2023 Japan Interchange Program
Please click on sessions listed below to view all presenters and topics within each session.
Session 1: Opening Plenary
Morning Break
Session 2: Second Plenary
Lunch Break
Session 3: Academic Research
Background:
The purpose of implementing CDISC standards is to function as part of an ecosystem formed by diverse stakeholders and resources that interact with each other to develop pharmaceuticals and medical devices. The submission of applications for drug and medical device approval is also part of this ecosystem.
Purpose:
CDISC standard SDTM is a standardized format for sharing and exchanging clinical trial data, and it plays an important role in building the ecosystem. In this study, we use current SDTM IG, TAS, and other CDISC standards-related papers to verify whether CDISC standards function correctly as part of the ecosystem when creating SDTM.
Methods:
SDTM modeling was conducted by three groups of 24 CJUG-SDTM members using protocols from investigator-initiated clinical trials for solid tumors, and the consistency and differences of SDTM modeling were verified.
Results:
Although it was only a partial case, there were cases where the modeling of all SDTMs did not match in the three groups. There were two main factors for the differences. One was the absence of cases that matched the reference materials for trial-specific items or disease-specific items. The other was the presence of multiple modeling cases for the same item in the reference materials.
Conclusion:
Currently, the development of pharmaceuticals and medical devices is in a transitional phase of expansion due to new technologies, systems, regulations, and market changes. As a result, new initiatives such as the application of CDISC standards to RWD and observational studies are also progressing. To function as part of the ecosystem of pharmaceutical and medical device development, continuous expansion of CDISC standards is necessary. When expanding, it is also necessary to consider compatibility and consistency with existing CDISC standards.
Background. In clinical trials, the Study Data Tabulation Model (SDTM) is widely used for the tabulation of clinical trial data for regulatory submission in the U.S., Japan, and other countries. However, the Clinical Data Acquisition Standards Harmonization (CDASH), which supports data collection in Case Report Form (CRF), is less widely used than SDTM. SDTM can be technically generated from any structure of CRF data but occasionally requires complex mapping and programming, especially when CDASH is not used. We suspect that one of the reasons why CDASH is not used so widely is due to a lack of understanding.
Objective. The purpose of this presentation is to introduce the benefit of adopting CDASH. We also introduce the website “CDISC eCRF Portal,” released by CDISC in 2021, and the REDCap Shared Library (RSL), released by CDISC and the REDCap project in 2022, as the latest free resources for adopting CDASH.
Methods. Our regulatory sub-team consisted of experts who have been involved in SDTM implementation in pharmaceutical companies, contract research organizations, software development companies, and academia. The member reviewed the previous reports regarding CDASH on the web and shared their opinions. Moreover, we investigated the CDISC eCRF portal and RSL, a shared CRF data library for the REDCap EDC system.
Results. We found the presentation entitled “SDTM and CDASH: Why You Need Both,” which summarized the differences between the models and is publicly available on the CDISC official website. According to this presentation, while most data is the same in both standards, each standard is designed for different purposes; therefore, differences exist both in philosophy and implementation. For example, SDTM assumes the data is clean, and missing data is assumed to be verified. CDASH only assumes that no data was received, which must be verified. This presentation also explained the differences in both models from the viewpoint of non-standard variables and human- versus machine-readable data. We then examined the “CDISC eCRF Portal” website, which shares ready-to-use, CDASH-compliant, annotated eCRFs in PDF, HTML, and XML format, to use as is or import into an EDC system for customization. The eCRF data are currently integrated into the library of the OpenClinica and REDCap EDC systems. By accessing the RSL and reviewing the eCRF data, we found that most eCRF data are nearly ready to use but need customization.
Discussions. Differences between CDASH and SDTM exist due to their different purposes, and CDASH supports more simply maps eCRF data into SDTM. For CROs, the adoption of CDASH by each sponsor should reduce the cost of developing eCRF and SDTM and provide benefits to sponsors. For most of academia, which has not yet adopted CDISC standards, a feasible first step would be to adopt CDASH first, rather than SDTM.
Conclusion: Both CDASH and SDTM are beneficial for standardizing clinical research data.
Phamaceuticals and Medical Devices Agency(PMDA) has been accepting electronic data for regulatory submission since October 1, 2016, and with the end of the transition period on March 31, 2020. Many pharmaceutical companies and Contract Research Organizations(CROs) have established their internal systems and process to comply with CDISC standards. However, the CDISC standard has not been spread widely among Academic Research Organizations(AROs) in Japan. One of the reasons is AROs do not have enough opportunities to be directly involved in the regulatory submission process, therefore they do not think they need to implement CDISC standards to go so far as devote resources. According to Nagai et al. (2022) responses from 41 institutions including public universities, private universities, and other institutions, the challenge to implement CDISC standards are “ human and economic effort required, lack of resources”, "Lack of human resources with CDISC knowledge" and "Not knowing how to implement the CDISC standard”. This result does not mean that AROs are not interested in CDISC, but rather that they would like to implement CDISC. Unfortunately, they do not have enough knowledge and skills on how to do so, and are therefore unable to take the step to do so. Therefore, we have established an "SDTM subteam for CDISC beginners" within the CDISC Japan User Group(CJUG) SDTM team, mainly consists of clinical research support staff affiliated with AROs, starting in December 2022. This main purpose of our subteam is to contribute to human resource development to promote CDISC implementation in AROs through knowledge and skills on CDISC by practicing SDTM data set creation through a mock protocol. In total, 65 people have inquired about the subteam's activities since December 2022, which is the establishment of the subteam. As of March 2023, the subteam consists of 39 members, indicating a high level of interest in the implementation of CDISC standards in the AROs. Currently, the subteam is working on the divided into two teams to construct a Trial Design Model and to create an eCRF. In this session, we would like to present the progression of the subteam's activities.
Afternoon Break
Session 4: Real World Data & Regulatory Presentations/Perspectives
Sharing of participant-level data from completed clinical trials has the potential to accelerate scientific progress, answer new lines of scientific inquiry and support reproducibility. Vivli is an independent, non-profit organization that has developed a global data-sharing and analytics platform. Patient-level data is available from 7,000 clinical trials that are provided by academic funders, pharmaceutical companies and charitable funders as part of the FAIR ecosystem. More than 80% of the data available in Vivli is formatted in the CDISC-SDTM standard. This talk will highlight the benefits of data sharing using case studies from data contributors who have using the Vivli platform to have researchers access their high-quality data and contribute to scientific discovery.
Evening Networking Event
Session 5: Data Science
As a required document in FDA submission package, annotated Case Report Form (aCRF) is helpful in data transcription from source documents. According to Version 2.0 of the Study Data Tabulation Model Metadata Submission Guidelines: Human Clinical Trials (SDTM-MSG-V2.0) published in April 2021, a new request for Table of Contents(TOC) for submitted aCRF has been raised. Over the past couple of years, though the time-consuming process led to the exploration of automation tools in the pharmaceutical industry, few papers have talked about the whole process of aCRF production from SDTM mapping to TOC, including automatically updating the page numbers in define specifications after the insertion of TOC. Specifically, in studies with multiple database modifications, the integrated process assures consistency and high quality of submission documents in that it links up the raw data, define specifications, CRF annotation and TOC together, and reduces manual issues through semi-automation process. In this paper, we will show the integrated process of semi-automatically generating an annotated CRF, with dual bookmarking and TOC that meets the SDTM-MSG-V2.0 standard, which involves commonly used software/tools in most pharmaceutical companies, mainly SAS, with the support of Visual Basic for Applications (VBA) and Adobe Acrobat.
Blockchain Technology is known as distributed ledger technology and it was developed as the core technology of Bitcoin.
Blockchain is a growing list of records called blocks that are linked using cryptography.
The characteristic of this technology is decentralized and immutable system.
In addition, it is possible to incorporate program called smart contract in Blockchain and it enables automation of pre-defined transaction.
In recent years, Blockchain technology is expected to be applied to the processes in various industries.
Similarly, in pharma industry, many ideas to make use of Blockchain are being considered.
In this presentation, I will introduce the basic technology overview of Blockchain, the use case of Blockchain in pharmaceutical industry and the overview of blockchain system considered by CJUG-SDTM (CDISC Japan User Group SDTM team) Blockchain sub team.
Our team considered that appropriate data access rights management in clinical data sharing is useful by combining Blockchain technology, encryption technology, and distributed database mechanisms, and proceeded to system verification.
Based on the results of this verification, we will present xthe advantages and challenges of this mechanism.
Morning Break
Session 6: Standards Implementation and Driving Automation
Automation of SDTM and ADaM in clinical trials is a popular topic in the pharmaceutical industry nowadays. Many companies have a well-established standards team, which is defining CDISC and regulatory authority compliant standards within their company. Naturally, standardization of clinical data leads to automation initiatives.
In this presentation, we will share Sanofi’s innovative effort in standardizing the derivation or transformation specifications for data creation, with the intent to facilitate machine automation. We will introduce pseudo-codes, as a simple yet standard structured way of writing specifications, so that the programmer (either person or a machine) can understand and interpret the specifications consistently. This presentation will provide examples of pseudo-code to demonstrate how it creates clear, specific, consistent, unambiguous derivation rules to enable automation. The end-to-end flow and the level of automation will also be discussed.
Defining and developing the standard specification is essential for automation to improve the quality of generated data, reduce programming time, and simplify mundane redundant tasks.
Keywords: standards, automation, pseudo-code, specification, derivation, transformation, ADaM, SDTM, metadata
Digital Data Flow (DDF) is the approach to accelerate start-up of clinical studies by creating digital study protocol and automating subsequent processes such as setting up clinical systems. CDISC Unified Study Definitions Model (USDM) is a standard model for the development of Study Definitions Repository (SDR), a repository of digital study protocols. DDF currently assumes to populate SDR from a study builder/authoring tool or TransCelerate eCPT (electronic Common Protocol Template).
Creating a study protocol using a study builder will require changes to current process where an author creates a study protocol as a document, and such changes could cause protocol authors less efficient authoring experience. Enabling population of SDR contents automatically from a study protocol document, although this is technically challenging, will help implementation of DDF while leaving as-is efficient authoring process unchanged.
This presentation introduces sample implementation of extraction of study definitions data from study protocol documents leveraging AI-based Natural Language Processing (NLP) techniques, and discusses how extracted data will help automating subsequent processes such as setting up EDC system, creating SDTM Trial Design datasets, and retrieving Biomedical Concepts for SDTM specification.
Lunch Break
Session 7: Global Regulatory Submissions
In BIMO, the sponsor submits data that describes the characteristics and outcomes of clinical investigations at the site level as part of the New Drug Application (NDA)/ Investigational New Drug (IND) Application.
Right now, the industry relies on BIMO Conformance Guide Submission of NDA, FDA-BIMO Checklist and various other clinical research papers as reference for such conversions. But there are still some areas that are grey. Along with conversion challenges there are also operational challenges to consider. Recently we were involved in a BIMO conversions and learned a few tips and tricks, including a template that could simplify the process for different type of studies. It is this experience we would like to share as part of this presentation.
As more pharmaceutical companies move towards open-source initiatives (especially R), a lot of training and programming efforts are in place to replace or augment similar activities traditionally done using SAS. For statistical programmers working on submission studies to FDA or PMDA, the requirement for Define-XML to provide details for the Analysis Results Metadata (ARM) section means programmers might encounter having to do this for the first time, without a lot of available resources or guidance to do so. In this paper, we will discuss a few methods to generate or update ARM using R, issues encountered, and some possible solutions or strategies to overcome them.
This presentation aims to raise awareness about the underutilized FDA guidance documents for CDISC implementers. In the presentation we will delve into key topics covered in these additional resources and we will share our experiences with the FDA, and in particular with some of their divisions, commenting on our data packages by referencing these guidance.
Afternoon Break
Session 8: Closing Plenary
- Akira Soma, Oracle
- Dr. Wenjun Bao, JMP Clinical
- Dr. Yuki Ando, PMDA
- Dr. Sam Hume, CDISC
- Rhonda Facile, CDISC
- Peter Van Reusel, CDISC