Sign in

Georg Heinze

@georgheinze.bsky.social
568 followers 93 following 48 posts

Biostatistician, prognosis researcher, beginner.

PostsRepliesMedia
Georg Heinze @georgheinze.bsky.social · 28/07/2026
Excellent!
020
Georg Heinze @georgheinze.bsky.social · 09/07/2026
Given that now some reviewers let AI do the job, how can we convince an AI reviewer that was prompted as „your role: grumpy reviewer 2“? (And who prevents them from uploading our ideas, for free, to OpenAI&co?)
050
Reposted by Georg Heinze
Tim Morris @timpmorris.bsky.social · 06/07/2026
This is a lot of people’s bugbear isn’t it
The ‘BATS AREN’T BUGS’ Calvin & Hobbes with class yelling ‘TARGET TRIAL EMULATION ISN’T A “STUDY DESIGN”!’
1174
Reposted by Georg Heinze
Frank Harrell @f2harrell.bsky.social · 25/06/2026
"Real world data" is an almost meaningless phrase. Our new @bmj.com paper pushes the more helpful phrase "routinely collected data" and explains the many limitations of such for research purposes. doi.org/10.1136/bmj-... #Statistics #StatsSky
24812
Reposted by Georg Heinze
Frank Harrell @f2harrell.bsky.social · 02/07/2026
Several great colleagues published an excellent paper on multiplicity adjustment. Wish it considered that "assertions are too easy" is miscast as multiplicity. See hbiostat.org/bayes/bet/mu... @sabinehoffmann11.bsky.social @gscollins.bsky.social @georgheinze.bsky.social
onlinelibrary.wiley.com
1133
Reposted by Georg Heinze
The BMJ @bmj.com · 17/12/2025
Doug Altman was an internationally renowned statistician who served as The BMJ’s chief statistical adviser. Read about life and work that made this statistician a "citation millionaire" #BMJChristmas www.bmj.com/content/391/...
06230
Reposted by Georg Heinze
Gavin Simpson @gsimpson.bsky.social · 12/11/2025
A new release of the mgcv #RStats 📦 is out on CRAN and Simon Wood (U Edinburgh) has added some significant new features despite the small bump in version number: 🌟 scasm() for estimating GAMs with shape constrained smooths. Can be used with any family & smoothness selection is via the EFS method
Model is:

b3 <- scasm(
  y ~ s(x0, bs = "bs", k= k) + s(x1, bs = "sc", xt = "m+", k = k) +
         s(x2, bs = "bs", k = k) + s(x3, bs = "bs", k = k),
  family=poisson, bs=200
)

The second smooth `s(x1) is a shape constrained smooth with a positive monotonicity constraint (xt = "m+").

The `bs = 200` arguments uses 200 boostrap samples, which generates bootstrap distributions for each coefficient in the model. These bootstrap samples respect the shape constraints, while the usual +/- 2 SE credible intervals may not.

The uncertainty in the partial effects is shown by two credible interval bands; a dark blue central band is a 68% Bayesian credible interval, while the lighter blue outer interval is a 95% Bayesian credible interval.

The background of each panel is light grey with white grid lines, in a similar style to ggplot2's default theme.
39524
Georg Heinze @georgheinze.bsky.social · 11/11/2025
They did not do an Initial data Analysis with the data they shared. They could have avoided a lot of trouble if they did. "Data sharing without regrets"... See some recommendations here: doi.org/10.1186/s128...
doi.org
Regression without regrets –initial data analysis is a prerequisite for multivariable regression - BMC Medical Research Methodology
Statistical regression models are used for predicting outcomes based on the values of some predictor variables or for describing the association of an outcome with predictors. With a data set at hand,...
031
Reposted by Georg Heinze
Darren Dahly @statsepi.bsky.social · 10/11/2025
You do not need to see the results in order to write ~2/3 of a quant. research paper in medicine/health.
1122
Reposted by Georg Heinze
Dan Lewer @danlewer.bsky.social · 29/10/2025
Idea for a Christmas BMJ paper: submit loads of Christmas BMJ papers, then do a survival analysis of time to rejection. Then submit the survival analysis as a Christmas BMJ paper the following year
2163
Georg Heinze @georgheinze.bsky.social · 06/11/2025
This commentary by John B. Carlin expresses many of the concerns that I have always had with 'prognostic factor (finding) studies' and the (weird) definition of 'independent predictors': doi.org/10.1016/j.jc...
doi.org
Redirecting
000
Reposted by Georg Heinze
Malika Ihle @malikaihle.bsky.social · 29/10/2025
There are lots of professional data repositories to share your data (see www.re3data.org) - if you want to share the code as well you can push to github and archive it on zenodo with a DOI.
re3data.org
Home | re3data.org
154
Reposted by Georg Heinze
Journal of Clinical Epidemiology @jclinepi.bsky.social · 18/09/2025
The distinction between causal, predictive, and descriptive research—there is still room for improvement - Journal of Clinical Epidemiology www.jclinepi.com/article/S089...
jclinepi.com
The distinction between causal, predictive, and descriptive research—there is still room for improvement
It has been proposed that medical research questions can be categorised into three classes: causal, predictive, and descriptive. This distinction was proposed to encourage researchers to think clearly...
1103
Georg Heinze @georgheinze.bsky.social · 18/09/2025
Great experience with a novel format of an invited session of #stratos TG2 @biometricsociety.bsky.social Austro-Swiss region meeting (ROeS 2025 Graz) (1/n)
142
Reposted by Georg Heinze
Stephen Senn @stephensenn.bsky.social · 25/08/2025
#ISCB46 very nice talk by Michael Kammer on COVID-19 prognostic models, which involved a ‘deep dive’ into the pond life of modelling. Bottom line: “If looking for reliable models, don’t hold your breath.”
084
Georg Heinze @georgheinze.bsky.social · 26/06/2025
Published: the paper 'On the uses and abuses of Regression Models: a Call for Reform of Statistical Practice and Teaching' by John Carlin and Margarita Moreno-Betancur in the latest issue of Statistics in Medicine onlinelibrary.wiley.com/doi/10.1002/... (1/8)
onlinelibrary.wiley.com
34818
Georg Heinze @georgheinze.bsky.social · 20/06/2025
This is taking statistical education to the next level: www.youtube.com/watch?v=pZTL... by Rafael de Andrade Moral, Maynooth University
youtube.com
Summary Song #9 - Overfitting (Stats Parody - Charlie Puth Attention)
YouTube video by The Raf
031
Reposted by Georg Heinze
ENAR @enar-ibs.bsky.social · 18/02/2025
Continuing our #BlackHistoryMonth series, we celebrate Sean Simpson, Professor of Biostatistics & Data Science, for his contributions to neuroimaging analysis, repeated measures, & covariance modeling. Outside of work, he enjoys time with his wife, statistician Felicia, and their daughter Sophia.
143
Georg Heinze @georgheinze.bsky.social · 24/03/2025
www.bmj.com/content/388/...
bmj.com
PROBAST+AI: an updated quality, risk of bias, and applicability assessment tool for prediction models using regression or artificial intelligence methods
The Prediction model Risk Of Bias ASsessment Tool (PROBAST) is used to assess the quality, risk of bias, and applicability of prediction models or algorithms and of prediction model/algorithm studies....
184
Reposted by Georg Heinze
Gordon Forbes @gforb.bsky.social · 05/03/2025
It's amazing how far back ideas go. This paper from 1984 by @f2harrell.bsky.social discusses the need for train/test data splits and the importance of assessing model calibration and discrimination and instability in variable selection methods. onlinelibrary.wiley.com/doi/epdf/10....
onlinelibrary.wiley.com
Regression modelling strategies for improved prognostic prediction
Regression models such as the Cox proportional hazards model have had increasing use in modelling and estimating the prognosis of patients with a variety of diseases. Many applications involve a larg...
1216
Reposted by Georg Heinze
Maarten van Smeden @maartenvsmeden.bsky.social · 07/03/2025
Five years ago @laurewynants.bsky.social and I started this living review. Within 1 year we reviewed >700 prediction models for COVID and we found many to be poor and decided to stop reviewing. Many external validation studies followed confirming many indeed were poor www.bmj.com/content/369/...
0277
Georg Heinze @georgheinze.bsky.social · 30/01/2025
I was just wondering what would have happened if the development had been the other way around. If machine learning had existed beforehand and regression analysis had only been invented afterwards.
240
Reposted by Georg Heinze
Frank Harrell @f2harrell.bsky.social · 08/01/2025
I can’t count how many of Ian White’s papers have improved my practice of #Statistics. #StatsSky
0112
Reposted by Georg Heinze
Laure Wynants @laurewynants.bsky.social · 20/12/2024
We have two vacancies for a postdoctoral and PhD researcher at Maastricht University. They will work on the development and application of value-of-information methods for the validation of clinical risk prediction models and medical AI. links below
vacancies.maastrichtuniversity.nl
PhD in Epidemiology on value-of-information from validating clinical prediction models and AI
PhD in Epidemiology on value-of-information from validating clinical prediction models and AI
12117
Reposted by Georg Heinze
Marc Henrion @marchenrion.bsky.social · 10/12/2024
Another exciting STRATOS session at #IBC2024ATL
131
Reposted by Georg Heinze
Kaspar Rufibach @numbersman77.bsky.social · 11/12/2024
Statistical Thinking Across Borders: Building Bridges and Expanding Horizons in Clinical Biostatistics at the 46th Annual Conference of the ISCB (International Society for Clinical Biostatistics), this time in the Trinational Eurodistrict of Basel, Switzerland, from August 24th to 28th, 2025.
274
Georg Heinze @georgheinze.bsky.social · 11/12/2024
Meeting @marchenrion.bsky.social at the poster session #IBC2024ATL #atlanta #internationalbiometricconference #STRATOS
030
Georg Heinze @georgheinze.bsky.social · 11/12/2024
After many online meetings, it feels great to meet fellow #Stratos colleagues in person #internationalbiometricconference #atlanta
061
Reposted by Georg Heinze
Marc Henrion @marchenrion.bsky.social · 09/12/2024
Very exciting to attend a Sander Greenland talk and being lectured on bad practices in epidemiology and biostatistics - and how to do better and change practice in the next generation of researchers. #IBC2024ATL
061
Reposted by Georg Heinze
Marc Henrion @marchenrion.bsky.social · 09/12/2024
Great STRATOS TG2 session with @georgheinze.bsky.social at #IBC2024ATL. Another session this afternoon at 2pm.
011
Georg Heinze @georgheinze.bsky.social · 29/11/2024
Two AIs talking about our paper on phases of methodological research in biostatistics doi.org/10.1002/bimj..., created with minimum effort - fascinating and a bit fearsome notebooklm.google.com/notebook/85f...
notebooklm.google.com
Sign in - Google Accounts
051
Georg Heinze @georgheinze.bsky.social · 15/11/2024
When imputing missing values with @f2harrell.bsky.social 's aregImpute better calibration of prediction models was achieved than with other methods: www.sciencedirect.com/science/arti...
sciencedirect.com
The performance of prognostic models depended on the choice of missing value imputation algorithm: a simulation study
The development of clinical prediction models is often impeded by the occurrence of missing values in the predictors. Various methods for imputing mis…
1296
Reposted by Georg Heinze
Achim Zeileis @zeileis.org · 16/09/2024
📢 PSA: Tired of #betareg in #rstats complaining when you have 0 and/or 1 in your response variable? 🎉 Good news: betareg can now capture this at the expense of a single extra parameter! 📄 New arXiv working paper with Ioannis Kosmidis. www.zeileis.org/news/xbx/
zeileis.org
265
Reposted by Georg Heinze
Frank Harrell @f2harrell.bsky.social · 19/09/2024
Glad to be in beautiful Leiden with Maarten at the STRATOS meeting. Anyone wanting to improve their statistical practice with respect to observational data analysis should get to know the many STRATOS resources. stratos-initiative.org
stratos-initiative.org
STRATOS initiative | The STRATOS initiative
1309
Georg Heinze @georgheinze.bsky.social · 14/08/2024
Happy to present this NEW PAPER of the STRATOS initiative on 'Regression without regrets: Initial data analysis is a prerequisite for multivariable regression' bmcmedresmethodol.biomedcentral.com/articles/10....
bmcmedresmethodol.biomedcentral.com
Regression without regrets –initial data analysis is a prerequisite for multivariable regression - BMC Medical Research Methodology
Statistical regression models are used for predicting outcomes based on the values of some predictor variables or for describing the association of an outcome with predictors. With a data set at hand,...
093
Georg Heinze @georgheinze.bsky.social · 02/08/2024
Happy to welcome three new predoctoral researchers at the Institute of Clinical Biometrics: Moritz Madern, Sinan Necdet Cevirme, Leonie Lenz-Seraphin, working on prediction and causal inference in health Credits to Theresa Ullmann for the picture!
000
Reposted by Georg Heinze
Gary Collins @gscollins.bsky.social · 16/04/2024
NEW PAPER out today in the BMJ TRIPOD+AI: reporting recommendations for studies developing or validating prediction models for use in healthcare that use #machinelearning methods bmj.com/content/385/... #ArtificialIntelligence #AIstandards #OpenAccess
24326
Reposted by Georg Heinze
Anne-Laure Boulesteix @boulesteixlaure.bsky.social · 27/02/2024
Hot tip for (stat/data science) methodological researchers (ctd.): Reviewer complains that your new prototype method does not seem to be ready for use and its evaluation is not extensive enough? You can answer them along the lines of our "phases paper": onlinelibrary.wiley.com/doi/10.1002/... 1/n
onlinelibrary.wiley.com
Phases of methodological research in biostatistics—Building the evidence base for new methods
The Biometrical Journal publishes papers on statistical methods and their applications to life sciences, encompassing medicine, environmental sciences & agriculture.
1156
Georg Heinze @georgheinze.bsky.social · 20/03/2024
another great paper doi.org/10.1002/sim.... @timpmorris.bsky.social !
111
Georg Heinze @georgheinze.bsky.social · 21/02/2024
Great to see this special collection published! onlinelibrary.wiley.com/doi/toc/10.1... @boulesteixlaure.bsky.social
041