F1000Research F1000Research 2046-1402 F1000 Research Limited London, UK 10.12688/f1000research.124082.1 Opinion Article Articles ‘Science by consensus’ impedes scientific creativity and progress: An alternative to funding biomedical research

[version 1; peer review: 2 approved with reservations]

 Düzgüneş Nejat Conceptualization Writing – Original Draft Preparation Writing – Review & Editing https://orcid.org/0000-0001-6159-1391 a 1 Department of Biomedical Sciences, University of the Pacific - San Francisco Campus, San Francisco, CA, 94103, USA nduzgunes@pacific.edu

No competing interests were disclosed.

 19 8 2022 2022 11 961 12 8 2022 Copyright: © 2022 Düzgüneş N 2022 This is an open access article distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The very low success rates of grant applications to the National Institutes of Health (NIH) and the National Science Foundation (NSF) are highly detrimental to the progress of science and the careers of scientists. The peer review process that evaluates proposals has been claimed arbitrarily to be the best there is. This consensus system, however, has never been evaluated scientifically against an alternative. Here we delineate the 15 major problems with the peer review process, and challenge the Science Advisor to the President, and the leadership of NIH, NSF, and the U.S. Academy of Sciences to refute each of these criticisms. We call for the implementation of more equitable alternatives that will not constrain the progress of science. We propose a system that will fund 80,000 principal investigators, including young scientists, with just half the current NIH budget, three-fold more than the current number of grants, and that will forego the cumbersome, expensive, and counterproductive peer review stage. Further, we propose that the success of the two methods over 5–10 years be compared scientifically.

 Peer review; grant applications;NIH;NSF;granting agency N/A The author(s) declared that no grants were involved in supporting this work. Introduction

The success rate for National Institutes of Health (NIH) grants is currently 20% ( NIH Report, 2022). The funding rate at the National Science Foundation (NSF) was 26% in 2021 ( National Science Foundation, 2022). The Gates Foundation does not even release its grant success rate information. In 2009 and 2010, NIH received more than 20,000 applications for its Challenge Grants funded through the American Recovery and Reinvestment Act; the success rate was only 4% (NIH Report, 2011). The ‘successful’ projects are those that have been deemed by the consensus of peers to be worthwhile pursuing. Despite these very low percentages that afflict the careers of the great majority of scientists, the peer review system has been claimed to be the best system there is to allocate funding for biomedical research. This consensus system, however, has never been evaluated scientifically against an alternative ( Düzgüneş, 1999).

Perhaps the earliest challenge to this system at NIH was made by John McGowan (1992), who was at the time the Director of Extramural Research at the National Institute of Allergy and Infectious Diseases (NIAID). He revealed that proposals to investigate human immunodeficiency virus (HIV) infections of macrophages had been rejected by a study section because “the literature does not support the hypothesis that HIV can grow in macrophages” ( McGowan, 1992). And this is untrue! Regrettably, study sections have had too much power over what projects should proceed and which ones should be scrapped. As we have stated before, “such ‘science by consensus’ is unhealthy for the unfettered and productive pursuit of biomedical science” ( Düzgüneş, 1999).

We challenge the Science Advisor to the President, and the leadership of NIH, NSF, and the U.S. National Academy of Sciences to refute each of the following 15 major problems with the current NIH and NSF grant systems. If they cannot, however, and we believe they cannot, we ask these institutions to implement more equitable alternatives that will not constrain the progress of science.

Problems with peer review

The NIH Peer Review document describes the mission of NIH to be seeking ‘fundamental knowledge about the nature and behavior of living systems and to apply that knowledge to enhance health, lengthen life, and reduce illness and disability.’ The document claims that the “NIH has a longstanding and time-tested system of peer review to identify the most promising biomedical research” ( NIH Peer Review, 2019). During the initial peer review, the scientific merit of a grant application is evaluated by the Scientific Review Group that comprises scientists with relevant expertise in the area. The second review is the responsibility of the National Advisory Councils or Boards that decide on funding a proposal as well as on research priorities. Despite the claims of NIH that this is a longstanding and time-tested review process, it has never been compared scientifically to an alternative system, with respect to scientific productivity and breakthroughs, new therapeutic modalities, patents and its psychological, personal and scientific impact on grant applicants who do not ‘succeed.’

Furthermore, NIH has to process over 80,000 applications a year, utilizing over 25,000 reviewers ( NIH Peer Review, 2019).

We have identified 15 major shortcomings and problems of peer review, which we delineate below.

Some major breakthroughs in biomedical sciences have not been funded by NIH or NSF. There have been several publicized cases of highly important research not being given grant support that have later gone on to be recognized as significant scientific discoveries. Nobel Prize winner Stan Prusiner was not able to obtain NIH funding for studying prions early on in his research. Craig Venter’s proposal to apply his whole-genome sequencing method to sequence a bacterial genome was not funded by NIH, and Nobel Prize winner Leon Cooper’s work on neural networks was not supported by either the NIH or NSF ( Bendiscioli, 2019). These examples should have been a history lesson for funding organizations like NIH and NSF ( Düzgüneş, 1998).

Grant reviewers are competitors of applicants. If they are truly ‘peers’, grant review panel members are very likely to be competitors of the grant applicant, even if not directly on the subject of the proposal. Thus, they will not be inclined to give the benefit of the doubt to an innovative research proposal that has not already been substantially carried out, particularly when they are struggling to procure funding themselves.

Discoveries are made before grant awards. The requirement for preliminary data in most grant applications indicates that a scientific discovery is expected to have already been made. Thus, the NIH and NSF may not be funding discoveries, but merely funding “mopping up operations,” in the words of Thomas Kuhn (1962), unless the preliminary data have been generated by a previous grant.

Reviewer critiques may be inaccurate, but without the responsibility of making inaccurate statements. Reviewers appear to have a mission to severely criticize applications to be able to weed them out, usually without the requirement to provide a published reference for any criticism. The reviewers are never accountable for their false statements or their scores ( Swift, 1996), even though they can derail scientific careers and the advancement of a field of science.

Criticism never ends. Grant applicants may re-apply after revising their proposal to respond to the written critique of the review panel. However, the panel may have new members at this later time and may then have entirely new criticisms. In essence, if the review panel does not want to fund an application it will not fund the application, revealing the whims of the individual reviewers.

Early career reviewers trained in a narrow area of science often think that valid science is what they are trained in. Thus, they prevent the progress of science that may otherwise produce significant insights or therapeutic approaches to treat diseases. This problem was emphasized by Costello (2010): “… the new generation of grant reviewers judge grant proposals through the myopic lenses of their specialties …. Important ideas and proposals that lie outside the current interest in molecular biology are unlikely to get a credible and knowledgeable review …”

Nonscientific, unpublished review criteria. Reviewers tend to use nonscientific criteria when making funding decisions. These include: (i) ‘probability of success’ which would favor projects proposing only incremental advances and no risk-taking; (ii) ‘level of enthusiasm’ which is highly subjective and depends on the reviewer’s mood at the time; and (iii) ‘grantsmanship’ which is essentially rendering grant-writing a game, expecting particular approaches to the project. Implying the nonscientific nature of the evaluation process, the study by Pier et al., (2018) has shown that there is very little agreement between reviewers evaluating the same NIH grant applications.

Translational projects may require long-term funding. Projects that need additional time and experimentation to translate basic findings and initial discoveries into therapeutics or diagnostics may be considered by reviewers not to be innovative, thereby precluding the rapid development of a product that could diagnose or treat diseases. An example of this problem with NIH peer review is our inability to obtain grant funding since the mid-2000s for our research to develop gene therapy for oral cancer based on our initial discoveries ( Neves et al., 2009), despite many applications.

Robbing Peter to pay Paul. Principal investigators may need to channel the funds of an existing grant to produce preliminary data for a new application in a new research area, instead of performing the funded experiments. Thus, experiments described in detail in applications may never be carried out and may essentially have been written only to convince reviewers to fund the grant application. In our view, this practice is unethical. It also demonstrates the absurdity of requiring preliminary data.

Precious scientist time is wasted on grant applications. Investigators spend a large proportion of their time on grant applications, which necessarily takes them away from their currently funded projects, if they are indeed grant recipients. This is not only counter-productive but may also be time paid by salary support from the granting agency, time that should have been spent on the funded project. A study by Kulage et al. (2015) calculated the cost of preparing a grant application. Principal investigators in this particular field spent between 70 to 162 hours per grant, and research administrators spent 34 to 66 hours, at a cost of USD $4,784 to $13,512. They estimated that, because funding rates are in the range 5–15%, a grant that is eventually funded would cost $72,460–$270,240. They concluded that “less costly and more efficient models of research funding are needed for the sustainability of the nursing profession” ( Kulage et al., 2015). Scientists who have spent years in training and in research should be spending their time on scientific research, not on bureaucracy.

Describing experiments to be performed in five years is unrealistic. The elaborate description of experiments that will be performed three or five years in the future in a grant application contradicts the true nature of scientific research. Thus, for reviewers to expect meticulous descriptions, as if this is how science advances, goes against the true nature of science. Science is driven by the insights of scientists and new discoveries, and often requires immediate changes in approach or direction.

Waiting for grant funding hinders scientific progress. Many fields advance rapidly while investigators are waiting for their grant applications to be evaluated and funded. If the investigator is not funded independently, the project barely moves forward. We cannot afford science to progress at this slow and saltatory rate with the uncertainty in grant funding.

The human and material cost of NIH peer review. The administration of approximately 80,000 applications and 25,000 reviewers per year ( NIH Peer Review, 2019) costs NIH and the research community both money and time that could have been used for actual research. For reviewers, evaluating grant applications is a chore performed for the sake of recognition and prestige, and perhaps to increase their own chances of obtaining funding. This can result in the compromise of objectivity by the reviewers, and even resentment, because of the inordinate amount of time required to complete a review. NIH officials conducting sessions at scientific meetings on how to write grants admit that reviewers may not be able to spend quality time on reviewing applications. Of course, this is never admitted in print, since peer review is supposed to be unquestionably the ideal system for funding science.

NIH scientists do not compete for grant funding. Although NIH provides extramural funds following grueling peer review of grant applications, its own scientists do not have to compete for this type of funding. Thus, NIH itself appears to have recognized the extreme drawbacks of the peer review system, enabling its intramural community to undertake long-term projects with stable funding and large laboratory groups. If peer review is such an indispensable system for funding science, why does NIH not implement this system for its own scientists? Why is the extramural scientific community considered second-rate citizens who must clamor for funding all their lives?

Review panel scores do not predict success. An analysis of 102,740 funded grants has shown that percentile scores generated by NIH review panels for the applications are poor predictors of publication and citation productivity ( Fang et al., 2016). Thus, the meticulous scoring process is essentially useless. Arturo Casadevall of Johns Hopkins University and the senior author of this study is quoted as saying “A negative word at the table can often swing the debate. And this is how we allocate research funding in this country” ( Johns Hopkins Bloomberg School of Public Health, 2016).

 Alternatives to peer review

We have previously proposed simple alternatives to the current peer review system ( Düzgüneş, 1999, 2007). This new system would provide continuous and stable funding for 10-year periods to scientists with a track record of solid publications ( Düzgüneş, 1999) and to young scientists starting their first independent positions in a university or a research institute ( Düzgüneş, 2007). Scientists opting for this mode of funding would merely submit a letter of intent with a one-page broad outline of their research direction. They could be chosen based on criteria including publications, citations and potential impact of their research field, by an international group of both established and young scientists who are not in a position to receive funding from NIH or NSF and are thus not competitors. About half of the NIH extramural funds could still be allocated to the current system of review, especially for projects requiring much larger budgets than what we are proposing below.

Under this new system, NIH grants to established scientists would be limited to $400,000 per year, and they would be phased in over several years, up to possibly 40,000 grantees. With indirect costs limited to 30%, these grants would cost $20.8 billion per year. Grants to young investigators would be set at $150,000 per year, with the same indirect cost rate. 40,000 such grants would cost NIH $7.8 billion. Thus, at a total cost of $28.6 billion the NIH could fund 80,000 such grants, with minuscule expenses for scientific review. Since the sum is only slightly more than half of the current NIH budget of $52 billion, the rest of the NIH budget could be allocated to about half (about 13,000) the current number of grants. Since this system will be phased in, and the NIH budget is likely to increase within the next five years, there will be no undue burden on the traditional grants and intramural funding. This system will result in the funding of 93,000 principal investigators, instead of the current approximately 26,000.

It is instructive to note the findings of Azoulay et al. (2011) in comparing Howard Hughes Medical Institute awardees and NIH grant recipients. They reported that “selection into the HHMI investigator program—which rewards long-term success, encourages intellectual experimentation, and provides rich feedback to its appointees—leads to higher levels of breakthrough innovation, compared with NIH funding—which is characterized by short grant cycles, predefined deliverables, and unforgiving renewal policies. Moreover, the magnitudes of these effects are quite large.”

Vaesen and Katzav (2017) analyzed the proposal to “distribute available funds equally among all qualified researchers, with no interference from peer review.” Their analysis indicated that “researchers could, on average, maintain current PhD student and Postdoc employment levels, and still have at their disposal a moderate (the U.K.) to considerable (the Netherlands, U.S.) budget for travel and equipment.” Our proposal combines this equitable distribution of funds with the option for scientists undertaking very expensive projects to apply for the remaining highly competitive funds.

 Evaluating the scientific success of grants obtained via peer review and the alternative system proposed here

The paradigm shift we are proposing does not end here. The scientific productivity of scientists in these two categories over a five-year and 10-year period will be analyzed, in terms of citations, significant discoveries, and development of therapeutics, per dollar amount spent. As we have indicated previously ( Düzgüneş, 1999), “The United States has expended enormous capital in the training of its scientists. The scientific potential of the more than 80 percent of biomedical scientists who are unable to procure grants is too precious a resource to waste.”

 Conclusions

While contemplating writing this section, we came across an e-mail sent to potential NIH grant applicants and a separate website aimed at academics and including advice on grant applications as part of an industry aimed at grant applicants for ‘winning’ reviews. The e-mail advertised that their program enabled the participants to ‘successfully write for reviewers’. If an applicant is writing to impress a particular reviewer, the detailed norms, supposed objectivity, and scoring system of NIH peer review becomes questionable. Another website gave the advice to involve the reviewers’ ‘reptilian brain’ and went on to say that the written review of a grant application may come from the rational, cerebral layer of the brain, but the decision on whether the grant is awarded or not actually comes from the most instinctual layer. What has become of the best method to review grant applications?

With all the problems of peer review of grant applications, we ask the Science Advisor to the President, and the leadership of NIH, NSF, and the U.S. National Academy of Sciences to implement more equitable alternatives that will not constrain the progress of science. A staring point is the very simple and highly cost-effective alternative we have proposed here.

 Data availability

No data are associated with this article.

 References Azoulay P Graff Zivin JS Manso G : Incentives and creativity: evidence from the academic life sciences. RAND J Economics. 2011;42(3):527–554. 10.1111/j.1756-2171.2011.00140.x Bendiscioli S : The troubles with peer review for allocating research funding: Funders need to experiment with versions of peer review and decision-making. EMBO Rep. 2019 Dec 5;20(12):e49472. 31680417 10.15252/embr.201949472 Costello LC : Perspective: is NIH funding the “best science by the best scientists”? A critique of the NIH R01 research grant review policies. Acad. Med. 2010 May;85(5):775–779. 10.1097/ACM.0b013e3181d74256 Düzgüneş N : History lesson. The Scientist. 1998;12(6):8. Düzgüneş N : Science by consensus: why the NIH grant review system must be changed. The Scientist. 1999;13(8):13. Düzgüneş N : A new paradigm for NIH grants. The Scientist. 2007;21(8):24. Fang FC Bowen A Casadevall A : NIH peer review percentile scores are poorly predictive of grant productivity. elife. 2016 Feb 16;5:e13323. 26880623 10.7554/eLife.13323 Johns Hopkins Bloomberg School of Public Health: Researchers: Peer Review System for Awarding NIH Grants Is Flawed. 2016.(accessed August 2, 2022). Reference Source Kuhn TS : The Structure of Scientific Revolutions. Chicago: University of Chicago Press;1962;264pp. Kulage KM Schnall R Hickey KT : Time and costs of preparing and submitting an NIH grant application at a school of nursing. Nurs. Outlook. 2015 Nov-Dec;63(6):639–649. 26566959 10.1016/j.outlook.2015.09.003 McGowan JJ : NIH peer review must change. J. NIH Res. 1992;4(8). National Science Foundation: Funding and support descriptions. 2022. (accessed July 4, 2022). Reference Source Neves S Faneca H Bertin S : Transferrin lipoplex-mediated suicide gene therapy of oral squamous cell carcinoma in an immunocompetent murine model and mechanisms involved in the antitumoral response. Cancer Gene Ther. 2009 Jan;16(1):91–101. 18690206 10.1038/cgt.2008.60 NIH Peer Review: grants and cooperative agreements:2019. (accessed July 30, 2022). Reference Source NIH Report: NIH ARRA funding. 2010. (accessed July 29, 2022). Reference Source NIH Report: Success rates: R01-equivalent and research project grants. 2022. (accessed July 4, 2022). Reference Source Pier EL Brauer M Filut A : Low agreement among reviewers evaluating the same NIH grant applications. Proc. Natl. Acad. Sci. U. S. A. 2018 Mar 20;115(12):2952–2957. 29507248 10.1073/pnas.1714379115 Swift M : Innovative research and NIH grant review. J. NIH Res. 1996;8(12):18. Vaesen K Katzav J : How much would each researcher receive if competitive government research funding were distributed equally among researchers? PLoS One. 2017;12(9):e0183967. 28886054 10.1371/journal.pone.0183967 10.5256/f1000research.136253.r200648 Reviewer response for version 1 Recio-Saucedo Alejandra 1 Referee https://orcid.org/0000-0003-2823-4573 National Institute for Health and Care Research (NIHR) Coordinating Centre, University of Southampton, Southampton, UK

Competing interests: No competing interests were disclosed.

 22 9 2023 Copyright: © 2023 Recio-Saucedo A 2023 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation approve-with-reservations

The author presents an interesting and important reflection on a debate that has been going on for quite some time: Is peer review the best mechanism ( e.g., in terms of efficiency, fairness, ability to detect ground-breaking ideas, resistance to bias) to make research funding decisions? Given the worldwide use of this mechanism for a number of activities within the research landscape (research funding, academic promotions, publications), peer review has been extensively studied many times with the aim of finding alternative mechanisms to fund research, some of them having been more or less successful than others.

In the article the author describes an alternative mechanism to distribute funds which would be relevant to US funding agencies. The potential impact of an efficient alternative to the use of peer review in funding decisions to the research practice is significant.

Some of the key points in the article indicate:

The author summarises peer review in research funding as ‘cumbersome, expensive, and counterproductive’

The author presents 15 issues with the current peer review system (science by consensus), including: missing major scientific breakthroughs, unfair/unbiased process driven by competition from researchers working in fields similar/the same to that of the applicant, requirements for preliminary data, lack of accountability for reviewers, poor understanding of the value of trans-disciplinary proposals, poor reviewer agreement, lack of long-term funding, and others.

The author proposes the alternative to be implemented with a view to evaluate its effect compared to the current peer review system in 5-10 years, also raising a challenge to senior management of major US research funders and leaders.

Overall strengths

The article is well structured, interesting, and contributes to the literature on how to increase research efficiencies, decision-making transparency, and impact, which keeps large funding organisations across the world preoccupied and involved in constant efforts for improvement.

Reading this led me to think about how I am indeed putting peer review into practice by writing my reflection on this article. At the same time I couldn’t avoid thinking that for all its challenges, peer review is a mechanism that encourages conversation, debate, and advances in scholarship, and how open peer review offers avenues to address some of the challenges associated with peer review discussed by the author.

Areas for improvement

The literature used to illustrate some of the research in peer review is not particularly the most recent. Comprehensive analysis of alternatives to peer review, which covers issues associated to peer review at all stages of funding processes, are available (see for example Peter Kolarz, Anete Vingre, Aaron Vinnik, Antonio Neto, Carlos Vergara, Claudia Obando, Rodriguez, Kalle Nielsen, Laura Sutinen, Review of peer review. Online report. June 2023, https://www.ukri.org/wp-content/uploads/2023/07/UKRI-060723-Review-of-peer-review-Final-report-revs-v2.pdf). Alternatives that substitute peer review at the funding decision stage have been published ( e.g., the use of random allocation mechanisms to allocate research funds: Avin S: Mavericks and lotteries. Stud Hist Philos Sci. 2019; 76: 13-23) as well as researchers’ views of some of these approaches (see for example:  Liu M, Choy V, Clarke P, Barnett A, et al.: The acceptability of using a lottery to allocate research funding: a survey of applicants.Res Integr Peer Rev. 2020; 5: 3).

The description of the alternative is summarised since it refers to previous publications, but the summary misses the key challenges and unintended consequences of implementing the proposed alternative. Identification of unintended consequences will allow for the development of relevant indicators that would generate data/evidence to evaluate the efficiency of the alternative in comparison to the traditional use of peer review. In addition, the alternative proposed  does not seem to offer a new way of allocation research funds, as it includes the use of peer review in the way that is currently used ( i.e., a panel of peers selects applications based on given criteria).

I appreciate that the author is not bringing in the positives of peer review to this article and that this strategy adds emphasis to the argument presented, however the narrative highlights the inefficiencies and issues surrounding peer review without acknowledging that no system is likely to be a one-size-fits-all solution to all the challenges surrounding allocation of research funds. Recognising unintended consequences of alternatives is key and recognising the pros of peer review would actually support the alternative proposed (since it utilises peer review).

In terms of the alternative proposed, there are some points that would benefit from clarification:

The author writes: “Review panel scores do not predict success. An analysis of 102,740 funded grants has shown that percentile scores generated by NIH review panels for the applications are poor predictors of publication and citation productivity (Fang et. al., 2016). Thus, the meticulous scoring process is essentially useless.”

Given the broad range of impacts that result from research, and the efforts on responsible use of metrics that are addressing known issues with research culture, the conclusion that a scoring system is not useful based on its ability to predict publication and citation productivity is to some extent reductive. I understand that the author is presenting more issues around peer review, but the emphasis on publication and citation is running counter intuitively to how research culture is evolving.

The proposed alternative reads: “They [applicants] could be chosen based on criteria including publications, citations and potential impact of their research field, by an international group of both established and young scientists who are not in a position to receive funding from NIH or NSF and are thus not competitors”

The alternative offered here is a variation of peer review, but does not offer a different method to assess funding applications that decreases potential bias in decisions or increase inter-rater agreement of scorers. As written earlier, some aspects of the alternative are not different from the current way of selecting applications based on given criteria. The author suggests that funding decisions be made by “established and young scientists who are not in a position to receive funding”. Finding established and young scientists who cannot receive funding by the selected organisations to conduct the review and/or make funding decisions decreases the pool of potential reviewers. Thus the workload on eligible individuals can become unmanageable very rapidly. Another aspect is training those young scientists in assessing applications for awards/grants that they do not know in depth. There seems to be underlying risks to the proposed alternative that run the risk of affecting Early Career Researchers (ECRs) disproportionately. In addition, the author suggests that funding panels consider potential impact of their research field. How can potential impact be objectively assessed by reviewers/funding panels? How would criteria that takes into account publications and citations be supportive of ECRs whose list of publications would not be competitive compared to senior researchers? Are ECRs inadvertently ineligible to the alternative?

“Thus, at a total cost of $28.6 billion the NIH could fund 80,000 such grants, with minuscule expenses for scientific review.”

If the process to identify reviewers and go through the peer review process for the other half of the NIHR extramural funds remain in place, how will costs decrease to the point that they become negligible?

Overall, this is a good paper that will benefit from including additional information on key challenges of a large-scale implementation of an alternative to the de facto mechanism used in the distribution/allocation of research funds, which currently also regulates all research and scientific activity (from funding to publication; recruitment and career progression).

Is the topic of the opinion article discussed accurately in the context of the current literature?

Partly

Are arguments sufficiently supported by evidence from the published literature?

Partly

Are all factual statements correct and adequately supported by citations?

Yes

Are the conclusions drawn balanced and justified on the basis of the presented arguments?

Partly

Reviewer Expertise:

I am Senior Research Fellow at the National Institute for Health and Care Research within the Research on Research group and have conducted research on peer review.

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.

References : Review of peer review. Online report. June 2023.2023; Reference source : From funding agencies to scientific agency: Collective allocation of science funding as an alternative to peer review. EMBO Rep .2014;15(2) : 10.1002/embr.201338068 131-3 24397931 10.1002/embr.201338068 : Recruitment of reviewers is becoming harder at some journals: a test of the influence of reviewer fatigue at six journals in ecology and evolution. Research Integrity and Peer Review .2017;2(1) : 10.1186/s41073-017-0027-x 10.1186/s41073-017-0027-x : Peer review of health research funding proposals: A systematic map and systematic review of innovations for effectiveness and efficiency. PLoS One .2018;13(5) : 10.1371/journal.pone.0196914 e0196914 29750807 10.1371/journal.pone.0196914 : Reducing bureaucratic burden in research, innovation and higher education. London .2020; Reference source : Grant Peer Review: Improving Inter-Rater Reliability with Training. PLoS One .2015;10(6) : 10.1371/journal.pone.0130450 e0130450 26075884 10.1371/journal.pone.0130450 : Alternatives to peer review in research project funding: 2013 Update. Santa Monica: RAND Corporation .2013; Reference source : The Air We Breathe: A Critical Look at Practices and Alternatives in the Peer-Review Process. Perspect Psychol Sci .2009;4(1) : 10.1111/j.1745-6924.2009.01105.x 40-50 26158832 10.1111/j.1745-6924.2009.01105.x : The acceptability of using a lottery to allocate research funding: a survey of applicants. Res Integr Peer Rev .2020;5: 10.1186/s41073-019-0089-z 3 32025338 10.1186/s41073-019-0089-z : Mavericks and lotteries. Stud Hist Philos Sci .2019;76: 10.1016/j.shpsa.2018.11.006 13-23 31558205 10.1016/j.shpsa.2018.11.006 Düzgüneş Nejat University of the Pacific, USA

Competing interests: No competing interests were disclosed.

 18 11 2023

Response to Dr. Recio-Saucedo

We thank Dr. Recio-Saucedo for her insightful review of our paper.

“At the same time I couldn’t avoid thinking that for all its challenges, peer review is a mechanism that encourages conversation, debate, and advances in scholarship, and how open peer review offers avenues to address some of the challenges associated with peer review discussed by the author.”

We agree that “ open peer review” has its advantages in encouraging debate on peer review and on advances in research. When individuals are willing to be identified in open peer review, they are less likely to be harshly critical of a scientific proposal. When “peer review” is anonymous, and it does not present published evidence to support the “review”, it can become irresponsible and vicious.

“The literature used to illustrate some of the research in peer review is not particularly the most recent. Comprehensive analysis of alternatives to peer review, which covers issues associated to peer review at all stages of funding processes, are available (see for example Peter Kolarz…June 2023,…”

We hope it is appreciated that our original article was submitted in August 2022, about 10 months earlier than the analysis mentioned by the Reviewer. Nevertheless, in the revised article, we cite and discuss this paper as well as some of the other publications the Reviewer has suggested. All the revisions in the manuscript are indicated with Track Changes.

“The description of the alternative is summarised since it refers to previous publications, but the summary misses the key challenges and unintended consequences of implementing the proposed alternative. Identification of unintended consequences will allow for the development of relevant indicators that would generate data/evidence to evaluate the efficiency of the alternative in comparison to the traditional use of peer review.”

The reviewer refers to “the key challenges and unintended consequences of implementing the proposed alternative;” but does not give any examples. Nevertheless, this is an important point. As an unintended consequence, we may think of cases where the track records of a few awardees were misidentified, and they do not generate any significant publications during the 10 years of funding. We would argue, however, that the loss of scientific progress in these circumstances should be compared to the loss of productivity by a laboratory when it cannot renew a 3-year grant and all the trained personnel have to leave the laboratory.

By contrast, we could envision a scientist who opted for the novel funding mechanism, but did not make the cutoff of scientific measures used to establish the track record. In this case, the scientist always has the option to apply for the competitive grants. This is the situation with the current peer review system in which about 80% of biomedical scientists find themselves. The frequency of such incidences could be quantified by the granting agency. Whether this analysis could be translated into the “efficiency of the alternate in comparison to the traditional use of peer review” begs the question, “is there any efficiency in the current peer review system?” Our answer would be “no.” We have added a section in the revised paper discussing the possible, but in our view minimal, unintended consequences of our alternative method of funding.

“In addition, the alternative proposed  does not seem to offer a new way of allocation research funds, as it includes the use of peer review in the way that is currently used ( i.e., a panel of peers selects applications based on given criteria).”

We stated in our paper that the alternative method of funding scientists would be based on the track record of solid publications of the scientists, a system which we had first proposed in 1999 (Düzgünes ̧, 1999). In 2007, we added the mechanism to provide funding to young scientists starting their first independent positions in a university or a research institute (Düzgünes ̧, 2007). The reviewer interprets the work of the “international group of both established and young scientists who are not in a position to receive funding from NIH or NSF and are thus not competitors” as a form of peer review, possibly because of our use of the word “chosen.” We should have given more details of what determines a “track record of solid publications” We did not intend this group of scientists to be involved in the choice of the scientists to receive funding, but merely to establish the track record criteria. Examples of such criteria are citations, publications, the h-index and the d-index (Di Caro et al., 2012). The international panel of scientists would only be involved in establishing the overall weight of such criteria. We have modified the manuscript to clarify this aspect of our proposed system.

“…however the narrative highlights the inefficiencies and issues surrounding peer review without acknowledging that no system is likely to be a one-size-fits-all solution to all the challenges surrounding allocation of research funds. Recognising unintended consequences of alternatives is key and recognising the pros of peer review would actually support the alternative proposed (since it utilises peer review).

One of our main contentions here is that a system that has been acclaimed for decades as being the best method to allocate research funding, has not even been tested side-by-side against an alternative, unbiased, and more rational system. As we have delineated above, we have recognized the positive aspect of peer review by proposing that the unequivocally objective criteria for the new system would be discussed and established by an international panel of junior and established scientists. Further pros of peer review may be seen in the review of manuscripts submitted to journals, where errors in the research or in the analysis of data may be identified. Even so, the rejection of breakthrough articles by Karikó and Weissman on mRNA by reputable journals, as reported in the NY Times, is not a plus for the reputation of peer review even for journal articles.

In terms of the alternative proposed, there are some points that would benefit from clarification:

The author writes: “Review panel scores do not predict success. An analysis of 102,740 funded grants has shown that percentile scores generated by NIH review panels for the applications are poor predictors of publication and citation productivity (Fang  et. al., 2016). Thus, the meticulous scoring process is essentially useless.”

Given the broad range of impacts that result from research, and the efforts on responsible use of metrics that are addressing known issues with research culture, the conclusion that a scoring system is not useful based on its ability to predict publication and citation productivity is to some extent reductive. I understand that the author is presenting more issues around peer review, but the emphasis on publication and citation is running counter intuitively to how research culture is evolving.

We are merely stating the conclusions of Fang et al. (2016), which may be reductive, but they provide strong support for our contention that the supposedly objective scoring system of grant peer review (which, by the way, has not been tested against an alternative system) does not lead to success in terms of measurable quantities, such as publications and citations. The Reviewer has not explained what she means by the thesis that these metrics are “running counter intuitively to how research culture is evolving.” Nevertheless, we believe this issue is outside the scope of our article.

The alternative offered here is a variation of peer review, but does not offer a different method to assess funding applications that decreases potential bias in decisions or increase inter-rater agreement of scorers. As written earlier, some aspects of the alternative are not different from the current way of selecting applications based on given criteria.

In the revised manuscript, we have clarified the involvement of the international panel of scientists. They do not act as peer reviewers of any grant applicant or application. They merely set the objective criteria by which scientists applying for the new system are ranked, without knowing anything about the scientist, their field of research or the general topic of the research the scientist wishes to pursue.

“Finding established and young scientists who cannot receive funding by the selected organisations to conduct the review and/or make funding decisions decreases the pool of potential reviewers.”

The majority of scientists live in countries outside of the United States. Although some non-U.S. scientists receive funding from NIH or NSF, their number is extremely limited, and they would not be permitted to participate in the international panel.

“Thus the workload on eligible individuals can become unmanageable very rapidly”

As we have clarified above and in the revised manuscript, the work of the international panel is minuscule compared to that of the NIH study sections or NSF reviewers. The panel merely determines the criteria of acceptance, such as the percentage weight of the d-index, the number of citations, the number of publications, and other criteria (e.g. the h-index) they may consider to be useful. If such ranking is too rigorous and eliminates too many of the applicants (which would be calculated almost instantaneously), the panel would relax the criteria.

“Another aspect is training those young scientists in assessing applications for awards/grants that they do not know in depth.”

As we have explained above, no such training would be required of any scientist. Nevertheless, the members of the review panel would have to become familiar with the d-index and the h-index, and perhaps other criteria.

“There seems to be underlying risks to the proposed alternative that run the risk of affecting Early Career Researchers (ECRs) disproportionately.”

On the contrary, in the scheme we have proposed, early career researchers beyond their post-doctoral training would be given 5-year grants to establish their own researach programs, without spending most of their time applying for grants, and most likely failing. This would help young scientists in obtaining tenure and keeping their laboratories and trained personnel. By contrast, in the current system, the fierce competition with established scientists puts early career researchers at a perilous disadvantage, and force a majority of them to close their labs and move to other universities, research institutions or industry.

In addition, the author suggests that funding panels consider potential impact of their research field. How can potential impact be objectively assessed by reviewers/funding panels? How would criteria that takes into account publications and citations be supportive of ECRs whose list of publications would not be competitive compared to senior researchers? Are ECRs inadvertently ineligible to the alternative?

We agree that considering the potential impact of the research field would be difficult; thus, we have deleted this expectation of the international panel. Regarding the ECRs, we have addressed this concern in the previous point above. In contrast to the Reviewer’s contention, ECRs are currently at an enormous disadvantage, because they are expected to compete with all the established investigators. In the system we are proposing, ECRs have the opportunity to set up their independent research with the help of the 5-year grant they will be given when they start their first independent position. Depending on the NIH and NSF budgets, the period and the amount of funding to the ECRs could be increased from those we are proposing here.

If the process to identify reviewers and go through the peer review process for the other half of the NIH extramural funds remain in place, how will costs decrease to the point that they become negligible?

The system we are proposing here would enable investigators whose research needs are much greater than the $400,000 per year, to compete for funds just like they have been doing for decades. The advantage of the split system we are proposing is that these investigators would not have to also compete with the 80,000 other applicants. We proposed the split system of funding with the expectation that the two segments could be compared at the end of a 5- or 10-year period, in terms of research productivity, breakthroughs and patents (Düzgüneş, 1999). The burden of peer review in study sections at NIH would be lowered, because there would be 80,000 fewer applications.

 10.5256/f1000research.136253.r149008 Reviewer response for version 1 Fang Ferric 1 Referee https://orcid.org/0000-0002-3243-110X Department of Laboratory Medicine and Pathology, University of Washington, Seattle, WA, USA

Competing interests: No competing interests were disclosed.

 12 9 2022 Copyright: © 2022 Fang F 2022 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation approve-with-reservations

This commentary is a critique of the approaches used by the U.S. National Institutes of Health and National Science Foundation to allocate research funding. The commentary lists what it describes as 15 “major problems with the peer review process.” In place of the current system, the commentary suggests awarding funds at fixed levels based on level of applicant experience, a one-page summary of research plans, and publication record. This would reduce administrative costs and substantially increase the number of funded researchers. Few of the issues with grant peer review or the solutions proposed in this commentary are new 1 , 2 , 3 .  Nevertheless, the topic remains timely, as low success rates for grant applications persist.

1. The 20% success rate for NIH grants is a misleading indicator of the current difficulty in obtaining research funding, as success rates count original and resubmitted grants that are received during the same fiscal year as a single submission and combine different types of grants that have variable success rates. A better indicator would be paylines, which are at 10-12% for R01 applications from established investigators at many institutes. Thus, the situation is even more dire than the commentary suggests.

2. There is some redundancy in the list of “major problems,” which essentially boil down to: (a) peer review is biased and unreliable; (b) the current funding system discourages innovation and long-term projects; (c) too much time is wasted on seeking funding or peer review instead of on research. Another concern, which is not mentioned, is that insufficient levels of public funding for research and development can have a disproportionate impact on female and underrepresented minority researchers 4 , 5 , 6 .

3. The commentary fails to consider potential disadvantages of the proposed funding strategy. For example, dramatically lowering the bar for research funding might result in an increase in lower quality research, because many labs whose work was not previously found to be sufficiently meritorious to warrant support would now be funded. Some institutions might be tempted to expand the size of their faculties in order to exploit the system, which would increase pressure on the research budget. Half of the extramural budget would be retained for competitive applications involving projects that require larger budgets, but this would essentially intensify competition among larger laboratories, which include some of the most productive research groups. The $400,000 per year budget falls far short of the $750,000 direct funding level that was found to be associated with peak measures of productivity in a 2006 NIH study 7 , and research costs have only risen since that time, suggesting that most labs would be underfunded under the new scheme.

4. The comparison of recipients of support from the NIH and Howard Hughes Medical Institute is invalid, because HHMI investigators represent a more highly selected subset of scientists with demonstrated high levels of productivity.

5. The proposed funding approach shares some features with the approach proposed by Pagano 1 , which should be cited.

6. A single paragraph suggests performing a 5- and 10-year formal study to compare the productivity resulting from the current and alternative funding approaches but fails to consider the many methodological challenges in performing such a study.

7. The commentary repeatedly refers to “we,” but only a single author is listed.

Although I share many of the frustrations expressed in this commentary, I have concerns about the practicality of the proposed solution.  Nevertheless, a dialogue on possible ways to address the inadequate current levels of research funding and the shortcomings of grant peer review should continue. This commentary may help to further that dialogue.

Is the topic of the opinion article discussed accurately in the context of the current literature?

Yes

Are arguments sufficiently supported by evidence from the published literature?

Yes

Are all factual statements correct and adequately supported by citations?

Yes

Are the conclusions drawn balanced and justified on the basis of the presented arguments?

Partly

Reviewer Expertise:

microbiology, molecular biology

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.

References : American Idol and NIH Grant Review. Cell .2006;126(4) : 10.1016/j.cell.2006.08.004 637-638 10.1016/j.cell.2006.08.004 : More time for research: fund people not projects. Nature .2011;477(7366) : 10.1038/477529a 529-31 21956312 10.1038/477529a : Research Funding: the Case for a Modified Lottery. mBio .2016;7(2) : 10.1128/mBio.00422-16 10.1128/mBio.00422-16 : Race, ethnicity, and NIH research awards. Science .2011;333(6045) : 10.1126/science.1196783 1015-9 21852498 10.1126/science.1196783 : Sex differences in application, success, and funding rates for NIH extramural programs. Acad Med .2011;86(6) : 10.1097/ACM.0b013e31821836ff 759-67 21512358 10.1097/ACM.0b013e31821836ff : Inequalities in the distribution of National Institutes of Health research project grant funding. eLife .2021;10: 10.7554/eLife.71712 10.7554/eLife.71712 : Study says middle sized labs do best. Nature .2010;468(7322) : 10.1038/468356a 356-357 10.1038/468356a Düzgüneş Nejat University of the Pacific, USA

Competing interests: No competing interests were disclosed.

 18 11 2023

Response to Dr. Ferric Fang

We thank Dr. Ferric Fang for his thorough review of our paper. We greatly appreciate his conclusion that “Nevertheless, a dialogue on possible ways to address the inadequate current levels of research funding and the shortcomings of grant peer review should continue. This commentary may help to further that dialogue.”

Our response to the Reviewer’s individual comments are given below. The changes in the manuscript are indicated by Track Changes.

1. The 20% success rate for NIH grants is a misleading indicator of the current difficulty in obtaining research funding, as success rates count original and resubmitted grants that are received during the same fiscal year as a single submission and combine different types of grants that have variable success rates. A better indicator would be paylines, which are at 10-12% for R01 applications from established investigators at many institutes. Thus, the situation is even more dire than the commentary suggests.

We have modified the Introduction to indicate this important point.

2. There is some redundancy in the list of “major problems,” which essentially boil down to: (a) peer review is biased and unreliable; (b) the current funding system discourages innovation and long-term projects; (c) too much time is wasted on seeking funding or peer review instead of on research. Another concern, which is not mentioned, is that insufficient levels of public funding for research and development can have a disproportionate impact on female and underrepresented minority researchers

We agree with the reviewer’s summary of the major problems with peer review; but we would prefer to retain the detailed exposé of peer review in our article. Regarding the disproportionate impact on female and minority resesarchers, we believe that our proposed system of grant allocation will alleviate the problem of recognition of the names of applicants by review panelists and the consequent bias against female and minority scientists. We have added a statement to this effect in the revised manuscript.

3. The commentary fails to consider potential disadvantages of the proposed funding strategy. For example, dramatically lowering the bar for research funding might result in an increase in lower quality research, because many labs whose work was not previously found to be sufficiently meritorious to warrant support would now be funded. Some institutions might be tempted to expand the size of their faculties in order to exploit the system, which would increase pressure on the research budget. Half of the extramural budget would be retained for competitive applications involving projects that require larger budgets, but this would essentially intensify competition among larger laboratories, which include some of the most productive research groups. The $400,000 per year budget falls far short of the $750,000 direct funding level that was found to be associated with peak measures of productivity in a 2006 NIH study 7 , and research costs have only risen since that time, suggesting that most labs would be underfunded under the new scheme.

We have added a section in the manuscript that discusses the unintended consequences of the proposed funding strategy, as also requested by Reviewer 2. Lowering the bar for research funding is highly unlikely to result in lower quality research, considering that a project that, for example, scores 10 points lower than the score of the funded grants is not expected to be of that much lower scientific quality. In support of this contention is the current inclination of some funding agencies to use a lottery to fund the grant applications in the top tiers.

The reviewer in essence admits that large laboratories obtain the large grants, leaving individual, creative scientists (who may not be good academic politicians), without funds. In the revised manuscript, we have cited the recent case of Nobel Laureate Dr. Katalin Karikó, who could not obtain funding during the period of her groundbreaking work on modified mRNA. The alternative system we are proposing would enable scientists like Dr. Karikó to do their work without having to become a junior member of a large lab.

The study reported in Wadman (2010), cited by the reviewer, regarding the high productivity of a lab with funding of $750,000 in direct costs is very misleading. The data presented in the article indicate that the productivity index of a lab with $400,000 is 6, whereas that of a lab with $750,000 is about 7.5. When expressed in terms of productivity per 1000 dollars, however, the $400,000 lab has a productivity index/1000 dollars of 0.015, compared to 0.01 for the lab with higher funding. This is 50% higher productivity per dollar!

4. The comparison of recipients of support from the NIH and Howard Hughes Medical Institute is invalid, because HHMI investigators represent a more highly selected subset of scientists with demonstrated high levels of productivity.

We cite this example only to illustrate the advantages of long-term funding, which allows for long-term and risky projects that may not produce results in a 3-year cycle. It is essential for policy-makers to implement such alternatives that will potentially produce results with high impact on understanding and curing diseases.

5. The proposed funding approach shares some features with the approach proposed by Pagano 1 , which should be cited.

We have cited Pagano (2006) regarding the impracticality of describing experiments to be done in 5 years, and the estimated total cost of writing and reviewing grant applications.

6. A single paragraph suggests performing a 5- and 10-year formal study to compare the productivity resulting from the current and alternative funding approaches but fails to consider the many methodological challenges in performing such a study.

The reviewer mentions the many methodological challenges in comparing the current and alternative funding approaches, but does not provide any examples. We propose that the international panel that sets the criteria for funding in the alternative system would use similar criteria, but weighted differently to also include patents, significant discoveries with high impact and developed therapeutics. We have also indicated provisions for center grants focusing on a particular disease.

7. The commentary repeatedly refers to “we,” but only a single author is listed.

The use of “we” is common practice in scientific writing. We have also cleared this with the Editorial Team at F1000.