IRBs Gone Wild

9 April 2009 at 9:41 am Peter G. Klein 10 comments

| Peter Klein |

We’ve noted before the strange behavior of university Institutional Review Boards. My own campus has a particularly prickly IRB, the result of an unpleasant incident a few years back involving the medical school. So, even social-science researchers must receive IRB training and have individual research projects — yes, every research project that involves “human subjects,” which includes research using secondary data — approved by the campus IRB.

My certification expired recently and I took an online test today to be re-certified. Some of you may find the questions interesting. Here is a selection. Keep in mind these are questions for an economist wishing to do research in economics and management, not for a pharmacologist or epidemiologist.

31. Phase 1 clinical drug trials include the initial introduction of an investigational new drug into humans.

True

False

32. The investigator is a 1/8th V.A. employee. She proposes to recruit MU outpatients into a study conducted exclusively at MU facilities. Which of the following groups must approve the research project before participants can be enrolled?

The MU Health Sciences Center IRB

The V.A. Research and Development Committee

Both a. and b.

Neither a. nor b.

33. All Phase 3 studies require a Data Safety Monitoring Board, and one may be required for Phase 1 and Phase 2 studies.

True

False

34. The investigator is an MU faculty member with courtesy privileges at the V.A. Hospital. He proposes to recruit V.A. outpatients to participate in a study conducted exclusively at MU facilities. Which of the following groups must approve the research project before participants can be enrolled?

The MU Health Sciences Center IRB

The V.A. Research and Development Committee

Both a. and b.

Neither a. nor b.

35. Conflicts of interest may include:

when a physician’s patient is also the physician’s research subject

when an investigator owns stock in the pharmaceutical company sponsoring her research

when an investigator receives unrestricted research funds from the corporation sponsoring his research

all of the above

Again, to do research, federally funded or not, I have to study for and pass this test, then have each project reviewed annually by the IRB. So, when I move up from interviewing entrepreneurs, handing out surveys to corporate executives, or analyzing Compustat data with Stata to conducting Phase I clinical trials, I will be well prepared!

Entry filed under: - Klein -, Education, Institutions, Methods/Methodology/Theory of Science.

Kaplan on Sutton Program for Searle Center Conference, “The Economics and Law of the Entrepreneur”

10 Comments Add your own

1. Per Bylund | 9 April 2009 at 9:59 am

every research project that involves “human subjects,” which includes research using secondary data

Not really, actually. I sat through the training session yesterday (spending one hour looking at boring powerpoint slides is equivalent of taking the test) and it was stated very clearly that IRB approval is required only if individuals in the data can be identified by the on-campus researcher. So IRB approval is only required for research on secondary data if the secondary data includes enough information for the researcher to identify the respective individuals. If there are no “identifiers” IRB approval is not required.

On the other hand, the rules aren’t perfectly clear and there are situations where some information can allow for identifying subjects even though they are not necessarily “identifiers,” so maybe one will have to contact the IRB to figure out exactly what and when certain data are considered “identifiers”…
2. REW | 9 April 2009 at 10:03 am

Peter, you could have been certified by attending a one-hour training session yesterday. Alas, it was at 8 AM!

My takeaway from the session is that our University IRB is more benign than many others, perhaps most others. Purdue’s IRB may have had the largest mission creep and it may trod more heavily on the necks of social scientists that all others I have heard about.
3. Peter Klein | 9 April 2009 at 10:09 am

Randy, you may be right. Ours is the most intrusive of the ones I’ve dealt with before, but that’s a small sample.

Per, you are right that the situation is better now than it was a few years back (come on, can’t you guys grant me just a little artistic license?!). In part, I’d like to think that’s because of a meeting with the IRB brass held in this very department about five years ago. Apparently it was the first time they had talked to social-science researchers. I was told, to my face, that even research on firms — e.g., using Compustat or CRSP data — fell within the “human subjects” category because, well, firms are populated by humans. I responded that by that criterion, research in, say, astronomy would be considered human-subjects research because the telescopes were constructed by humans.
4. REW | 9 April 2009 at 10:48 am

How does Sancho Panza EVER keep up with you?
5. samir | 9 April 2009 at 10:51 am

I Feel Your Pain.
6. Per Bylund | 9 April 2009 at 10:56 am

I was told, to my face, that even research on firms — e.g., using Compustat or CRSP data — fell within the “human subjects” category because, well, firms are populated by humans. I responded that by that criterion, research in, say, astronomy would be considered human-subjects research because the telescopes were constructed by humans.

Well, there are obviously humans living on planets (at least one), which should be enough for astronomy research to be IRB approved.

What I got from the discussion yesterday, all the IRB cares about is if there are human beings involved (directly or indirectly) and, if so, if they are in any way identifiable by the researcher. If individuals are not, then you do not need approval.

This also means that you need approval only for the surveying part of a research project in which you construct a database where you have destroyed all identifiers and work only with that database. So you can construct a project and get IRB approval for the data gathering and cataloging in which you destroy all identifiers, and then use the data in whatever projects you think them applicable for as long time as you like without approval.

But I also understood from the discussion yesterday that whenever you are not completely sure that all the data you are using are totally anonymous or anonymified (is that a word?), you should contact the IRB “just in case.” Most social science research should fall in their least regulated category, so all you need to do is fill out a short form and the approval process could take but a couple of hours.

I’m not saying the IRB approval requirement is a good thing, but they seem to try hard not to make it too intrusive or take too much time of the researchers’ time.
7. Fred Thompson | 9 April 2009 at 11:08 am

IRB’s have a legitimate function: to ensure the protection of human and non-human animals who participate in research.”

In many cases, however, IRB charters are ill specified. In some cases it is not clear what the scope of their responsibility is, who delegated the IRB the authority to carry out their responsibilities, or that the authorizing body had the authority to do so. IRB’s should have charters authorized by the institution’s board, clearly specifying the scope and limits of their authority. Second, many boards have no formal procedures governing standard setting/rule making. Lacking formal procedures, IRBs frequently announce rules and interpretations that are both arbitrary and capricious. Moreover, in many instances, they go beyond the guidelines laid out by the NSF, with respect to review criteria, when and how informed consent is required and secured, and frequency of review. Moreover, once they have promulgated a rule, rule-making, or decision there is seldom if ever a satisfactory appeal mechanism in place. It is not surprising that they frequently over reach.
8. jeremy hunsinger | 9 April 2009 at 12:54 pm

I did the 1 hour course last week, certificate should arrive sometime soon. We basically need approval, either waiver, expedited, or full board review for anything with human subjects… However, that does not include documents, which is an interesting distinction. I’m guessing most icpsr data sets fall into the document category, as there is no way for the research, unless she is the originator, to find the represented folks. I on the other hand am doing ethnographic interviews of science professionals for the second stage of my science in virtual worlds study, so i need to get this certificate, bleh.
9. Mike Sykuta | 10 April 2009 at 11:24 am

Ahh, IRB.

Per, while a simple form can be filled out and should only take a couple hours to review and process, that assumes it is actually looked at in a short amount of time. I must say that our office has improved greatly in that respect, but there is still an occasional problem of the LILO production flow. Even if you are asked to make a simple clarification, your form gets put on the bottom of the stack as if it were a completely new submission. Thus what should be a couple hour process can take days depending on how many other requests are in the pipe.

However, Peter’s frustration seems not so much with the role or existence of IRB necessarily, but the kinds of questions social scientists are expected to answer in order to be certified to conduct human subjects research. Easily 1/3 of the questions have absolutely no relevance to social science research, thus social scientists must either a) invest in learning regulations that are completely useless and irrelevant to their field, and b) develop other ways of answering a sufficient number of questions accurately without learning the material.

I won’t elaborate on the last point, but I suspect it is more the rule than the exception.
10. Training Day « Institutional Review Blog | 10 April 2009 at 1:07 pm

[…] Peter Klein of the Organization and Markets blog offers a sad account of what it takes for a University of Missouri economist […]