How can we better manage hospital readmissions?

1. Design a tool to monitor what's happening
2. Integrate predictive models for targeted intervention
3. Test interventions to prescribe action

www.zajichekstats.com/post/managin...

It's still impossible to figure out what is meant by the blanket "AI" term.

Apparently, it's solid data collection, infrastructure, analytical workflows and statistical rigor (good). OR, it's prompt engineering with ChatGPT.

Please stop conflating these.

youtu.be/52qT1Q8-Qcs?...

Please explain what is meant by "AI".

Great book for anyone interested in the divergence and convergence of science and faith.

Assemble your team early, have clear goals, think about action, define the end product, and check the utility.

www.centralstatz.com/posts/ways-t...

A quick example why statistical significance doesn't tell the whole story and should not be the sole basis for decision making or action.

www.centralstatz.com/posts/simple...

Proportional-odds logistic regression is a common approach for modeling likert-like data. It's an extension of standard logistic regression for a binary response, except the outcome has more than two possibilities, but they have a known order.

www.centralstatz.com/posts/how-to...

There are tons of tools out there that enable you to begin building and deploying custom web applications, interactive documents, and more. Easily and at no cost.

www.centralstatz.com/posts/low-co...

A few thoughts on why building and maintaining a blog is a useful endeavor for those in data science, especially students and other learners.

www.centralstatz.com/posts/you-sh...

Here we discuss how Bayesian statistics enables you to construct a valid, and potentially useful, model before any data is collected.

m.youtube.com/watch?v=bUer...

Must read. Comes down to accountability and incentives that enable this practice to persist. Why journals still publish the same way despite seemingly widespread agreement about the issues with p-values. Tells you there’s flaws in the whole scientific enterprise.

www.nature.com/articles/506...

Splines are useful for modeling nonlinear relationships in regression models, but can be tricky to understand what's going on behind the scenes. Here we show how to reconcile the regression equation when there's a spline term in the model.

www.centralstatz.com/posts/reconc...