Hacker News new | ask | show | jobs
by Balgair 3220 days ago
> But what purpose is having a review board made up of people who can think for themselves if their behavior is going to be indistinguishable from a program running through a checklist without any care whatsoever for whether or not the checklist makes sense in that context?

In general, most IRBs are there to help out, not to get in the way. It may feel like getting in the way, at the time, but they mostly are there to help.

For example: the wordings of your consent forms. FDA says that they need to be at an 8th grade level, +/- 0.2 grade levels. You can game MS word to get that to an 8.2 grade level but have it be 140+ pages, and a lot of researchers do that. This is especially prevalent in clinical drug trials where many Pharam companies try to make the consent form a contract filled with legalese. That egregious example is non typical, but highlights the issues with consent forms being readable and an actual consent form a person. This is where a decent IRB would help out and tell you that you need to re-write the consent form to be an something that is not gibberish. (Side-note: Most newspapers are written at a 6th grade level it turns out).

Another example is International research. There was a case I heard of where a group was going to go collect HIV data in Tanzania. They were just going to go out and collect blood samples and sexual health questions alongside. The part of Tanzania they were collecting from is rural and believes in very strange medical practices like eating albino people or having sex with virgins to cure HIV, though I forget exactly what. Balancing the local beliefs with data collection can be tricky in that case. Typically, you would like a local IRB in that region, but there was none. So the US based IRB decided that the researchers needed to go out to the religious and government centers in that part of Tanzania and ask them what to do and what questions they could ask, etc. This added a lot of time and expenses to the study to spend in country before data collection, but the US IRB held firm on that. This was a good idea, as the local peoples had a lot of reservations and wanted to see all the data themselves before it was sent out of country, so as to go off and then kill women that were HIV positive. If the US researchers had found this out beforehand, a lot of deaths may have occurred. They subsequently then left off a lot of questions and data gathering.

Researchers are human too and we forget things all the time and have a lot of pressures on us as well. IRBs should be there to help out and not be a road-block. If the IRB you are working with does turn out to be that, talk to other researchers and find if they have the same issues. If you all do, go to the IRB and explain to them that they are unreasonable, most of them do want to know that they are holding things up unnecessarily. If it persists, go to the hospital/university admin as a group and make the case. These advancements need to be triple checked, but they still need to go through to begin with. IRBs are not evil, sclerotic maybe, but not evil. Work hard to get them to be better, we are all in this together.

1 comments

So the original study planned on sharing de-anonymized data of who had HIV with the local populace? And the IRB prevented that?
Essentially, yes. The naivete of the team was staggering. They thought, at first, that they local people just wanted to see the data and have a copy of it, for pure curiosity's sake. Nope, murder.
Woah, I assumed I misread that. Just common sense/decency says you don't share who has HIV with anyone who does not need to know. True in the first world, especially true in the third world where stigmas can run even stronger.
Yes, this is why we have IRBs, for the exact reasons like this one. When dealing with health and medical stuff, the stakes are usually not this high, but they can be very easily.