Shionogi & Co., Ltd - 2025 Customer Recognition Awards: Innovative Problem Solver
Shionogi & Co., Ltd
Contact: Satoshi Sakai
Country: Japan
Award Category: Innovative Problem Solver
Tell us about the business problem you were trying to solve.
The problem that the SHIONOGI Data Science Department is addressing is the efficiency and transparency of Real World Data (RWD) analysis to generate Real World Evidence(RWE).
Ensuring this efficiency and transparency is a challenge not only for SHIONOGI but for all companies that perform RWD analysis.
This is because, unlike clinical trials, data limitations, quality levels, and a wide variety of data should be handled for analysis, and technology is required to fairly evaluate and fuse them.
And in the RWD analysis process, third-party oversight is generally lacking, which leads to suspicions of biased analysis and the resulting perception of low-quality RWE.
FDA requires transparency in RWD analysis for regulatory submissions.
To accelerate the generation of RWE that meets FDA standards, significant issues related to efficiency and transparency must be addressed.
Therefore, we have taken the following approaches to each of these issues.
1. Efficiency
Although we place great importance on how to quickly collect and process the purchased RWD while ensuring its quality and connecting it to analysis, our existing on-premise integrated analysis environment did not have enough machine power or storage, making it inefficient.
Therefore, in February 2024, we built an integrated analysis environment in the AWS cloud environment, introduced Snowflake, and seamlessly linked it to SAS Viya to create an environment capable of mass-producing high-quality evidence.
In addition to building an analysis environment, we are also taking advantage of our accumulated knowledge of RWD analysis to build a system called "AI-SAS for RWE" that semi-automatically creates analysis materials and analysis programs, aiming to significantly improve efficiency (launching soon).
2. Transparency
By semi-automatically creating analysis materials and linking with GitHub, we have achieved automatic recording of the creation of analysis materials and the analysis process.
This allows us to record that analysis materials have been created in advance and when and what analysis results were obtained, building a system that contributes to improving the transparency of research and ultimately the quality of research.
Ensuring this efficiency and transparency is a challenge not only for SHIONOGI but for all companies that perform RWD analysis.
This is because, unlike clinical trials, data limitations, quality levels, and a wide variety of data should be handled for analysis, and technology is required to fairly evaluate and fuse them.
And in the RWD analysis process, third-party oversight is generally lacking, which leads to suspicions of biased analysis and the resulting perception of low-quality RWE.
FDA requires transparency in RWD analysis for regulatory submissions.
To accelerate the generation of RWE that meets FDA standards, significant issues related to efficiency and transparency must be addressed.
Therefore, we have taken the following approaches to each of these issues.
1. Efficiency
Although we place great importance on how to quickly collect and process the purchased RWD while ensuring its quality and connecting it to analysis, our existing on-premise integrated analysis environment did not have enough machine power or storage, making it inefficient.
Therefore, in February 2024, we built an integrated analysis environment in the AWS cloud environment, introduced Snowflake, and seamlessly linked it to SAS Viya to create an environment capable of mass-producing high-quality evidence.
In addition to building an analysis environment, we are also taking advantage of our accumulated knowledge of RWD analysis to build a system called "AI-SAS for RWE" that semi-automatically creates analysis materials and analysis programs, aiming to significantly improve efficiency (launching soon).
2. Transparency
By semi-automatically creating analysis materials and linking with GitHub, we have achieved automatic recording of the creation of analysis materials and the analysis process.
This allows us to record that analysis materials have been created in advance and when and what analysis results were obtained, building a system that contributes to improving the transparency of research and ultimately the quality of research.
What SAS products did you use and how did you use them?
We adopted SAS Viya, which allows for easy implementation of machine learning and deep learning, to build our system.
We recognized the importance of pre-defining the contents of the analysis in a protocol or Statistical Analysis Plan(SAP) before the analysis, creating a report of the analysis results, and ensuring consistency between document creation and analysis timing for transparency.
As our future scope, we plan to utilize generative AI technology, which strongly supports researchers in creating protocols and SAPs from their research questions, to improve the efficiency of document creation.
For the implementation of analytical tasks, we are leveraging our know-how in AI-SAS, which semi-automatically generates programs from past specifications and mock-ups.
We decided to record these processes using GitHub within the system.
This approach has made it possible to achieve both efficiency and transparency.
We recognized the importance of pre-defining the contents of the analysis in a protocol or Statistical Analysis Plan(SAP) before the analysis, creating a report of the analysis results, and ensuring consistency between document creation and analysis timing for transparency.
As our future scope, we plan to utilize generative AI technology, which strongly supports researchers in creating protocols and SAPs from their research questions, to improve the efficiency of document creation.
For the implementation of analytical tasks, we are leveraging our know-how in AI-SAS, which semi-automatically generates programs from past specifications and mock-ups.
We decided to record these processes using GitHub within the system.
This approach has made it possible to achieve both efficiency and transparency.
What were the results or outcomes?
1. Optimizing data distribution with Snowflake and streamlining analysis with SAS Viya
1) As an example of accelerating digital transformation within SHIONOGI, various data held by the value chain is processed by data scientists so that they can be analyzed immediately, then stored in Snowflake, where it is analyzed using SAS Viya and other tools, and the results are quickly fed back, greatly contributing to data-driven business.
2) The combination of Snowflake and SAS Viya has significantly improved analysis speed by up to 70% compared to the conventional on-premises environment, speeding up analysis output.
This effort led to the company winning the Data Driver of the Year award, the highest award in the data cloud field in Japan, at the DATA DRIVERS WINNER 2024 sponsored by Snowflake Japan.
3) Due to the significant improvement in analysis speed, many analysis results using SAS Viya have been linked to business decision-making, and in terms of academic achievements, many analysis-related and clinical research results using SAS Viya have been accepted as peer-reviewed papers.
2. Building AI-SAS for RWE (to be announced at SAS Innovate 2025)
1) By converting various RWDs with different table structures and data formats into a common table structure and data format, we succeeded in standardizing the data.
2) We built a system that semi-automatically creates analytical materials and programs from a simple analysis setting, generates RWE from standardized data and automatically records them.
3) By using this AI-SAS for RWE for analysis, we expect to be able to reduce the time required for conventional RWD analysis by approximately 50%.
1) As an example of accelerating digital transformation within SHIONOGI, various data held by the value chain is processed by data scientists so that they can be analyzed immediately, then stored in Snowflake, where it is analyzed using SAS Viya and other tools, and the results are quickly fed back, greatly contributing to data-driven business.
2) The combination of Snowflake and SAS Viya has significantly improved analysis speed by up to 70% compared to the conventional on-premises environment, speeding up analysis output.
This effort led to the company winning the Data Driver of the Year award, the highest award in the data cloud field in Japan, at the DATA DRIVERS WINNER 2024 sponsored by Snowflake Japan.
3) Due to the significant improvement in analysis speed, many analysis results using SAS Viya have been linked to business decision-making, and in terms of academic achievements, many analysis-related and clinical research results using SAS Viya have been accepted as peer-reviewed papers.
2. Building AI-SAS for RWE (to be announced at SAS Innovate 2025)
1) By converting various RWDs with different table structures and data formats into a common table structure and data format, we succeeded in standardizing the data.
2) We built a system that semi-automatically creates analytical materials and programs from a simple analysis setting, generates RWE from standardized data and automatically records them.
3) By using this AI-SAS for RWE for analysis, we expect to be able to reduce the time required for conventional RWD analysis by approximately 50%.
Why is this approach innovative?
We believe that automating and standardizing the analysis process using AI-SAS for RWE will increase both efficiency and transparency and ensure compliance with regulatory standards.
Addressing these challenges means improving the quality of RWE itself, which is important for the entire healthcare industry.
Therefore, we do not intend to make this service exclusive to SHIONOGI, but to quickly roll it out to the entire healthcare industry.
We believe that rolling out this service will accelerate the creation of pharmaceutical evidence.
As a result, we are confident that it will lead to improved quality of medical care and the contribution it will make to patients and their families, and this is a very innovative initiative.
Addressing these challenges means improving the quality of RWE itself, which is important for the entire healthcare industry.
Therefore, we do not intend to make this service exclusive to SHIONOGI, but to quickly roll it out to the entire healthcare industry.
We believe that rolling out this service will accelerate the creation of pharmaceutical evidence.
As a result, we are confident that it will lead to improved quality of medical care and the contribution it will make to patients and their families, and this is a very innovative initiative.
