A simple and rapid focal seizure screening tool in resource-limited setting

Abstract

Background & Objective: A substantial proportion of patients were managed by general clinicians in resource-limited settings. Therefore, we aim to develop a simple and rapid focal seizure screening tool that can assist non-neurologists in seizure classification. Methods: We conducted a self-administered qualitative questionnaire study on seizure manifestation developed based on the validated comprehensive questionnaires by Reutens et al. for clinical diagnosis of seizures. Logistic regression analyses were conducted to determine the features of seizure which form the items of a focal seizure screening tool. A receiver operating characteristic (ROC) curve analysis was conducted to determine the optimal cut-off score. Results: Among the 199 subjects, 145 (73%) had focal seizures. Three categories of symptoms, i.e., (i) aura (fear, déjà vu, and epigastric aura), (ii) unilateral motor phenomena, and (iii) oral automatism, are significantly associated with the diagnosis of focal seizure. The logistic regression model containing these 3 items / features was statistically significant, χ2 (3, N=199) = 22.93, p<0.001, and correctly classified 71.4% of cases. A focal seizure screening tool was developed by assigning a score of 1 to each category of symptoms. The area under the curve of ROC curve is 0.706 (95% CI: 0.629-0.783), p<0.001. A score of ≥2 is the most optimal cut-off with a sensitivity of 50.3%, specificity of 85.2%, positive predictive value (PPV) of 91.0%, and negative predictive value (NPV) of 41.8%. Conclusion: A simple and rapid self-administered focal seizure screening tool was established with high specificity and PPV.