| Summary: | 碩士 === 國立陽明大學 === 生物醫學資訊研究所 === 104 === Because the price for genome sequencing has lowered drastically, "precision medicine" becomes feasible. To reach the goal, it is necessary to integrate the bio- and clinical information into a knowledge network. De-identification is a typical method to protect the privacy of a donor. We have established an information platform, which serves as an honest broker in the data integration process. This platform uses a global unique identifier (GUID) approach to de-identify data. The sub-primed GUIDs of the same person are different in different centers. These identifiers are converted to the original GUID and then use the same key to create a new sub-primed GUID on the honest broker platform. These new identifiers are unique for each person. This method has been successfully used to discover three specimen donors, who donated specimen in two centers of a multi-centered “bio-specimen bank on women’s cancers” in Taiwan.
Such a knowledge network may facilitate the translation of genomic discovery to medical applications. The effect of any medical applications needs to go through the clinical trial to collect evidence for regulatory approval. In October 2016 and 2017, Japan and the U.S. regulatory agencies require submittal of data must comply with CDISC standards, respectively. The U.S. Food and Drug Administration (FDA) provided a set of rules to verify the integrity of submittal data. Therefore, we have classified these rules and separated the compliance checking in stages. When FDA releases new rules in the future, these rules will be classified and put into the rule engine. The first stage of the compliance checking is to make sure the case report form included all the standard-required fields. This step ensures that all dependent information will be collected. The second stage is to check whether the filled data are complied with the standards. The third stage is a final check of integrated data. This stage-wise approach uses standards from the start of a trial, which reduces the efforts of time consuming data clean up and compliance checking at the end of a trial.
This study established a mechanism to integrate de-identified data. Besides, the compliance of CDISC standards in a clinical trial was achieved by using a step-wise approach to shorten the process of preparing data for regulatory approval.
|
|---|
