View from the Summit - PHASTAR Life Science Summit 2020 in review

PHASTAR’s inaugural Life Science Summit took place on 30th June and 1st July and welcomed over 350 attendees from 40 countries to two days of engaging, insightful statistical and data science discussions. We were delighted to have such well respected speakers such as Frank Harrell, Janet Wittes, Stephen Senn and Thomas Jaki join us and share such in-depth knowledge with all those who attended.

The impact of the COVID-19 pandemic has been far reaching in every facet of peoples’ lives, both personally and professionally. PHASTAR decided that there was a demand for a forum to discuss statistical issues relating to both COVID-19 and non-COVID-19 studies, where participants from across the globe could join, with ease, and participate in discussions from industry leaders, as well as offering those in pharma the opportunity to deliver their own content of interest.

Presentations were separated into three streams – Statistics in Clinical Trials, Data Science (sponsored by AstraZeneca) and a dedicated Young Statisticians’ Stream which offered those who are newer to the industry to fine tune their presentation skills and showcase their knowledge. 31 presentations were delivered alongside two plenary discussions. Topics discussed included Data Handling and Reporting for Ongoing Clinical Trials during COVID-19, Ordinal Outcomes in Severe Respiratory TRT Trials, An AI Approach for Addressing Clinical Data Quality and Trends and in and Challenges with Machine Learning in Healthcare. 

The Plenary sessions delivered during the PHASTAR Life Science Summit generated a huge amount of interest. Our day 1 Plenary, Challenges of Designing Clinical Trials During a Pandemic, delivered by Thomas Jaki, focussed on the unique challenges on conducting clinical trials in a novel disease, and drew from Prof. Jaki’s personal experiences of being involved in the current COVID-19 outbreak. Prof Jaki discussed several trials including the LOTUS trial, an early trial evaluating remdesivir in mild, moderate and severe cases of COVID-19 and the ongoing RECOVERY trial, which is a basket trial to evaluate four active treatments on day 28 mortality with streamlined data collection. 

Prof. Jaki then spent time discussing the challenges in choosing an appropriate endpoint and the trade-off between using an ordinal endpoint and time to event outcomes. Difficulties with ordinal outcomes include choosing the appropriate time point over which to measure the response and difficulties with interpretation. Challenges when using time to clinical improvement include deciding how to handle deaths and less power (compared with ordinal outcomes).

Prof. Jaki concluded by stating how RCTs remain the gold standard even during a pandemic but that such trials should be adaptive with a focus on re-purposing drugs, dropping ineffective treatments early and trade off in choosing the appropriate endpoint.

Time for questions and answers was tight, however Prof. Jaki provided his answers post conference and can be reviewed here.

Our day 2 plenary was a dynamic discussion on The Best Laid Plans…Navigating Uncertainty during COVID-19, led by industry leaders Stephen Senn, Janet Wittes and Frank Harrell. Each panellist gave a short presentation, followed by a lively Q&A session chaired by PHASTAR’s Head of Statistical Research, Prof. Jennifer Rogers. 

Janet Wittes kicked off day 2 delivering a candid assessment of how the COVID-19 pandemic has changed her thinking on a variety of aspects of clinical trials while pondering on if, and how, future clinical trials will ever look post-COVID-19. Highlighting that presently there are over 500 COVID-19-related clinical trials and counting, Wittes focussed on how we might use this unfortunate situation to learn and adapt our approaches so in the post-COVID-19 setting, clinical trials might be more efficient and better designed.

Following Wittes presentation, Stephen Senn opened his talk, Clinical trials: quo vadis in age of COVID-19? by remarking that “any damned fool can analyse a randomized clinical trial, and frequently does”. He began by discussing trade-offs between one-and two-sample tests in the analyses of a single arm study - one sample test nominates a target response rate based on historical data whereas the two-sample test finds a set of historical controls and treats them as if they were a concurrent control. The second part of Senn’s presentation covered examples with relevance to COVID-19 touching on some recent studies coming out in the press. These include the single arm HCQ study that has been critiqued for incorrect use of statistical methods, the RECOVERY trial and a 2009 study about use of face masks on virus transmission. Senn concluded that not all studies during COVID-19 have been done well – careful thinking, control and decision making are required in all current and future COVID-19 research.

Frank Harrell was the final of the three plenary speakers on day 2, speaking on Bayes for Flexibility in Urgent Times. Harrell began with his view of the ‘big picture’ which is that efficacy is not a hypothesis but a matter of degree, hypothesis testing and thresholds hurt science, and that probabilities conditioning backward in time or information are often not actionable. He asks, ‘would we rather know the chance of making an assertion of efficacy when the treatment has no effect, or the chance the treatment is effective?’. Directly referencing COVID-19, Harrell spoke about the fact that in the beginning very little was known about the virus, which lead to the need for rapid decision making for researchers and DSMBs. This led  to his insightful presentation on the advantages of Bayes particularly in sequential or adaptive trial designs.

Questions and answers from our day 2 plenary can be found here.

On the Statistics in Clinical Trials stream, Ziad Taib delivered a comprehensive presentation on Statistical Validation of Biomarkers for use in Early Clinical Development & Biotechnology, considering biomarker validation methods based on type of biomarker and discuss several associated pitfalls from an early clinical biometrics perspective. Ziad discussed how biomarker validation, viewed as a confirmatory process aiming at validating a specific biomarker for a certain purpose, should ideally be based on proper statistical models and methodology. 

Olivier Collignon from GSK presented on Collaborative platform trials to fight COVID-19: methodological and regulatory considerations for a better societal outcome.  Olivier posed that in the current pandemic, drug developers are facing a challenging situation where, while speed is of the essence, quality of the evidence remains of primary importance and the rush to initiate trials. 

Olivier discussed how trials designed in isolation are more likely to fail to provide robust statistical  evidence and how potential alternatives for a better societal outcome. In this respect a collaborative Phase 2 platform trial comparing several treatments to a shared control arm appears as an attractive solution, and we will discuss possible adaptive designs allowing to drop inefficacious treatments as data accrue. Olivier emphasised certain difficulties such as the heterogeneity of the target population, the potentially high number of false positive treatments progressed to Phase 3 and a time-varying standard of care.

In the Data Science stream, Sheila Diamond of Acorn AI, by Medidata, a Dessault Systemes Company, spoke on how analysing clinical trials across the globe allows Acorn AI to see the impact of COVID-19 on halted enrolments, missed visits and explore the financial implications from cancelled studies and delayed milestones. Focusing on the importance of understand the evolving situation, Sheila noted that the first step is understanding impact on trials in real-time and gaining insights into when and where to plan for recovery.

Paul Metcalf of AstraZeneca delivered a discerning presentation on Trends in and Challenges with Machine Learning in Healthcare – discussing how machine learning is changing the world around us, so why not healthcare?  Paul noted that comparatively simple problems have a lot of machine learning behind them, but the human body is a more complex problem. Data are increasing and becoming more standardised, but machine learning is a tool that needs proper handling and constant updating. He discussed how it must be remembered that whilst machine learning is a tool, it needs proper handling in terms of good quality data and is in need of constant monitoring and updating – for example, we need to ensure the data has proper randomisation, and to carefully check any hidden stratification.

PHASTAR’s Life Science Summit’s Young Statisticians Stream was very popular over the two days – with several alternate conferences cancelled or digital with a reduced agenda, our summit offered young statisticians the opportunity to delivery presentations and share their knowledge with a large, global audience. 

Natasha Viglianti of AstraZeneca kicked off the Young Statistician’s stream with her presentation on  Oncology Umbrella Studies – Learnings from a Study of Triplet Combinations,  outlining her experiences working on an exploratory umbrella study establishing the recommended dose of multiple triplet combination treatments. Ellie Grainger, also of AstraZeneca, led an informative presentation on The Go/No Go Framework, where she explored an overview of the framework, its underlying parameters and what its resulting outcomes can mean for the decision. 

Identifying an issue and coming with a quality solution is one of the fundamental tenets of PHASTAR. We applied this to what is typically a very busy ‘conference season’ that was far quieter than usual in the wake of the COVID-19 pandemic and produced what was an excellently attended and informative summit. Videos of the presentations delivered at the PHASTAR Life Science Summit can be found here.