Latent profile analysis of emergency department presentation characteristics in patients with acute ischemic stroke.
Frontiers in neurology · 2026-01-01 · Observational study
Abstract
OBJECTIVE: This study aimed to identify potential profile types characterizing emergency department presentations among patients with acute ischemic stroke (AIS), analyze differences in demographic, clinical features, and clinical outcomes across profiles, and explore associations between presentation characteristics and prognosis (modified Rankin Scale, mRS) to inform precision intervention strategies. METHODS: Using convenience sampling, 220 AIS patients admitted to our hospital from January to December 2024 were enrolled. The Latent Profile Analysis (LPA) model was expanded to include eight indicators capturing multidimensional prehospital behavioral constructs: time from onset to symptom recognition, time from symptom recognition to decision-making, time from onset to presentation, and Stroke Knowledge Awareness Score, emergency medical services (EMS) utilization, living alone status, residential area, and hospital transfer pathway. Latent Profile Analysis (LPA) was performed using Mplus 8.0 software incorporating all eight indicators. Demographic characteristics, clinical features, and clinical outcomes were compared across profiles using χ2 tests and Kruskal-Wallis H tests. Spearman correlation analysis assessed the strength of associations between presentation indicators and mRS scores. Multivariate ordinal logistic regression was conducted to examine the independent prognostic value of latent profile membership for mRS, adjusting for age, baseline NIHSS score, comorbidity burden, and onset-to-presentation time. RESULTS: LPA successfully identified four latent profiles of emergency department presentation characteristics among AIS patients, with optimal model fit (AIC = 7354.379, BIC = 7452.794, Entropy = 0.976). Based on the expanded indicator set, these profiles were recharacterized as: rapid ems-activated profile (31.36%, n = 69), delayed recognition-self-presentation profile (27.73%, n = 61), family involvement-cautious decision-making profile (16.82%, n = 37), and multiple barriers-delayed presentation profile (24.09%, n = 53). Significant differences existed among the four profiles in age, education level, residence, living alone status, number of comorbidities, onset time, and NIHSS score (p < 0.05). Regarding clinical outcomes, the rapid decision-making-emergency transport group exhibited the highest reperfusion therapy rate (63.77%) and the highest favorable prognosis rate (72.46%). while the multiple barriers-delayed presentation group had the lowest reperfusion rate (7.55%) and highest poor prognosis rate (86.79%) (p < 0.001). Correlation analysis revealed significant positive correlations between mRS scores and time from onset to presentation (r = 0.440), time from onset to symptom recognition (r = 0.446), and time from symptom recognition to decision-making (r = 0.458; all p < 0.001). Stroke knowledge scores showed a significant negative correlation with mRS scores (r = -0.431, p < 0.001). Multivariate regression confirmed latent profile membership as an independent prognostic factor. Compared with the Rapid EMS-Activated Profile, the Multiple Barriers-Delayed Presentation Profile was associated with higher odds of poor prognosis (OR = 4.87, 95%CI:2.92-8.13, p < 0.001), independent of age, NIHSS, comorbidity burden, and onset-to-presentation time. CONCLUSION: Emergency presentation patterns among AIS patients exhibit significant heterogeneity, allowing classification into four latent profiles with distinct behavioral characteristics. Patients across profiles demonstrate marked differences in presentation efficiency, stroke knowledge levels, and clinical outcomes. Latent profile classification independently predicts functional prognosis beyond traditional clinical confounders. Developing stratified intervention strategies based on core profile characteristics can help reduce presentation delays, enhance treatment efficacy, and provide critical guidance for optimizing emergency management of AIS.