How to help statisticians help you

You’re alone in a dark office, surrounded by empty coffee cups, staring at a messy spreadsheet. Deadlines are creeping up, but urgency doesn’t help the numbers make sense. You think back to those statistics lectures from your undergraduate years, or maybe when you started your research. But that isn’t helping today – you need more.

Working effectively with statisticians can save a lot of headache.

At some point, every clinical researcher needs statistics to interpret their data. Sometimes this is simple. But not everyone has the extensive knowledge and training required for some of the more complex analyses, and this is where statisticians can help. Particularly for large clinical trials with many variables and endpoints, their expertise can help untangle the seemingly impenetrable web of patient demographics, risk factors, disease characteristics, and so on. For anyone who isn’t immersed in statistics on a daily basis, experts can guide you out of a rut and on with your project . . .or prevent you falling in that rut in the first place.

Dr Hannah Johns is a statistician at the Florey Institute of Neuroscience and Mental Health who uses her expertise to improve stroke rehabilitation. In the AVERT trial, which involved over 2000 participants from eight countries, her statistical methods helped discover findings that affect stroke guidelines worldwide. She is now working on AVERT-DOSE, where she is responsible for the adaptive randomisation module used, and provides developmental support.

If you haven't got much statistics experience yourself, it's sometimes hard to know how to get help. So how can you get the most out of your statistician?

 Hannah Johns

‘Be prepared to bridge the knowledge gap,’ says Dr Johns. ‘As overwhelming and unknowable as a lot of statistics can seem from another researcher’s perspective, the same holds true for the research context from a statistical point of view.’

So, what does a statistician want to know about your research in order to best help you? ‘It’s important that both statisticians and other researchers share a common understanding of what specific questions are being asked. Be sure to allow for time for the statisticians you’re working with to understand your research problem, as well as any logistical challenges you may have around data collection and data quality.’

And it’s okay to admit when you don’t understand what the statistician is talking about. ‘If you don’t understand the statistical methods being suggested, ask questions until you do,’ Dr Johns suggests. ‘Making sure that you’re on the same page is critical to getting the most benefit when working with a statistician.’

Dr Hannah Johns is an early-career researcher in the AVERT Early Rehabilitation Research Group. Her work focuses on the development and application of new statistical and visualisation methods to understand complex clinical trial data.