Potentially Tainted Alzheimer's Research Exposes Flaws in NIH Funding Process
The incentive structures for both researchers and the NIH deserve greater scrutiny
Last week, a bombshell feature in Science magazine showed that the foundation for much of the funding for a particular theory of Alzheimer’s disease research may have been fraudulent. The entire article is worth the time to read but here is a summary (without going into the technical details of the disease because I’m not an expert on the disease):
There are several competing hypotheses on the causes of Alzheimer’s disease, the devastating dementia that afflicts tens of millions around the globe, but the most prominent is the amyloid hypothesis. Despite billions in research funds from the National Institutes of Health (NIH) and pharmaceutical drug companies, many drugs and therapeutics developed based on this hypothesis have failed to demonstrate any level of efficacy. As of February 2022, the only one of 15 drugs that made it out of phase III trials by demonstrating any efficacy, and therefore approved by the FDA, generated so much controversy that a number of FDA advisory committee members resigned in protest that the data didn’t support the efficacy decision to approve it. That the amyloid hypothesis is the dominant among many, and therefore being the object of affection of NIH, is due mostly to an influential 2006 paper that was led by two (and six other authors) respected scientists now based at the University of Minnesota—Karen Ashe and Sylvian Lesne—who “discovered” a subtype of plaques that caused dementia in rats. Since then, Alzheimer’s research along this line has received significant NIH funding to the chagrin of proponents of other theories. Between 2006 and 2021, NIH funding increased from zero to over a quarter of a billion dollars.
Ashe and Lesne’s discovery, published 2006 in Nature, has been very influential to say the least. It is the fifth most cited paper on basic Alzheimer’s research, and its authors have won awards. Concerned about the approval of drugs of disputed efficacy, another scientist also contracted by a legal firm, Matthew Shragg, and two other independent researchers contracted by Science, found evidence of results tampering, not just in the 2006 Nature paper but in several other papers by Lesne and Ashe.
The University of Minnesota is investigating, and so I will be cautious and not conclude that there is fraud. However, I want to focus on NIH’s grant-making process. NIH is the one agency that both major political parties compete to fund. The agency is flush with cash, with more than $41 billion in discretionary appropriations in 2021 alone. In his excellent book, the Innovator’s Prescription, Clayton Christensen and his co-authors address a problem of the NIH’s review process (pp 371-373): the tendency for reviewers of proposed research to favor those with which they are most familiar or extend the knowledge within their domain. The side effect is that proposals that are outside the domain of reviewers may not receive funding. In the Alzheimer’s story, several researchers who focus on other hypotheses complain about being neglected. A timely illustration: In May 2022, Lesne received a major NIH grant, and guess who the grant officer was? One of the six co-authors of the 2006 study. This isn’t to allege fraud or even inappropriate influence, but it shows the extent to which the granting process revolves around a narrow group of researchers on any topic. As siloed research domains keep getting more narrow and highly specialized, only a few researchers have the depth of knowledge required to effectively review a research proposal, and as scholars of innovation have noted, breakthroughs often come from “unconventional intersections” of scientific disciplines.
It is quite remarkable that despite billions in research funding for Alzheimer’s disease there’s been hardly any breakthrough. We may not fully understand the reason but judging from the Science article, we need to review the incentive structures for both researchers and the NIH.