Why the study?
Successful acute stroke intervention depends on early hospitalization, motivating examination of the factors influencing the time interval from stroke onset to hospitalization and whether expanding the treatment time window translates into more treatment.
Observational (n=724)
Random division into training and validation sets at a 7:3 ratio
Yes
Effect estimate: AUC 0.959 (95% CI 0.926-0.992)
Absolute Event Rate: 0.959% vs 0.935%
p-value: p=<0.05
XGBoost model shows high accuracy for severe post-stroke dysphagia; hypothesis-generating and requires prospective validation before clinical use.
OBJECTIVE: Successful acute stroke intervention depends on early hospitalization. The time interval between stroke onset to hospitalization was examined to identify the factors influencing the interval and also to determine whether treatment time window expansion will translate into more treatment. METHODS: Data was obtained from Takashima Stroke Registry covering approximately 55,000 residents in Japan. During 1988-2002, the interval between stroke onset and hospitalization was available for 1,446 registered patients (men 761, women 685). Multivariate regression analyses were performed to evaluate the factors influencing early and late admission. RESULTS: The proportions of the patients who arrived within 3 h of onset were 61.4%. Only 8.7% of the patients arrived during a 3- to 6-hour interval. A substantial proportion (29.9%) of patients arrived after 6 h. Early stroke hospitalization was significantly associated with the severity of the stroke event and time of stroke onset. CONCLUSION: Clinically more severe stroke patients were hospitalized earlier and nocturnal strokes had delayed admission. Even expanding the therapeutic time window from </=3 to </=6 h, there are possibilities that a substantial proportion of patients would not benefit from acute intervention. Future research should focus not only on developing therapies for expanding the treatment time window, but also place emphasis on reducing the interval between onset and hospitalization.
No takes yet. Share an insight, caveat, or question.
Turin et al. (2009) conducted an observational in Dysphagia after ischemic stroke (n=724). XGBoost machine-learning model vs. SVM, RF, DT, KNN, and LGBM machine-learning models was evaluated on Prediction of severe dysphagia (Water Swallow Test Grades IV-V) in the external test set (AUC 0.959, 95% CI 0.926-0.992, p=<0.05).
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: