Prognostic Prediction of Pediatric DHF in Two Hospitals in Thailand
6
Issued Date
2023-01-01
Resource Type
ISSN
03029743
eISSN
16113349
Scopus ID
2-s2.0-85164003286
Journal Title
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume
13897 LNAI
Start Page
303
End Page
312
Rights Holder(s)
SCOPUS
Bibliographic Citation
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) Vol.13897 LNAI (2023) , 303-312
Suggested Citation
Haddawy P., Yin M.S., Meth P., Srikaew A., Wavemanee C., Niyom S.L., Sriraksa K., Limpitikul W., Kittirat P., Malasit P., Avirutnan P., Mairiang D. Prognostic Prediction of Pediatric DHF in Two Hospitals in Thailand. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) Vol.13897 LNAI (2023) , 303-312. 312. doi:10.1007/978-3-031-34344-5_36 Retrieved from: https://repository.li.mahidol.ac.th/handle/123456789/87897
Title
Prognostic Prediction of Pediatric DHF in Two Hospitals in Thailand
Other Contributor(s)
Abstract
Dengue virus infection is a major global health problem. While dengue fever rarely results in serious complications, the more severe illness dengue hemorrhagic fever (DHF) has a significant mortality rate due to the associated plasma leakage. Proper care thus requires identifying patients with DHF among those with suspected dengue so that they can be provided with adequate and prompt fluid replacement. In this paper, we use 18 years of pediatric patient data collected prospectively from two hospitals in Thailand to develop models to predict DHF among patients with suspected dengue. The best model using pooled data from both hospitals achieved an AUC of 0.92. We then investigate the generalizability of the models by constructing a model for one hospital and testing it on the other, a question that has not yet been adequately explored in the literature on DHF prediction. For some models, we find significant degradation in performance. We show this is due to differences in attribute values among the two hospital patient populations. Possible sources of this are differences in the definition of attributes and differences in the pathogenesis of the disease among the two sub-populations. We conclude that while high predictive accuracy is possible, care must be taken when seeking to apply DHF predictive models from one clinical setting to another.