In March 2017, I was working as a research fellow in Professor Willie Hamilton’s Discovery Group in the Medical School. My focus was on quantifying the impact of updated national cancer referral guidelines on cancer outcomes in the UK. I had recently completed my PhD, and had a keen eye on my progression as a researcher. Up popped an email about seed corn project opportunities for “discipline-hopping” at the Centre for Biomedical Modelling and Analysis (CBMA). It looked like a golden opportunity to develop my interdisciplinary skills and build my CV, all working towards submitting a fellowship application in 2018. I quickly contacted Chrissie Walker, and it all took off from there.
I began a 6-month placement at the CBMA in July 2017, working on a seed corn project proposed by Lisa Price and Jeff Lambert from the Department of Health and Sports Science. The project is a key stepping stone towards designing future studies to determine the “dose” of activity associated with an improvement in mood in people with depression. It uses secondary data from Jeff’s eMotion study; namely, physical activity measured using wrist-worn accelerometers, and self-reported mood.
I was given two main tasks… The first was to write an R script to process raw accelerometer bin files tailored specifically to the project’s needs. The second was to explore methods for modelling the temporal patterns between mood and physical activity in people with depression. Low levels of activity are a feature of depression, and any analysis method needs to take account of this added complexity.
While I am pretty slick at using Stata, I was definitely an “R newbie”. I started to find my feet, with help from Ryan Ames and various websites – this being a favourite: aRrgh: a newcomer’s (angry) guide to R (with the dispiriting opening lines “R is a shockingly dreadful language…”). Six months on, I have improved massively and am even able to write my own (rudimentary) functions. I am close to finishing the bespoke R script for Lisa and Jeff.
I did lots of reading around potential modelling methods, and received much support from researchers at the Living Systems Institute and Maths Department. Thanks to Lauric Ferrat, Mohammod Mostazir, Hossein Mohammadi, Eder Zavala, Piotr Slowinski, Petroula Laiou, Prof Salvador Olmos, Margaritis Voliotis, Mark Kelson and others for all the valuable discussions.
I am learning time series analysis methods, which will be useful when I return to working in Willie Hamilton’s group in the medical school. I’m also working with Piotr, applying multidimensional scaling to determine a person’s individual motor signal.
The secondment came with a training budget, which I used to attend a Bayesian analysis course at the MRC Biostatistics Unit. This was fantastic – a huge benefit of the placement.
Later on, I started developing a fellowship application. The CBMA arranged a seminar on approaches to public involvement, which is massively important to funders of clinically oriented research. We were encouraged to apply for Wellcome Trust engaged research funding, so I did and was awarded £3,000. My stage 1 application for a National Institute for Health Research fellowship went in on 1 December 2017. Public involvement is the main focus of the second stage of the application, so I will be well placed if I get that far.
I’m sad that the secondment is drawing to a close, but will remain in contact with the CBMA after my return to the medical school. If you’re thinking about applying for a secondment, I’d say definitely go for it. They are hard work, and do require a strong sense of initiative, but the rewards are plentiful!
