CORe Advanced Statistics Course
Causal inference in neurology
5-7 February 2020
Mantra Lorne, Australia
Jonathan Sterne, NIHR Bristol Biomedical Research Centre, University of Bristol, UK
Tony Blakely, Melbourne School of Population and Global Health, University of Melbourne
Damjan Vukcevic, School of Mathematics and Statistics, University of Melbourne
Charles Malpas, CORe, University of Melbourne
Gareth Ball, Murdoch Children’s Research Institute, University of Melbourne
Tomas Kalincik, CORe, University of Melbourne
Aim of the course
The mission of the CORe Advanced Statistics Course is to deliver theoretical and practical training in advanced topics in statistics and epidemiology, with applications to clinical research in neurology, to clinical neuroscientists who have fundamental knowledge of statistical methods.
In November 2019, we contacted leading national and international researchers and clinicians with an invitation to attend the 2020 CORe Advanced Statistics Course. We received an overwhelmingly positive response, with applications received from many states around Australia, as well as France, Germany, Switzerland and Italy. Those wishing to attend were all required to submit a brief letter outlining their interest in the course and the anticipated benefit, as well as a CV, in order to facilitate selection of participants based on their appropriate existing knowledge of statistics. From this pool, 35 participants were provided with first-round offers to attend. Following this, seven second-round offers were made when additional places became available. The final breakdown of attendees was 9 CORe members, 5 speakers, 6 industry representatives and 32 applicants.
The CORe Advanced Statistics Course (CASC) was held at the Mantra Lorne, Victoria, Australia between the 5th-7th February, 2020. The course was endorsed by the European Committee for Treatment and Research in Multiple Sclerosis (ECTRIMS). The course covered a variety of timely analytical topics relevant to the present and emerging clinical research in neurology, with focus on the main theme of the course – causal inference. The presentations included lectures, seminars and discussions on moving beyond p-values, using frequentist and Bayesian approaches, role for machine learning in causal inference, personalised medicine, model selection, confounders and comparative effectiveness analyses. The level was tailored to audience with existing substantial knowledge of statistical theory and methods. Within this framework, some of the teaching blocks provided introduction to advanced analytical topics, whereas other blocks focused fully on advancing the knowledge in highly specific topics in causal inference.
The course consisted of 9 interactive blocks distributed over two days, with total teaching time of 13 hours. The mode of delivery was varied, including lectures, discussions, questions & answers, illustrative examples, critical appraisal of published research work, and practical problems with demonstration of solutions.
In order to help quantify the outcomes of the CORe Advanced Statistics Course, participants were asked to complete surveys at the beginning and end of the two days. Overall, the results indicated that the course was successful in meeting its goals, both as an educational tool and as an event. From a learning perspective, the participants responded that there was an increase in the mean understanding and confidence in interpreting for each of the 10 topics covered. In particular, there was a large positive shift in the mean group scores observed for Bayesian approaches, machine learning in causal inference and analytics for personalised medicine.
Overall, the feedback received was just highly encouraging. All attendees were asked to rate the quality of the presentations, the location of the event, the complexity of the material and the length of the course on a scale ranging from 1 (low) to 5 (high). In all of these categories, the average scores fell between 4 and 5, highlighting the high regard with which participants viewed the speakers, the content they covered and the organisation of the event.
Importantly, participants were also asked to score how valuable they thought the lessons they learned at the course would be for their research. Once again, the average score received for this category was in excess of 4.5. Attendees strongly believed that their participation at the course would have a significant positive influence on their approach to neurological research projects over the coming year. Further emphasising this point, we also enquired about whether those who attended this course would participate in a future event that covered different topics and if they would recommend it to their colleagues. Pleasingly, the scores for these questions were also both in excess of 4.5. In our eyes, this not only reflects benefit that the inaugural CORe Advanced Statistics Course brought to the participants, but also the need for further iterations of his event, with the aim of addressing further key topics inherent in the research of observational data in neurology.
Conclusion and future direction
The overwhelming feedback received from participants, both in-person at the event and through their feedback surveys, shows that the inaugural CORe Advanced Statistics Course was a success. This would not have been possible without the generous support of our sponsors: Biogen, Celgene, Merck, Novartis, Roche and Sanofi-Genzyme. We value active participation of the delegates from our sponsors as attendees at the course. The course was free from any commercial bias.
Based on the positive feedback from the attendees at the course, we have decided to offer another CORe Advanced Statistics Course in the future. The topic of the next course will be new and complementary to the first edition of the course. We will advertise the course and its theme to the potential participants later in 2020.