Simon Gates

Very happy for anyone to put counter-arguments to me or show me where I'm wrong.

This may be because they have legitimate uses in oncology (especially), so people are used to seeing them, leading to belief that it's a legitimate design and use in inappropriate situations.

To clarify: my contention is that they are often used inappropriately, and their results often don't mean what is claimed.

I wrote some stuff about single-arm trials

simonelgato.github.io/2025/06/10/s...

New presentation: www.fharrell.com/talk/gdesign/ #Statistics #StatsSky #rct #clinicaltrial #bayes

My brain reads it as "sesevenen"

Always annoys me when they write that film's name with a "7" in the middle. It looks nothing like a v.

A Not-So-Fond Farewell to Vinay Prasad The Steve Urkel of COVID Twitter turned out to be incompetent at both drug regulation and being an authoritarian apparatchik

Don't let the door hit you in the ass on the way out of CBER, you fascist-enabling, incompetent, contrarian twerp.

open.substack.com/pub/rasmusse...

We are receiving desperate messages of starvation from #Gaza, including from our colleagues.

Food prices have increased 40 fold.

UNRWA has enough food for the entire population for over three months, stockpiled just outside Gaza.

Lift the siege and let aid in safely and at scale.

We often now hear the term “Bayesian borrowing” in trials. But what is Bayesian borrowing and what are the pros and cons around its use? 1/6
#MethodologyMonday #122

How hazard ratios can mislead and why it matters in practice - European Journal of Epidemiology Hazard ratios are routinely reported as effect measures in clinical trials and observational studies. However, many methodological works have raised concerns about the interpretation of hazard ratios ...

How hazard ratios can mislead and why it matters in practice. Elise Dumas, Mats J. Stensrud. European Journal of Epidemiology. link.springer.com/article/10.1...

Just astonishing levels of wrongness

No

Show Your Stripes Visualising how the climate has changed for every country across the globe

Here are the warming stripes for Cardiff - very striking visual representation of what is happening. #ShowYourStripes

showyourstripes.info/s/europe/uni...

No the great thing is the money they make for their creators. It's magic!

The accuracy of his predictions is astounding.
Just not in a good way.
Incredible that (a) this guy is now in the House of Lords and (b) anyone still takes any notice of anything he says.

I recall when he started being "contrarian" over covid someone called something he wrote something like "a job application for a leadership position in a Trump administration".
(On Xwitter so I haven't looked for it)

Following a visit to the GP I am now putting "shared care" into the same category as "rail replacement bus."

My suggestion would be the sort of analysis that Frank Harrell has advocated - use Bayesian methods to calculate the probabilities of different treatment effects. So we can get a probability that IV is worse, prob that it's 10% better. 20% better etc.

And that's not to say that the results aren't valuable - in the absence of other trials, of course they are.

I get that doing this trial at all was a fantastic achievement - but that doesn't give us licence to misinterpret the results.

I'm going to be critical here: conclusion was "no evidence of a difference in mortality at 96 hours was noted"
But the trial was way too small to find anything other than a huge difference (272 participants - assumption of sample size was 17% absolute difference in death)

Indeed

"Single arm nonsense"
I'm right with you there.

There are cases where single arm trials are a good idea. Most of the examples you see are not that.

Are they Fragile or are they Close to the Edge?

(that might be a bit niche!)

This is an interesting one. I guess if there was a benefit you'd expect it to be small, as (if I've understood correctly) you're treating patients that you think are cancer-free, presumably to make sure. Question is whether benefits outweigh harms

Any oncologists want to comment?

This week saw the publication of the accompanying editorial from Rachel Phillips and Victoria Cornelius providing a really useful overview of learning from the series 3/7
@vcornelius.bsky.social
@rachpips.bsky.social
trialsjournal.biomedcentral.com/articles/10....

Look forward to that!

So what I was going to say was that I've never heard of this "zone of bias" thing, though I've been involved in evidence-based medicine (not capitalised) for a long time.

But now I want to know more.