Carsten Henneges, Keith R Abrams, Imran Hossain  

Data Science is a growing area of expertise in the pharmaceutical industry complementing traditional statistical methods that are more well-established. In this webinar, we explore the application of data science methods in medicine, taking in a range of perspectives. Three speakers from a pharmaceutical, CRO and academic perspective talk about what Data Science in medicine means to them. The session will include examples from their work and a panel discussion.

The Clinical Global Impression Scale is a 7-point scale to rate patient's illness. Lorenz Uhlmann is presenting various graphical displays proposed this months. Visualisations are available on the Wonderful Wednesday blog. 

The very clean bar chart is helping to concentrate on the main message and an example how to avoid clutter in graphical displays. Line graphs help to see relationships and imply a certain data structure. To observe proportions at certain cut-offs the cumulative distribution plot has clear advantages. And the combination of triple dot plots and forest plots gives a comprehensive overview. The new challenge is on US disease data. See the Wonderful Wednesday homepage for more detail.

Wonderful Wednesdays are brought to you by the Visualisation SIG. The Wonderful Wednesday team includes: Bodo Kirsch, Alexander Schacht, Mark Baillie, Zachary Skrivanek, Lorenz Uhlmann, Rachel Phillips, David Carr, Steve Mallett, Rhys Warham, Lovemore Gakava, Zara Sari, Paolo Eusebi, Greg Ball.

This months challenge was to display diary data on six symptoms to track disease progression over time. Zak Skrivanek is presenting the proposed visualisations and discusses the impact of colour selection and arrangement of multiplots. Visualisations are available on the Wonderful Wednesday blog. 

The first plot is a stacked bar chart showing the different symptoms on top of each other. With six symptoms the coloring gets more difficult. Zak highligths ways to increase accessibility for color-blind users. And with 100 time points the plot is actually a stacked area chart. The next plot uses vertical time axes to align the plots for the two treatments and added a plot for treatment differences. The symptoms are combined as multiplot. In the arrangemnent of the plots enclosure and proximity can improve readability. The new challange is on a simple CGI data set. See the Wonderful Wednesday homepage for more detail.

Wonderful Wednesdays are brought to you by the Visualisation SIG. The Wonderful Wednesday team includes: Bodo Kirsch, Alexander Schacht, Mark Baillie, Zachary Skrivanek, Lorenz Uhlmann, Rachel Phillips, David Carr, Steve Mallett, Rhys Warham, Lovemore Gakava, Zara Sari, Paolo Eusebi, Greg Ball.

Moderator: Judith Anzures-Cabrera (Roche). 
Presenters: Nikhil Kamath (Roche), Sue McKendrick (Thermo Fisher Scientific), Antonia Morga (Astellas), David Wright (AstraZeneca)

This is the fourth in the series of webinars from The Estimands Academy for Trial Teams. The estimand framework (ICH E9 (R1)) is applied to a case study in heart failure from the perspective of multiple stakeholders. Topics covered will include using different strategies for handling intercurrent events to construct estimands for different stakeholder questions of interest.  The estimands bring transparency in the treatment effects we will estimate from the protocol through to product labels. 

The targeted learning outcomes are as follows:
· To apply the estimand framework to a challenging realistic setting
· To appreciate the different stakeholder views when determining estimands
· To understand the relative merits of different strategies for intercurrent events
· Understanding implementation challenges and possible solutions.  

Target audience (no prior knowledge of estimands required): Clinicians, Regulators, Investigators, Academics, members of Ethics Committees, Health technology assessors, Statisticians.

Presenters: Christine Fletcher (GSK), Christy Mazzarisi (Mersana) and Paul Frewer (AstraZeneca).

In this session we discuss the need for a conscious evolution of leadership styles as remote working becomes commonplace. 

Whilst there has been a lot of debate on office-based vs. remote workers,  there has been less discussion on what 'good' leadership looks like in the new working world. This panel discussion explores how leaders in our industry have had to adapt, adjust and evolve in order to remain effective in their roles.

Els Pattyn.

Development of an adaptable yet fully automated end-user tool that allows a harmonized and regulatory compliant calculation and evaluation of immunogenicity cut-points This will be followed by Q&A.

Steve Mallet is presenting ways to better visualize data from different studies. The starting point was a graph published summarzing the results of 12 studies on quality of life measured by DLQI. The improved visualisations are available on the Wonderful Wednesday blog. 

The estimand framework is considering different aspects of measuring the study outcome. This includes for example intercurrent events. We're discussing proposals to present those aspects visually to easily understand their impact. The visualisations are available on the Wonderful Wednesday blog. 

The visualisation of study results should also reflect the defined estimand. Different graphs are presented for continuous and binary/TTE outcomes. We see charts that include information on the occurrence of intercurrent events and the amount of missing data. The next challenge will be to improve a given visualisation of DLQI results across different studies. See the Wonderful Wednesday homepage for more detail.

Wonderful Wednesdays are brought to you by the Visualisation SIG. The Wonderful Wednesday team includes: Bodo Kirsch, Alexander Schacht, Mark Baillie, Zachary Skrivanek, Lorenz Uhlmann, Rachel Phillips, David Carr, Steve Mallett, Rhys Warham, Lovemore Gakava, Zara Sari, Paolo Eusebi, Greg Ball, Martin Brown, Martin Karpefors, Benjamin Lang

Presenters: Khadija Rantell (MHRA); Rima Izem (Novartis); Rackeh Williams (GSK), Luis Garcia-Gancedo (GSK); David Wright (AstraZeneca); Magnus Jörnten-Karlsson (Astra Zeneca); Magalie Hilton (roche); Yashoda Sharma (DiMe) 

In a decentralised clinical trial, patients’ assessments are (partly) transferred from the clinical trial site to the patients’ home and might occur without any face to face interactions between patient and study personnel. In particular during the Covid-19 pandemic, decentralising clinical trials attracted considerable attention as one way to continue conducting trials during lockdowns and curfews. In this online event, experts in the field discuss the opportunities and challenges of decentralised clinical trials for the development of drugs and the promise they hold for modernising the way trials are conducted.

The presentations in part 2 of this meeting were as follows:

Towards more patient-centric trials: opportunities and challenges of decentralized clinical trials. (Khadija Rantell)
The COVID-19 pandemic has necessitated flexibility in trial conduct and accelerated changes in the clinical trial landscape. Decentralised clinical trials (DCTs) introduce new approaches to the conduct of clinical trials that aim to make clinical trials more easily accessible and convenient for participants to take part in. Moving forward, it will be important to leverage the use of more innovative tools and approaches to trial design, and conduct, including adoption of remote approaches such as electronic patient-reported outcomes, home visits by health staff, follow-ups over phone, telemedicine etc. This talk will cover key considerations for suitability of use of decentralised clinical trial model, reflecting on the opportunities and challenges of implementing decentralised elements in clinical trials.

Decentralized clinical trials: scientific considerations through the lens of the estimand framework. (Rima Izem)
While the industry and regulators’ interest in decentralized methods is long-standing, the Covid-19 pandemic accelerated or broadened the adoption and the experience with these methods in clinical trials. Potential benefits of decentralizing studies include broadening access of clinical trials to a more diverse population, reducing the burden of participation in trials, or using innovative endpoints. As decentralization is moving from a necessity to a choice, researchers need to consider the actual value added and implications of these designs beyond the operational aspects of their implementation.
Our presentation illustrates how the estimand framework can clarify our thinking about those strategic decisions around decentralization. Those questions go beyond whether or not to decentralize, to how much and what to decentralize. That is, going through questions related to each attribute of the estimand, and especially population, treatment and variable, can guide the teams through a systematic thought process about the scientific opportunities, assumptions and potential risks associated with a possible use of decentralized components in the design of a trial. Illustrations from decentralized trial case studies will support the presentation and discussion of various considerations from this framework. Using this approach, teams can more easily calibrate the design decentralization complexity (from hybrid to fully decentralized) to the value added in answering scientific questions that could not be solely answered by an all on-site approach.

Developing Digital Biomarkers: Application to Rheumatoid Arthritis. (Rachel Williams and Luis Garcia-Gancedo)
Data collected in clinical trials present an incomplete view of disease activity due to daily symptoms fluctuation and heterogeneity. Digital technology can enable remote continuous monitoring and the addition of objective of patient assessments. Rheumatoid arthritis (RA) is a chronic condition with fluctuating symptoms allowing for feasibility assessment of digital biomarkers. First, we performed the PARADE study as the first in pharma to use the Apple Researchkit® for fully remote enrollment and data collection via the iPhone, including baseline information, symptoms, wrist movement, and a walk test. Based on this experience, we performed the weaRAble PRO study, enrolling patients through a clinic, and adding an Apple Watch as well as the iPhone for data collection. These studies indicate that digital technology allows for remote data collection, increased measurement frequency, a richer picture of disease activity, and a focus on patient centricity.

Panel discussion, with presenters from both days.
Discussion on the topics raised during the two half-days of the meeting, including input from the audience. 

Presenters: David Wright (AstraZeneca); Magnus Jörnten-Karlsson (Astra Zeneca); Magalie Hilton (roche); Yashoda Sharma (DiMe).

In a decentralised clinical trial, patients’ assessments are (partly) transferred from the clinical trial site to the patients’ home and might occur without any face to face interactions between patient and study personnel. In particular during the Covid-19 pandemic, decentralising clinical trials attracted considerable attention as one way to continue conducting trials during lockdowns and curfews. In this online event, experts in the field discuss the opportunities and challenges of decentralised clinical trials for the development of drugs and the promise they hold for modernising the way trials are conducted.

The presentations in part 1 of this meeting were as follows:  

What are Decentralised Trials? What are the key opportunities and risks with their use? (David Wright)
This talk gives an introduction to Decentralised Trials and outlines some of the opportunities and risks they bring if conducted instead of site based/centralized trials.  

Bringing true value to patients, site and sponsor through digital transformation of clinical trials. (Magnus Jörnten-Karlsson)
To successfully support decentralized clinical trials we need to think holistically about designing a toolkit for patients and sites. In this talk I will touch on how we define, evaluate and select technology (medical devices); current progress on the tools, moving away from isolated silo tools that only deliver one thing to broader platforms; and the impact that is having on our sites and patients.

ALpha-T: a Pre-Pandemic Decentralized Trial in Oncology. (Magalie Hilton)
ALpha-T is a Phase II, Open-Label, Single Arm trial with a decentralized (DCT) home-based approach. Patients with locally-advanced or metastatic ALK-positive non-NSCLC are enrolled. The collaborations involving the patients, the caregivers, the sequencing partner, the telemedicine company, the regulators and the Sponsor are key for the implementation of home-based solutions. Based on this case study example, top challenges and opportunities of a DCT are presented.

Digital Medicine - Opportunities for Advancing Health Outcomes and Health Equity. (Yashoda Sharma)
Founded in 2019, the Digital Medicine Society (DiMe) is a global non-profit and the professional home for all members of the digital medicine community. Together, we drive scientific progress and broad acceptance of digital medicine to enhance public health. Digital medicine offers the potential to redefine healthcare, solving some of the most pressing and persistent challenges to good health for all. The digitization of healthcare presents a fleeting opportunity to move away from a system designed to treat the sick and toward one that is defined by each individual’s health. Core to our work of driving the development and adoption of digital health measurements is advancing health equity, to fuel the effectiveness of healthcare.
Challenges to effective and equitable clinical trials cannot be solved with individual entities working in isolation; multi-stakeholder, interdisciplinary expertise is required. DiMe partnered with the FDA and organizations across the healthcare continuum to prioritize inclusion in digital health measurement product development and deployment, and diversity, equity, and inclusion in digitized clinical trials. With this work we developed toolkits that healthcare and research organizations can incorporate into their current workflows to drive the adoption and implementation of digital and decentralized clinical trials, and best practices to ensure health resources are delivered with a health equity lens.

In this webinar, you will hear an introduction on biomarkers and learn more about recent advances in biomarker-based designs, machine learning for biomarkers and data repositories for biomarker use cases. 

The webinar starts with an introduction of the Biomarkers Special Interest Group and a whistle-stop tour on biomarkers.

Presentation 1: Enrichment designs with predictive biomarkers
Most biomarker-driven trial designs are based on the assumption that the biomarker is predictive of response to treatment.
Would you rather verify this assumption during the trial?
If so, which population and subpopulations would you test?
What regulatory issues might you encounter?
How would you setup your hypotheses?

At the PSI Biomarkers SIG, we are addressing these issues within the setting of adaptive enrichment designs for Phase II trials in oncology.

Presentation 2: Who said Machine Learning for Biomarkers?
What do you expect from an inspiring  presentation on Machine Learning for Biomarkers? 
1. Excitement, because you have learned about all the sophisticated Machine Learning methods out there?
2. Being impressed by success stories on Machine Learning & biomarkers?
3. Actionable guidance so that you can incorporate Machine Learning for Biomarkers into your own work?
We, the PSI Biomarkers SIG, want to give it a try and share our thoughts on the past, present and future of Machine Learning for Biomarkers with you.

Presentation 3: But where is the data?
You've discovered this cool new Machine Learning technique but have no data to try it on?
Or you have heard of a new groundbreaking kind of biomarkers?
Or you are like: it is time for me to get my hands dirty with RNA-seq data!

We’ve all been in such situations but then comes the harsh reality: finding publicly available biomarkers data is hard. Finding publicly available biomarkers data along with corresponding treatment response is even harder. 
Changing this was the motivation of the PSI Biomarkers SIG to have a focus on data repositories. 
During this presentation, we will share with you the fruit of this work! 

We're looking back to Wonderful Wednesday Webinars of 2022. What graphical displays stood out and what made them special? The visualisations are available on the Wonderful Wednesday blog. 

In 2022 a lot of impressive visualisations have been submitted to solve very different challenges. The best ones are being revisited in this webinar to highlight how visual displays can be used most effectively. The next webinar will address the visualisation of estimands. See the Wonderful Wednesday homepage for more detail.

Wonderful Wednesdays are brought to you by the Visualisation SIG. The Wonderful Wednesday team includes: Bodo Kirsch, Alexander Schacht, Mark Baillie, Zachary Skrivanek, Lorenz Uhlmann, Rachel Phillips, David Carr, Steve Mallett, Rhys Warham, Lovemore Gakava, Zara Sari, Paolo Eusebi, Greg Ball