Many of us in the New South Wales branch will remember the winter of 2021 for being locked down due to a stubborn COVID-19 Delta variant outbreak. As of late August, the lockdown was still in place and the August branch meeting was carried out totally using the Zoom tele-meeting application. We have some bright spots: spring is around the corner and accelerating vaccination rates are pointing to gradual returns to normality. Another bright spot was a pandemic-era seminar by Dr Nancy Briggs, who is a manager and senior statistical consultant at Stats Central, University of New South Wales. Around 30 people signed in for the event.

A major theme throughout the talk was the impact of the COVID-19 pandemic, starting with how it has changed where we work, how we work and the kind of work that we can do. Statistical issues concerning, for example, conclusions in a COVID-19 world compared with those pre-pandemic were discussed.

A running example throughout the talk concerned a randomised clinical trial on parenting of children aged between 14 and 24 months. The trial had three treatment arms: parent-child interaction therapy-toddler, circle of security-parenting and waitlist control. Some of the outcomes were parenting sensitivity and stress. The writer of this article was pleased to see that generalised linear mixed models played a central role in the analyses. The study was designed before anyone knew that a pandemic was about to occur, but the analyses were carried out in the early 2020s. After it was clear that the pandemic was having a big effect on the many randomised clinical trials going on, authors Cro et al. published a 2020 paper titled "A four-step strategy for handling missing outcome data in randomised trials affected by a pandemic". Speaker Briggs told us that she found this paper very useful for her recent research. One issue was whether data, missing or observed, should be treated differently depending on whether participants are directly affected by COVID-19 (either by infection or changes in treatment). Pandemic-related papers by Degtyarev et al. (2020) and Meyer et al.(2020) were also mentioned.

The main body of the talk was a detailed look at the toddler parenting randomised clinical trial by following the four steps recommended by Cro et al. (2020), as well as advice in Degtyarev et al. (2020) and Meyer et al. (2020). These latter papers recommend that, for trials that started pre-COVID-19, a pandemic-free estimand be chosen since this is what the trial was established to do. The Cro et al. Step 2 is "Establish what data are missing for the chosen estimand". For the toddler parenting clinical trial Nancy explained that there are missing data due to ceasing data collection. An example of a recommendations from these references is to summarise study population characteristics before and after pandemic onset.

Even though we are all looking forward to getting this pandemic behind us, it was interesting to see how it is impacting applied statistics.

Matt Wand University of Technology Sydney