Welcome to The Effective Statistician Conference!

The following speakers have already agreed to present:

1. Heidi Seibold (1:00pm CET) – Open and Reproducible Data Science Trainer & Consultant
   
   Title: Why Reproducible Research?

   
   Abstract: Reproducibility is a minimum standard when working with data: when the same analysis is conducted on the same data we expect the same results. Yet, making analyses reproducible is something most people struggle with. In this talk I will explain why reproducibility is so crucial and share a few tips for and insights on reproducible research.

2. Kaspar Rufibach (1:25pm CET) – Expert Statistical Scientist within the Methods, Collaboration and Outreach group at Roche

   
   Title: Some challenges with implementing estimands in real life

   Abstract: Ever planned a trial and struggled with features such as potential cure, multiple treatment decision points, or complex interventions such as transplantations or CAR-T? Or been interested in overall survival and wondered how to account for treatment switching after a drug has been approved based on progression-free survival?Have you been involved in discussions around how to assess and model the impact of Covid or the war on Ukraine on a clinical trial?Have you ever been asked to run analyses for “subgroups” generated by post-randomization variables, such as e.g. occurrence of some toxicity or dose modification?
   And the question of all questions: Can you explain in mathematical terms why we randomize?|Often, these type of questions are answered using simplified analyses that are only valid under strong assumptions.
   Using examples we will provide answers to the above questions, by discussing the application of the ICH E9(R1) addendum to clinical trials. To make implicit assumptions transparent we will discuss estimands, estimation methods, and impact on data collection. To explain principal stratification we will give a very brief introduction into potential outcomes and causal inference. This is essential to understand why we actually randomize in clinical trials.
   More on the oncology estimand WG: http://www.oncoestimand.org

3. Thomas Neitmann (1:50pm CET) – Senior Data Scientist
    

   Title: The Power of Open-Source for Clinical Trial Data Analysis and Reporting

   Abstract: The Pharmaceutical Industry has for decades relied on commercial software to analyse and report its clinical trial data in order to gain market approval for novel therapeutics. Over this period there has been negligible improvement in terms of analysis and reporting capabilities. This is in stark contrast to Data Science outside the clinical trial realm which evolved at a rapid pace. What’s the reason for this discrepancy? This talk will argue that Open-Source Software is a main driver of innovation in Data Science at large which has the potential to profoundly improve the way clinical trials are analysed and reported. In particular the collaborative nature of Open-Source software development and its positive implications for sponsor companies will be discussed. Ultimately, the case is made that the industry as a whole should adopt and collaboratively build open-source tools for all aspects of statistical programming.

4. Gary Sullivan (2:15pm CET) – Leadership Expert/Statistical Scientist at Espirer Consulting.
   

   Title: Courage, Curiosity and Statistical Leadership

   Abstract: Some simple definitions of leadership include “influence without authority” or “doing the right things”. Other definitions may include phrases like “inspiring others” or “driving change” or “achieving results”. It’s hard to find a definition with the word “courage” but one could argue that courage is a common characteristic of exceptional leaders. Although courage may conjure images of heroes physically saving lives, a more realistic definition might include words like sacrifice, risk, or facing fear.
   In this talk, I will discuss what courage means for statisticians, the many opportunities statisticians have to be courageous, and how curiosity can be a stepping-stone to courage. I will also address fear – how you should process fear and why rising leaders need to be mindful of it.

5. Stefan Walzer (2:40pm CET) – CEO, President & Founder at MArS Market Access & Pricing Strategy

   

   Title: Beyond drug development – possibilities for Medical Statisticians in the digital health landscape

    Abstract: Historically, most health care statisticians focus on drug development in the pharmaceutical world. However, besides fields such as medical devices and diagnostics there is a quite new field of application available: The digital health application.

   Since 2020, it has been possible to apply to the Federal Institute for Drugs and Medical Devices (BfArM) in Germany for reimbursement by the statutory health insurance funds for DiGA. DiGA are digital medical devices that can be prescribed. The BfArM guideline “The Fast Track Procedure for Digital Health Applications” provides information on the specific requirements, design and approval of DiGAs for manufacturers.
   The examination of the requirements and the decision on provisional or permanent inclusion takes place within three months of the application. If a positive supply effect has not yet been sufficiently proven, but all other requirements have been met, the manufacturing company can apply for provisional inclusion in the DiGA list. In this case, the necessary proof of a positive supply effect must be provided within 12 (in exceptional cases 24) months in a specific DiGA study. For such studies the input and experience of statisticians is obviously needed.
   Other countries such as France, Austria and others are currently also developing processes similarly to the DiGA process in Germany.

6. Kimberley Hacquoil  (3:05pm CET) – Chief Data Scientific Officer at Exploristics
   
   

   Title: Fake it to make it: Informing strategy with simulation

   Abstract: In this new digital era, where we are already inundated with information and data, why would we want to create more data? Doesn’t more data mean more noise, more confusion, and more uncertainty?
   I would argue that if it’s the right data then more data brings transparency, clarity and drives clearer decision-making. In this talk I will be covering why it’s so critical to embrace uncertainty through simulation of appropriate data to inform the design and analysis of clinical trial programs.

7. Miguel Pereira (3:30pm CET) – Statistical Consultant, Cogitars

   

   Title: Phase I Clinical Trials using the Bayesian Optimal Interval (BOIN) Design – a “How to” guide

   Abstract: The Bayesian Optimal Interval, or BOIN, design is a alternative to the still commonly used 3+3 dose escalation design in phase I clinical trials. It is a model-assisted method that combines the superior performance of model-based designs with the simplicity of algorithm-based designs. The BOIN is based on estimating dose escalation and de-escalation boundaries derived from a pair of pre-specified toxicity probability thresholds for underusing and overdosing.
   In this talk, we will focus on the methodology behind the BOIN design and run a demo applying it in a clinical trial setting using a well-known online tool.

8. Jenny Devenport (3:55pm CET) – People & Product Leader, Product Development Data Sciences at Roche
   
   Title: Is Medical Affairs the Wild West of Statistics? (Part II)- Never bring a knife to a gunfight

   
   Abstract: Drug development is a highly regulated process, designed to ensure that market authorizations are granted only after a medicine is demonstrated to be safe and effective in the target patient population.  But believe it or not, market authorization is just the start—getting new medicines to appropriate patients in clinical practice is the finish line and this involves addressing additional stakeholder questions in a somewhat less structured and potentially more resource-constrained environment. Statisticians working in the access and medical affairs arenas must have a thorough understanding of the strengths and limitations of the data leading to health authority approval, curiosity about the needs of different stakeholders, familiarity with the treatment landscape (ie., current and future treatment options), tolerance of ambiguity, and creativity to address evidence gaps.  In this session I will introduce the needs of different stakeholders, their evolving evidence demands, and opportunities for statisticians to add tremendous value.

9. Greg Ball (4:20pm CET) – ASAP Expert

   

   Title: How to work together on a multi-disciplinary team – insights from over 20 years as a biostatistician and 8 years of dedicated work in safety evaluation.

   Abstract: We all work on projects with people from other functions, but no other area depends on cross-functional communication and mutual understanding more than safety. Safety, clinical, epidemiology, statistics, and other functions work together to get to a deep understanding of the product safety profile.
   Safety assessment is fundamentally different from efficacy analysis and we cannot continue to simply repurpose the same approaches established for efficacy. Adjusting for multiplicity is one such example.
   In this presentation, Greg will share his rich experience of cross-disciplinary scientific engagement working on safety management teams. However, not only safety statisticians will benefit from this.

10. Alexander Schacht (4:45pm CET) – CEO at Sanevidence

    
    Title: How to get more time to work to create great data visualisations!

    Abstract: The number #1 reason for not creating better data visualisations I hear is a lack of time. And I can understand this very well given the increasing number of responsibilities statisticians have.
    Still, great data visualisations will help you communicate your results much more effectively than tables or the usual bar and line charts. And they will make you stand out from the crowd.
    In this presentation, I will speak about ways how you can get more support and create data visualisations more effectively.

11. Alex Dmitrienko (5:10pm CET) – President at Mediana LLC
    
    

    Title: Key multiplicity issues in clinical trials

    Abstract: This presentation will provide an overview of key multiplicity issues in confirmatory clinical trials. The importance of the general topic of multiplicity has been emphasized in the recently published regulatory guidelines (EMA, 2017; FDA, 2022). Multiplicity problems arising in clinical trials with multiple objectives, including the analysis of multiple endpoints, dose-control comparisons and patient populations, will be presented. Software implementation of popular multiplicity adjustment methods and sample size/power calculations in clinical trials with complex objectives will be discussed using R packages (Mediana and MedianaDesigner).