Detecting Dengue fever in children: using sequencing symptom patterns for an online assessment approach

Abstract

Background: Dengue fever (DF) is an important health problem in Asia. We examined it using its clinical symptoms to predict DF.
Methods: We extracted statistically significant features from 17 DF-related clinical symptoms in 177 pediatric patients (69 diagnosed with DF) using (1) the unweighted summation score and (2) the non-parametric H^T person fit statistic, which jointly combine (3) the weighted score (yielded by logistic regression) to predict DF risk.
Results: Six symptoms (Family History, Fever ≥ 39°C, Skin Rash, Petechiae, Abdominal Pain, and Weakness) significantly predicted DF. When a cutoff point of −1.03 (p = 0.26) suggested combining the weighted score and the H^T coefficient, the sensitivity was 0.91 and the specificity was 0.76. The area under the ROC curve was 0.88, which was a better predictor: specificity was 5.56% higher than for the traditional logistic regression.
Conclusions: Six simple symptoms analyzed using logistic regression were useful and valid for early detection of DF risk in children. A better predictive specificity increased after combining the non-parametric H^T coefficient to the weighted regression score. A self-assessment using patient smartphones is available to discriminate DF and may eliminate the need for a costly and time-consuming dengue laboratory test.