You can download the slides of the talks below from this page.
Closing keynote speaker
Thursday, August 27
11.00 - 12.00
Emmanuel Lesaffre - Catholic University of Leuven, Belgium
Errors in evaluating errors: Misclassification in clinical studies
Abstract: Measurement error occurs in basically each clinical study. When it occurs on a categorical variable, it is called misclassification. To assess the degree of misclassification directly a gold standard and a validation data set are needed. When no gold standard is available then misclassification can only be evaluated indirectly via the (dis)agreement between different scorers. In practice, too often only the misclassification is reported without further action. Unbiased estimates of model parameters requires correction for misclassification, nevertheless some argue that correction might not be worthwhile. We show the effect of ignoring misclassification in a regression analysis. In addition, in this talk we review frequentist and Bayesian methods for correcting for misclassification. Admitted, an efficient correction for misclassification requires reliable correction terms. This may not be easy in practice. Here we discuss practical difficulties in deriving reliable correction terms from validation studies. But, we also show in some cases correction for misclassification can be done without a validation study. Finally, we show that even when correcting for misclassification cannot be achieved validation studies are useful for improving the measurement process, namely to explore the factors that cause misclassificaton. We consider the case of misclassification of binary, nominal and ordinal variables used as response or covariate in a regression model. Illustrations are taken from several medical studies conducted in Belgium and the Netherlands.
Emmanuel Lesaffre will be the closing keynote speaker on thursday morning. He has done his undergraduate degree in Mathematics at the University of Antwerp and Doctorate in Science at KU Leuven (both in Belgium). He was full professor and chairing the Biostatistical Centre at KU Leuven from 1998 to October 1, 2007. From 1 October 2007 until 30 September 2014, he was head of the department of Biostatistics at Erasmus Medical Centre in Rotterdam (the Netherlands), and had a part-time appointment at L-Biostat of KU Leuven and at the University of Hasselt (Belgium). From 1 October 2014, he is full professor at L-Biostat, visiting professor at the University of Hasselt and honorary professor at Erasmus University in Rotterdam.
Emmanuel is the co-founder of the local Royal Statistical Society group “Three Country Corner”. He has been the Secretary and the President of the International Society of Clinical Biostatistics (ISCB) in 2007 and 2008, the founding chair of the Statistical Modelling Society (2002-4). In addition, he has been a member of various ISCB subcommittees, is a ISCB honorary member, is an elected member of the Task Force on Design and Analysis, and also a ASA and ISI fellow. He is also the initiator and (co)-organiser of the International Meeting on Methodological Issues in Oral Health (Leuven, 2004, Ghent, 2006, Gargnano, 2008, Istanbul, 2010, Graz, 2012, Adelaide, 2014), organized the Bayes2013 meeting in Rotterdam and is part of the organizing committee of this annual meeting on Bayesian methods. Furthermore, he has served for various statistical, medical and dental journals as editor, associate editor or referee.
Emmanuel is the author of about 350 peer-reviewed papers in (inter)national peer reviewed journals in statistical journals like JASA, JRSSB, Biometrika, Biometrics, SIM, Applied Statistics, Statistica Sinica, etc. but also in medical/dental journals like the JDR, Community Dentistry and Oral Epidemiology, Dental Materials, Lancet, JAMA, etc. He is the (co)-author of 6 books, among which principal editor of the Wiley book: “Methodological and Statistical Issues in Oral Health Research” together J. Feine, B. Leroux, and D. Declerck that appeared in 2009. He is the first author of the Wiley book “Bayesian Methods in Biostatistics” together with Andrew Lawson (2012). He is also the co-editor of the Springer book “Understanding Evidence-Based Rheumatology” with H. Yazici and Y. Yazici as other editors. He is the co-author of a forthcoming CRC book “Analysis with Interval-Censored Data: A Practical Approach with R, SAS and BUGS”. Furthermore, he has been a (co)promoter of about 40 doctoral students. Finally, he has been an invited speaker at numerous statistical and medical meetings.
Emmanuel was the (co)-promoter in about 20 grants both statistical, medical as well as dental and has been involved in more than 50 clinical trials either as a member of the Data and Safety Monitoring Board, or as a member of the Steering Committee or as the principal statistician.
He has taught elementary courses on biomedical and dental statistics to medical and dental students, Bayesian statistics, multivariate analysis, regression analysis, repeated measurements analysis as part of master programs in (Bio)statistics at KULeuven, UHasselt, Erasmus MC at Rotterdam and the University of Leiden. Apart from these courses he has taught advanced statistics courses at various conferences and universities.
President's invited speaker
Monday, August 24
9.30 - 10.30
Robert Hemmings - Medicines and Healthcare products Regulatory Agency, London, UK
Evolutions to support the next era of medicines' development
Abstract: Much has changed in the development and licensing of medicines over the last 15 years. For example, our understanding of human biology and pharmacology permits targetted therapeutics and gene therapies and the pharmaceutical industry comprises a greater number of small and medium size drug development enterprises supporting fewer ‘me-too’ drugs with a greater focus on rare diseases. The regulatory environment has also changed with the introduction of legislation focussed on medicines for children and pharmacovigilance, and the impact of Health Technology Appraisal has been critical in determining patient access. Another change has been the demand for increased transparency and engagement with regulatory agencies, with increased involvement of patients in discussions and decisions, proposals for access to clinical trial reports and individual patient data, and greater opportunities for multi-stakeholder dialogue and engagement in the drug development process. The talk will present an overview of this evolution and highlight some of the impacts on clinical drug development, clinical trial methodology, the pharmaceutical industry and the regulator. I will highlight areas where the evolution of standards appears to challenge well-established statistical principles and gives rise to opportunities for methodological research. This evolution is ongoing and I will indulge in a little crystal ball gazing.
Robert Hemmings will be the president's invited speaker on monday morning. Rob has been with the Medicines and Healthcare products Regulatory Agency for 15 years and heads the group of medical statisticians and pharmacokineticists. The group provides technical and methodological expertise to the assessment of marketing authorization applications, ‘scientific advice’ to drug developers and input to methodological and therapy area regulatory guidance documents. Rob is a co-opted member of CHMP, a committee responsible for preparing the European Medicines Agency's (EMA) opinions on questions concerning medicines for human use. He is also the chair of the CHMP’s Scientific Advice Working Party (SAWP) and a member of CHMP’s Biostatistics Working Party and the EMA Modelling and Simulation Working Group.