RPACT reposted this
Four conference days in Freiburg, plenty of ideas for RPACT Code, and some very welcome reunions. At ISCB GMDS 2026, my main focus was AI in statistical practice: how can we make it useful while keeping assumptions, results, and responsibility clear? I’ve shared my highlights, photographs, and takeaways for rpact and RPACT Code on our blog: 👉 https://lnkd.in/ewP_8JB2 Three themes stayed with me: • Guardrails for AI-based sample size tools - especially relevant to our work on RPACT Code. • Reproducible statistical code and the importance of explicit analysis assumptions. • Hallucination detection, and why a confident answer still needs checking. The conversations between sessions were just as valuable. It was lovely to reconnect with my PhD supervisor, Andreas Ziegler, and with Anika Großhennig, Michael Brendel, and Silke Szymczak from our PhD days at IMBS in Lübeck. I also enjoyed catching up with Marcel Wolbers from Bayer, the regular exchanges with Lukas Widmer from Novartis, and meeting Werner Brannath, co-author of the book underpinning rpact. Hearing from people who use rpact in their work is always valuable to me. With Carolin Herrmann, whose methodological work has already contributed to rpact, we renewed our interest in exploring whether we might implement further methods together in the future. I’m heading home with new ideas and questions to explore - and feeling grateful for the people who make these conferences worthwhile. #ISCBGMDS2026 #Biostatistics #ClinicalTrials #AI #rpact