BACKGROUND
With the increasing variety of drugs, the incidence of Adverse Drug Events (ADEs) is increasing year by year. Massive ADEs are recorded in Electronic Medical Records and Adverse Drug Reaction (ADR) reports which are important source of potential ADRs information.
OBJECTIVE
Meanwhile, it is essential to make latent ADR information to be available automatically for better post-marketing drug safety reevaluation and pharmacovigilance. This present study describes how to identify ADR-related information from Chinese ADE reports.
METHODS
Our study established an efficient automated tool, named BBC-Radical (BBC-Radical is model that consists of three components – Bidirectional Encoder Representations from Transformers (BERT), Bidirectional Long-Short Term Memory (Bi-LSTM), and Conditional Random Field (CRF)) model to identify ADR-related information from Chinese ADR reports. Token features and radical features of Chinese characters are used to represent the common meaning of a group of words. BERT and Bi-LSTM-CRF were novel models combined these features to conduct Named Entity Recognition (NER) tasks in the free-text section of 24,890 ADR reports of Jiangsu Province Adverse Drug Reaction Monitoring Center from 2010 to 2016. Moreover, the Man-Machine comparison experiment on the ADE records of Drum Tower hospital was designed to compare the NER performance between BBC-Radical model and manual method.
RESULTS
The NER model achieved relatively high performance of Precision of 96.41%, Recall of 96.03%, and F1 score of 96.22%. It was indicated the performance of the BBC-Radical model (Precision: 87.17%, Recall: 85.69%, and F1 score: 86.43%) is much better than that of the manual method (Precision: 86.1%, Recall: 73.8%, and F1 score: 79.5%) in the recognition task of each kind of entity.
CONCLUSIONS
The proposed model shows competition in ADR related information extraction from
ADE reports and the results suggested that the application of our method in the information extraction of ADR related information is of great significance in improving the quality of ADR reports and post-marketing drug safety evaluation.