Akyon Seyma Handan, Akyon Fatih Cagatay, Yılmaz Tarık Eren
Family Medicine Department, Ankara City Hospital, University of Health Sciences, Ankara, Türkiye.
Graduate School of Informatics, Middle East Technical University, Ankara, Türkiye.
Front Med (Lausanne). 2023 Mar 8;10:1029198. doi: 10.3389/fmed.2023.1029198. eCollection 2023.
The main complications of polypharmacy, which is known as the simultaneous use of more than five drugs, are potentially inappropriate medicines(PIMs), drug-drug, and drug-disease interaction. It is aimed to prepare an auxiliary tool to reduce the complications of polypharmacy and to support rational drug use(RDU), by evaluating the patient with age, drugs, and chronic diseases in this study.
In the first phase of this study, as methodological research, an up-to-date and comprehensive auxiliary tool as a reference method was generated with a database containing interaction information of 430 most commonly used drug agents and chronic diseases in geriatrics in the light of current and valid 6 PIM criteria for geriatric patients, and medication prospectuses, relevant current articles, and guidelines. Then, an artificial intelligence(AI) supported web application was designed and developed to facilitate the practical use of the tool. Afterward, the data of a cross-sectional observational single-center study were used for the rate and time of PIM and drug interaction detection with the web application. The proposed web application is publicly available at https://fastrational.com/.
While the PIM coverage rate with the proposed tool was 75.3%, the PIM coverage rate of EU(7)-PIM, US-FORTA, TIME-to-STOPP, Beers 2019, STOPP, Priscus criteria in the web application database respectively(63.5%-19.5%) from the highest to the lowest. The proposed tool includes all PIMs, drug-drug, and drug-disease interaction information detected with other criteria. A general practitioner detects interactions for a patient without the web application in 2278 s on average, while the time with the web application is decreased to 33.8 s on average, and this situation is statistically significant.
In the literature and this study, the PIM criteria alone are insufficient to include actively used medicines and it shows heterogeneity. In addition, many studies showed that the biggest obstacle to drug regulation in practice is "time constraints." The proposed comprehensive auxiliary tool analyzes age, drugs, and diseases specifically for the patient 60 times faster than the manual method, and it provides quick access to the relevant references, and ultimately supports RDU for the clinician, with the first and only AI-supported web application.
多重用药是指同时使用五种以上药物，其主要并发症包括潜在不适当用药（PIMs）、药物相互作用和药物与疾病相互作用。本研究旨在通过对患者的年龄、药物和慢性病进行评估，制备一种辅助工具，以减少多重用药的并发症并支持合理用药（RDU）。
在本研究的第一阶段，作为方法学研究，根据针对老年患者的6项现行有效PIM标准、药品说明书、相关近期文章和指南，利用一个包含430种老年医学中最常用药物和慢性病相互作用信息的数据库，生成了一个最新且全面的辅助工具作为参考方法。然后，设计并开发了一个人工智能（AI）支持的网络应用程序，以方便该工具的实际使用。之后，将一项横断面观察性单中心研究的数据用于通过该网络应用程序检测PIM和药物相互作用的发生率及时间。所提出的网络应用程序可在https://fastrational.com/上公开获取。
所提出工具的PIM覆盖率为75.3%，而网络应用程序数据库中EU(7)-PIM、美国FORTA、TIME-to-STOPP、2019年Beers标准、STOPP、Priscus标准的PIM覆盖率分别为63.5% - 19.5%，从高到低排列。所提出的工具包含了用其他标准检测到的所有PIM、药物相互作用和药物与疾病相互作用信息。全科医生在没有网络应用程序的情况下平均需要2278秒来检测患者的相互作用，而使用网络应用程序时平均时间降至33.8秒，这种情况具有统计学意义。
在文献和本研究中，仅PIM标准不足以涵盖实际使用的药物，且显示出异质性。此外，许多研究表明，实践中药物监管的最大障碍是“时间限制”。所提出的综合辅助工具针对患者对年龄、药物和疾病进行专门分析的速度比手动方法快60倍，它提供了对相关参考文献的快速访问，最终通过首个也是唯一的AI支持的网络应用程序为临床医生支持合理用药。
