3 ms·
Why go directly to double blind trial? Shouldn't you first do some open label tests and see if there appears to be any difference, and only then prove it with
by ars 12y ago
Why go directly to double blind trial?
Shouldn't you first do some open label tests and see if there appears to be any difference, and only then prove it with a double blind?
Seems like that approach would save money.
- cperciva 12y agoTime is probably a significant constraint. Better to waste a million dollars on a trial which could have been avoided than to delay the discovery of a cure by two years.
- zaroth 12y agoDouble-blind makes no sense to me either? T1D doesn't just go away. If the treatment works, it would be obvious... Verapamil was approved by the FDA in 1982, so I would assume safety is well established.
- cperciva 12y ago> Double-blind makes no sense to me either? T1D doesn't just go away. If the treatment works, it would be obvious... They're not just interested in whether T1D "goes away". If insulin requirements drop or A1c %s improve, those would also be useful results (either innately, or as a "look, there's something going on here, we just need to fine tune the dosage"). You can't get good data on those if patients' behaviour is influenced by unblinded treatment. > I would assume safety is well established. You can't assume that when you're dealing with smaller populations. For example, short courses of corticosteroids are very safe in the general population, but can cause life-threatening hyperglycaemia in diabetics. So they have to look for side effects; and if they're looking, they need the trial to be blinded in order to avoid seeing a side effect which doesn't really exist.
- zaroth 12y agoI remember there was a great post a few weeks back about double-blind controlled trials being used in neonatal care, and the difficulty in getting the medical community to accept double-blind randomized trials over the last several decades. So I don't want to discount the validity of the method. [1] But just to belabor the point, couldn't you detect the impact on A1C or a dropping dosage pretty quickly? Everything's logged on the meter/pump/CGM, so you would have an extensive baseline to compare to.... Oh I just realized they want to test this on newly diagnosed patients. Hmmm, in that case I can see more the need for this. I wonder how much the "honeymoon" period would confound the results -- with my son we saw vast changes in required dosages throughout the first year. Not to mention it takes a while to get a handle on how to deal with it, and in the meantime... errors in handling, dosing, mis-counting carbs, etc... combined with honeymooning, I don't know how you could make heads or tails of a newly diagnosed T1D's numbers in just one year, unless the result is very significant. I wonder why they don't test on people with well controlled T1D. I hope it doesn't imply it's something like AAT where it only possibly works if started soon after onset. I guess it's reasonable to expect early treatment after onset could be more effective, so maybe that figures into their strategy. [1] - https://news.ycombinator.com/item?id=8490442 https://news.ycombinator.com/item?id=8490442
- cperciva 12y agoThey're doing this on newly diagnosed patients because that's where they hope to see results: This isn't expected to be a cure so much as a "honeymoon extension" solution -- they're looking at preventing further beta cell destruction, not creating new beta cells.
- ejstronge 12y agoI'm looking up the clinical trial description (check out ref. 1) but I think these researchers are moving directly to a double-blind study because this drug is already FDA-approved (ref. 2). Typically, clinical trials occur in three phases (ref. 3). Since this drug's side effects have been studied, the researchers are moving directly to test the drug's effects. As a final point, open label studies suffer from bias due to a person's expectations of improved health. A double-blind study minimizes this. Why spend money on a biased test that you'll need to redo instead of just doing the right test the first time? [1] clinicaltrials.gov [2] http://www.fda.gov/Safety/MedWatch/SafetyInformation/ucm172949.htm http://www.fda.gov/Safety/MedWatch/SafetyInformation/ucm1729... [3] http://clinicaltrials.gov/ct2/help/glossary/phase http://clinicaltrials.gov/ct2/help/glossary/phase