Beyond Face Arm and Speech Test (Beyond FAST), a predicted probability-based stroke identification scale

Published Date: 15th May 2024

Publication Authors: Abdul-Rahim. A, Maducolil. JE, Hume. C, Iijima. A

Poster presentation for the ESOC 2024. Full conference programme available here: https://apps.congrex.com/esoc2024/en-GB/pag/

 

Background and aims
Despite its widespread usage in stroke identification, the FAST test exhibits modest accuracy. This study assesses the effectiveness of integrating additional clinical parameters to distinguish strokes and stroke mimics.

Methods
A retrospective analysis of 39 clinical parameters was undertaken, involving patients suspected of having a stroke and brought to two hyperacute stroke units in the UK via '999' calls from 2021 to 2023. Clinical notes were reviewed, and final diagnoses of stroke/TIA or mimic were independently coded by stroke consultants. Data from Whiston Hospital was used for model development, with Countess of Chester Hospital data reserved for validation. Logistic regression was utilised to determine model-predicted probabilities.

Results
The study encompassed 273 patients from Whiston Hospital. Six parameters – increased age, new onset of speech loss, unilateral arm weakness, absence of loss of consciousness, absence of seizures, and presence of neurological signs – were statistically significant in the multivariate model (p<0.05). The validation dataset from the Countess of Chester Hospital included 398 patients. The probability-based model exhibited high sensitivity at the lower cutoff, where patients may not need to travel to comprehensive stroke centres, and high specificity at the higher cutoff level, where calls from the paramedics can be taken directly by the stroke nurses, as illustrated in Table 1, Figures 1 and 2.

Conclusions
Employing a predictive probability scale rather than a binary threshold scale could more effectively guide patients to suitable stroke services. This approach has the potentials to reduce pre-hospital delays and improve stroke identification accuracy.

 

Harfoush, A; Abdul-Rahim, A; Maducolil, JE; Hume, C; Iijima, A et al. (2024). Beyond Face Arm and Speech Test (Beyond FAST), a predicted probability-based stroke identification scale. Conference poster from 10th European Stroke Organisation Conference, 15-17 May 2024, Basel, Switzerland.

 

« Back