Dimitris Rizopoulos

A new version of JMbayes2 with many new features is on CRAN. It supports zero correlations for the covariance matrix of the random effects, IPCW for the AUC and the Brier score, Super Learning, and fitting interval-censored competing risks data: drizopoulos.github.io/JMbayes2/ #rstats

Excited to be teaching a course at the IBC in Seoul on Dynamic Predictions for Longitudinal and Survival Data, with applications in R using JMbayes2. Want to learn how to build and interpret dynamic predictions from joint models? Check course SC02 at ibc2026.org/events/shortco…

New feature in 📦JMbayes2 supporting different specifications of the baseline hazard function - learn more about it here: drizopoulos.github.io/JMbayes2/artic… #rstats

A new version of JMbayes2 is out with new implementations of the AUC and Brier score for Cox models, and internal validation using cross-validation and the Bootstrap. Read more about it here: drizopoulos.github.io/JMbayes2/artic… #rstats

I’m honored to be a candidate for the Executive Board of @IBSstats. I hope to contribute to our society’s efforts to strengthen our field’s connections with other disciplines while promoting the responsible use of statistical methods. Please consider voting for me by January 10th

Our first invited speaker @drizopoulos gave a very insightful talk on how longitudinal data can be modelled via joint models to derive dynamic predictions. His approach is illustrated by estimating metastasis risk for prostate cancer patients based on their PSA history.

Heading to Salzburg 🏞️ to present at #SAM2024. Join us in Invited Session 16 (July 10, 09:00) for insights on efficiently integrating high-dimensional data in joint models for longitudinal and time-to-event outcomes. Joint work with @drizopoulos and @ERandrinopoulou.

Special Announcement! We're pleased to welcome Sander Greenland as our Keynote Speaker for IBC2024. Dr. Greenland is Emeritus Professor of Epidemiology and Statistics at the University of California, Los Angeles. Registration is Open! Visit ibc2024.org for more info!

A great meeting at @ErasmusMC today hosting the @BMS_ANed Spring Meeting on the occasion of the Hans van Houwelingen, with @DeliuNina, @ThomasKlausch and @zwyang96 speaking about Statistical and Machine Learning Methods for Sequential Decision Making.

@f2harrell Very nice paper. Can these models handle unbalanced data in the sense that each patient provides measurements at completely different time points? In these cases, the “previous-measurement” covariate has a different meaning per patient.

This new #Statistics in Medicine tutorial by our outstanding PhD student Max Rohde is in my view an amazing introduction to the Markov longitudinal ordinal model, with detailed examples

Interested in applications of machine learning and statistics for sequential decision making? Join us at the BMS-ANed spring meeting: vvsor.nl/biometrics/eve… #epitwitter #statstwitter

@ericnovik When computing speed matters and you have a complex model, writing your own sampler will be more efficient (though much harder to do) than relying on general-purpose software (e.g., STAN, JAGS, etc.).

@drizopoulos Rolled your own sampler in 2024 -- respect! (this is not sarcasm, I mean it)

A new version of JMbayes2 with many new features is on CRAN. It supports zero correlations for the covariance matrix of the random effects, IPCW for the AUC and the Brier score, Super Learning, and fitting interval-censored competing risks data: drizopoulos.github.io/JMbayes2/ #rstats

@DocEd Yes, I want to find the time to do it.

@ericnovik We use Gibbs sampling.

@drizopoulos This is awesome but can can I ask which MCMC sampler you are using under the hood? Is this Gibbs, NUTS, something else?

Five pre-conference courses @KellyVanLancker @drizopoulos @BabakOskooei @CNakas and many others +invited sessions covering the stat of the art in biostatistics + an exciting social program (Mount Olympus, ancient Vergina, swimming etc).

Want to learn how to improve the accuracy of dynamic predictions? Have a look at our new super learning procedure for joint models: doi.org/10.1002/sim.10… Example code available on JMbayes2's website: drizopoulos.github.io/JMbayes2/artic…

@CNakas @mgpolitis Raclette?

A new vignette on the website of JMbayes2 showcasing how time-varying coefficients (non-proportional hazards) for the longitudinal outcomes are implemented in the package: drizopoulos.github.io/JMbayes2/artic… #rstats #statstwitter #epitwitter

Now at #ISCB44 @drizopoulos presenting JMbayes2, an R package to fit extended joint models for longitudinal and time-to-event data.

Great talk by @drizopoulos on the causal effects of salvage therapy for radical prostatectomy #prostatecancer #ISCB44