Data that IRBs should collect about themselves (created 2009-05-22)

This page is moving to a new website.

Somone on the IRBForum (TS) asked about what type of reports that an IRB should provide. There were a lot of good comments. I encouraged a data centric approach to reporting. Here's what I wrote.

I'm not on an IRB, though I have worked closely with one. While many of the comments so far are good, I wonder if most IRBs would benefit from collecting data directly from their customers. This would involve a survey of researchers asking for qualitative data on their interactions with the IRB. The logistics would be more complex, but perhaps a survey of research volunteers would be helpful also.

I also have to mention a topic which is not of direct concern to IRBs, but which, perhaps, should be. In my experience, many researchers overpromise and underdeliver on the time frames in which they will complete their studies. This raises both financial and ethical issues--research delayed is research denied.

With the IRB at my old job, I conducted a survey of 125 prospective human IRB approved studies and found that 90% failed to include an appropriate time frame for completion of the study. An IRB that approves a study without a definitive completion date is signing a blank check. Of the studies that did report a completion date, 73% took longer than planned, with one study taking almost three years rather than the planned 9 months.

A closely related problem is sample size shortfalls. More than half of the studies fell short of the planned sample size. The average shortfall among these studies was 55%. Almost every IRB insists on a sample size justification in order to assess the risk/benefit ratio. But when a study falls well short of its target sample size, does that invalidate the IRB's risk benefit calculation?

The full details are not published yet, but are available on my old website:

I suspect that many IRBs have the same problem with slipped deadlines and sample size shortfalls. I'd love to talk to anyone who has collected the type of data I'm describing here or who would like to collect this type of data. .