I'm highlighting the bottom section of a blog this week on the big-data-driven PHOENIX criteria for pediatric sepsis. I'm posting the novel statistic here, so it can be found via a scan of blog titles.
##
For the original blog on big data and sepsis, here:
https://www.discoveriesinhealthpolicy.com/2024/06/sepsis-rationale-definition-newly.html
I rarely mention statistics, but this popular and important JAMA paper hinges on a statistic called "AUPRC" which was new and confusing to me. Here's my short text about it, from my own "Phoenix" blog.
###
Nerd Note - Novel Statistics Right Here in JAMA
The Phoenix paper makes much use of the AUPRC metric - area under the precision recall score, which is an alternative to the far, far more common "AUROC" or area under the receiver operating curve. I'm not sure I've got my head around this yet.
- Whereas the virtue of the AUROC is that it looks at clean simple sensitivity and specificity as if in a vacuum, and AUROC deliberately leaves base rate and PPV-NPV for a later day, the AUPRC builds PPV into itself, so it's only valid for a population with a similar base rate.
Read about AUPRC in the Phoenix paper, and for more about AUPRC, they cite a "super paper," Saito & Rehmsmeier 2015, in PLOS One. Saito has nearly 3000 citations.
Beyond reading Saito, I will leave a public link to a long Chat GPT dialog where I tried to teach myself more about PRC - here.
You can also take yourself to a YouTube video - here.