My First Year as a Pre-Doctoral Training Fellow
In October 2019 I started working at Imperial Clinical Trials Unit (ICTU) as part of a 2-year NIHR pre-doctoral training fellowship. Previously I had worked as a statistician in the media sector performing analysis on HR and finance datasets to provide business insights for senior management relating to hires and acquisitions; I also performed client-side analysis assessing how our advertising strategy was benefitting them, as well as delivering pitches to potential clients. However, my interests had always lain in medical statistics, so changing career paths had been on my mind for a while. Moving from the fast-moving world of instant deadlines, demanding clients and repetitive quarterly data collections, it was a culture shock to enter a world of long projects and self-driven research. After a year working at ICTU, I felt it was a good time to reflect and share my experiences to help anyone thinking about applying for a pre-doctoral fellowship or making the move into medical statistics.
I had wanted to make the move to a career in medical statistics for a while, so in 2016 I embarked on a part-time masters at Sheffield University whilst continuing to work in the media sector. Once I’d completed my masters, I started applying for statistical roles. Whilst applying for a role at ICTU the idea of a pre-doctoral training fellowship was raised. It seemed like a great opportunity to gain the applied skills and experience needed to introduce me to the field whilst also setting me up for further progression. I decided to apply and worked with my supervisors to develop the application and project plan. As someone just embarking on an academic career, this was my first time having to produce anything like this and I was surprised to see it was far more about selling myself as a potential researcher and putting together a training plan for me to undertake than about a specific project. However, I received great assistance and advice from ICTU in putting it together and was successfully awarded the fellowship in August 2019.
The first few weeks at ICTU were a quick plunge into medical statistic, I spent my time working my way around SOPs filled with best practice and unfamiliar acronyms, as well as learning a new programming language (Stata) that was to become my main language for analysis moving forward, and I was also given the opportunity to attend my first academic conference. I was soon able to apply many of my learnings from my masters, whereas in my media role the analysis was mainly time-series and regression modelling in Excel or R. It was quite a learning curve, but with training, on-going practical application and support from my supervisors the lingo and programs soon became familiar and comfortable to use.
After settling into ICTU I was excited to start work on the research project I had developed for the course of the fellowship. The research project (an investigation into how Bayesian methods can be used to overcome issues in trials of rare disease due to limited sample sizes) incorporated both a systematic review of rare disease trials and a re-analysis of a current trial using Bayesian methodology. This gave me the opportunity to experience both methodological research and applied analysis that medical research involves. Throughout the year, as I have learnt more, we have found opportunities for the project to expand into more areas, including plans for multiple Bayesian elicitation meetings and a qualitative review about the process. Undertaking such a range of research projects, I have been able to better understand where my interests lie and which I struggle with. Whilst searching through hundreds of papers for a systematic review can feel repetitive and tiresome, it helped me understand the importance of setting up and undertaking thorough searches so as to avoid repetition and to better understand how to identify gaps in existing research. I highly enjoyed the applied work, attributing new methods to real trial datasets, and it was thrilling to see whether methods I had read about and ideas I had could indeed make a real impact. Being able to see the bigger picture was important to stay motivated and not get lost within the research, without which I wouldn’t have been able to accomplish the applied work.
Whilst my fellowship has focused on mainly research methods, I have had the opportunity to take part in multiple areas of ongoing applied trial work. This has ranged from helping to develop grant applications to performing final analysis for clinical trial reports. For this, I have been able to shadow colleagues working on grant work ups and replicating other statisticians’ analysis. This has given me the chance to see what different stages of applied trial work entail and to practice and apply my skills. This has helped demonstrate that this is something that I can do and something that really interests me and has helped inform my potential career decisions.
As the CTU is attached to a university I have also gained the opportunity to get involved in teaching, something that would not have been possible in my old job in the private sector or even in medical statistics roles in industry. Initially this seemed like a daunting task, with my only other teaching experience many years prior coaching trampolining and giving a few lessons of Latin to primary school children. I was worried I would be unable to help or at worst end up confusing students further. However, with interested Masters students who were keen to learn the experience was really enjoyable and I was able to successfully help with issues and queries that arose. Helping with the teaching of students really allowed me to solidify my own understanding of the statistics I was using and my Stata programming skills as well as opening my thoughts to furthering myself in teaching.
Unfortunately, as I felt I was getting into the stride of things, 2020 took a turn and we (along with much of the country) were forced to start working from home. Working from home after only 5 months in a new industry was an unnerving experience, but the support of the team and the efforts to keep the team connected helped greatly. I continued to work on multiple projects across multiple trials in collaboration with others at ICTU, whilst some ongoing trials and in person projects had to be paused, all analysis work was able to continue as planned. This movement towards mainly analytical aspects gave me the opportunity to get involved in more applied work alongside my research project. A distinct change because of lockdown was shifting a Bayesian elicitation meeting to an online endeavour. This was a large undertaking logistically and practically; and overcoming all obstacles to produce a successful meeting that was engaging to all clinician attendees is something I am immensely proud of.
One of the more major drawbacks of working from home this year has been the reduction in in-person training courses and conferences. Whilst online seminars have taken off, I have found it harder to engage as fully with them but the extra practical work I have been involved in has helped to implement my learnings and to ensure they are ingrained. It has also meant that self-learning through reading textbooks and keeping up with research papers is more key than ever.
The world of medical statistics has been vastly different from any work I have done previously, the rigid structure of hierarchy and client requests switched for a free flow of ideas and self-led projects. Whilst feelings of accomplishment when learning or understanding a new area is a real enjoyment, there has definitely been times where progress is slow. I had previously enjoyed the accomplishment of looking back on a list of completed tasks and projects at the end of a day or week, whereas research can often feel fruitless and with what feels like limited progress made for days on end. Therefore, adding in analysis of applied projects alongside has helped me to feel more accomplished in times when there were slow periods in my research.
I don’t believe I could have found a better entry into medical statistics. The fellowship has allowed me to cover multiple aspects of research and given me an understanding of the multiple facets of how research can exist as a stand-alone entity or can be tied into specific trials. Whilst working for ICTU I have also been able to branch out beyond methodological research; looking at all aspects of trials analysis and the role of statisticians at every stage of a trial, as well as teaching within a university. The support I have received in my first year at ICTU has been fantastic, and it has been great to have a group of people around me of whom I can ask for support at any time and tap into their different areas of expertise. I would definitely recommend fellowships like these, especially based within university CTUs, to anyone starting a career in medical statistics who has not yet had a chance to fully experience everything the field has to offer, or has a research proposal they wish to undertake but also want to ensure they are staying well rounded in applied research.