster neurologist. It is a protocol that does in parallel what most departments still do in sequence.

This page is educational and describes published quality-improvement strategies; it is not individualized clinical advice. Eligibility, agent choice, and dosing always follow your institution's protocol and current guidelines, and specifics vary between centers.

Why minutes are the whole game

In a Get With The Guidelines–Stroke analysis of more than 58,000 thrombolysis-treated patients, faster onset-to-treatment time in 15-minute increments was associated with lower in-hospital mortality, less symptomatic intracranial hemorrhage, more independent ambulation at discharge, and more patients discharged home.1 The benefit of thrombolysis is steeply time-dependent: the modeled trade-off is roughly the equivalent of additional disability-free life for every block of time saved, which is why shaving even single-digit minutes off a workflow is worth the effort.

Critically, faster treatment was not bought at the cost of safety — in that registry, shorter times went together with less symptomatic hemorrhage, not more.1 Speed and safety are aligned, not in tension, when the acceleration comes from removing handoffs rather than skipping checks.

Bedside pearl

The single highest-yield lever is the EMS pre-notification call. A structured pre-arrival report (last-known-well, suspected deficit, glucose, anticoagulant use, contact for consent) lets the team assemble, pull up the order set, and stage tenecteplase or alteplase before the patient crosses the threshold. The clock has not started, but the work has.

The parallel-processing playbook

Every minute saved comes from converting a serial step into a parallel one. The interventions below are the backbone of the Helsinki model and the AHA/ASA Target: Stroke initiative, and they reinforce each other:24

  • EMS pre-notification with structured data — the team is at the door, not paged after arrival.
  • Single-call activation of the entire stroke team (neurology, nursing, CT tech, pharmacy) to the CT suite at once.
  • Direct-to-CT from the EMS stretcher — bypass the ED bay; the first stop is the scanner table, not a triage room.
  • Run everything at once — imaging, point-of-care glucose and coagulation testing, NIHSS, focused history, weight estimate, and consent happen in parallel, not one after another.
  • Thrombolytic drawn and ready at the bedside — tenecteplase as a single weight-based bolus, or alteplase pre-mixed, so the drug is in hand the instant hemorrhage is excluded.
  • Automated imaging post-processing for non-contrast CT, CTA, and perfusion so interpretation does not wait on manual reconstruction.
  • Treat in the scanner — give the bolus on the CT table once non-contrast CT excludes hemorrhage, rather than transporting back to a bed first.

The 2019 AHA/ASA guideline endorses this systems-of-care approach across the prehospital and in-hospital phases, with a DTN goal of ≤60 minutes as a floor — not a ceiling — for what a well-run center should achieve.5

Proof it transfers: the Helsinki model

Skeptics assume sub-30-minute times require a national health system, an in-house neuro-ED, and resources most hospitals lack. The evidence says otherwise. When the Royal Melbourne Hospital adopted the three core Helsinki components — ambulance pre-notification with patient details, transfer directly onto the CT table from the ambulance stretcher, and tPA delivered in CT immediately after imaging — the in-hours median DTN fell from 43 to 25 minutes within four months, and the change took hold without a dedicated neurologic ED or electronic records.4 The model has since been reproduced in other centers with ordinary, real-world staffing.

At the national scale, the AHA/ASA Target: Stroke initiative was built around a defined set of best-practice strategies — EMS pre-notification, single-call activation, rapid imaging acquisition and interpretation, premixing the thrombolytic, a team-based approach, and rapid data feedback.2 A later survey of Get With The Guidelines–Stroke hospitals identified multiple strategies independently associated with shorter DTN; if a center implemented all of them, the modeled cumulative saving was on the order of 20 minutes.3

Pitfall

In-hours numbers flatter you. In the Melbourne data, the entire DTN improvement came during business hours; out-of-hours times barely moved.4 Reporting a single overall median hides a slow night shift. Audit in-hours and after-hours times separately, and build the protocol so the resident covering at 3 a.m. runs the same parallel pathway as the daytime team.

Don't let the thrombectomy candidate slow the needle

For patients with a suspected large-vessel occlusion at a center without thrombectomy, the two clocks run together, not one after the other. Drip-and-ship means starting thrombolysis locally for eligible patients while images are transferred to the comprehensive stroke center in parallel and transport is arranged. The thrombolytic decision should not wait on the thrombectomy decision; running them concurrently protects DTN without delaying the path to the angiosuite. Local transfer agreements, image-sharing platforms, and pre-arranged acceptance protocols are what make this parallelism possible. Specific transfer criteria and imaging requirements vary by network and should follow your regional stroke system's protocol.

Measure, feed back, repeat

None of this is sustained by enthusiasm. The centers that hold sub-30-minute times treat DTN as a continuously audited metric: timestamps captured at door, CT, and needle; case-by-case review of any treatment over target; and rapid feedback to the whole chain, including EMS.23 The improvement is iterative — each delayed case is a defect to diagnose, not a person to blame. When the data return to the team that generated them, the median falls and stays down.