Prehospital triage of patients with a suspected acute stroke

Every year, more than 30,000 patients with a stroke are hospitalized in the
Netherlands. Treatment with intravenous alteplase (IVT) is an effective
treatment for patients with ischemic stroke, but the effectiveness depends on
the speed of administration. Approximately 23% of patients with acute ischemic
stroke have a proximal intracranial arterial occlusion. The effect of IVT in
this group of patients is small. For patients with a proximal intracranial
occlusion, endovascular thrombectomy (EVT) is an effective treatment. This is a
complex treatment that is only available in specialized intervention centers.
Like with IVT, there is a strong time-dependent treatment effect, with
treatment being more effective the earlier it is initiated. An important cause
of delay is a transfer between hospitals. This occurs when patients with an
intracranial occlusion are first transferred to a hospital where EVT is not
possible. A transfer can be prevented by recognizing patients with a proximal
intracranial occlusion in the ambulance and taking them immediately to an
intervention center. This can be done through prehospital triage. To this end,
prehospital stroke scores have been developed. These scores can be used to
estimate the risk of an intracranial occlusion. Recently, several of these
prehospital stroke scores have been prospectively validated, showing that some
scores can predict the presence of an intracranial occlusion with good
accuracy. Of these scores, the RACE scale shows good results in several
studies. This was also the case in the PRESTO study, conducted within our ROAZ
region. The RACE scale appeared to perform best of all examined prehospital
scales, with an area under the receiver operating characteristic curve (AUROC)
of 0.83.
Prehospital scores have now been implemented in a number of other regions in
the Netherlands and beyond. The following procedure is used in these regions:
when the total score of the prehospital score is higher than a defined cut-off
point, the decision is made to transport patients directly to the intervention
center. When the score falls below this cut-off point, the patient is
transported to the nearest hospital. This method does not take into account
essential regional characteristics. Previous research showed that not only the
probability of having an intracranial occlusion, but also the driving times to
hospitals, the workflow times of hospitals and the possible transfer times
between hospitals are important factors in determining the optimal
transportation strategy. We bring the aforementioned factors together in a
multivariable decision model, which is made available online in the form of the
Stroke Triage App for ambulance staff to determine the best transport strategy
for each individual patient.

The overall objective of the PRESTO-2 study is to evaluate the implementation
of the Stroke Triage App, a decision support tool for individualized
prehospital triage of patients with potential stroke.
The specific objectives are the following:
1) Evaluation of the effectiveness of the Stroke Triage App implementation in
reducing the time to EVT.
2) Evaluation of the consequences of the implementation of the Stroke Triage
App for the number of ambulance transportations and the distribution of stroke
patients across hospitals in the region.

1. Before-and-after study design with difference-in-differences analysis.
2. Cohort study with simulation.

Implementation of the Stroke Triage App in the intervention region.

The burden on participants is minimal. The study procedures take place in the
prehospital setting and consist of a short neurological examination (2 extra
items compared to current clinical practice) and transport to a hospital. There
is a risk that standard treatment with intravenous thrombolysis will be
minimally delayed, but the time gained to standard treatment with endovascular
thrombectomy is expected to outweigh this. (For more detailed explanations and
percentages, please refer to the answers to E2 (with regard to patient burden)
and to E1, E1a, E9, E9a and E12a (with regard to risk assessment)).