Influence of Distance to Scene on Time to Thrombolysis in a Specialized Stroke Ambulance
Background and Purpose—Specialized computed tomography–equipped stroke ambulances shorten time to intravenous thrombolysis in acute ischemic stroke by starting treatment before hospital arrival. Because of longer travel-time-to-scene, time benefits of this concept are expected to diminish with longer distances from base station to scene.
Methods—We used data from the Prehospital Acute Neurological Treatment and Optimization of Medical Cares in Stroke (PHANTOM-S) trial comparing time intervals between patients for whom a specialized stroke ambulance (stroke emergency mobile) was deployed and patients with conventional emergency medical service. Expected times from base station to scene had been calculated beforehand using computer algorithms informed by emergency medical service routine data. Four different deployment zones with–75% probability–expected arrival within 4, 8, 12, and 16 minutes and total population coverage of ≈1.3 million inhabitants were categorized for stroke emergency mobile deployment. We analyzed times from alarm-to-arrival at scene, to start of intravenous thrombolysis and from onset-to-intravenous thrombolysis.
Results—Corresponding to the size of the respective catchment zone, the number of patients cared increased with distance (zone 1: n=30, zone 2: n=127, zone 3: n=156, and zone 4: n=217). Although time to stroke emergency mobile arrival increased with distance (mean: 8.0, 12.5, 15.4, and 18.4 minutes in zones 1–4), time from alarm-to-intravenous thrombolysis (mean: 41.8 versus 76.5; 50.2 versus 79.1; 54.5 versus 76.6; and 59.3 versus 78.0 minutes, respectively; all P<0.01) remained shorter in the stroke emergency mobile group across all zones.
Conclusions—In a metropolitan area such as Berlin, time benefits justify a specialized stroke ambulance service up to a mean travel time of 18 minutes from base station.
Intravenous thrombolysis (IVT) with tissue-type plasminogen activator has proven beneficial within 4.5 hours of onset.1 Shortening symptom onset-to-treatment (OTT) times improves the outcomes of stroke patients2 and has become the aim of several strategies in stroke systems of care.3–5
Specialized computed tomography–equipped stroke ambulances reduce OTT by starting treatment before hospital arrival as shown in the Prehospital Acute Neurological Treatment and Optimization of Medical Cares in Stroke (PHANTOM-S) study.3,6 With growing distance and travel time from base station to scene, the benefits of this concept are expected to diminish, especially in cities with specialized stroke units available at short distance for regular ambulances. In this study, we investigate whether time benefits are sustained within the different catchment zones used in PHANTOM-S.
This is a secondary analysis of the PHANTOM-S study that was approved by the Charité Internal Review Board Ethics Committee.3 Precise details of the methodology have been published earlier.7,8 In brief, a specialized stroke ambulance (stroke emergency mobile [STEMO]) was implemented and based at a fire station close to the center of Berlin, Germany (Figure). Staff included a neurologist trained in emergency medicine, a paramedic and a radiology technician. A senior neurologist and a neuro-radiologist were able to connect using telemedicine. At dispatch level, a specialized algorithm was used to identify emergency calls of patients with suspected strokes within 4 hours of onset or unclarified time of onset.9 If STEMO was activated, a regular ambulance was usually sent to scene simultaneously. The STEMO deployment area covered ≈1.3 million inhabitants in Berlin with an average citywide population density of 3908 inhabitants per km2. Four different operation zones with expected arrival—75% probability—within 4, 8, 12, and 16 minutes from base station were categorized for STEMO deployment (Figure). Expected times from base station to scene had been calculated beforehand using computer algorithms acquired by emergency medical service routine data of the Berliner Feuerwehr (ie, Berlin fire brigade). When STEMO had finished a previous emergency deployment but had not yet arrived at base station, it could be deployed on its way back home.
If STEMO was not available (during control weeks, in operation or maintenance), patients received conventional care.
For this analysis based on PHANTOM-S study data, we used the same patients and the baseline parameters as in the original study. We analyzed times from alarm-to-arrival at scene, from alarm-to-hospital arrival, from alarm-to-imaging, from alarm-to-IVT, and from onset-to-IVT in the 2 groups of patients for whom STEMO was deployed or patients who were cared by conventional emergency medical service.
We provide data in descriptive analyses and used IBM SPSS Statistics version 22 (IBM SPSS). We used the Pearson χ2 test or Fisher exact test to compare categorical variables and the Mann–Whitney U test or the Kruskal–Wallis test (if >2 groups) to test for non-normally distributed data. A 2-sided significance level of α=0.05 was used. No correction for multiple testing was applied.
Between May 1, 2011 and January 31, 2013, 530 patients received IVT. Of those, 330 patients were cared by conventional emergency medical service versus 200 patients with STEMO deployment. Results are listed in Table. We found no significant differences in demographics or comorbidities between patients receiving conventional care versus STEMO care, but stroke severity measured with the National Institutes of Health Stroke Scale (NIHSS) was lower in conventional care.
The numbers of patients treated in the 4 catchment zones increased with distance and population served in each zone (zone 1: n=30; zone 2: n=127; zone 3: n=156; and zone 4: n=217). Although times from alarm-to-first-emergency medical service-arrival at scene were similar comparing conventional and STEMO care in all catchment zones, mean times from alarm-to-STEMO-arrival increased from 8.0 to 18.4 minutes from zone 1 to 4 (P<0.01).
The proportions of stroke codes in the dispatch center that were activated during ongoing STEMO operations increased from zone 1 to 4 (36.9, 38.6, 43.0, and 47.5, respectively; P<0.01).
All treatment relevant times remained shorter in the STEMO group compared with conventional care with mean time differences in alarm-to-treatment time between 34.7 minutes in zone 1 and 18.7 minutes in zone 4. Mean time differences in OTT time showed a sharper decrease from zone 1 (62.9 minutes) to zone 4 (8.1 minutes).
Our study shows a significant time advantage of STEMO over conventional care for a deployment within a mean travel radius of ≤18 minutes from base station. Within this area, mean alarm-to-treatment times remained at least 19 minutes shorter compared with conventional care. The smaller OTT time benefit in more distant areas may be caused by different population profiles in the deployment areas with lower average inhabitant’s socioeconomic status in some areas of zone 4.10 In addition, patients with very long onset-to-alarm times could only receive thrombolytic treatment when cared by STEMO, whereas they would not have received IVT in conventional care.
Berlin has 15 stroke units distributed in a city of ≈3.5 million inhabitants. Our data suggest that within this metropolitan area, the STEMO concept is unlikely to reduce alarm-to-treatment times by a relevant degree in deployment areas beyond 18 minutes mean travel time from base station. However, in areas with fewer deployments during ongoing operations or with centralized systems of care where wider distance to scene are associated with wider distance to the stroke center, this cutoff may not be applicable. As shown in the Helsinki area, a centralized approach can minimize door to needle times down to 20 minutes (median) with widely unchanged onset to hospital arrival time (median, 89 minutes).5 In some urban/suburban areas, the STEMO concept may be complemented by a meet and greet system. In such systems, patients are transported by conventional ambulances from scene to a meeting point between STEMO base station and scene, there the patient could be transferred to STEMO. However, evaluations of this approach are currently not available.
Our study has limitations. First, as a post hoc analysis of the PHANTOM-S study, it was not designed to determine the optimal base-to-scene distance. Second, geographical areas and medical systems are very different. Our results may therefore not be generalizable to other regions or other systems of care. Third, with deployments on the way back to base station, the distances to scene sometimes differed from the originally calculated distances. Fourth, the STEMO project was conducted before the publication of thrombectomy trials; endovascular treatment times are currently being assessed.
Saver et al2 emphasized the importance of shortening OTT. In zone 1, half of all STEMO patients received IVT within 60 minutes while patients in conventional care received treatment on average more than 1 hour later than that. However, when it comes to total time savings, with the higher numbers of patients treated in the outer areas, cumulative minutes saved for all patients by the STEMO approach were highest in zone 4 (1520 minutes) compared with 455 minutes in zone 1, 1450 minutes in zone 2, and 1254 minutes in zone 3. These results endorse the chosen deployment radius of the STEMO service.
In metropolitan areas such as Berlin, relevant time savings can be achieved with STEMO even in relatively remote zones with mean travel times of 18 minutes from base station.
Sources of Funding
Prehospital Acute Neurological Treatment and Optimization of Medical Cares in Stroke (PHANTOM-S) was funded by the Zukunftsfonds Berlin and the Technology Foundation Berlin with European Union cofinancing by the European Regional Development Fund (ERDF).
Dr Audebert received speaker honoraria from Boehringer Ingelheim (manufacturer of Alteplase; but not involved in any form in the trial). Drs Audebert and Endres have received institutional funding by the German Federal Ministry for Education and Research within the Center for Stroke Research Berlin grant. The other authors report no conflicts.
- Received March 6, 2016.
- Revision received April 29, 2016.
- Accepted May 17, 2016.
- © 2016 American Heart Association, Inc.
- Fonarow GC,
- Zhao X,
- Smith EE,
- Saver JL,
- Reeves MJ,
- Bhatt DL,
- et al
- Ebinger M,
- Kunz A,
- Wendt M,
- Rozanski M,
- Winter B,
- Waldschmidt C,
- et al
- Ebinger M,
- Lindenlaub S,
- Kunz A,
- Rozanski M,
- Waldschmidt C,
- Weber JE,
- et al
- Krebes S,
- Ebinger M,
- Baumann AM,
- Kellner PA,
- Rozanski M,
- Doepp F,
- et al
- 10.↵Regionaler Sozialbericht Berlin-Brandenburg 2013. Amt für Statistik Berlin-Brandenburg Potsdam, Germany, 2015. https://www.statistik-berlin-brandenburg.de/instantatlas/interaktivekarten/sozialbericht2013/atlas.html. Accessed February 2, 2016.