Prediction of adverse drug reactions using demographic and non-clinical drug characteristics in FAERS data

被引:1
作者
Farnoush, Alireza [1 ]
Sedighi-Maman, Zahra [2 ]
Rasoolian, Behnam [3 ]
Heath, Jonathan J. [4 ]
Fallah, Banafsheh [3 ]
机构
[1] Univ South Carolina, Darla Moore Sch Business, Columbia, SC 29208 USA
[2] Georgetown Univ, McDonough Sch Business, Washington, DC 20057 USA
[3] Auburn Univ, Dept Ind & Syst Engn, Auburn, AL 36849 USA
[4] St Bonaventure Univ, Sch Business, Washington, DC 20057 USA
来源
SCIENTIFIC REPORTS | 2024年 / 14卷 / 01期
关键词
Adverse drug reactions; Deep learning; Random forest; Machine learning; Demographic features; Non-clinical features; POLAR SURFACE-AREA; AGE; HOSPITALIZATIONS; LIPOPHILICITY; DISCOVERY; EVENTS; GENDER; RISK;
D O I
10.1038/s41598-024-74505-2
中图分类号
O [数理科学和化学]; P [天文学、地球科学]; Q [生物科学]; N [自然科学总论];
学科分类号
07 ; 0710 ; 09 ;
摘要
The presence of adverse drug reactions (ADRs) is an ongoing public health concern. While traditional methods to discover ADRs are very costly and limited, it is prudent to predict ADRs through non-invasive methods such as machine learning based on existing data. Although various studies exist regarding ADR prediction using non-clinical data, a process that leverages both demographic and non-clinical data for ADR prediction is missing. In addition, the importance of individual features in ADR prediction has yet to be fully explored. This study aims to develop an ADR prediction model based on demographic and non-clinical data, where we identify the highest contributing factors. We focus our efforts on 30 common and severe ADRs reported to the Food and Drug Administration (FDA) between 2012 and 2023. We have developed a random forest (RF) and deep learning (DL) machine learning model that ingests demographic data (e.g., Age and Gender of patients) and non-clinical data, which includes chemical, molecular, and biological drug characteristics. We successfully unified both demographic and non-clinical data sources within a complete dataset regarding ADR prediction. Model performances were assessed via the area under the receiver operating characteristic curve (AUC) and the mean average precision (MAP). We demonstrated that our parsimonious models, which include only the top 20 most important features comprising 5 demographic features and 15 non-clinical features (13 molecular and 2 biological), achieve ADR prediction performance comparable to a less practical, feature-rich model consisting of all 2,315 features. Specifically, our models achieved an AUC of 0.611 and 0.674 for RF and DL algorithms, respectively. We hope our research provides researchers and clinicians with valuable insights and facilitates future research designs by identifying top ADR predictors (including demographic information) and practical parsimonious models.
引用
收藏
页数:11
相关论文
共 72 条
[1]  
Alberti P, 2014, METHODS MOL BIOL, V1175, P301, DOI 10.1007/978-1-4939-0956-8_12
[2]   Factors affecting the development of adverse drug reactions (Review article) [J].
Alomar, Muaed Jamal .
SAUDI PHARMACEUTICAL JOURNAL, 2014, 22 (02) :83-94
[3]   Adverse drug reaction monitoring: support for pharmacovigilance at a tertiary care hospital in Northern Brazil [J].
Alves de Araujo Lobo, Marcia Germana ;
Botelho Pinheiro, Sandra Maria ;
Diaz Castro, Jose Gerley ;
Momente, Valeria Gomes ;
Pranchevicius, Maria-Cristina S. .
BMC PHARMACOLOGY & TOXICOLOGY, 2013, 14
[4]  
[Anonymous], 2003, Exploratory data mining and data cleaning
[5]  
Arnott J.A., 2013, J. Appl. Biopharm. Pharmacokinet, V1, P31, DOI DOI 10.14205/2309-4435.2013.01.01.6
[6]   The influence of lipophilicity in drug discovery and design [J].
Arnott, John A. ;
Planey, Sonia Lobo .
EXPERT OPINION ON DRUG DISCOVERY, 2012, 7 (10) :863-875
[7]   A curated and standardized adverse drug event resource to accelerate drug safety research [J].
Banda, Juan M. ;
Evans, Lee ;
Vanguri, Rami S. ;
Tatonetti, Nicholas P. ;
Ryan, Patrick B. ;
Shah, Nigam H. .
SCIENTIFIC DATA, 2016, 3
[8]  
Bhuiyan P., 2001, ICH Harmonised Tripartite Guideline: guideline for good clinical practice
[9]   Random forests [J].
Breiman, L .
MACHINE LEARNING, 2001, 45 (01) :5-32
[10]   Side effects of opioids during short-term administration: Effect of age, gender, and race [J].
Cepeda, MS ;
Farrar, JT ;
Baumgarten, M ;
Boston, R ;
Carr, DB ;
Strom, BL .
CLINICAL PHARMACOLOGY & THERAPEUTICS, 2003, 74 (02) :102-112