Thiago

Magic 🪄🎩!

The hype/buzzy word is too strong. There are quite a few papers saying that it is a TTE without any information about the "target trial". They consider that it's tte only because using IPTW or matching. :/

However, that's not a problem with the cohort study design. It is thresearchers' "self-inflicted" harms. The TTE try to prevent it, but I don't see TTE providing any additional benefit compared to a well designed cohort

Any bias that exists in "cohort studies" will be in TTE. I would say that TTE is just a more standardised way of conducting cohort studies.

Any good schematics for target trial emulation framework? I see a lot of schematics mixing analysis (e.g. CCW) with the framework/idea
#episky
#rstats

Maybe because it isn't in favour of AI hype? (LinkedIn from Microslop)

Job losses at research-intensive universities double in two years - Research Professional News Exclusive: Scale of redundancies revealed branded a “disaster” for UK research capacity

Job losses at UK research-intensive universities double in two years.

Exclusive: Scale of redundancies branded a “disaster”.

www.researchprofessionalnews.com/rr-news-uk-u...

Not enough.
Wait until the whole morning for hazard ratios

What Does the Odds Ratio Estimate in a Case-Control Study? Abstract. The use of the term ‘odds ratio’ in reporting the findings of case-control studies is technically correct, but is often misleading. The meaning o

academic.oup.com/ije/article-... (g)old paper about it.

In case-control, you can estimate ORs. But those ORs *are* different measures, such as rate ratio (but they are OR from logistic reg). It depends on the control selection.
The interpretation as other measures is fine, but the ppl need to know what they are doing.

I received two PRs from recent submissions that appear to be LLM-generated. One of them was reasonably okay (a lot of repetition with some real points), the second was horrible, 90% hallucination. It's crazy to think a lot of people consider it a good "substitute"

"We are seeing readily available data assets, such as... the UK Biobank, NHANES, and many other databases, being weaponized to generate manuscripts with minimal scholarly value and utility"

Roll two 🎲 🎲 to find a new association in the Nurse Health Study

(cdc) The overall burden of influenza (flu) for the 2023-2024 flu season was an estimated 40 million flu-related illnesses, 18 million flu-related medical visits, 470,000 flu-related hospitalizations, and 28,000 flu-related deaths.

Thanks!

What would be viewed as post-phd independence?

One on left is a black dog and above it the words “Reality”. Below it is “I chased a squirrel” One the right is a black dog and above it says “LinkedIn”. Below it says, Proud to announce that I effectively executed a rapid-response squirrel displacement strategy to mitigate potential yard intrusions. Humbled by the unwavering support of my family and local stakeholders. This experience reinforced the importance of vigilance, ownership, and continuous improvement. Looking forward to scaling this impact in future engagements.

😂

And the 10x engineer (any number)? Where do they fit? ,

It looks very academic-focused, all those 31 items.

Palantir and NHS: BMA to tell members to limit use of tech system Doctors should not engage with any aspect of the NHS Federated Data Platform (FDP) outside direct patient care because of its links to the controversial US tech giant Palantir, the BMA says. The unio...

#BMJExclusive: Doctors should not engage with any aspect of the NHS Federated Data Platform outside direct patient care because of its links to the controversial US tech giant Palantir, the BMA says
https://bit.ly/3OnZxcG

This one even has a Johnson–Neyman that I remember @mattansb.msbstats.info talking about...

bsky.app/profile/matt...

Another one:
"To emulate a target trial estimating the effect of remimazolam on postoperative outcomes... generate two hypothetical scenarios using a static intervention framework..." (no "protocol" word in the paper)

I just googled "target trial emulation" and looked at the last papers. Look at this. "Target Trial Emulation Flow. Target Trial Emulation Flow". No table showing the target trial, how to emulate, nothing. Just a nice figure about IPTW. 😉

A lot of TTE that isn't TTE hehe.
I just want to reference something about the appropriation /misuse of the terms

@pwgtennant.bsky.social @georgiatomova.bsky.social Do you guys have any paper/comment about misuses of "target trial emulation"? I'm writing a piece about misuses in causal inference. Like use of buzzy words just for the hype

Because it doesn't count for funders/promotions etc. The problem is more with the game than the player

I think the same thing (a lot are fake) could be said about the burden of diseases (GBD) 🙃. Not fake, but not grounded in any reality

In the past, it sometimes bugged with updates on either side (pandas, numpy or the reticulate package).

Why not save / re-read? (parquet)

I think random allocation (after some selection) would be a good thing, as it reduces inbreeding in academia.

Ps:Seems this AI is only to reformat your grant proposal to a new funder. Similar to journals... funders should use standard templates, we have better things to do than "play" with the text