Thread by @DrSherryPagoto

I teach a grad course on clinical trial methodology for behavioral interventions.

If this is a topic you’d like to learn more about, here are the main take homes from my class (with recommended readings!) in a thread! 🧵

Here goes! 1/x

First, you must understand the phases of behavioral intervention development and testing. I can’t tell you how much I see these phases confused in grant applications and papers.

The ORBIT model is a fantastic resource delineating the phases:
ncbi.nlm.nih.gov/pmc/articles/P…

2/x

Next, intervention development starts with a conceptual model that delineates the key drivers of the behavior you are targeting. To understand those drivers, your assignment is to delve into the basic behavioral science and observational study literature on the topic. 3/x

Ready for a pilot feasibility trial? Yay! The biggest mistake investigators make in designing pilot trials is picking the wrong outcomes.

This paper spells out what your feasibility outcomes should and should not be:
ncbi.nlm.nih.gov/pmc/articles/P…
4/x

Hot tip: don't test feasibility & efficacy in the same trial. Cramming them together is like an airplane pilot running tests to see if the plane is OK to fly while the plane is in the air. Better to test that before taking flight not after!

Live scene from my pilot trial:

5/x

Warning: Many pilot trials in the literature are riddled with bias. Here is a great piece on 9 sources of bias in pilot trials (feat @briemcgrievy).

Be sure to avoid these! ijbnpa.biomedcentral.com/articles/10.11… 6/x

Now for control groups! This is the biggest decision you will make. Different control groups answer different questions, so make sure your control group answers the right question.

Love these two papers on control groups:
researchgate.net/publication/26…

pubmed.ncbi.nlm.nih.gov/28396067/
7/x

Hot tip: In pilot trials, testing the feasibility of your control group is often just as important as testing the feasibility of the intervention.

How will people respond to it? Will they drop-out? Will it have an intervention effect? How much?

8/x

What about attention control groups? (They are way over-utilized) We should only use these if there's reason to believe "attention" could affect the primary outcome.

Here is a fantastic piece on whether attention control is right for your trial:
pubmed.ncbi.nlm.nih.gov/23362502/
9/x

Hot tip: Attention control groups may seem benign but they can cause harm! We once saw worse outcomes in an attention control relative to usual care. 😲

Worst part is we only added the AC bc a grant reviewer told us to! 😩
ncbi.nlm.nih.gov/pmc/articles/P…

10/x

Now for efficacy vs effectiveness. They aren't synonyms. It's important to know the distinction. Efficacy trials ⬆️ internal validity and effectiveness trials ⬆️ external validity. Here are some resources on the difference:
ncbi.nlm.nih.gov/pmc/articles/P…
sciencedirect.com/science/articl…
11/x

Now for treatment fidelity--the degree to which the intervention was delivered as intended. For ex, did your health coaches cover all of the protocol objectives or end up doing their own thing?

Here’s a 'go to' resource on treatment fidelity:
pubmed.ncbi.nlm.nih.gov/15367063/
12/x

Onto the cardinal sin of randomized trials--- violating the intent-to-treat principle. 😱

“Once randomized, always analyzed” is the clinical trial pledge of allegiance.

Here is a great resource on why ITT is important:
pubmed.ncbi.nlm.nih.gov/27733397/

13/x

Finally, be sure to design your behavioral intervention with implementation and dissemination in mind. Where will the intervention “live”? How will you get it into real world settings?

Check out this fantastic piece on D&I science by @drrachelshelton
pubmed.ncbi.nlm.nih.gov/32060805/

Do you have some go-to articles for behavioral clinical trial methodology?

Add them to this thread! I could always use great resources for class!

Happy trails and happy trials! 😄

14/x