Thread by @VPrasadMDMPH

Our new paper in @EJCI_News argues that Randomized trials are necessary in medicine & PH for interventions w putative benefit & at best MED to LG effect size.

Parachutes & smoking are not good counter examples

Here is the explanation 🧵
onlinelibrary.wiley.com/doi/abs/10.111…

Some people argue that b/c we did not need RCTs to know smoking is harmful or Parachutes are life saving, we don't need them to test cloth masking, or the Impella, or some new cancer drug, or HCQ, or <insert ur favorite practice>

But this is based on misunderstanding

There is a huge range of things we can do to someone that might hurt them or save them, imagine the spectrum (below)

Let's start on the harms side

At most extreme, you could shoot someone in the chest at point blank range or throw them off a cliff

There has never been an RCT that doing that is harmful.

Wow, who knew?!?

But we know it is harmful. The effect size is massive. Near certain death.

Now, look at the right most edge

There is no RCT that pulling someone out of the way of a speeding truck is life saving, & no (non-humorous) RCT of parachutes but again the effect size is massive.

Visible to naked eye

Now look to the middle left.

Smoking, pollutants in the water, carcinogens in food-- none of these have RCTs showing they are harmful, and we generally do no not run RCTs for putative harms

We draw upon risk factor epi & make a determination that mitigation is reasonable.

If we wish, we can subject smoking CESSATION strategies to an RCT. You can power trials for smoking reduction, but there is no rule that says you can't power them for all cause mortality

By doing that, you immediate move to the right side of the mid-point; the green arrow

you have an intervention that possibly offers a modest to marginal effect size.

Turns out that is where most of biomedicine lies

Or what we call "The RCT Zone"

Here, RCTs are desperately needed to separate true effect from hope & wishful thinking & propaganda

Early in the pandemic, some opposed cloth mask RCTs saying that cloth masks were like a parachute

Seemed farfetched to me, and now
That was a terrible decision

Of course their effect on the primary endpoint in Bangladesh was 0%. Cloth failed. Surgical had 11% RRR (but open Qs)

Regardless, RCTs were not only possible, they were desperately needed; the effect size was at best modest but possibly null, and RCTs work well to separate

True effects from wishful thinking

If we had more RCTs of masking-- particularly kids in school--- we would end a bitter debate that is driven by low credibility data

medpagetoday.com/opinion/vinay-…

Of course many believe that high quality observational studies like 'target trials' will emulate RCTs

I talk about that more in this thread & paper last year

https://twitter.com/vprasadmdmph/status/1407185384342056961

Closing thoughts:

Most of the time people say you can't do an RCT-- they have little conceptual clarity on what they are saying.

For interventions with at best med-lg effect sizes (not massive effect size), you usually need to do an RCT

We discuss much more in our new paper

Check it out here:
onlinelibrary.wiley.com/doi/abs/10.111…

And if you enjoyed this, follow @vkprasadlab for more updates regarding our research