Visualization of Clinical Research-related Laws Using Co-occurrence Network

Abstract

Current clinical research is required to plan and pursue in accordance with the following three laws: “GCP Ordinance,” “Clinical Trials Act,” and “Ethical Guidelines for Medical and Health Research Involving Human Subjects.” However, it is not easy to understand these laws and regulations completely while managing a busy clinical practice. For this reason, many studies have recently focused on methods to explain, such as classification of each point of laws related to clinical research. It is expected that enough comprehension by simplifying and visualizing law and regulations related clinical research. In the field of law informatics, the use of natural language processing to analyze legal texts and visualize the structure of consisting words lead to the effects such as understanding the relation between laws and reducing management cost. The aim of this study is to visualize the ministerial ordinances, laws, and guidelines related to clinical research. In this paper, the purpose is capturing the outline of laws and regulations related clinical research by visualizing the relation of words constituting letter of law by use of co-occurrence network. The word combinations with particularly strong co-occurrence and common words for several laws were obtained, furthermore the words were focused that common words are co-occurring. As a result, it was predicted at the word level that, the important words for each law, and how are the common words used in, in each law. The evaluation accuracy improvement is expected by applying these relations to evaluation of similarity between laws. Furthermore, by using the visualization results, it is expected the possibility of understanding these laws and regulations related to clinical research from a different perspective until now.