Research and Funding
Inside NIA: A Blog for Researchers

A footnote to our funding line: The AD PARs

A footnote to our funding line: The AD PARs

Posted on March 22, 2017 by Robin Barr, Director of the Division of Extramural Activities. See Robin Barr's full profile.

A recurring question from readers after I post a new funding line blog post is: Does that line apply to my application? It usually does, bringing good news to those blessed with applications within that line. The normal caution applies—the line means that we expect to pay awards. I have been at NIA long enough to remember the occasional sad tale of expectation not becoming reality. In one instance, a university went bankrupt before we could make the award!

Still, the funding line doesn’t apply to everything. In particular, we have “set-asides”—funds allocated to particular initiatives. We can’t post information about decisions on these set-asides because the information would make it too easy to work out who did and didn’t receive an award.

What are the AD PARs?

And that brings me to what we call the “AD PARs.” These are a group of 10 funding opportunity announcements—Program Announcements with Special Review—that cover the gamut of disciplines and fields within Alzheimer’s disease and Alzheimer’s-related dementia (AD/ADRD) research. We did set aside some funds for this group and issued them in the fall of 2015, when we first became aware that substantial additional funds for AD/ADRD research were likely to be made available in FY2016. They remain active now and, across the 10 FOAs, we continue to receive more than 100 applications each Council round.

Are these applications covered in the funding line post? No. And here’s why!

We arranged for the applications to obtain percentile ranks in review, as well as the raw impact rating. In principle, we could combine the applications from the 10 FOAs into one list and draw a funding line at the 18th percentile as we have done for all our other AD/ADRD applications. But we didn’t do that.

Using different methods for percentiles

The problem is that, for good and varied reasons, the scientific review officers responsible for organizing review of the AD PARs chose their own methods to percentile the applications. For the effects of using different percentile ranges on evaluation, imagine yourself on an elevator with a basketball team or with a group of elementary students. You feel differently about your height in these two cases. The base for comparison matters.

Some applications were reviewed by members of a regular panel convened as a special study section. Because the same scientists reviewed the AD PAR applications as reviewed the regular panel applications, it made sense to percentile the AD PAR applications using the regular panel applications as the base. (Think basketball players compared to basketball players though from a different team.)

Other applications were reviewed by scientists who were specially recruited for these reviews. For these applications, no parent panel existed to percentile against and so the review officer used the all-CSR average to percentile them. (Think basketball players or elementary school students compared to everyone who was ever on the elevator.)

Following so far? There’s one more twist. One particular PAR has received so many applications that the review officer chose to percentile them against themselves. (Who is the tallest, and who is the shortest, basketball player?)

In summary, the AD PARs were percentiled against themselves; or against a larger set of applications reviewed by the same reviewers; or against a much larger set of applications reviewed by everyone who had completed reviews on any panel in that round.

Making funding decisions

We didn’t think it was reasonable to combine these percentiles into one list. That seemed a little too arbitrary to use for funding decisions.

Instead, we separated the applications into individual groups that matched the particular PAR to which they responded, as well as the review committee which reviewed all the applications for that PAR. Then we made decisions for each list. In the end, we followed general guidance around that 18th percentile line, but we saw individual variation within each PAR.

If you applied to any of these AD PARs, contact your program officer for information on funding. And if you cannot reach your program officer, contact me.

In case you’re wondering what fate will befall your application—that one you submitted to our second round of AD FOAs that we issued in the fall of FY 2016? We have only recently received the applications responding to these FOAs. Review teams are being assembled and applications are being assigned to committees. The reviews will arrive over the summer and we will consider these for funding at the end of the fiscal year.

And, if you want to comment on the challenges of percentiling (or the value of tall basketball players) or the wisdom of continuing AD PARs, then the floor is open to you!

2 Comments
Share this:
Email Twitter Linkedin Facebook

Posted by Paul Bryant on Mar 29, 2017 - 8:20 am

Hello Dr. Barr, I heard yesterday that current K23 and R21 awards are scheduled to take a 10% reduction this year (both are AD-related). I just submitted an RPRR, and the next is due in a few weeks. I've been searching to confirm or locate any information regarding these cuts. Can you please point me in the right direction? Many Thanks as always -

Posted by Robin Barr on Mar 29, 2017 - 11:33 am

Our current NIH policy is to pay all noncompeting awards at 90% of their commitment for this year during the continuing resolution. That is probably where the 10% cut rumor originated. We make that cut regularly during a continuing resolution as we do not know what our final budget will be. In past years, we have been able to restore or partially restore these cuts when we received our full year budget. I hope that will also be true this year.

Post a Comment

We review comments before publication, as described in our comment policy. If you provide an email address, we will email you when your comment is published. Your name or nickname will be posted. Your email address will not be posted.