What is an N-of-1 trial?

A quick video on the interesting N-of-1 clinical trial design

🔗 LINKS MENTIONED:

OTHER CHANNEL LINKS

[TheRealBuilder]

This solves a huge problem for me. Ive wondered how to measure the effect of certain things in my life (waking up early, drinking coffee, etc) to better understand the effects. Problem is how to determine the effects of interactions efficiently lol. Would take too much time.

[s.k_]

Because of videos like this I'm falling in love with statistics.
Thank you so much.🎉

[marcoantonio]

The topic of statistics for causality seems really interesting

[adiaphoros]

Let's say a pharmaceutical company wanted to splurge and combined the n-of-1 and RCT methods.

For example:
1. Randomly select the participants¹.
2. Randomly group the participants into 2 equal sized groups A and B.
3. For the 1st period, assign group A to the placebo, and group B to the actual treatment.²
4. For the 2nd period, assign group B to the placebo, and group A to the actual treatment.²
5. Repeat Steps 3 and 4 til desired.³

¹ The number of participants is preferably at least a thousand.
² The length of each period is preferrably at least a month.
³ The steps are preferrably repeated at least 3 times.

There are no constraints to the budget and time. Obviously, this process is obscenely expensive and inefficient, but would the data say something more about the treatment? If so, what?

[khaduopha]

How do N-of-1 trials account for long term variations? Like maybe a treatment have some unknown long term side-effects that "contaminate" the effects observed during later rounds of placebo. We could take average over large population because it's reasonable to assume that one person is unrelated to another but it's probably not reasonable to assume later treatment rounds are independent from earlier rounds.

[renitadykstra]

Would it be possible to separate blood types and find a common denominator 1st 🤔 There is a common root problem, possibly in how blood types absorb basic nutrients 🤔

[davidmoore]

Interesting! I've heard of the statistics of N=1 events in physics or when considering stuff like the anthropic principle. Some neutrino detections with absurdly high energies, the "oh my god particle", or the observation that intelligent life evolved on earth, are N=1 events. In the case of "the existence of intelligent life on earth" I think the arguments are doomed to be more for-fun than for-science, but with neutrinos and cosmic rays you actually want to do some science with them!