Polypharmacy is prevalent among older adults, often leading to adverse drug events (ADEs). In the context of polypharmacy, certain ADEs have a low incidence during the use of a single drug, but are triggered only by specific drug combinations, thereby manifesting as rare events.  While traditional statistical models suffer from limited assumptions and data processing capacity, machine learning significantly improves prediction efficacy, particularly for drug-drug interactions. This review aims to summarize recent advances in applying machine learning to predict ADEs associated with polypharmacy, focusing on case studies at both the molecular and population levels. At the molecular level, graph convolutional networks (Decagon) predict drug side effects via multimodal biological networks, while tensor factorization (SimplE) substantially reduces training time. The dual-view substructure learning network for drug-drug interaction prediction further enhances DDI prediction by integrating atomic substructures and interaction relationships. Population-level studies using electronic health records employ random forests algorithms and graph neural networks for ADE prediction across drug classes. However, this field still faces many challenges, such as insufficient model interpretability, strict requirements for data quality, and barriers to cross-institutional data sharing. In the future, causal inference and machine learning technology can be integrated to achieve accurate evaluation of the safety of personalized treatment, thereby effectively reducing the risk of ADE associated with polypharmacy.

1.Masnoon N, Shakib S, Kalisch-Ellett L, et al. What is polypharmacy? A systematic review of definitions[J]. BMC Geriatrics, 2017, 17(1): 230. DOI: 10.1186/s12877-017-0621-2.

2.Pazan F, Wehling M. Polypharmacy in older adults: a narrative review of definitions, epidemiology and consequences[J]. Eur Geriatr Med, 2021, 12(3): 443-452. DOI: 10.1007/s41999-021-00479-3.

3.Mehta RS, Kochar BD, Kennelty K, et al. Emerging approaches to polypharmacy among older adults[J]. Nat Aging, 2021, 1(4): 347-356. DOI: 10.1038/s43587-021-00045-3.

4.Wang X, Liu K, Shirai K, et al. Prevalence and trends of polypharmacy in U.S. adults, 1999-2018[J]. Glob Health Res Policy, 2023, 8(1): 25. DOI: 10.1186/s41256-023-00311-4.

5.《医养结合机构衰弱老年人多重用药安全管理中国专家共识（2022版）》编写组, 中国老年医学学会医养结合促进委员会. 医养结合机构衰弱老年人多重用药安全管理中国专家共识（2022版）[J]. 中国心血管杂志, 2022, 27(5): 403-410. [Writing Group of Chinese Expert Consensus on Safety Management of Polypharmacy for Frail Elderly People in the Institutions of Combination of Medical and Senior Health Care (2022 version), Committee for the Promotion of Combination of Medical and Senior Health Care of Chinese Geriatrics Society. Chinese expert consensus on safety management of polypharmacy for frail elderly people in the institutions of combination of medical and senior health care (2022 version)[J]. Chinese Journal of Cardiovascular Medicine, 2022, 27(5): 403-410.] DOI: 10.3969/j.issn.1007-5410.2022.05.002.

6.Stevenson JM, Williams JL, Burnham TG, et al. Predicting adverse drug reactions in older adults; a systematic review of the risk prediction models[J]. Clin Interv Aging, 2014, 9: 1581-1593. DOI: 10.2147/CIA.S65475.

7.Rozenblum R, Rodriguez-Monguio R, Volk LA, et al. Using a machine learning system to identify and prevent medication prescribing errors: a clinical and cost analysis evaluation[J]. Jt Comm J Qual Patient Saf, 2020, 46(1): 3-10. DOI: 10.1016/j.jcjq.2019.09.008.

8.Vo TH, Nguyen NTK, Kha QH, et al. On the road to explainable AI in drug-drug interactions prediction: a systematic review[J]. Comput Struct Biotechnol J, 2022, 20: 2112-2123. DOI: 10.1016/j.csbj.2022.04.021.

9.张倩, 李沭, 李朋梅, 等. 美国老年医学会老年人潜在不适当用药Beers标准2023版解读[J]. 中国全科医学, 2023, 26(35): 4372-4381. [Zhang Q, Li Shu, Li PM, et al. Interpretation of the 2023 AGS Beers Criteria: Potentially Inappropriate Medication Use in Older Adults[J]. Chinese General Practice, 2023, 26(35): 4372-4381.] DOI: 10.12114/j.issn.1007-9572.2023.0336.

10.Sommer J, Viviani R, Wozniak J, et al. Dealing with adverse drug reactions in the context of polypharmacy using regression models[J]. Sci Rep, 2024, 14(1): 27355. DOI: 10.1038/s41598-024-78474-4.

11.Badillo S, Banfai B, Birzele F, et al. An introduction to machine learning[J]. Clin Pharmacol Ther, 2020, 107(4): 871-885. DOI: 10.1002/cpt.1796.

12.Greener JG, Kandathil SM, Moffat L, et al. A guide to machine learning for biologists[J]. Nat Rev Mol Cell Biol, 2022, 23(1): 40-55. DOI: 10.1038/s41580-021-00407-0.

13.Ngiam KY, Khor IW. Big data and machine learning algorithms for health-care delivery[J]. Lancet Oncol, 2019, 20(5): e262-e273. DOI: 10.1016/S1470-2045(19)30149-4.

14.Hu Q, Tian F, Jin Z, et al. Developing a warning model of potentially inappropriate medications in older Chinese outpatients in tertiary hospitals: a machine-learning study[J]. J Clin Med, 2023, 12(7): 2619. DOI: 10.3390/jcm12072619.

15.Li YF, Zhao WC, Bo D, et al. Research on adverse drug reaction prediction model combining knowledge graph embedding and deep learning[C]. Zhuhai, China: 2024 4th International Conference on Machine Learning and Intelligent Systems Engineering (MLISE), 2024: 322-329. DOI: 10.1109/MLISE62164.2024.10674360.

16.Zhang S, Tong H, Xu J, et al. Graph convolutional networks: a comprehensive review[J]. Comput Soc Netw, 2019, 6(1): 11. DOI: 10.1186/s40649-019-0069-y.

17.Lukashina N, Kartysheva E, Spjuth O, et al. SimVec: predicting polypharmacy side effects for new drugs[J]. J Cheminform, 2022, 14(1): 49. DOI: 10.1186/s13321-022-00632-5.

18.Xue R, Liao J, Shao X, et al. Prediction of adverse drug reactions by combining biomedical tripartite network and graph representation model[J]. Chem Res Toxicol, 2020, 33(1): 202-210. DOI: 10.1021/acs.chemrestox.9b00238.

19.Wang Y, Liu J, Zhao Y, et al. Multi-relational hierarchical embedding with multi-graph neural networks for drug-drug interaction prediction[C]. Rochester, NY: Social Science Research Network, 2025: 1. DOI: 10.2139/ssrn.5136595.

20.Gao Y, Zhang X, Sun Z, et al. Precision adverse drug reactions prediction with heterogeneous graph neural network[J]. Adv Sci (Weinh), 2024, 12(4): 2404671. DOI: 10.1002/advs.202404671.

21.Zitnik M, Agrawal M, Leskovec J. Modeling polypharmacy side effects with graph convolutional networks[J]. Bioinformatics, 2018, 34(13): i457-i466. DOI: 10.1093/bioinformatics/bty294.

22.Stage TB, Brøsen K, Christensen MMH. A comprehensive review of drug-drug interactions with metformin[J]. Clin Pharmacokinet, 2015, 54(8): 811-824. DOI: 10.1007/s40262-015-0270-6.

23.Nováček V, Mohamed SK. Predicting polypharmacy side-effects using knowledge graph embeddings[J]. AMIA Jt Summits Transl Sci Proc, 2020, 2020: 449-458. https://pubmed.ncbi.nlm.nih.gov/32477666/.

24.Masumshah R, Aghdam R, Eslahchi C. A neural network-based method for polypharmacy side effects prediction[J]. BMC Bioinformatics, 2021, 22(1): 385. DOI: 10.1186/s12859-021-04298-y.

25.Lloyd O, Liu Y, Gaunt TR. Fast polypharmacy side effect prediction using tensor factorization[J]. Bioinformatics, 2024, 40(12): btae706. DOI: 10.1093/bioinformatics/btae706.

26.Li Z, Zhu S, Shao B, et al. DSN-DDI: an accurate and generalized framework for drug-drug interaction prediction by dual-view representation learning[J]. Brief Bioinform, 2023, 24(1): bbac597. DOI: 10.1093/bib/bbac597.

27.Mullard A. 2020 FDA drug approvals[J]. Nat Rev Drug Discov, 2021, 20(2): 85-90. DOI: 10.1038/d41573-021-00002-0.

28.Hansen PW, Clemmensen L, Sehested TSG, et al. Identifying drug-drug interactions by data mining: a pilot study of warfarin-associated drug interactions[J]. Circ Cardiovasc Qual Outcomes, 2016, 9(6): 621-628. DOI: 10.1161/CIRCOUTCOMES. 116.003055.

29.Chen J, Wu G, Michelson A, et al. Mining reported adverse events induced by potential opioid-drug interactions[J]. JAMIA Open, 2020, 3(1): 104-112. DOI: 10.1093/jamiaopen/ooz073.

30.Dara ON, Ibrahim AA, Mohammed TA. Advancing medical imaging: detecting polypharmacy and adverse drug effects with Graph Convolutional Networks (GCN)[J]. BMC Medical Imaging, 2024, 24(1): 174. DOI: 10.1186/s12880-024-01349-7.

31.Vimbi V, Shaffi N, Mahmud M. Interpreting artificial intelligence models: a systematic review on the application of LIME and SHAP in Alzheimer's disease detection[J]. Brain Inform, 2024, 11(1): 10. DOI: 10.1186/s40708-024-00222-1.

32.Ponce-Bobadilla AV, Schmitt V, Maier CS, et al. Practical guide to SHAP analysis: explaining supervised machine learning model predictions in drug development[J]. Clin Transl Sci, 2024, 17(11): e70056. DOI: 10.1111/cts.70056.

33.Rodríguez-Terol A, Caraballo MO, Palma D, et al. Quality of interaction database management systems[J]. Farm Hosp, 2009, 33(3): 134-146. https://pubmed.ncbi.nlm.nih.gov/19712597/.

34.Kanehisa M, Goto S, Furumichi M, et al. KEGG for representation and analysis of molecular networks involving diseases and drugs[J]. Nucleic Acids Res, 2010, 38(Database issue): D355-D360. DOI: 10.1093/nar/gkp896.

35.Feuerriegel S, Frauen D, Melnychuk V, et al. Causal machine learning for predicting treatment outcomes[J]. Nat Med, 2024, 30(4): 958-968. DOI: 10.1038/s41591-024-02902-1.

36.Wang YR, Li HX, Zhu MQ, et al. Causal inference with complex treatments: a survey[J]. J ACM, 2023, 37(4): 111. https://doi.org/10.48550/arXiv.2407.14022.

37.Hernán MA, Robins JM. Using big data to emulate a target trial when a randomized trial is not available[J]. Am J Epidemiol, 2016, 183(8): 758-764. DOI: 10.1093/aje/kwv254.

38.Bica I, Alaa AM, Lambert C, et al. From real-world patient data to individualized treatment effects using machine learning: current and future methods to address underlying challenges[J]. Clin Pharmacol Ther, 2021, 109(1): 87-100. DOI: 10.1002/cpt.1907.

39.Jiao L, Wang Y, Liu X, et al. Causal inference meets deep learning: a comprehensive survey[J]. Research (Wash D C), 2024, 7: 0467. DOI: 10.34133/research.0467.

40.Hung A, Kim YH, Pavon JM. Deprescribing in older adults with polypharmacy[J]. BMJ, 2024, 385: e074892. DOI: 10.1136/bmj-2023-074892.