Detection of overdose and underdose prescriptions-An unsupervised machine learning approach

• Published in: PloS one Vol. 16; no. 11; p. e0260315
• Main Authors: Nagata, Kenichiro, Tsuji, Toshikazu, Suetsugu, Kimitaka, Muraoka, Kayoko, Watanabe, Hiroyuki, Kanaya, Akiko, Egashira, Nobuaki, Ieiri, Ichiro
• Format: Journal Article
• Language: English
• Published: United States Public Library of Science 19.11.2021; Public Library of Science (PLoS)
• Subjects: Adolescent; Adult; Aged; Aged, 80 and over; Algorithms; Analysis; Anomalies; Biology and Life Sciences; Child, Preschool; Clinical medicine; Comparative analysis; Computer and Information Sciences; Data Analysis; Data Collection - statistics & numerical data; Data Management - statistics & numerical data; Databases, Factual - statistics & numerical data; Drug development; Drug dosages; Drug Overdose - diagnosis; Drug stores; Drugs; Electronic health records; Electronic Health Records - statistics & numerical data; Electronic medical records; Errors; Humans; Infant; Influence; Lagrange multiplier; Learning algorithms; Machine learning; Medical errors; Medicine and Health Sciences; Mental Recall; Middle Aged; Modelling; Outliers (statistics); Overdose; Pediatrics; Performance evaluation; Pharmaceuticals; Pharmacy; Physical Sciences; Prescription drugs; Prescription Drugs - adverse effects; Prescription writing; Prescriptions - statistics & numerical data; Recall; Research and Analysis Methods; Software services; Support Vector Machine - statistics & numerical data; Support vector machines; Unsupervised learning; Unsupervised Machine Learning - statistics & numerical data; Young Adult; Japan
• DOI: 10.1371/journal.pone.0260315
• ISSN: 1932-6203

Overdose prescription errors sometimes cause serious life-threatening adverse drug events, while underdose errors lead to diminished therapeutic effects. Therefore, it is important to detect and prevent these errors. In the present study, we used the one-class support vector machine (OCSVM), one of the most common unsupervised machine learning algorithms for anomaly detection, to identify overdose and underdose prescriptions. We extracted prescription data from electronic health records in Kyushu University Hospital between January 1, 2014 and December 31, 2019. We constructed an OCSVM model for each of the 21 candidate drugs using three features: age, weight, and dose. Clinical overdose and underdose prescriptions, which were identified and rectified by pharmacists before administration, were collected. Synthetic overdose and underdose prescriptions were created using the maximum and minimum doses, defined by drug labels or the UpToDate database. We applied these prescription data to the OCSVM model and evaluated its detection performance. We also performed comparative analysis with other unsupervised outlier detection algorithms (local outlier factor, isolation forest, and robust covariance). Twenty-seven out of 31 clinical overdose and underdose prescriptions (87.1%) were detected as abnormal by the model. The constructed OCSVM models showed high performance for detecting synthetic overdose prescriptions (precision 0.986, recall 0.964, and F-measure 0.973) and synthetic underdose prescriptions (precision 0.980, recall 0.794, and F-measure 0.839). In comparative analysis, OCSVM showed the best performance. Our models detected the majority of clinical overdose and underdose prescriptions and demonstrated high performance in synthetic data analysis. OCSVM models, constructed using features such as age, weight, and dose, are useful for detecting overdose and underdose prescriptions.