Ramanbhai Patel College of Pharmacy, Charotar University of Science and Technology, Vadodara, India.
Meteroic Biopharmaceuticals, Ahmedabad, India.
Rev Recent Clin Trials. 2021;16(3):242-257. doi: 10.2174/1574887115666201202110919.
Immediately after the outbreak of nCoV, many clinical trials are registered for COVID-19. The numbers of registrations are now raising inordinately. It is challenging to understand which research areas are explored in this massive pool of clinical studies. If such information can be compiled, then it is easy to explore new research studies for possible contributions in COVID-19 research.
In the present work, a text-mining technique of artificial intelligence is utilized to map the research domains explored through the clinical trials of COVID-19. With the help of the open-- source and graphical user interface-based tool, 3007 clinical trials are analyzed here. The dataset is acquired from the international clinical trial registry platform of WHO. With the help of hierarchical cluster analysis, the clinical trials were grouped according to their common research studies. These clusters are analyzed manually using their word clouds for understanding the scientific area of a particular cluster. The scientific fields of clinical studies are comprehensively reviewed and discussed based on this analysis.
More than three-thousand clinical trials are grouped in 212 clusters by hierarchical cluster analysis. Manual intervention of these clusters using their individual word-cloud helped to identify various scientific areas which are explored in COVID19 related clinical studies.
The text-mining is an easy and fastest way to explore many registered clinical trials. In our study, thirteen major clusters or research areas were identified in which the majority of clinical trials were registered. Many other uncategorized clinical studies were also identified as "miscellaneous studies". The clinical trials within the individual cluster were studied, and their research purposes are compiled comprehensively in the present work.
新冠疫情爆发后，许多临床试验被迅速注册用于研究 COVID-19。目前注册数量急剧增加，要理解在如此庞大的临床试验群体中，哪些研究领域正在被探索是极具挑战性的。如果能够获取这些信息，就可以方便地探索新的研究领域，为 COVID-19 研究做出可能的贡献。
本研究利用人工智能的文本挖掘技术，对 COVID-19 临床试验所探索的研究领域进行图谱绘制。借助开源、基于图形用户界面的工具，我们对 3007 项临床试验进行了分析。数据集来自世卫组织的国际临床试验注册平台。借助分层聚类分析，根据共同的研究，将临床试验进行分组。通过手动分析这些聚类的词云，了解特定聚类的科学领域。基于此分析，对临床试验的科学领域进行了全面的综述和讨论。
通过分层聚类分析，超过 3000 项临床试验被分为 212 个组。通过对这些聚类进行手动干预，利用其各自的词云，可以识别出在 COVID19 相关临床试验中探索的各种科学领域。
文本挖掘是探索大量已注册临床试验的简单快捷方法。在我们的研究中，确定了 13 个主要的聚类或研究领域，其中大多数临床试验都在这些领域中注册。还确定了许多其他未分类的临床试验作为“杂项研究”。对各个聚类中的临床试验进行了研究，并在本工作中对其研究目的进行了综合编译。
