CSM medical statistician Alex Ooms won best oral presentation at the 2020 Young Statisticians’ Meeting. Alex shares his experience of the online meeting. Two days of statistics online: what could go wrong?
The Young Statisticians’ Meeting (YSM) is an annual conference for early-career statisticians and students to meet similar-minded folk, discuss budding ideas, and present their work in a new city. YSM 2020 was due to be held in Manchester but because of You Know What had to be moved online. The new city was replaced by my sofa and the busy hive of conversation snubbed in favour of a chat box. I did still get to present, despite some obligatory technical glitches. Not only was this my first online conference, but the first conference I’ve attended and presented at as a medical statistician. I didn’t quite know what to expect.
The two-day conference (29-30 July) was run via Blackboard Collaborate. Each session had its own web link, and only the session host and presenter could talk or change slides, minimising the need for reminders to mute and the sound of small children wailing in the background. With 14 contributed talks, 11 posters, and 3 keynote speakers, there was a lot to get through. Fortunately, the organisers were fantastic at providing helpful resources ahead of the conference and at dealing with issues as they arose in this new format, so I always felt like I was in very safe hands.
My 15-minute talk was called “Exploring the use of Bayesian techniques in a surgical trauma randomised control trial” and went through some advantages of using Bayesian techniques in late-phase trials, rather than the more common frequentist methods. For example, Bayesian analyses are simpler to interpret because the resulting posterior distributions allow direct probability statements about parameters of interest based on the observed data. You can find the probability of observing that your tested treatment makes a minimum clinically important difference in your outcome, given the trial data, for instance. Bayesian methods also allow historical data to be incorporated where appropriate. The talk was based on a real trial I’ve been analysing this summer, DRAFFT2, where I’ve used its predecessor, DRAFFT, to make assumptions about certain parameters (called informative prior distributions) to contribute to the analysis.
The excitement of presenting was doubled when the person before me had audio problems and then my slides froze, forcing us to abandon our Blackboard session and prompting a quick shift to an improvised Zoom call. After the hiccups had been resolved, the talk went smoothly, and I got plenty of engaging questions. I reckon there may even be a few more closet Bayesians lurking in some UK universities now. Much to my surprise, I even managed to win best oral presentation for my efforts.
Despite the challenging circumstances, the coordinators managed to achieve a friendly and encouraging atmosphere, as best you can virtually. Two full days of very eclectic topics, ideas, and speakers were taxing, amplified by the inability to change setting. But it was also incredibly rewarding and inspiring to see such deep thought being put into important and relevant problems. The keynote speakers were particularly stimulating with Green Templeton College’s Professor Denise Lievesley speaking brilliantly about the battle against misinformation and the role of statistics in the fight.
Overall, I had a fantastic experience at YSM 2020 and would highly recommend attending, or better yet, giving a talk or presenting a poster – hopefully in person next time where less can go wrong. I believe this is the third year in a row that a CSM representative has taken home a prize, so there’s also our winning streak to keep alive!