PSI Conference Webinar: Career Young Statistician Session

Add to:

Sponsored by rochelogo_10p_2


Date: Tuesday 9th June 2020      

Time: 10:00 - 11:30
Speakers: Jack Keeler (IQVIA), Ruth Owen (LSHTM), Ines Reis (MHRA) & Georgios Nikoladis.


This webinar is part of our 2020 Conference Webinar Series. Further information including details of other webinars that are included in the Conference package can be found here.
Members receive all webinars in the conference series for free.
Non-members receive all webinars in the conference series for £100+VAT, which includes complimentary membership* of PSI until the 31st December 2020. 

To register your place for this event, and others in the Conference webinar series, please click here.

Speaker Details:



Session Abstract

Jack Keeler

Is currently a new starter to the statistical industry having completed an MSc in Statistics at Lancaster University at grade Distinction in 2019. The focus of my dissertation was examining how exactly what this abstract concerns, Adaptive Enrichment Trials when concerning survival trials, achieving a grade of 76%. It was supervised by Dr Fang Wan of Lancaster University. Now working for IQVIA, I have begun to get to grips with how expansive the pharmaceutical industry is, and how the theories of university apply to trials being conducted in reality.

Enrichment Designs with Survival Data.

In modern medicine, it is becoming more apparent of patient and disease heterogeneity, and this can have consequences in clinical trials that do not take this into consideration. For example, Ellis and Taylor (2002) mention that ACE inhibitors are less effective in African American patients than White patients when treating heart failure. Many trials do not have sufficient information from early phase trials to definitively predict treatment effects for differing subgroups and some trials explore subgroups as exploratory endpoints, but exploratory conclusions are not always considered concrete. A solution for dealing with the differing biomarkers is to use adaptive trial designs, namely, enrichment designs. Survival trials may benefit greatly from such adaptations as survival trials are some of the longest trials conducted. Using Magnusson and Turnbull’s (2013) Group Sequential Design incorporating Subgroup Selection (GSDS), a trial can use the first interim analysis as a means of discontinuing treatments for subgroups who are not experiencing the desired effect from the treatment. This is making the trial ethically sound, as patients in certain subgroups are saved from ineffective treatments. GSDS does not follow the same boundary calculation rules as normal group-sequential designs, due to the selection criteria, but the conduct of the trial is very similar, making it feel familiar to statisticians.Enrichment trials are currently rare, so an example trial, using simulated survival data, will demonstrate how these trials could perform in reality, and examine the advantages and disadvantages of such designs.

Ruth Owen

Ruth Owen is a research fellow in the Medical Statistics Department at the London School of Hygiene and Tropical Medicine. She has been in this role since completing her MSc in Medical Statistics at the school in September 2018 and is currently working in a team of statisticians working under Professor Stuart Pocock. Her projects are mainly in the cardiovascular disease area and have included an international clinical trial (SECURE) which investigates the effects of the polypill in elderly patients with cardiovascular disease, several projects using data from a multi-national cardiovascular registry (TIGRIS) and consultancy work with a team of cardiologists at the Royal Brompton and Harefield NHS Trust. She has also previously worked for one year as a medical statistician in the Unit of Medical Statistics at King’s College London.

Methods to Evaluate the Benefit-Risk Trade-Off in Individual Patients.

For many RCTs the efficacy of a new treatment is accompanied by safety concerns. While overall results may demonstrate a favourable risk-benefit trade-off there may be individuals where the harm outweighs the benefit.

Methods: We describe methods to predict the individual patient’s absolute benefit and risk based on multivariable models using patient baseline characteristic. Taking account of the relative clinical importance of the respective benefits and harms we develop an algorithm for clinical use whereby rapid decisions can be made on the preferred treatment strategy for each individual patient.

Results: We illustrate this approach with findings from three major cardiovascular studies:
1) the SPRINT trial of intensive versus standard blood pressure lowering, where ischaemic benefits are accompanied by some major adverse events
2) the TIMI 50 trial of vorapaxar versus placebo post-myocardial infarction, where ischaemic benefits are accompanied by increased risk of major bleeding
3) a meta-analysis of 7 studies in coronary patients receiving a stent, with the goal of identifying which patients at high risk of bleeding need a shorter duration of effective dual anti-platelet drugs.

Conclusions: Our findings illustrate how quantitative methods can help identify those individual patients for whom the risk of harms outweighs the benefits of a new treatment.

Inês Reis

Inês Reis has been a Statistical Assessor in the Statistics and Pharmacokinetics Unit of the Medicines and Healthcare products Regulatory Agency (MHRA) since 2018. She has previous experience working in the Biostatistics and Methodology Support Office at the European Medicines Agency. Inês has been collaborating with the ICH E9(R1) expert working group on estimands since 2016 and was an important contributor to the training material presentations accessible in the ICH website. She holds an MPharm and a Post-graduate Diploma in Biostatistics, both from the University of Lisbon.

The Young Statistician's Guide to regulatory statistics.

Successful and safe devolvement, licensing and marketing of medicines cannot happen without intense cooperation and dialogue between regulators and pharmaceutical companies. On the regulatory side, the Medicines and Healthcare products Regulatory Agency (MHRA) has decades of experience in medicines and medical devices regulation, during many of which statisticians have been deeply involved. Not only at the level of licensing of medicines, but also in the pharmacovigilance and medical device areas, statisticians play an important role in the Agency's activities, not forgetting their involvement in real-world data collection and analysis (CPRD) and characterisation, standardisation and control of biological medicines (NIBSC).

In this talk you will learn about the MHRA, how we work, and the statistician's roles in the system, as well as some hints of current hot topics in regulatory statistics such as estimands and real-world data. You will also discover the types of interactions that can be held with regulators at the UK (MHRA) and European (EMA) levels, the different types of regulatory procedures and how statisticians from both sides of the table can contribute to such dialogue, ultimately helping their companies navigate through the regulatory system more smoothly.

Georgios Nikoladis, 

Georgios is a Research Fellow (Health Economics) in the Centre for Health Economics (CHE) in the University of York working on the development of evidence synthesis and decision modelling methods for Health Technology Assessment and on evaluating manufacturers submissions to the National Institute for Health and Care Excellence (NICE) in the UK.Until October 2019, he was a PhD student in CHE affiliated with the Team for Economic Evaluation and Health Technology Assessment where he undertook a thesis entitled “Borrowing strength from indirect evidence in HTA: methods and policy implications”. His work has been presented in multiple conferences including the Royal Statistical Society conference (2018), Health Economics Study Group (2018), and Health Technology Assessment International (2017).

Georgios graduated with distinction from the MSc in Health Economics in the University of York and holds an MPharm from the Aristotle University of Thessaloniki.Georgios has a genuine interest in statistical (network) meta-analytic methodologies and, specifically, in Bayesian methods for multi-parameter evidence synthesis. George is also interested in methods for decision analytic modelling, exploring the value of further research, and evaluating diagnostics.

Borrowing strength from indirect evidence in HTA.

Sparse relative effectiveness evidence is a common problem in Health Technology Assessment (HTA). For example, evidence on a paediatric population may be limited. Where evidence directly pertaining to the decision problem is sparse, one option is to expand the evidence-base and include studies that relate to the decision problem only indirectly; for instance, a decision on children may borrow strength from evidence on adults. Usually, in HTA, such indirect evidence is either included by ignoring any differences (`lumping`) or is completely disregarded (`splitting`). However, more sophisticated methods exist in the literature which, rather than `lumping` or `splitting`, impose more moderate, perhaps more appropriate, degrees of information-sharing.

We developed network-meta analytic methods for the combination of, aggregate-level, binary, direct and indirect evidence. These can be categorized into functional-, exchangeability-based, prior-based and correlation-based relationships. The estimates produced with each method were subsequently used in a case-study that evaluated the cost-effectiveness and value of information of intravenous-immunoglobulin (IVIG) for adults with severe sepsis and septic shock.

Results: Depending on the information-sharing method used, Incremental cost-effectiveness ratios (ICERs) varied between around 20,000 –50,000 £ per Quality-Adjusted Life Year (QALY), and the optimal sample size of a future trial ranged between 3500 patients and 0 patients (i.e. no further trial is needed).

Information-sharing method choice can lead to different adoption and further research recommendation decisions. It is, hence, important to scrutinize methods’ underlying assumptions and create a transparent, systematic, process that analysts can use when facing evidence sparsity problems.




Upcoming Events

Latest Jobs