Boland Mary Regina, Miotto Riccardo, Weng Chunhua
Department of Biomedical Informatics, Columbia University, New York, NY, USA.
Stud Health Technol Inform. 2013;192:481-5.
Clinical trial eligibility criteria define fine-grained characteristics of research volunteers for various disease trials and hence are a promising data source for disease profiling. This paper explores the feasibility of using disease-specific common eligibility features (CEFs) for representing diseases and understanding their relatedness. We extracted disease-specific CEFs from eligibility criteria on ClinicalTrials.gov for three illustrative categories - cancers, mental disorders and chronic diseases - each including seven diseases. We then constructed disease-specific CEF networks to assess the degree of overlap among the diseases. Using these automatically derived networks, we observed several findings that were confirmed in medicine. For example, we highlighted connections among schizophrenia, epilepsy and depression. We also identified a link between Crohn's disease and arthritis. These observations confirm the value of using clinical trial eligibility criteria for identifying disease relatedness. We further discuss the implications of CEFs for standardizing clinical trial eligibility criteria through reuse.
临床试验资格标准定义了各类疾病试验中研究志愿者的细粒度特征，因此是疾病概况分析的一个有前景的数据源。本文探讨了使用特定疾病的通用资格特征（CEF）来表征疾病并理解它们之间关联性的可行性。我们从ClinicalTrials.gov上的资格标准中提取了特定疾病的CEF，用于三个说明性类别——癌症、精神障碍和慢性病——每个类别包括七种疾病。然后，我们构建了特定疾病的CEF网络，以评估疾病之间的重叠程度。利用这些自动生成的网络，我们观察到了一些在医学上得到证实的发现。例如，我们突出了精神分裂症、癫痫和抑郁症之间的联系。我们还确定了克罗恩病和关节炎之间的关联。这些观察结果证实了使用临床试验资格标准来识别疾病关联性的价值。我们进一步讨论了CEF通过重复使用对标准化临床试验资格标准的影响。
