We recently spoke to Mohammed A Mohammed, Director of the NHS-R Community, Professor at the University of Bradford and Principal at the Strategy Unit (NHS), about his work supporting the flourishing R community and its impact in the NHS.
The NHS-R Community was born in 2018 following a successful funding bid. Today the Community is a grassroots tour-de-force, spearheading NHS data analytics into the 21st Century by liberating analysts and data scientists based in the NHS to collaborate across their organisation’s boundaries. The aim of the NHS-R Community is to promote the use of R in the NHS.
Mohammed said: “The meteoric rise of NHS-R is a testimony to the amazing people in the NHS-R Community, who volunteer their time, skills and expertise to develop open source R solutions for the NHS.”
In this interview blog, Mohammed shares his journey and tips for setting up an R community – so if you are early in your journey or a seasoned campaigner, this might offer some useful insights.
Can you start by telling us more about your role in the NHS and your relationship with the R community?
“My involvement with the NHS began a few decades ago in academia where my interests were measuring and improving the quality and safety of healthcare. I used R in my academic work to analyse data and produce beautiful graphs and it was clear from the academic literature that R was being widely used by research teams in academia. The NHS is one of the largest and best healthcare systems in the world. It was launched in 1948 with the guiding principle of being free at the point of delivery – a kind of crowd funded, open-source freeware equivalent of healthcare. Likewise, R was conceived in 1992 as a free open-source statistical programming environment. In 2018 I had the opportunity to propose an arranged marriage between my two loves - the NHS and R.”
Why was the timing in 2018 right for your community development?
“The Health Foundation, a charity, fund quality improvement research and development in the health and care sector. For the first time, in 2018, they put out a funding call aimed at data science in healthcare. This was my opportunity to seek funding for the marriage for the happy couple - NHS and R. All the ducks were now lined up! Our bid was successful. It was based on a straightforward case to support the proposed marriage.
The NHS in England deals with about 1 million people every 36 hours and is continually generating vast amounts of data. This data is one of the most precious, yet under tapped, resources in the NHS. Drilling and mining NHS data could improve the NHS. But mining these mountains of data is a colossal task. This is where R comes in. R is a data science tool widely used in industry and academia. The absence of R at scale in the NHS, means that the NHS is unable to take advantage of the huge benefits of R, including cutting-edge visualisation, statistical tools, automation, teamworking, knowledge dissemination along with a worldwide R community, which freely shares learning and resources.”
Why did you choose R?
“The NHS is a crowdfunded, open-source freeware in the sense that we all pay for it, with its knowledge being freely available to anybody in the world. Likewise, R is a free, open-source, leading data science language with a worldwide community that shares its knowledge freely. Also, my personal academic experience with R made it a good choice, although I did find that R could be hard work especially for data manipulation. So, another key development was the arrival of RStudio and the tidyverse approach – both of which were gamechangers! This now meant that R could be introduced at scale in the NHS, on steroids.
We are still predominantly R based but we are committed to supporting open-source solutions and are delighted to support a growing Python community in the NHS.”
Take us back to the early challenges of setting up the community.
“As an academic I knew zilch about setting up a community or network, but I knew a wee bit about how to develop engaging workshops. So my early challenges centred around two major questions: ‘How do we reach analysts in the NHS?’ and ‘How do we get them interested in learning and using R?’
For the first question: Along with an early website presence, I approached the newly formed Association of Professional Healthcare Analysts (ApHA) who helped me network with groups and individuals across the UK who also aligned with the vision of promoting R in the NHS.
For the second question: I offered hands-on workshops for NHS analysts to learn how to use R. These workshops were free, friendly and open to all with no programming experience required. I was very keen that anyone with an interest should feel welcome and supported to learn R. I had no idea of what people’s capabilities were – or how passionate they would be - but I set out a series of healthcare related problem-orientated workshops which were designed to appeal to analysts in their day-to-day roles. I focused on real life examples they might come across, such as demand and capacity problems in the health service and how they might evaluate complex interventions. This helped tremendously to surface the level of expertise and determine the onwards support required by the community. A key technical challenge in the early days was installing R on people’s laptops. Inevitably the first hour of every workshop was spent trying to fix installation problems. This was frustrating. Today we use RStudio Cloud which took away the installation headaches – another gamechanger!
So that is how we slowly and steadily started to build our community and then with some trepidation I announced the first ever NHS-R conference in 2018. Initially I thought no one would come, and then I feared that the venue would not be big enough!
The first NHS-R conference was a sell-out - attended by 119 delegates. Despite lots of organisational hiccups (and lots of practical lessons for me to learn – eg inadequate WiFi in the venue) the conference was a pivotal catalyst for the community. There clearly was an appetite and our 2019 NHS-R Conference was booked out in two-and-a-half hours which required a change of venue to accommodate over 300 delegates. The NHS-R conference is now the premier data science conference in the NHS”.
Have you kept to a plan for your community development, or has it naturally evolved?
“There was an overall vision and a plan which was really useful initially, but much of the growth of the community has been organic, without formal planning. Just light touch centralised coordination and support alongside decentralised execution where people with passion are liberated to experience the joy of learning, sharing, and solving together to help improve the NHS.”
How important has it been to facilitate a culture of belonging?
“First and foremost, we wanted analysts to have a safe place to learn, play and share. Explore without needing permission from HQ because it’s disruptive. I think an open-source community recognises these values - like the wider R community, we’re all very friendly, open and welcoming. There’s no hierarchy and people feel like they belong. And equally important is a mindset that if it works, it works. If it doesn’t – we just take lessons from it and move on – no blame games, no performance management, no targets, no KPIs, no bottom line.
Like the NHS it’s meant to be accessible to everyone, irrespective of their ability (to pay). I also wanted the NHS-R Community to mirror this. So, we made sure that anyone interested in promoting the use of R in the NHS could join by recognising two types of contributions – people who promote the aims as champions but are not data scientists alongside people who are data scientists.
We note that the NHS does not work alone and that it has support from public health and social care and likewise there are other open source solutions (eg python) and so we continually seek ways to embrace wider stakeholders.”
What’s made you most proud of NHS-R Community?
The NHS-R Community has touched the professional lives of data analysts working in the NHS. The personal stories are inspiring! Our journey has led to members of the NHS-R community being proud of their connection with NHS R. So much so, it is held in high esteem and regarded as a marker of quality. NHS-R is now a badge of honour to put on a CV – delightful!.
The NHS-R Community has developed solutions that did not exist before and has worked in ways it has never before. A case in point is the development of the “plot the dots” package which was developed remotely by members of the NHS-R Community, across different organisational boundaries, who did not seek permission, they just self-organised, used Github, worked as a team and just got on with it, for the joy of giving and sharing.
The NHS-R Community is owned by its members who have the freedom to continue to spawn new ideas – such as podcasts, drop-in help sessions with R and incorporating a python track at the NHS-R conference. Helping people see that they can be part of their own success has been inspirational.”
Professor Mohammed’s advice for others looking to develop their own communities is simple: “Apart from the obvious things, such as it takes time - often more than first envisaged - my key advice, is articulate a vision which touches the soul and then seek out and build relationships with passionate soulmates whilst giving them the freedom to self-organise by removing barriers and then – relax, trust and let go - and wait to see what amazing things happen when people are liberated!”