PhUSE EU Connect 2019 was held in the beautiful city of Amsterdam between the 10th and 13th of November. This clinical data science conference comprised 19 Streams, including 150 papers, 24 posters and 3 engaging data scientists as keynote speakers. The event was well attended and had several interesting and innovative presentations. Caroline Terrill, Associate Director of Statistical Programming at Cytel UK, conducted a session “No Place Like Home: Managing Remote Programmers Remotely” and stood out as the winner in the Personnel Management category. Based on 5 years' experience of managing remote programmers, Caroline’s paper gives guidance on issues to be considered and traps to be avoided if you are managing people who work remotely.
In this blog, we share the presentations from our Statistical Programmers and summarize some of the sessions that our team members attended.
Contributed Sessions Featuring Cytel Programmers
SAS Outputs in Excel workbook using ODS EXCEL
Choosing the right path to follow when integrating ADaM
Magalie Gallet and Laura Phelan
No place like Home: Managing remote programmers
Detriment index-based safety ranking of Opioid drugs
Click the button to receive downloadable copies of their slides.
Session summaries from the team
Caroline Terrill, Associate Director Statistical Programming at Cytel UK, shares her experience.
Multiple imputation as a valid way of dealing with missing data, presented by Vadym Kalinichenko, Intego Group, LLC
In many studies, missing values are not always dealt with in the most effective way and this can lead to bias. This presentation interested Caroline because she wanted greater insight into how missing value imputation should be considered for more studies. This session provided the code for performing the imputation and imparted information on how to analyse the data after imputation. Another interesting aspect this session covered was how many SAS procedures use complete case analysis by default and this could mean losing some valuable data. The presenters used SAS and R procedures for creating multiple imputations for incomplete multivariate data, and then analysed and compared results from multiple imputed data sets.
Missing value imputation should be considered in more studies. However, the implementation of this and the computing time needed for the iterations could add additional time to the analysis.
Early experiences using SDTM to organize clinical data collected using a physical activity tracker, presented by Martin Gram and Gianluca Mortari, Novo Nordisk A/S
Real-world evidence data is being discussed widely. But, when it has to be submitted to the FDA, the standard structure of SDTM should be followed. Caroline was interested to learn about the challenges in doing this.
Wearable technologies and real-world evidence data are playing a growing role in clinical trials. The data collected through these sources can have a large potential for new clinical endpoints, but the structure of this data is unlike traditional data collected at clinical sites. The presenters shared their experience of mapping this type of data to SDTM standards.
SDTM does not appear to be ‘fit for purpose’ for this new type of datasets because of the nature and volume of data collected. SDTM also contains a lot of repeated information that only adds to the size of the files, without really adding any additional information.
Using ODS LAYOUT to Create ABSOLUTE-ly Stunning Patient Profiles, presented by Michael Allie, AstraZeneca
Caroline has programmed many patient profiles and knows how it can be time consuming to get them exactly as the reviewers would like, which is with minimal blank space especially as the amount of information per subject can be quite different.
This session talked about how clinical reviewers really value getting all important information for one subject in one report, and ideally on one page. But the layout from SAS procedures typically puts each new topic (demography, AEs, Conmeds) on separate pages. The presentation explored ODS LAYOUT ABSOLUTE, a tool that resolves this issue as it can be used to put information from multiple procedures on a single page.
It is possible to generate the type of profiles clinical reviewers like by using ODS LAYOUT with standard SAS procedures. Although some extra effort is needed initially to get this to work dynamically for each subject, the resulting profile looks effective.
Guillaume Herve, Senior Statistical Programmer at Cytel, summarizes the interesting talks he attended.
Define-XML Review: Failing to Plan is Planning to Fail, presented by Parveen Kumar, GCE Solutions
The presentation covered the approach for fool proof planning and execution of Define-XML review. Define-XML is the Table of Contents of the submission package and the most useful document for describing the structure and the content of the data submitted. The session defined a convenient step by step approach for detecting the minutest of errors. Guillaume found this subject interesting as Define-XML is a complex document to produce and it is important in our field of work.
A Define-XML document with errors can delay the approval. This session teaches additional checks that can be performed to improve the efficiency of the Regulatory Review process.
Lessons Learned from an FDA Real-time Oncology Review Pilot Programme: A Novartis & Roche Perspective, presented by Delphine Kerzerho, Novartis and Hiren Naygandhi, Roche
In 2018, FDA Oncology Center of Excellence launched Real Time Oncology Review (RTOR) pilot program. RTOR aims to improve the efficiency of the review process for supplemental applications by allowing for the submission of key efficacy and safety tables/figures and datasets prior to the complete dossier submission. In this session, Novartis and Roche together present their experiences of going through FDA RTOR. Senior Statistical Programmers at Cytel, Guillaume Herve and Adrien Vallee, attended this presentation to learn about this innovative submission approach. They both found it interesting to see how Novartis and Roche collaborated with the FDA on this new and not very well-known process.
The pilot program process and challenges faced by both the companies were well explained. FDA can have varying approaches depending on the study design or format of data. As this program aims at accelerating the submission process of a drug, a common piece of advice was to manage priorities, be flexible and anticipate the production of submission documents in earlier stages, even if it occurs before database lock.
Adrien Vallee, Senior Statistical Programmer at Cytel, shares his insights.
Finding the Right Balance - Data Management Surveillance in a Fully Outsourced Partnership Model, presented by Konstanze Morgenroth, Merck Healthcare
Merck Healthcare conducts a majority of its clinical studies in partnership with Contract Research Organizations (CROs). However, as a sponsor, Merck is still responsible for the quality of the study and also the quality of the data collected. Adrien attended this session to find out how Merck managed its CROs.
Merck Healthcare has implemented several data quality processes and tools to ensure inspection readiness. These tools are used both internally (by the data management team) and cross-functionally for joint data checks. Additionally, the CROs have tools in place that help Merck to perform its supervisory role.
Organizations like Merck Healthcare that operate on a fully outsourced partnership model with their preferred CROs need to have an established oversight program to ensure efficiency. Sponsors are ultimately responsible for the quality of the study and hence need to maintain a certain level of control.
Meet the Cytel team at PhUSE US Connect 2020
We are heading to PhUSE US Connect 2020 in Orlando, March 8-11. Schedule an expert consultation and stop by Cytel's booth to speak with one of our expert Statistical Programmers.