Joint Statistical Meetings (JSM) 2025

Earlier in August, I had the opportunity to attend the Joint Statistical Meetings (JSM) 2025 conference, held in Nashville, Tennessee. JSM is one of the largest gatherings of statisticians, data scientists, and researchers in the world, featuring multiple sessions simultaneously on a wide range of topics. Many academics, industry professionals, and students from all over the world attend the conference for a week of presentations, workshops, and discussions on a large variety of the latest developments in statistical methodologies. 

My main reason for attending was to present a poster created by the researchers at Generable. The poster, titled “Personalized Dosing Decisions Using a Bayesian Exposure-Hazard Multistate Model”, summarized the results of a novel Bayesian Exposure-Hazard MSM.

In summary, the objectives of the work were two main ones:

• To build a Bayesian multistate model that incorporates competing risks and prior event history, thereby improving event risk prediction and enabling personalized dosing.
• To develop a computationally efficient strategy that integrates exposure estimation with hazard prediction, making the approach more practical for clinical application.

We will dig into the model details in the later posts.

The poster attracted some attention, with more than 15 people asking for a deeper explanation — a reasonably high number of questions, in my own experience with conferences. Attendees showed interest in multiple aspects of the poster: the mathematical details, the application and its potential extensions and collaboration, the code results and accompanying R package, the main challenges encountered, and more. 

‍

‍

Sessions and Highlights

Outside of my presentation, I attended a variety of sessions covering everything from new statistical modeling techniques to advances in computational methods. Many sessions were happening at the same time, and it was noticeable that the conference organizers were concerned about having a wide diversity of topics. 

In addition to this, one of the best parts of JSM is the people you meet. I had the opportunity to reconnect with colleagues I hadn’t seen in years and to meet new people, which served as a friendly reminder of how conferences help keep our professional networks alive.

Exploring Nashville

Of course, a trip to Nashville wouldn’t be complete without experiencing a bit of the city itself. In the evenings, I had the chance to visit Broadway Street, listen to live music in typical Nashville bars, and enjoy some incredible southern food—the mix of academic exchange by day and Nashville’s energy by night made for a memorable balance.

Reflections and Takeaways

Presenting the poster and attending sessions expanded my perspective on the field, once again. Most importantly, I had the opportunity to present Generable's work to the world, which clearly got some interest from multiple people from different backgrounds of the broader statistical community.

A big thank you to the conference organizers, the people who stopped by the poster, everyone who shared ideas and insights throughout the week, and to the Generable team who created such fantastic work!