Video-on-Demand Library


Collection
Keyword or phrase

Topic

16 September 2020

Watch this Journal Club webinar on “Longitudinal Data”. Florian Lasch (Hannover Medical School, Germany) and Mutamba Kayembe (Maastricht University, Netherlands) presented their recent work. The webinar was chaired by Michael O’Kelly (IQVIA).

Read more...

Florian Lasch (Hannover Medical School, Germany):- Empirical evaluation of the implementation of the EMA guideline on missing data in confirmatory clinical trials: Specification of mixed models for longitudinal data in study protocols
Authors: Sebastian Häckl, Armin Koch & Florian Lasch
Pharmaceutical Statistics, 2019; volume 18, Issue 6, Pages 636-644.
Access the slides here.

Mutamba Kayembe (Maastricht University, Netherlands):- Imputation of missing covariate in randomized controlled trials with a continuous outcome: Scoping review and new results
Author: Mutamba T. Kayembe, Shahab Jolani, Frans E. S. Tan & Gerard J. P. van Breukelen
Pharmaceutical Statistics. Early view. 
Access the slides here.

09 September 2020

Alexander Schacht, Mark Baillie, Daniel Saure, Bodo Kirsch, Zachary Skrivanek, Lorenz Uhlmann, Rachel Phillips, Markus Vogler, David Carr, Steve Mallett, Abi Williams and Mike Greenwood.

Read more...

How to display safety data? This month's challenge has shown there are very different ways to visualize adverse event data. Although the example data set was from a two-arm study and relatively simple, the display of type of AE, frequency, timing, severity and seriousness is not easily combined in one plot.

03 September 2020

Caroline Caudan, Paulo Eusebi & Michael O'Kelly

Read more...

This webinar features presentations from 3 speakers on the topic of Using Visualisations to Help Make Decisions: Caroline Caudan presents ‘Interactive statistical monitoring to optimize review of potential study issue with R-Shiny’, Paolo Eusebi presents ‘Effective visualization of uncertainty – Where we are and where to go’ and Michael O’Kelly presents ‘Subgroup analysis: a look at the SEAMOS approach (Standardised Effects Adjusted for Multiple Overlapping Subgroups)’. These presentations were originally planned as part of the 2020 PSI conference in Barcelona, and have been reorganised as a webinar.

12 August 2020

Alexander Schacht, Mark Baillie, Daniel Saure, Bodo Kirsch, Zachary Skrivanek, Lorenz Uhlmann, Rachel Phillips, Markus Vogler, David Carr, Steve Mallett, Abi Williams and Mike Greenwood.

Read more...

In this month’s webinar, we discussed an exacerbation example data set. The data is based on the RISE study for patients with moderate COPD. The primary endpoint is the number of exacerbations during a six month treatment period. Event data – but patients can (and do) have multiple exacerbations. Statistical analysis used a Negative Binomial model. The dataset also included other variables which (may) effect the exacerbation rate: % Predicted Normal of Forced Expiratory Volume in 1 second (FEV1), Exacerbations in previous year, and Region / Gender.

The challenge was to produce a data visualisation incorporating the information on the number of exacerbations observed. The discussed visualisations included a straightforward and clear presentation of the number of exacerbations, a Power BI app, longitudinal plots, and a display of patient-level outcomes.

15 July 2020

Bruno Boulanger

Read more...

4 key learnings attendees will take away from this webinar:
- The value of good design of experiments
- Consider Bayesian statistics to answer your question
- P-values is not always what you’re looking for
- Adopt a life-cycle over the long-run, not just study by study.

For 15 years the scientific literature has repeatedly underlined the important lack of reproducibility and replicability of studies in biomedical research. As a consequence, several scientific organisations (journal, scientific societies, universities, national agencies) have identified some root causes to this issue and proposed good research practices to improve the replicability of the results. The misuse of statistical concepts, from design of studies to analysis of data to decision-making is at the heart of the crisis, even if not the only cause. In this webinar we will explain the how to understand the sequences of issues and how to fix it in order to drastically improve the replicability of the results. Through example the presentation will cover concepts such as OFAT vs DoE, p-values and Bayesian statistics and Power vs Assurance as easy opportunities to improve robustness of decisions.

08 July 2020

Alexander Schacht, Mark Baillie, Daniel Saure, Bodo Kirsch, Zachary Skrivanek, Lorenz Uhlmann, Rachel Phillips, Markus Vogler, David Carr, Steve Mallett, Abi Williams and Mike Greenwood.

Read more...

This month's challenge was to summarise changes in haemoglobin (Hgb) concentration over time in patients with anaemia associated with Chronic Kidney Disease (CKD). An experimental medicine was compared with a control group, but in addition to demonstrating efficacy, there was a special interest in visualising intra-individual variability, as large changes in Hgb are a potential safety concern. The visualisations presented ranged from grouped line plots over time, Sankey-diagrams and line plots with quantile bands for summary views and a lasagna plot for the display of individual data.
Some general hints were given on how to define suitable color palettes for graphs depending on the type of data. Presented by Steve Mallett.

11 June 2020

How can the Estimand Framework Help Us to Address Impacts of the COVID-19 Pandemic? This webinar focused on how the estimand framework can help us to pose and answer clinically relevant questions in the light of the COVID-19 pandemic. This session brings together colleagues from the EFPIA/EFSPI estimand implementation working group and the Estimands in Oncology ESIG.

Read more...

Chrissie Fletcher, David Wright, Kaspar Rufibach.

We will consider impacts on ongoing case studies in chronic obstructive pulmonary disease (COPD) and oncology. The COVID 19 pandemic causes disruption of healthcare systems leading to inaccessibility of sites, treatment interruptions and discontinuations, missed or delayed visits, use of unexpected concomitant medications and deaths. The scale of the impact will vary depending on our target population, therapeutic area and the timing of study recruitment relative to the course of the COVID-19 pandemic.

We will show how the estimand framework helps us to:
• revise the treatment effects of interest considering all five estimand attributes including accommodation of unexpected intercurrent events;
• propose sensitivity analyses to explore robustness of departures from assumptions (including missing data assumptions);
• propose supplementary analyses to more fully investigate and understand the trial data.

Our case studies will illustrate how pre-specified estimands are impacted by the COVID-19 pandemic, and how methods for estimation including sensitivity analyses may need to be reconsidered to address missing data issues.

11 June 2020

The 2019 Prize winners were; Mike Smith (Pfizer) for excellence in research around population modelling in drug development and pioneering practical Bayesian methods with drug development in the context of optimising decision making within trials and modelling dose response with pharmacokinetic/pharmacodynamic modelling and Graeme Archer and Jacquie Christie (GSK) for their work on Quantitative Decision-Making for clinical development, which was deemed to be at the cutting edge of revolutionising data driven decision making within the industry.

Read more...

Mike Smith, Jacquie Christie

Mike Smith (Pfizer) talks on innovation, play and failing. What is innovation? How do I innovate? Is it a skill that I can learn or improve on? In this presentation I will attempt to define what I mean by innovation and break down what I consider to be drivers of innovation. Some of these are things that you CAN definitely work on. It is true that some are environmental or cultural but even these are things that you can influence or change. I will illustrate how these elements have come together to help me and my colleagues deliver on projects, innovate and break out of “tried and tested” methods, including those cited in my RSS/PSI Statistical Excellence in the Pharmaceutical Industry award.

Jacquie Christie (GSK) talks about Quantitative Decision-Making for Clinical Development. Low productivity and late-stage attrition are well-known pharmaceutical R&D problems; one cause is poor decision-making at key milestones. Failure to consider the impact of current design choices on future investment options increases uncertainty when the decision-point arrives, and consequently increases the risk of a wrong decision. This talk will outline an approach to quantitative decision-making in a completely Bayesian setting. Prior distributions are used to evaluate proposed study designs and decision rules against not only the probability of success of the current study, but that of the subsequent study or set of studies. Termination and progression criteria and associated risks are then clear, allowing more informed decision-making. The approach is applicable to all phases of R&D.

10 June 2020

Alexander Schacht, Mark Baillie, Daniel Saure, Bodo Kirsch, Zachary Skrivanek, Lorenz Uhlmann, Rachel Phillips, Markus Vogler, David Carr and Steve Mallett.

Read more...

This month the discussion is about visualizations of the satisfaction data using cumulative distribution functions, a tree plot, staggered bar plots, a power BI dashboard and two interactive funnel plots using plotly and RShiny. These examples nicely demonstrate that different displays support different purposes. Some visualizations are telling a story about the data and others help to explore the data. Presented by Bodo Kirsch.

10 June 2020

Inferring what is likely to be true based statistical analysis of data is a very difficult task and certainly not as easy as computing a p-value for declaring H0 to be true or false based on an arbitrary cut-off value. This session will include the leaders from the two most prominent statistical societies in the world – The RSS (Deborah Ashby) and the ASA (Ron Wasserstein) – and moderated by Stephen Ruberg. Listen to their perspectives and sharpen your statistical thinking.

Read more...

Stephen Ruberg, Deborah Ashby, Ron Wasserstein. 

Inferring what is likely to be true based statistical analysis of data is a very difficult task and certainly not as easy as computing a p-value for declaring H0 to be true or false based on an arbitrary cut-off value. Much has been written over the last eight decades on hypothesis testing, p-values, Bayesian inference with the conversations (or sometimes acrimonious debate) continuing. This session will include the leaders from the two most prominent statistical societies in the world – The Royal Statistical Society (Deborah Ashby) and the American Statistical Association (Ron Wasserstein) – and moderated by Stephen Ruberg. Listen to their perspectives and sharpen your statistical thinking. This session of the 2020 PSI Conference Webinar Series was kindly sponsored by Astrazeneca.

10 June 2020

Christen Gray, Kirsty Hicks and Elizabeth Williamson each speak on their work related to real world data. Topics include Comparing the impact of unmeasured confounding due to selection bias in external comparator studies using RWD, Advanced Analytics of Digital Data: A focus on sensor data, and how we can use RWD to emulate trials.

Read more...

Elizabeth Williamson, Christen Gray, Kirsty Hicks.

Elizabeth’s talk will explore different approaches to trial emulation using RWD using two examples both using data from the UK Clinical Practice Research Datalink (CPRD), a large database of UK primary care records. In the first example, data from the CPRD was used to emulate a trial of macrolide antibiotics on all-cause mortality prior to the relevant trial being conducted. In the second example, data from CPRD was used to generalise the results of the TORCH COPD trial to a patient group less represented among the original trial participants – those with mild COPD.

Christen will speak about augmentation of the control arm of a randomized controlled trial (RCT) with external data, which has been proposed in recent years where standard RCTs face enrolment restrictions. Using real-world data (RWD) for external controls is a natural next step. However, in order to do so, there need to exist accessible methods for researchers which can minimize the risk from unmeasured confounding in this setting. Bayesian borrowing methods, which discount the external data dependent upon the similarity of the outcomes to the internal controls, have been applied when the external data is prior control arms of clinical trials. The simplest of these approaches is the power prior. In using RWD, greater variation in the underlying population and measured variables is expected.

Finally Kirsty speaks on digital data and how it can be used to collect information on sleep patterns, respiration rate, step count and continuous monitoring of heart rate and energy expenditure. Collecting data through digital devices also increases patient engagement and can provide real time compliance monitoring. The regulations for use of digital technologies in clinical trials are still evolving, and current recommendations for the analysis of such data are limited. As a statistician initial questions that spring to mind include ‘How can this volume of data be analysed or represented visually?’ Statistical methods that can be applied to this data are under investigation and this presentation will introduce some of these. This session of the 2020 PSI Conference Webinar Series was kindly sponsored by Amgen.

09 June 2020

This session covers the essential principles that statisticians need to be aware of regarding clinical data and patient privacy (e.g. GDPR) and how data can be re-used responsibly.

Read more...

Rebecca Sudlow, Janice Branson, Katherine Tucker.

This session covers the essential principles that statisticians need to be aware of regarding clinical data and patient privacy (e.g. GDPR) and how data can be re-used responsibly. What has changed? Why is data now an asset? Over the last decade we have seen a massive change with regard to data transparency from clinical trials - increase in sharing through CT.gov, EUDraCT as well as many companies and institutions actually allowing access to individual level patient data upon request from researchers. More recently EMA and other Health Authorities have rolled out policies to make documents such as anonymised clinical study reports publicly available. Within Pharma companies we see many initiatives to try to make better use of the wealth of data the companies have and even initiatives to leverage data across companies and from health records in an attempt to have smarter, shorter and more ethical drug development programs. This is all needs to be done taking into account new legal requirements coming from GDPR for example as well as working in the framework of informed consent or anomymizing the data or even exploring newer areas like synthetic data.
Page:

Upcoming Events

Latest Jobs