The work of a statistician

Ellie Curnow is a statistician in Statistics and Clinical Studies, based in Stoke Gifford, Bristol. She is also a part-time PhD student in Population Health Sciences at the University of Bristol and her blog describes some of the work she does for NHSBT.

I’m aware that statistics can seem mystifying, if not down-right terrifying, so I wanted to give you an idea of a statistician’s role in NHSBT.  Here is some of what I did in the second week of January.

On Monday,  I had a meeting with my PhD supervisors.  I’m currently researching methods for handling missing acute graft-versus-host disease (GvHD) data.   Acute GvHD is very common after stem cell transplant and can be fatal, so it’s really important to make sure we are making the best use of the info we have available.

On Tuesday, I had a telecon with a colleague in the Components Development Lab about a red blood cell (RBC) wash methods study.  A big part of my role involves understanding study requirements.  There’s no point in doing a complicated analysis if it doesn’t answer the right questions, so good listening skills are key.  We sorted out the analysis I needed to perform and the rest of the day was spent running and writing up the results.

On Wednesday,  I had another telecon, this time with the rest of the team for the TREATT trial: a clinical trial of tranexamic acid to help prevent bleeding in blood cancer patients.  The trial team is spread out across the country (and Australia!) and discussion can be on a wide variety of topics, from recruitment of study participants to drug supply to clinical research.

I like hearing all the updates, even the bits that don’t seem very statistical at first glance – it’s surprising what can end up affecting the analysis.

On Thursday, the day started with a check-in (another telecon!) for a study of granulocytes usage.  This is not a clinical trial, so we’re not changing patients’ care in any way; instead we’re observing what happens in usual practice.  Although this means there can be a bit more flexibility with the study design and analysis, it can be harder to achieve good quality data.  After the telecon, I prepared to be on the panel of a statistical drop-in session.  This session provides an opportunity for members of the department to bring complex statistical queries for wider discussion. The session itself was a bit of a statistical workout, but it’s great to share ideas with other colleagues; usually I’m the lone statistical voice for each of my projects.

On Friday, I was wrestling with some donor lymphocyte infusion (DLI) data.  DLI is a therapy offered after stem cell transplant to ‘boost’ the donor stem cells and reduce the chance of a patient’s disease returning.  Tracking down the DLI data in NHSBT’s databases is not easy.  It then needs to be combined with patient data from an external registry, which involves careful checking that the two datasets match and are consistent.  This stage requires good detective skills and always takes longer than I think it will, but it’s very satisfying when all the problems are untangled.  As usual, my day ended abruptly at 2pm with a mad dash for a train and school pick-up.

Looking back on the week, I see that I discussed, listened, planned, explained, wrote, investigated, sorted…and spent a small fraction of my time actually doing statistics too!  But it’s the variety of people I work with, projects, and range of skills I get to use in my job that makes it so interesting and makes the Statistics and Clinical Studies Department such a special place to work.

If you liked this blog, leave a comment for Ellie in the box below.

Leave a Reply

Your email address will not be published. Required fields are marked *

One comment

  1. 1

    Great blog Ellie – thanks for explaining what you do. I am always pleased to hear the number of times you ‘touch’ the operational businesses and how you can influence what we do by looking at the data effectively – amply demonstrated by your week

You Might Also Like