Paper Info
Title
iDrug - Pediatric Drug Interaction Modeling and Risk Evaluation Leveraging Prescription Big Data.
Abstract
Drug interaction caused by irrational co-prescribed drugs have become a serious problem in pediatric groups, due to the vulnerability of childrenu0027s biological and psychological status and the lack dedicated pediatric drug instructions and standards. Traditional prescription review systems fails to provide accurate drug interaction risk evaluation when applied to the pediatric groups, since the pediatric drug information database used in these systems is usually incomplete and outdated. In this work, we propose iDrug, a data-driven approach to pediatric drug interaction modeling and risk evaluation. First, we propose a graph model to capture the drug co-prescription patterns from historical prescription data, and augment it with the existing drug instruction database to build a drug interaction graph. Then, we exploit a community detection algorithm to identify the intrinsic structures of the graph. Finally, we propose a drug interaction risk evaluation metric based on the community weights of the graph. We evaluate our framework with large-scale, real-world anonymized prescription big data collected from a tertiary hospital for two years. Results show that the risk evaluation approach achieves an accuracy above 84.2% in the test prescription set, and outperforms other baseline methods.
Year
DOI
Venue
2019
10.1109/SmartWorld-UIC-ATC-SCALCOM-IOP-SCI.2019.00155
SmartWorld/SCALCOM/UIC/ATC/CBDCom/IOP/SCI
DocType
Citations 
PageRank 
Conference
0
0.34
References 
Authors
0
9
Name
Order
Citations
PageRank
Yunting Shao100.34
Linghong Hong200.34
Tianqi Xie300.34
Binbin Zhou400.34
Jinzhun Wu500.34
Ming Cheng65413.93
Cheng Wang711829.56
jonathan li811.03
Longbiao Chen900.34