1School of Information Management, Nanjing University, Nanjing 210023, China 2Institute of Medical Information, Chinese Academy of Medical Sciences, Beijing 100020, China 3Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, Beijing 100730, China
【目的】在线健康社区为药物不良反应信号监测提供了新的信息源,从患者对不良反应的评论数据中识别药物不良反应信号有助于为抗糖尿病类药物不良反应的监测预警提供参考。【方法】以Ask a Patient网站中患者的抗糖尿病药物不良反应评论为数据源,结合自然语言处理技术和UMLS、MedDRA等词表对患者评论数据进行规范化处理和映射,构建药物-不良反应共现矩阵,采用PRR法识别符合信号监测阈值的药物-不良反应对并对抽取结果进行专家判读,最后以Drugs.com作为金标准对方法的有效性进行评估。【结果】共计识别出539组药物-不良反应对,整体识别准确率达85%,金标准整体吻合率达82%,说明该方法具有一定的有效性。【局限】 因MedDRA中纳入了部分检查、手术操作、社会环境等类型的与ADR无关的术语,对ADR术语的识别准确率产生一定影响。【结论】本研究采用的药物不良反应信号识别思路,丰富了药物不良反应信号监测渠道和方法,为药物不良反应信号监测提供了参考。
[Objective] Online health communities provide new information for detecting adverse drug reaction (ADR) signals. This study identifies ADR signals from patients’ reviews and generates early warnings for potential side-effects of antidiabetic drugs. [Methods] First, we retrieved patients’ reviews (adverse reactions) on antidiabetic drugs from Ask a Patient website. Then, we combined natural language processing techniques and lexicons (UMLS and MedDRA) to normalize and map these reviews. Third, we constructed a drug-ADR co-occurrence matrix and used the PRR method to identify drug-ADR pairs meeting the signal detection threshold. Finally, we invited expert to interpret the extracted results, which were evaluated with Drugs.com standards. [Results] A total of 539 drug-ADR pairs were identified, with an overall identification accuracy of 85% and recall of 82%. [Limitations] The accuracy of identifying ADR terms was affected by the inclusion of non-ADR terms, such as examination, surgical operation, and social environment from MedDRA. [Conclusions] The proposed model enriches the data sources and methods of ADR signal detection.
郭进京, 夏光辉, 黄奇, 何丽云, 张化冰. 基于在线健康社区的药物不良反应信号识别方法研究*[J]. 数据分析与知识发现, 2022, 6(7): 70-86.
Guo Jinjing, Xia Guanghui, Huang Qi, He Liyun, Zhang Huabing. Detecting Signals of Adverse Drug Reactions with Data from Online Health Community. Data Analysis and Knowledge Discovery, 2022, 6(7): 70-86.
Drug reaction with eosinophilia and systemic symptoms
10.04
7
Bone spur
8.60
3
Dislocation
8.60
3
Gingival pain
7.17
5
Malignant breast neoplasm
5.73
4
Muscle stiffness
4.81
47
Sciatica
3.32
11
Stiffness
3.24
48
Decompression sickness
3.19
15
Atrial fibrillation
3.13
6
Joint stiffness
2.87
10
Laceration
2.37
12
3
瑞舒伐他汀-CRESTOR
Tietze’s Syndrome
15.89
3
Electric shock
15.89
3
Radiating pain
5.30
4
Gout
3.71
7
4
辛伐他汀-ZOCOR
Oliguria
15.20
4
Lymphadenopathy
7.60
4
Bursitis
4.84
7
Mg++ increased
3.80
5
Heterosexuality
3.62
10
Shock
2.80
7
5
左甲状腺素钠-SYNTHROID
Dry skin
35.62
48
Racing thoughts
28.94
3
Premenstrual syndrome
24.12
5
Dry hair
16.08
5
Carpal tunnel syndrome
9.65
6
Endometriosis
9.65
5
Hashimoto’s disease
9.65
3
Photophobia
8.04
5
Suicide
6.43
4
Hypertrichosis
4.70
39
Drying
4.48
13
Dryness of eyes
4.29
12
Upper respiratory infection
3.57
10
Cerebration impaired
3.38
7
Libido decreased
2.89
9
6
二甲双胍-GLUCOPHAGE
Taste metallic
32.64
18
Thirst
13.30
10
Arsenical keratosis
2.78
6
Hypertrichosis
2.45
13
7
依折麦布-ZETIA
Pulmonary congestion
19.65
3
Diverticulitis
13.10
3
Inflammation mucous membrane
7.86
3
Cardiac flutter
6.55
5
Night sweats
4.37
11
Middle insomnia
4.37
6
Anxiety attack
4.09
5
Bursitis
3.74
4
8
吡格列酮-ACTOS
Severe pre-eclampsia
14.71
8
Increased appetite
7.03
5
Dyspnea
6.66
44
Mg++ increased
5.98
3
Epistaxis
5.98
3
Renal pain
5.12
3
9
非诺贝特-TRICOR
Slurred speech
7.48
3
Skin greasy
5.96
5
10
利拉鲁肽-VICTOZA
Rhinorrhea
6.90
3
Ataxia Telangiectasia
5.67
4
Xerostomia
3.92
5
11
普伐他汀-PRAVACHOL
Nephritis
6.07
6
12
艾塞那肽-BYETTA
Anorexia
13.78
3
Ataxia Telangiectasia
7.07
4
13
唑来膦酸-RECLAST
Exostosis
119.83
3
Paronychia
11.41
4
Inguinal pain
8.56
3
Death NOS
4.56
4
Heterosexuality
4.28
3
14
罗格列酮-AVANDIA
Sluggishness
30.32
4
Heart fluttering
27.29
3
15
西格列汀-JANUVIA
Rhinorrhea
21.04
3
Poor weight gain
15.30
3
Table 8 部分需要关注的ADR警戒信号
[1]
魏巍. 药物不良反应知识发现与利用模型研究[D]. 武汉: 武汉大学, 2017.
[1]
( Wei Wei. Study on the Model of Knowledge Discovery and Utilization in Adverse Drug Reaction[D]. Wuhan: Wuhan University, 2017.)
[2]
Yang C C, Yang H D, Jiang L. Postmarketing Drug Safety Surveillance Using Publicly Available Health-Consumer-Contributed Content in Social Media[J]. ACM Transactions on Management Information Systems, 2014, 5(1): 1-21.
[3]
Pappa D, Stergioulas L K. Harnessing Social Media Data for Pharmacovigilance: A Review of Current State of the Art, Challenges and Future Directions[J]. International Journal of Data Science and Analytics, 2019, 8(2): 113-135.
doi: 10.1007/s41060-019-00175-3
[4]
World Health Organization. The Importance of Pharmacovigilance[EB/OL]. [2021-10-25]. https://apps.who.int/iris/bitstream/handle/10665/42493/a75646.pdf?sequence=1&isAllowed=y.
[5]
World Health Organization. Guidelines for Good Clinical Practice(GCP) for Trials on Pharmaceutical Products[EB/OL]. [2021-10-25]. https://digicollections.net/medicinedocs/printable-whozip13e#d/whozip13e.
[6]
World Health Organization. WHO Guidelines on Safety Monitoring of Herbal Medicines in Pharmacovigilance Systems[EB/OL]. [2021-10-25]. https://apps.who.int/iris/bitstream/handle/10665/43034/9241592214_eng.pdf.
[7]
Han J, Kamber M, Pei J. Data Mining: Concepts and Techniques[M]. The 3rd Edition. Waltham: Elsevier, 2012: 5-8.
[8]
Karapetiantz P, Bellet F, Audeh B, et al. Descriptions of Adverse Drug Reactions are Less Informative in Forums than in the French Pharmacovigilance Database but Provide More Unexpected Reactions[J]. Frontiers in Pharmacology, 2018, 9: 439.
doi: 10.3389/fphar.2018.00439
pmid: 29765326
[9]
Government of Canada. About GPHIN[EB/OL]. [2021-10-25] https://gphin.canada.ca/cepr/aboutgphin-rmispenbref.jsp?language=en_CA.
[10]
Leaman R, Wojtulewicz L, Sullivan R, et al. Towards Internet-Age Pharmacovigilance: Extracting Adverse Drug Reactions from User Posts to Health-Related Social Networks[C]// Proceedings of the 2010 Workshop on Biomedical Natural Language Processing. 2010: 117-125.
[11]
Sarker A, Ginn R, Nikfarjam A, et al. Utilizing Social Media Data for Pharmacovigilance: A Review[J]. Journal of Biomedical Informatics, 2015, 54: 202-212.
doi: 10.1016/j.jbi.2015.02.004
pmid: 25720841
[12]
Sarker A, Ginn R, Nikfarjam A, et al. Mining Twitter for Adverse Drug Reaction Mentions: A Corpus and Classification Benchmark[C]// Proceedings of the 4th Workshop on Building and Evaluating Resources for Health and Biomedical Text Processing. 2014.
[13]
Tang Y X, Yang J S, Ang P S, et al. Detecting Adverse Drug Reactions in Discharge Summaries of Electronic Medical Records Using Readpeer[J]. International Journal of Medical Informatics, 2019, 128: 62-70.
doi: 10.1016/j.ijmedinf.2019.04.017
( Zhao Mingzhen, Lin Hongfei, Xu Bo, et al. Potential Adverse Drug Reactions Discovery from Social Networks[J]. Journal of Chinese Information Processing, 2017, 31(5): 194-202.)
[15]
Rezaei Z, Ebrahimpour-Komleh H, Eslami B, et al. Adverse Drug Reaction Detection in Social Media by Deep Learning Methods[J]. Cell Journal, 2020, 22(3): 319-324.
doi: 10.22074/cellj.2020.6615
pmid: 31863657
[16]
Yogita, Sangma J W, Anal S R N, et al. Clustering-Based Hybrid Approach for Identifying Quantitative Multidimensional Associations Between Patient Attributes, Drugs and Adverse Drug Reactions[J]. Interdisciplinary Sciences, Computational Life Sciences, 2020, 12(3): 237-251.
doi: 10.1007/s12539-020-00365-9
[17]
Harpaz R, DuMouchel W, Shah N H, et al. Novel Data-Mining Methodologies for Adverse Drug Event Discovery and Analysis[J]. Clinical Pharmacology & Therapeutics, 2012, 91(6): 1010-1021.
[18]
Karapetiantz P, Audeh B, Lillo-Le Louët A, et al. Signal Detection for Baclofen in Web Forums: A Preliminary Study[J]. Studies in Health Technology and Informatics, 2018, 247: 421-425.
pmid: 29677995
( Lin Xin, Guo Jinjing, Ren Huiling. Analysis of the Building of Adverse Drug Reaction Term System[J]. Journal of Medical Informatics, 2019, 40(6): 60-65.)
[22]
Hammond I W, Rich D S, Gibbs T G. Effect of Consumer Reporting on Signal Detection: Using Disproportionality Analysis[J]. Expert Opinion on Drug Safety, 2007, 6(6): 705-712.
pmid: 17967159
( Sun Yalin, Li Yongchang, Du Wenmin, et al. Analysis of Application of Disproportional Measures in Pharmacovigilance[J]. Chinese Journal of Pharmacoepidemiology, 2009, 18(3): 147-150.)
[24]
Evans S J W, Waller P C, Davis S. Use of Proportional Reporting Ratios(PRRS) for Signal Generation from Spontaneous Adverse Drug Reaction Reports[J]. Pharmacoepidemiology and Drug Safety, 2001, 10(6): 483-486.
pmid: 11828828
[25]
van Puijenbroek E P, Bate A, Leufkens H G M, et al. A Comparison of Measures of Disproportionality for Signal Detection in Spontaneous Reporting Systems for Adverse Drug Reactions[J]. Pharmacoepidemiology and Drug Safety, 2002, 11(1): 3-10.
pmid: 11998548
[26]
Bate A, Lindquist M, Edwards I R, et al. A Bayesian Neural Network Method for Adverse Drug Reaction Signal Generation[J]. European Journal of Clinical Pharmacology, 1998, 54(4): 315-321.
pmid: 9696956
[27]
DuMouchel W, Pregibon D. Empirical Bayes Screening for Multi-Item Associations[C]// Proceedings of the 7th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2001: 57-76.
[28]
International Diabetes Federation. Diabetes Around the World in 2021[EB/OL]. [2021-11-19]. https://diabetesatlas.org/idfawp/resource-files/2021/11/IDFDA10-global-fact-sheet.pdf.
[29]
International Diabetes Federation. Diabetes in Western Pacific in 2021[EB/OL]. [2021-11-19]. https://diabetesatlas.org/idfawp/resource-files/2021/11/IDF-Atlas-Factsheet-2021_WP.pdf.
( Tang Wenlu, Wang Yongming, Du Wenmin, et al. Analysis of Adverse Drug Reaction of Antidiabetic Agents According to Documents within Forty-Two Years in China[J]. Chinese Journal of New Drugs and Clinical Remedies, 2002, 21(12): 753-758.)
( Yin Chengxia, Xu Fengquan. Analysis of the Use and Adverse Reaction of Diabetic Drugs[J]. China Practical Medicine, 2020, 15(1): 179-180.)
[32]
James f, Kasia L. Metformin in 2019[J]. The Journal of the American Medical Association, 2019, 321(19): 1926-1927.
doi: 10.1001/jama.2019.3805
[33]
Flory J H, Hennessy S, Bailey C J, et al. Reports of Lactic Acidosis Attributed to Metformin, 2015-2018[J]. Diabetes Care, 2020, 43(1): 244-246.
doi: 10.2337/dc19-0923
[34]
Dungan K M, Povedano S T, Forst T, et al. Once-Weekly Dulaglutide Versus Once-Daily Liraglutide in Metformin-Treated Patients with Type 2 Diabetes(AWARD-6): A Randomised, Open-Label, Phase 3, Non-Inferiority Trial[J]. The Lancet, 2014, 384(9951): 1349-1357.
doi: 10.1016/S0140-6736(14)60976-4
[35]
Álvarez-Villalobos N, Treviño-Alvarez A M, González-González J. Liraglutide and Cardiovascular Outcomes in Type 2 Diabetes[J]. New England Journal of Medicine, 2016, 375(18): 1797-1799.
doi: 10.1056/NEJMc1611289
[36]
Holman R R, Bethel M A, Mentz R J, et al. Effects of Once-Weekly Exenatide on Cardiovascular Outcomes in Type 2 Diabetes[J]. The New England Journal of Medicine, 2017, 377(13): 1228-1239.
doi: 10.1056/NEJMoa1612917
pmid: 28910237
[37]
Andersen A, Lund A, Knop F K, et al. Glucagon-Like Peptide 1 in Health and Disease[J]. Nature Reviews Endocrinology, 2018, 14(7): 390-403.
doi: 10.1038/s41574-018-0016-2
pmid: 29728598
[38]
FDA Approved Drug Products[EB/OL]. [2021-10-25]. https://www.accessdata.fda.gov/drugsatfda_docs/label/2020/022341s036lbl.pdf.