The rapid development of open source technology is consistently amazing. The landscape has changed so dramatically in just the last few years. What amazes me most is how quickly the interest has grown within the clinical world. Four years ago, content around R or Python at pharmaceutical conferences was sparse – but today we’re seeing a significant number of presentations on R, Python, and even Julia.
It’s easy to get swept up in all of this – the beauty of the graphics coming out of R is astounding, and Shiny brings interactive capabilities in the hands of R programmers instead of web developers or stand-alone products. There are countless forward-looking initiatives within the industry that are embracing this technology and accelerating the already rapid progress further.
The first challenge my team accepted after the founding of Atorus was, on the surface, a fairly simple task: use R to replicate the CDISC Pilot Submission Package that was released in 2007. Embarking on this adventure was eye opening because we found that RTF support for table creation was lacking. It wasn’t that the capabilities weren’t out there, just that they didn’t support what we do within the conventional clinical data pipeline, which includes SDTM, ADaM, and tables, listings, and figures. Instead of creating some bespoke solution for Atorus, we decided to develop a package, pharmaRTF, and release it into the open source community.
This showed us an area of opportunity: though the open source landscape is vast, the specific needs of the traditional clinical data pipeline tend to get overlooked in the solutions offered. This isn’t surprising because this is work that has been done almost universally in SAS for decades. R remains a novel tool to many working in this arena. With this in mind, we at Atorus have started dedicating attention to solutions specifically geared towards these challenges, first with pharmaRTF and then our second package release, Tplyr.
But we haven’t stopped there. In fact, we’re increasing these efforts even further by collaborating with GlaxoSmithKline (GSK). In this partnership, both Atorus and GSK are dedicating resources to open source development. Initially, we will continue to focus on expanding the capabilities of R within the conventional clinical data pipeline, both to support and improve existing solutions, as well as develop new solutions where opportunities are identified. All products of this collaboration will furthermore be released under open source licenses, enabling the industry as a whole to benefit from the solutions that we create. Through this effort, we additionally aim to address another problem we’ve seen within the industry: individual organizations creating their own bespoke solutions to problems which realistically are universally shared across the industry. We share many challenges as an industry and we all stand to gain by broadcasting solutions and sharing what we learn. We hope that the solutions released by the Atorus and GSK collaboration will be reusable, generalizable, and adaptable into most organizations’ processes.
It has been exciting to watch the paradigm shift of open source acceptance sweep across the industry over the last few years. At Atorus, we’ve deemed it essential that we embrace this and embed it into our company culture. I’m thrilled that GSK has chosen to join us in this venture, as they bring a wealth of analytical experience in clinical trial reporting. Furthermore, they have demonstrated their commitment to leveraging the power of open source software in the creation of submission-ready clinical analyses. Stay tuned – exciting things to come.