The RSS Sheffield local group held an online seminar on 17 June 2021, organised by Lyn Taylor and Lauren Cowie with speaker Professor Jen Rogers of PHASTAR.
About the speaker
Jen was an associate professor and director of statistical consultancy services at the University of Oxford before moving to PHASTAR in 2019. She is now the vice president for statistical research and consultancy at PHASTAR, providing specialist Biometrics research services globally. She has also held various positions at the RSS including vice president for external affairs. Jen regularly works with journalists to improve the reporting of statistics in the media. She is a popular statistics presenter, making appearances on radio and TV. Throughout the pandemic Jen has written numerous blog posts on current Covid-19 clinical research.
Seminar summary
The Covid-19 pandemic has thrust data and statistics into the limelight in a way it has never been before. With Boris Johnson stating that decisions surrounding easing of restrictions would be made based on 'data, not dates', many people were undoubtably monitoring the daily updates released by the government and anxiously waiting for ONS announcements. With so much data being released, how is the general public expected to make sense of it all? In this talk, Jen led the audience through just how difficult it can be for even a statistician to digest the sheer amount of information being released, and how answering even the simplest of questions is not a trivial task.
Starting with the simple question of 'How many people in the UK have Covid-19?', Jen demonstrated various estimation methods which led to very different views of the peaks and troughs of the pandemic prevalence over time. Initially using the number of positive Covid-19 tests divided by the number in population, we were shown the number of cases changing over time, peaking in January 2021. However, when this was adjusted for the number of people being tested over the same period, the picture was very different, showing a likely highest peak in April 2020. This radical change in the prevalence estimates is partially due to the low frequency of tests back in April 2020, making the true estimate difficult to quantify. This change in testing frequency, coupled with other changes in testing policies (ie from testing in hospital only to substantial home testing), makes true estimation of the number of people with Covid-19 extremely complex.
Jen went on to discuss exponential growth, and nicely highlighted that it’s not just the rate of the growth (doubling, tripling) but also the timeframe in which the growth occurred (over a year vs over a day is very important!).
Jen concluded the talk by looking at some of the news reports released over the past year, reporting data on the Covid-19 vaccines and comparisons of them. A take-home message for everyone was the common mis-conception: A 95% efficacy profile for a vaccine does not mean that if you have had the vaccine, you only have a 5% chance of getting Covid. Instead, we should think of it as of 100 people who have Covid-19, the vaccine would have prevented 95 of them! Hence, even post-vaccine, continuing to be careful and take precautions, is very important to stop the spread.
Watch the meeting here: (View on YouTube)
Author
Dr Lyn Taylor (associate director of statistics at PHASTAR, on behalf of the Sheffield Local group RSS committee).