B_PROUD Trial: Mobile Stroke Unit Dispatch vs Conventional Ambulance in Berlin
In suspected acute ischemic stroke dispatched in Berlin's EMS system, does mobile stroke unit dispatch improve 90-day disability outcomes compared with conventional ambulance alone? (Allocation by MSU availability — not patient-level randomization.)
Ebinger M, et al. (JAMA 2021;325(5):454–466) · doi:10.1001/jama.2020.26345 · 1543 patients
Population
Included
- Suspected acute ischemic stroke, dispatched as emergency call
- Within Berlin MSU catchment area (3 base stations)
- MSU operating hours: 7am–11pm (daytime only)
- Final diagnosis of acute ischemic stroke or TIA
- At least 18 years old
- Emergency call prompted a mobile-stroke-unit dispatch code during the MSU operating hours (7 AM to 11 PM, Monday to Sunday)
- Onset of symptoms to dispatch within 4 hours (allowing a minimum of 30 minutes for travel to the emergency site and assessment within the 4.5-hour window for thrombolysis)
- Within the catchment area of a Berlin mobile stroke unit (3 base stations)
- Ambulatory before the stroke (a rough proxy of modified Rankin Scale score 3 or less)
- Final hospital-based diagnosis of acute ischemic stroke (ICD-10 I63) or transient ischemic attack (ICD-10 G45, except G45.4) without symptom resolution before arrival in the emergency vehicle
- Considered eligible for thrombolysis or thrombectomy
Excluded
- Stroke onset outside catchment or operating hours
- Stroke mimics (final diagnosis other than ischemic stroke/TIA)
- ICH on imaging
- No randomization. Allocation by MSU availability (quasi-experimental)
- Absolute contraindications against both thrombolysis and thrombectomy (for example, patients without large-vessel occlusion taking oral anticoagulation, or nonmedication-induced hemophilia)
- Malignancy or other severe primary disease with life expectancy of less than 1 year
- Major surgery within the past 4 weeks
- Previous stroke within 3 months
- No deficits at emergency medical service arrival
- Symptom onset or last-seen-well to alarm more than 4 hours
- Not transported to a hospital, or transported to a hospital without a stroke unit
- Younger than 18 years or did not have a final diagnosis of acute cerebral ischemia
- Symptoms resolved before emergency department medical service or at hospital arrival, or arrived at the hospital more than 6 hours after symptom onset
- Patient opted out (opt-out informed-consent design)
Source: Ebinger M et al., JAMA 2021;325(5):454–466; PMID 33528537· Retrieved 2026-06-09
Primary Outcome — Ordinal mRS Shift
Quasi-experimental — interpret as association, not causation: Allocation was by MSU availability, not patient-level randomization. Per clinical-trial-audit, NNT is not displayed for observational/registry designs because residual confounding prevents causal absolute-risk-difference interpretation. Primary analysis used ordinal mRS shift (common OR 0.71 for worse mRS).
Small absolute difference — interpret with caution
mRS 0–3 or living at home (coprimary)
Study Arms
- Agent
- Mobile stroke unit (MSU) dispatch
- Route
- Prehospital, ambulance-based
- Co-interventions
- Simultaneous dispatch of a conventional ambulance; transport to the nearest stroke-unit hospital (or nearest thrombectomy-capable hospital if large-artery occlusion detected on CT angiography)
MSU equipped with a CT scanner allowing angiography, telemedicine-enabled connections to radiologists, and a point-of-care laboratory (international normalized ratio, blood cell count, glucose, creatinine, electrolytes). Staffed by a paramedic, a radiology technician trained in emergency medicine, and a neurologist trained in emergency medicine. When criteria were met, thrombolysis was started prehospital on board the MSU. Dispatch code triggered simultaneous MSU and conventional-ambulance dispatch only when an MSU was available.
- Agent
- Conventional ambulance transport
- Route
- Prehospital, ambulance-based
- Co-interventions
- Transport to a hospital stroke unit; thrombolysis and thrombectomy delivered in hospital
Conventional ambulances staffed with a paramedic and an ambulance technician. Patients allocated to this group when no MSU was available at the time of the emergency call. In-hospital thrombolysis and reperfusion decisions made by the receiving stroke team.
Safety
Symptomatic secondary intracranial hemorrhage
3.2%
2.8%
Adjusted OR 1.20 (95% CI 0.66–2.19). No significant safety penalty despite more frequent prehospital thrombolysis.
Death at 90 days
7.1%
8.8%
Numerically lower with MSU dispatch in the coprimary tier; primary analysis is the ordinal mRS shift, not mortality.
Trial Design
Type
- Prospective, nonrandomized controlled intervention study (quasi-experimental)
- Allocation by MSU availability. Not patient-level randomization
- MSU dispatch plus ambulance vs conventional ambulance alone
- Pragmatic Berlin stroke system evaluation (3 MSU base stations, 24/7 staffing, paramedic + radiology tech + emergency neurologist)
Timeline
Berlin, Germany; February 1, 2017 to October 30, 2019 (final inclusion target reached May 8, 2019)
N
1543
Enrollment
1,543 patients in Berlin (3 MSU base stations, 7am–11pm operating hours). Enrolled Feb 2017 – Oct 2019. Published JAMA 2021;325(5):454–466.
ClinicalTrials.gov
NCT03027453Bedside Pearl
In Berlin's urban EMS system, MSU dispatch shifts the ordinal mRS distribution toward less disability (common OR 0.71 for worse mRS) compared with conventional ambulance. Thrombolysis use rose (60% vs 48%) and dispatch-to-tPA shortened by ~26 min. Quasi-experimental allocation by MSU availability: association, not causation.
See also