Stroke recognition in medical emergency calls: A novel sensitivity definition as a basis for developing artificial intelligence decision support
Bibliographic record
Abstract
Abstract Background The sensitivity of emergency medical communication centers (EMCC) for stroke detection varies widely. However, few studies offer detailed insights into the entirety of prehospital pathways in patients with stroke. Therefore, this study aimed to lay the foundation for artificial intelligence (AI) decision support tools in EMCCs by exploring their ability to detect strokes in medical emergency calls, describe a novel method for stroke sensitivity calculation in the EMCC, and identify factors associated with stroke recognition during a call. Methods In total, 1,164 patients with stroke in the catchment area of Bergen EMCC in 2018 and 2019 were included, and a dataset from the EMCC was established manually and linked with data from the Norwegian Stroke Registry (NSR) for analysis. Descriptive statistics, Chi-square test for categorical variables, Mann–Whitney U test for continuous variables, and multivariate logistic regression (LR) were performed on data obtained from patients primarily assessed by EMCC (n=838). Results Using a novel method, we found a stroke detection sensitivity of 76.8% in our study, compared to the 63.4% when using the traditional sensitivity detection method. LR analysis showed a positive association between stroke suspicion and ischemic strokes (odds ratio [OR]=0.317 [0.209–0.481]; p<0,001, with ischemic stroke as the reference) and wake-up strokes (OR=1.716 [1.110–2.653]; p=0.015). Among the NSR symptoms, only aphasia/dysarthria was positively associated with stroke suspicion (OR=1.600 [1.087–2.353]; p=0.017), while leg paresis (OR=0.609 [0.390–0.953]; p=0.009) and vertigo (OR=0.376 [0.204–0.694]; p=0.002) were negatively associated. Conclusions This study introduced a novel and more accurate method for calculating EMCC stroke sensitivity, which is relevant for developing decision support tools, such as AI. Moreover, we identified factors of particular interest for future EMCC research that are relevant to developing AI decision-support tools. Clinical trials https://clinicaltrials.gov/study/NCT04648449
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.005 | 0.008 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".