> the process has become more important than the problem you’re trying to solve.

This holds true in almost every professional field for which life is on the line. Medicine, class 3 electronics, aviation.

We have a word for this, which roughly translates to "rule of paperwork". Bureaucracy.

And I think there is a reason why the bureaucracy exists in these cases. "Move fast and break things" doesn't work very well there.

What happens if your study clearly hurts people? What happens if your study clearly helps people? You find out in the first few weeks, what do you do? How do you ensure you collected enough of a sample of a general population to make your study representative? How do you ensure your patients properly consented to the study (past shameful human experiments aside, you likely need many institutions participating, so you can't control everything yourself).

Do I keep going or is the IRB approval process clearer now? There is a reason it exists.

I think once again - when the process becomes the metric it’s insane. What time things are being administered is already random and not regulated or organized. “What if it hurts” isn’t relevant for something like this because the reasoning is that the baseline is that “when” doesn’t matter, you’re still giving the same dosage. “What if it clearly helps?” What if. Then you publish a paper or give a talk at a conference and try to better mobile the medical community. Or see if the administrators are willing to help scale this up further.

> How do you ensure you collected enough of a sample of a general population to make your study representative?

You don’t need to. This would be a pilot study to check whether there’s maybe a there there before you do it larger scale to measure predictive power at population level.

> Do I keep going or is the IRB approval process clearer now? There is a reason it exists.

I think you’re completely failing to engage with the argument that this particular case about time shifting delivery of a drug should not need meaningful IRB engagement other than “I’d like to change the time I deliver the drug for 2 more patients because we had one patient respond positively and this isn’t believed to be a factor” “ok cool yup”.

You’ve jumped from no IRB to full IRB without considering the context of the problem being solved which is why I said when the process becomes the goal vs the problem you’re trying to solve - you’re imaging the worst and most complicated situations possible for a case that would never demand it.

You are approaching things from software development perspective of "what's the worst that can happen? I rollback". In the topic discussed, you cannot rollback. While you might have a reasonable suspicion that changing the time will improve some outcomes in most, you cannot be sure that it won't greatly reduce positive outcomes in many. The IRB is often in place not to stop positive outcomes, but to reduce negative ones.

No I'm not. I'm pointing out the time of administration literally is already under the discretion of the hospital. There's literally no recommendation one way or the other and hospitals administer randomly based on what's convenient for staffing (i.e. not a medical decision). A) there's nothing dangerous about taking something your doing randomly anyway and systematizing it. B) there's no plausible way this is even remotely dangerous.

> The IRB is often in place not to stop positive outcomes, but to reduce negative ones.

Research can literally be IRB exempt if it provides minimal or no risk to patients which is literally what this is. Even if you put this in the "minimal risk" category which would be extreme that's still minimal IRB oversight and approval takes ~1-3 weeks.

You're imagining IRB is something it's not even intended to be and then saying it's a reasonable bottleneck in general because of real problems it prevents and thus justified for this specific experiment (where it wouldn't be relevant).

This is top to bottom a failure to follow up - doctor's are overworked & fail to follow up on potential research results because they act more like mechanics.

We can appreciate that process is important, but at some point you're falling down a slippery slope here, surely?

We're talking about a factor that no one has previously had reason to consider important.

Of course, I don't know hard it truly is to undertake a study. I have to imagine for something like this you could write up a basic study protocol in fairly short order.