PHUSE EU Connect 2020

This year’s annual PHUSE event will be held the 9th - 13th of November in the virtual world. 2020 marks a special 16th year for PHUSE and this conference will be a unique and inspiring experience.

PHUSE will also be looking ahead to the ever-changing landscape, and the implications for the industry, as well as the challenges and benefits which this presents. Looking beneath the tip of the clinical trial iceberg, there will topics such as patient recruitment, electronic health records, the cloud, collaboration, analytical software and remote trials.

Six of our consultants, Sofia Vale, Joske van Schijndel, Kai Wanke, Pournima Prabhakaran, Louella Schoemacher 
and Fatima Kassim will present at the digital conference.

Join our statistical programmers, project managers and other staff members at the digital conference! Or contact us for more information at any other time that suits you best.

Sofia Vale (MSc)

"Posters - Defining Standards: Challenges and Solutions in the Development of Define-XML for SDTM and ADaM"

Define-XML is an integral part of any submission package of clinical studies to regulatory agencies, providing the metadata for all datasets being submitted. Therefore, it is essential that it is accurate and complete.

However, its development usually requires extensive manual input, being time-consuming and prone to errors. There are solutions in the market, but these are often expensive or incomplete.

Thankfully, due to its repetitive structure, it is possible to automate the development of Define-XML. This way, much of the manual input can be avoided, and conventions can be implemented consistently across several studies. 

Our solution integrates several tools and information already available, such as datasets or shells already developed and the standards present in the CDISC Library, in a simple and intuitive way for the end user. It turns the development of Define-XML into a fast and more accurate process, reducing the manual input as much as possible.

Joske van Schijndel (MSc)

"Standards Implementation - Pooling lab data to support FDA submission: keep it clear and simple"

In the conversion of more than 25 studies to CDISC standards, we came across hundreds of lab tests from different specimen types such as blood, urine and tissue. The lab analyses on these specimens were carried out by several laboratories each having their own standards. This resulted in combinations of lab tests with different lab metadata (lab test names, categories, specimens, methods and units) for 662 lab test concepts.

To overcome this challenge a metalab library was created to align the metadata including the assignment of the standardised SI unit, for which no cross-industry agreement yet exists.

This presentation will describe the benefits of having the metalab in place and will also focus on the challenges faced in the process of generating and maintaining it.

Kai Wanke (MSc/PhD)

"Application Development - Breaking boundaries: Working across different workspaces in LSAF"

In LSAF it is common practice to create a parent job that runs multiple child jobs consecutively. The parent and child jobs will create and act in their own, separate workspaces. While often it is sufficient to act within these boundaries, there can be cases when it is desirable that one job accesses the workspace of a another one, for example when a child job needs to access a status report continuously updated by its parent.

Here we present an approach developed by Janssen Pharmaceutica and OCS Life Sciences that allows a child job access to the parent workspace by overwriting existing job parameters in the child , that is modified on-the-go to contain the parent job’s workspace location.

This approach can be further modified to create shared workspaces to facilitate the direct interaction and cooperation between parent and child jobs.

Pournima Prabhakaran (MSc)

"Data Visualisation - Template based graph generation in Python and its maintenance in Git"

Graphical visualization is essential in understanding key relationships and summarizing patterns in clinical data. Ad-hoc analyses help in identifying possible covariates, data issues and help in improving the quality of clinical data reporting.

We will explore how Python can be used for data processing and plotting. Python offers several plotting libraries like Matplotlib and data processing libraries like Pandas, which could be adapted for this purpose.

Using Python together with a version control system such as Git can help different users add features and specific data processing functions. Template graphs generated using this collaborative approach will be shown in this presentation.

We will also briefly discuss about the integration of Python with the SAS Viya Platform. SAS Viya empowers users to work in their preferred coding language, accelerates processing of large analytics data and facilitates collaboration.

Louella Schoemacher and Fatima Kassim (MSc)

"Data Standards and Governance - Let food be thy medicine and medicine be thy food"

“Dear diary, This morning I ate two sandwiches with jam. I drank one glass of grapefruit juice and one glass of water for my medication.”

The person who wrote this diary is exposed both to nutrition and pharmaceuticals. However, the way data would be collected and standardised in clinical trials is very different. How do these differences affect data standardisation across these industries?
The challenges lie not only in the way data is collected, but there are also stricter laws governing one over the other. The nutrition TAUG was published in September 2019, so will this lead to better standardisation practices?
In this paper we present viewpoints on this subject and hopefully guide others to a meaningful discussion on the implementation of standards for nutritional data. What is the added value of using CDISC standards for nutritional data and what can be learned from pharma and the other way around?”