Our paper on the statistical open-source software revolution in pharma

paper
software
Author

Daniel

Published

February 5, 2026

Summary

We are excited to announce the open-access publication of our new paper “The statistical software revolution in pharmaceutical development: challenges and opportunities in open source” in Drug Discovery Today! (Sabanés Bové et al. 2026)

This is a major milestone, not just because of the high impact factor of this journal (7.5). Most importantly, this paper represents a comprehensive overview of the transformative impact that open-source software is having on statistical practices within the pharmaceutical industry. Our paper makes this visible to a broader audience outside of the statistics and programming communities, which are already well aware of these changes. The authors of the paper come from academia, industry and regulatory agencies, reflecting the broad basis of expertise and perspectives that informed this work.

Acknowledgments

First and foremost, I would like to thank all of my co-authors for their teamwork and dedication over several years that made this paper possible: Heidi Seibold, Anne-Laure Boulesteix, Juliane Manitz, Alessandro Gasparini, Burak K. Günhan, Oliver Boix, Armin Schüler, Sven Fillinger, Sven Nahnsen, Anna E. Jacob, and Thomas Jaki. Thanks for all your contributions, discussions, and patience!

The paper would not have been published in this form without the motivating and optimistic input from Jenny Devenport. We are grateful for her encouragement to tailor the content to a broader audience and for her insightful suggestions on how to achieve this. Isaac Gravestock also provided helpful review comments.

The beautiful illustration in Figure 1 in the paper was drawn by Gaelle Klingelschmidt. Thanks Gaelle for capturing an important message from our paper in such a delightful way!

We would like to thank the Federal Institute for Drugs and Medical Devices (BfArM) in Germany for funding the open access publication of this paper, and in particular our co-author Armin Schüler for making this possible. Given that the paper is on the topic of open source, it is even more important that it is freely accessible to everyone and under a Creative Commons License (CC-BY 4.0).

Finally, Thomas Jaki received funding from the UK Medical Research Council (grant number MC_UU_00040/03) that supported his contribution to this work.

Main Messages

Of course, you should read the paper yourself to get the full picture. Here, I just want to highlight the main messages from my perspective:

  1. Open-source statistical software has become an integral part of statistical practice in pharmaceutical development, offering significant advantages in terms of flexibility, transparency, and innovation.
  2. The adoption of open-source tools presents both challenges and opportunities for statisticians, regulators, and the broader pharmaceutical community. Key challenges (and solution strategies) include:
    1. Opposition from management (yes, you need statistical software engineers for applying cutting edge statistical methods!)
    2. Validation concerns and silos within the organization (yes, regulatory agencies accept open-source software when developed and used properly!)
    3. Proprietary mindsets (remember that the medicine is the product, not the software!)
  3. Important ingredients for successful adoption of open-source software in pharma include:
    1. Good usability by design
    2. High quality through accuracy, reliability, and reproducibility
    3. Collaboration through communities for efficiency and reliability (highlighting here of course in particular openstatsware)
  4. There are many success stories of open-source statistical software in pharma, some of them are mentioned in the paper:
    1. rbmi for missing data imputation for trials with continuous longitudinal outcomes
    2. crmPack for model-based dose escalation designs
    3. bonsaiforest for subgroup treatment effect estimation through shrinkage
    4. tern for common tables, listings, and graphs used in clinical trials
    5. rpact for fixed and adaptive clinical trial design, simulation, evaluation and analysis
  5. Building dedicated research software engineering teams within organizations can greatly facilitate the adoption and development of open-source software solutions.

The Story Behind the Paper

The journey started about 4 years ago, at the beginning of 2022, when Daniel and Heidi organized a panel discussion about Research Software Engineering in Clinical Biostatistics for the International Society of Clinical Biostatistics (ISCB) conference in Newcastle in September that year. I wrote about this even in an earlier LinkedIn post, which includes more details and slides (unfortunately, the video recordings of the panel discussion are no longer available).

For that conference, a special issue was planned in the Biometrical Journal. So I had the idea to put together a paper on this topic, and invited the panel participants and other colleagues from our network to contribute. We missed the deadline for the special issue, but finished the first manuscript version and submitted it to another statistics journal in 2023. Unfortunately it was rejected after a longer review time, with the main reason being that one of the two reviewers thought that this was all nothing new, because it had been written about before in other contexts and application areas. This version is still worth reading today, as it has more details on the “opportunities” topic.

Then for a longer time we did not pursue the paper further, until Jenny Devenport in 2025 came up with the idea to rewrite the paper for a broader audience, and to submit to a non-statistics journal instead. We took this challenge, restructured and rewrote the paper substantially, and submitted the second manuscript version to Drug Discovery Today in September 2025. Fortunately the review process was much faster here, and a supportive reviewer’s comments were helpful to further improve the paper (the first section about what open-source software actually is was added based on their feedback).

After the acceptance of the paper, Armin had the great idea to apply for funding for open access publication from his employer, the BfArM in Germany, which worked out well and we are grateful for their support.

So it was a long story, but with a happy end! (By the way, I think the pdf version of the paper looks best, see here.)

References

Sabanés Bové, Daniel, Heidi Seibold, Anne-Laure Boulesteix, Juliane Manitz, Alessandro Gasparini, Burak K. Günhan, Oliver Boix, et al. 2026. “The Statistical Software Revolution in Pharmaceutical Development: Challenges and Opportunities in Open Source.” Drug Discovery Today 31 (2): 104613. https://doi.org/10.1016/j.drudis.2026.104613.