We are fast approaching the summer months and the PHUSE projects have already published six deliverables this year, along with numerous posters and endless abstracts for our PHUSE events. Many more deliverables are to be finalised and published in the coming weeks. Recent participation in our CSS event at Silver Spring has helped projects close out, as well as new ideas being formulated to fill the gaps in industry.

With over 150 new members joining our projects since the beginning of 2019, our Working Groups continue to grow at a rapid pace. Our blog will provide further details on each Working Group’s activities and achievements.

Data Transparency:

PHUSE successfully transitioned two guidance papers from TransCelerate BioPharma Inc. under its umbrella. These guidance documents address clinical data transparency and the de-identification and anonymisation concerning individual patient data. The first paper, Protection of Personal Data in Clinical Documents – A Model Approach, reflects the latest thinking on protecting privacy, including the European Medicine Agency’s Policy 0070 guidance issued in March 2016 to support the publication of clinical documents. The second paper, De-identification and Anonymisation of Individual Patient Data in Clinical Studies – A Model Approach (V2.0), highlights which techniques are currently available to conform to existing directives and regulatory guidance and to respond to research requests while safeguarding the privacy of individuals for sharing in a non-public environment.

Data utility is a vague concept that sponsors are asked to take into consideration and justify in their Anonymization Report, as per EMA Policy 0070 External Guidance and Health Canada Public Disclosure of Clinical Information Guidance. The PHUSE Data Transparency Working Group has explored the different objectives researchers may follow when using regulatory data for secondary purpose analyses. The final deliverable will detail the Working Group’s findings, the data utility scales that have been defined and its recommendation on how to use it in the context of a regulatory public data disclosure.

Standard Analyses & Code Sharing:

The FDA will be creating a Type C project group, with the specific purpose of reviewing integrated safety analysis plans for submissions.

Pilot with FDA on Sharing Interactive Displays: The Proof of Concept with CDISC pilot data has now been completed and the project is currently running a pilot with real submissions (html volcano plot).

The Communications, Promotion and Education team has created a front-end for Script Repository and Safety Analytics. Feedback on how to develop our site is welcome.

Code Sharing (Repository) has identified four targets from white papers and is seeking FDA input. The team will be presenting to the FDA on 17th July. In addition, this project is looking at R and Julia initiatives.

Nonclinical Topics:

The nSDRG team continues to collaborate with the FDA, collecting feedback on the usefulness of section 5 to SEND dataset consumers at the FDA. The team is looking to develop a new version of the nSDRG incorporating this feedback.

Data Visualisation as an Enabler for Nonclinical Safety Signal Detection presented a Poster at the CSS which explored how data visualisation can be used as a means for enabling safety signal detection within the scope of SEND-formatted, nonclinical data.

The SEND survey team presented its 2019 results in the form of a poster and to the wider Nonclinical team during a breakout session at the CSS. The results of the survey suggested overall SEND readiness across the industry; however, challenges such as the ability for data collection software and process to respond to changing needs remain. This emphasised the importance for sustained efforts by the Nonclinical group to help overcome these challenges.

The Demystifying Define-XML Codelist Handling for Nonclinical Studies project has provided recommendations for specific problems/questions encountered, when fulfilling the Define-XML codelist section for a nonclinical study.

Educating for the Future:

From empathy to testing, the Design Thinking project presented findings during the closing session of the CSS by sharing the prototype it had created, following feedback and input from members that attended its breakout session. The project is now looking to collaborate with the FDA and will present at one of our upcoming webinars.

Data Science is the newest project within Educating for the Future, which focusses on a training repository where materials will be consolidated and shared on our Education platform. The team will look to provide use cases of data science in the healthcare sector, as well as other industries, quantitative and advanced analytics and machine learning. Outputs from this project will be displayed at the upcoming EU Connect Poster Session.

Data Engineering presented a Poster at the CSS on progress to date and what the next steps are. The project will further explore other industry use cases, automation software and multiple data sources, as well as develop content on our Education platform.

Emerging Trends & Technologies:

The official release and publication of framework version 4 in May serves as the foundation to enter the cloud. The Cloud Adoption in the Life Sciences Industry team is now focusing on developing a new appendix on how companies that have adopted external cloud can keep up with the accelerated model, and add these incremental services.

The Machine Learning project ran a successful Workshop at the CSS, which taught attendees how to conduct simple machine learning projects. All four sub-teams were represented during the breakout sessions. The most recent sub-team is looking to collect training datasets from different companies to use within the SDTM mapping sub-team. The SDTM mapping team will use this data to explore and develop the mapping algorithms using NLP, ML and NN.

Tim Williams hosted the ever-popular Let’s Make a Knowledge Graph Workshop at the CSS, where participants experienced Linked Data for the first time. Attendees created a "white board" model of their own clinical trial, then exported that diagram to Linked Data.

Optimizing the Use of Data Standards:

CDISC Implementation Primer: The project reached out to the CSS audience to seek input on topics that should be included in its final primer deliverable. Ideas captured are currently being reviewed by the project team to see how these can be incorporated. The primer will be available for public review in Q3.

SDTM ADaM Implementation FAQ: Following public feedback, the Best Practices for Submission of Event Adjudication sub-team is finalising its white paper. This will be published by the end of July.

In addition, while there are a variety of scenarios that will arise during the preparation of the integrated databases, in support of the Integrated Summary of Safety (ISS) and the Integrated Summary of Efficacy (ISE), the proposed white paper by the ISS/ISE sub-team will focus on the following two scenarios that represent those most commonly seen in the industry:

• ADaM-only integration

• SDTM and ADaM integration

Feedback from the FDA is imminent on the ADRG package, and the project is looking to publish its update, together with example templates, this quarter.

Define-XML Completion Guidelines v2.0 has now been published as a final deliverable. You can review the documentation here.

Should you wish to join any of our projects, contact workinggroups@phuse.eu.