Alternative models of funding

Research Funding: the Case for a Modified Lottery | mBio

https://www.templetonworldcharity.org/blog/why-we-decided-pick-25-million-grants-out-hat

Letter to NIH - send Tuesday 11 Feb ,2025

Suggestion to make health research better, cheaper and faster

Changes to the grant review process could hugely increase innovation and speed of research, lower admin costs, and support experts around the country not just in select universities.

Problem: the current NIH review process leads to boring incremental science, limited to narrow topics in a few elite institutions.

Who am I? I’m a professor at an R1 university who has had several NIH grants and served on several study sections. So I’ve benefited from the current system, but I think it’s poor.

Suggestions

One page letter of intent, full proposal upon invitation

Use 1-page letter of intent (LOI) before asking for full proposal. Have 15 day turn-around after deadline. NIH staff review the LOI for fit with funding priorities and possible impact.

The 1-page specific aims is the critical page anyway, for writers and reviewers, so why get people to waste time on the other 100-150 pages of an R01 if the topic is not a funding priority? Wastes resources writing it, wastes people’s time reading it.

Additionally, scientists are poor at matching applications to funding priorities, so it’s best they are not involved. Staff can be well-versed in funding priorities, and responsive to government policy.

Cut full scientific proposal to 6 pages + simple budget (1 page max)

As a reviewer I don’t need the pages and pages of fluff. Institutions have IRB’s so we don’t need Human Subjects. We don’t need 5 pages of biosketch. We don’t need made-up detailed budget rationale down to pennies on a $2 million grant. We don’t need resources. Let just the science be reviewed by scientists.

If the proposal is selected for funding, then provide a detailed budget and anything else essential. For something like Data Sharing Plan, just point the PI to the NIH policy and tell them they have to adhere.

A further advantage is the review will take way less time. At present it’s 6-9 months for a funding decision (sometimes longer), but with this reduced proposal size the review & decision period could be cut to less than 2 months.

Scientific review should simply be: is this rigorous science that will let the aims be met?

Scientists are biased and inaccurate at predicting groundbreaking work. However, they are good at assessing scientific rigor. Let them do that job.

Proposals that are highly rated for rigor are passed to the next step.

Step after scientific review: can these researchers at this institution do this work?

Proposal goes back to the NIH staff, and they do due diligence to make sure the institution can support the work, and that the research team is capable of doing the work well.

Rationale: scientists are highly biased towards the top R1 institutions and long-established researchers who have done the same work. This pattern leads to incremental research at the same few R1 universities.

Top applications go into a lottery to select the ones to fund

Evidence and experience shows that scientists cannot distinguish between high quality proposals. Remove them from that process, and let exciting ideas get funded. Most dramatic discoveries are serendipitous anyway.

An added bonus, this would be much faster than council rounds at present, where PO’s fight to get their favorite proposals funded.

Exceptions for clinical trials of specific high-need interventions: use RFAs (don’t just pretend to use them)

There will be areas of high need, especially in later stage clinical research, where we need data to inform healthcare practice. The high need studies can be identified in the first step (LOI) where researchers can respond to an RFA.

High need areas should be identified in an RFA, and only applications that directly match the RFA should be passed to the full proposal stage. At present, RFA’s seem to be ignored by researchers and councils when making funding decisions.

Ban pilot data

Pilot data makes for boring incremental science by a small number of researchers at well-off institutions. Pilot data is not needed - what is important is the impact of meeting the aims, the rigor of the science, and whether the team can perform those studies.

Pilot data means the work has been started, and for longitudinal studies, only people and institutions with substantial resources can collect such data. When reviewers ask for more pilot data, the researcher has to find the time and resources to collect more data, then analyze and present the new findings. My colleague today told me her R01 reviewers want to see more pilot data for her 6 month intervention - that means she has had to put in a new IRB, then will need to find money to collect the data, analyze it, and resubmit - a minimum 8 month wait till resubmission, and a delay of over a year for the revised proposal to be reviewed. To repeat myself, only a small number of established, well-resourced researchers can respond to such a request.

Another problem with pilot data is that reviewers expect it to show findings that align with the specific aims. But this is foolish, as chance means some pilot data will not align with aims. If the aims are worth studying, who cares what the pilot data show?

Background data and findings are fine, but data matching the aims should be automatic rejection.

Take-home

These changes would empower the NIH to support the government’s priorities, allow exciting ground-breaking research to flourish, speed up research, massively cut administrative waste at the NIH and in universities, and spread great research across the country, engaging great researchers everywhere, not just those who got into the elite R1s.