Cheng-Hao Wang
ABSTRACT
In this paper1, we investigated the feasibility of applying deep learning methods to the detection of adverse drug reactions (ADRs) using spontaneous reporting systems (SRS) data. We adopted Convolutional Neural Network (CNN) to extract automatically appropriate features from the FAERS data with the help of a well-known ADR knowledge base, SIDER, to establish a model for future ADR detection for newly marketed drugs. Seven kinds of drugs not listed in SIDER that may cause myocardial infarction from FDA's safety report were considered. We use the instances that recorded these seven drugs as testing sets and detect them by our proposed CNN models. Our results show that if we consider adverse reactions in HLT level of MedDRA, the ADR signals detected by our models were far earlier than the FDA's alerts, also ahead of the time detected by conventional statistics-based approaches.
- A. Bate, M. Lindquist, I.R. Edwards, S. Olsson, R. Orre, A. Lansner, and R.M. De Freitas. 1998. A Bayesian Neural Network Method for Adverse Drug Reaction Signal Generation. European Journal of Clinical Pharmacology 54, 4 (1998), 315--321.Google Scholar
Cross Ref
- N.V. Chawla, K.W. Bowyer, L.O. Hall, and W.P. Kegelmeyer. 2002. SMOTE: Synthetic Minority Over-Sampling Technique. Journal of Artificial Intelligence Research 16, 1 (2002), 321--357.Google ScholarCross Ref
- R. Collobert and J. Weston. 2008. A Unified Architecture for Natural Language Processing: Deep Neural Networks with Multitask Learning. In Proceedings of the 25th International Conference on Machine Learning (ICML'08).Google Scholar
- S. Evans. 2003. Sequential Probability Ratio Tests Applied to Public Health Problems. Controlled Clinical Trials 24, (2003), 67S.Google Scholar
- S.J. Evans, P.C. Waller, and S. Davis. 2001. Use of Proportional Reporting Ratios (PRRS) for Signal Generation from Spontaneous Adverse Drug Reaction Reports. Pharmacoepidemiology and Drug Safety 10, 6 (2001), 483--486.Google ScholarCross Ref
- S.A. Goldman. 1998. Limitations and Strengths of Spontaneous Reports Data. Clinical Therapeutics 20, supp. 3 (1998), 40--44.Google ScholarCross Ref
- A. Krizhevsky, I. Sutskever, and G.E. Hinton. 2012. Imagenet Classification with Deep Convolutional Neural Networks. In Proceedings of the 25th International Conference on Neural Information Processing Systems (NIPS'12), 1097--1105.Google ScholarDigital Library
- W. Y. Lin, H. Y. Li, J. W. Du, W. Y. Feng, C. F. Lo, and V. W. Soo. 2012. iADRs: Towards On-Line Adverse Drug Reaction Analysis. SpringerPlus 1, 1 (2012), arc. 72.Google Scholar
- W. Y. Lin and J. W. Du. 2013. Towards On-Line ADR Detection: An Innovative Data Cube-based Approach. In Proceedings of the 14th World Congress on Medical and Health Informatics (Medinfo 2013), Denmark.Google Scholar
- K. J. Rothman, S. Lanes, and S. T. Sacks. 2004. The Reporting Odds Ratio and its Advantages over the Proportional Reporting Ratio. Pharmacoepidemiology and Drug Safety 13, 8 (2004), 519--523.Google ScholarCross Ref
- E. P. van Puijenbroek, A. Bate, H. G. Leufkens, M. Lindquist, R. Orre, and A. C. Egberts. 2002. A Comparison of Measures of Disproportionality for Signal Detection in Spontaneous Reporting Systems for Adverse Drug Reaction. Pharmacoepidemiology and Drug Safety, 11, 1 (2002), 3--10.Google ScholarCross Ref
- FDA Adverse Event Reporting System (FAERS), Available: http://www.fda.gov/cder/aers/default.htmGoogle Scholar
- MedDRA, Available: https://www.meddra.orgGoogle Scholar
- MedWatch: The FDA Safety Information and Adverse Event Reporting Program, Available: https://www.fda.gov/safety/medwatch-fda-safety-information-and-adverse-event-reporting-programGoogle Scholar
- SIDER4.1, Available: http://sideeffects.embl.deGoogle Scholar
Index Terms
- Deep learning from spontaneous reporting systems data to detect ADR signals
Recommendations
Rough-set-based ADR signaling from spontaneous reporting data with missing values
Display Omitted We inspected the feasibility of rough-set-based ADR detection approach.We proposed twelve rough-set-based measuring methods, showing only six are feasible.We conducted a comparative analysis of the six methods in terms of accuracy.We ...
Mining for Adverse Drug Events on Twitter
IC3K 2014: Proceedings of the International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management - Volume 1At the post-marketing phase when drugs are used by large populations and for long periods, unexpected adverse events may occur altering the risk-benefit relation of drugs, sometimes requiring a regulatory action. These events at the post-marketing phase ...
A High Performance Agent-Based System for Reporting Suspected Adverse Drug Reactions
ITNG '12: Proceedings of the 2012 Ninth International Conference on Information Technology - New GenerationsAdverse drug reactions (ADRs) represent a serious worldwide problem. Current post-marketing ADR detection approaches largely rely on spontaneous reports filed by various healthcare professionals such as physicians, pharmacists et.al.. Underreporting is ...
Comments