Analysis of Adverse Drug Reactions Identified in Nursing Notes Using Reinforcement Learning
1
조회
0
다운로드
1
인용
0
참고문헌
키워드
초록
Objectives: Electronic Health Records (EHRs)-based surveillance systems are being actively developed for detecting adverse drug reactions (ADRs), but this is being hindered by the difficulty of extracting data from unstructured records. This study performed the analysis of ADRs from nursing notes for drug safety surveillance using the temporal difference method in reinforcement learning (TD learning). Methods: Nursing notes of 8,316 patients (4,158 ADR and 4,158 non-ADR cases) admitted to Ajou University Hospital were used for the ADR classification task. A TD(λ) model was used to estimate state values for indicating the ADR risk. For the TD learning, each nursing phrase was encoded into one of seven states, and the state values estimated during training were employed for the subsequent testing phase. We applied logistic regression to the state values from the TD(λ) model for the classification task. Results: The overall accuracy of TD-based logistic regression of 0.63 was comparable to that of two machine-learning methods (0.64 for a naïve Bayes classifier and 0.63 for a support vector machine), while it outperformed two deep learning-based methods (0.58 for a text convolutional neural network and 0.61 for a long short-term memory neural network). Most importantly, it was found that the TD-based method can estimate state values according to the context of nursing phrases. Conclusions: TD learning is a promising approach because it can exploit contextual, time-dependent aspects of the available data and provide an analysis of the severity of ADRs in a fully incremental manner.
Abstract
Objectives: Electronic Health Records (EHRs)-based surveillance systems are being actively developed for detecting adverse drug reactions (ADRs), but this is being hindered by the difficulty of extracting data from unstructured records. This study performed the analysis of ADRs from nursing notes for drug safety surveillance using the temporal difference method in reinforcement learning (TD learning). Methods: Nursing notes of 8,316 patients (4,158 ADR and 4,158 non-ADR cases) admitted to Ajou University Hospital were used for the ADR classification task. A TD(λ) model was used to estimate state values for indicating the ADR risk. For the TD learning, each nursing phrase was encoded into one of seven states, and the state values estimated during training were employed for the subsequent testing phase. We applied logistic regression to the state values from the TD(λ) model for the classification task. Results: The overall accuracy of TD-based logistic regression of 0.63 was comparable to that of two machine-learning methods (0.64 for a naïve Bayes classifier and 0.63 for a support vector machine), while it outperformed two deep learning-based methods (0.58 for a text convolutional neural network and 0.61 for a long short-term memory neural network). Most importantly, it was found that the TD-based method can estimate state values according to the context of nursing phrases. Conclusions: TD learning is a promising approach because it can exploit contextual, time-dependent aspects of the available data and provide an analysis of the severity of ADRs in a fully incremental manner.
참고문헌 (16)
- Jensen PB, "Mining electronic health records : towards better research applications and clinical care", Nat Rev Genet, 13(6), pp.395~405 (2012)
- Oliveira JL, "The EU-ADR Web Platform : delivering advanced pharmacovigilance tools", Pharmacoepidemiol Drug Saf, 22(5), pp.459~467 (2013)
- Kho AN, "Practical challenges in integrating genomic data into the electronic health record", Genet Med, 15(10), pp.772~778 (2013)
- Ahn HJ, "Adverse-drug-event surveillance using narrative nursing records in electronic nursing records", Comput Inform Nurs, 31(1), pp.45~51 (2013)
- Lee S, "Standard-based comprehensive detection of adverse drug reaction signals from nursing statements and laboratory results in electronic health records", J Am Med Inform Assoc, 24(4), pp.697~708 (2017)
- Park MY, "A novel algorithm for detection of adverse drug reaction signals using a hospital electronic medical record database", Pharmacoepidemiol Drug Saf, 20(6), pp.598~607 (2011)
- Jagannatha AN, "Structured prediction models for RNN based sequence labeling in clinical text", Proc Conf Empir Methods Nat Lang Process, 2016, pp.856~865 (2016)
- Hughes M, "Medical text classification using convolutional neural networks", Stud Health Technol Inform, 235, pp.246~250 (2017)
- Sutton RS, "Temporal credit assignment in reinforcement learning", pp.1984
- Van Seijen H, "True online temporal-difference learning", J Mach Learn Res, 17(1), pp.5057~5096 (2016)
- Kim Y, "Convolutional neural networks for sentence classification", Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pp.1746~1751 (2014)
- Ng AY, "Policy invariance under reward transformations: theory and application to reward shaping", Proceedings of the Sixteenth International Conference on Machine Learning (ICML), pp.278~287 (1999)
- Hochreiter S, "Long short-term memory", Neural Comput, 9(8), pp.1735~1780 (1997)
- Trinh TH, "Learning longerterm dependencies in RNNs with auxiliary losses", Proceedings of the 35th International Conference on Machine Learning (ICML), pp.4972~4981 (2018)
- Bahdanau D, "An actor-critic algorithm for sequence prediction", Proceedings of the 5th International Conference on Learning Representations (ICLR) (2017)
- Lopez-Gonzalez E, "Determinants of under-reporting of adverse drug reactions : a systematic review", Drug Saf, 32(1), pp.19~31 (2009)