Thank you everybody. It's absolutely a delight to be here today. I'm still in awe of the work of the Center for Evidence-Based Medicine. When I got the invitation to come, I figured who can give up the chance to talk about epistemology? That's the most important topic that moves our world, of course. I don't want to harp on this too much, but I think this sets the stage for what the stakes are. What we saw during the COVID pandemic was an absolute fight primarily over epistemology. The question was: what standards should we use to decide — should we recommend masking at scale? Should we mandate a vaccine for the entire population? Should we recommend closing schools, lockdowns, all of that? Each fight was over a particular topic, but ultimately, fundamentally at heart, it was a fight over epistemology. How do we decide what's true? What are the standards we use to decide what's true? And who gets to decide what's true? That fight was at the heart of every single controversy during the COVID pandemic and it led to all kinds of issues.
Now, I'm the director now of the National Institute of Health, but hopefully for not terribly long. I'm also the acting director of the US CDC, the Center for Disease Control. And what I've seen firsthand is the tremendous pressure that comes to public health and to governments when there are crises. The idea is that you must make decisions quickly or else people will suffer or die. And the push to have evidence standards that are below what I think the CEBM would say are right are tremendous. And so what I want to talk about today is that balance — that question of how do we maintain high evidence standards in the face of tremendous pressure to do something.
Okay. So I wanted to be as fair as I could, and I was subject to all kinds of controversies. I suspect many of the people in the room were during the COVID pandemic. But the question is: what strengths were there? And there were some, right? You saw the world scientific community mobilized to focus on essentially one thing in many different aspects. You saw huge collaborations across disparate fields. I learned about the Center and its work primarily through the COVID pandemic. I think I would never have learned of the Center at all given where I was in my research career. And I think this rapid collaboration has been quite productive.
And the ability to generate knowledge at scale. If you look at the literature, there's a COVID-ization of the scientific literature — the fraction of scientific papers focused on COVID just absolutely shot through. I think it's only just now starting to decrease in scale. All that I guess are strengths. But there's also huge vulnerabilities, right? There are problems with, as I said, getting reliable evidence under pressure. And I think let me just say another thing about this: there's also tremendous difficulty when people are scared. They're looking for answers from people in authority and power. But if you think about what science really is, it cannot be the case that people in authority and power get to decide what's true or false. The first scientific revolution was an inversion of epistemological power if it was anything. Instead of a few high ecclesiastical authorities getting to decide what's true or false, instead it's a few clever people with telescopes get to decide whether the moons of Jupiter move. And it's that democratization of the epistemological power that's at the heart of science.
What we saw during the COVID pandemic was that it turns out we're not that different than the situation was before the first scientific revolution. A relatively few people with a tremendous amount of authority and power do get to decide what's true. Second, I think there are tremendous difficulties in communicating uncertainty. I would routinely read the scientific literature, say to myself, okay, well, this evidence isn't very strong. I was trained in these — I'll show you the pyramid in just a second. Everyone's seen the pyramid. It's the heart and soul of the evidence pyramid. I was trained in that. And I thought to myself, well, boy, we really don't know. And yet I would see authorities go on TV and just with absolute conviction say, you must keep your kids out of school or else everybody will die.
You must mask your toddlers. The UK didn't have masking toddlers, did you? The US — we recommended masking kids as young as two years old. So I looked for the literature on masking toddlers. It turns out there was none. Anyone mask their toddler? Hopefully I'm not insulting anybody in the room. But the point is that the literature didn't recommend it and yet it was adopted at scale in the United States. Why was that? Again, it was an epistemological pressure. Someone needed to decide what was true. Does masking toddlers work? Is it worth doing? Authority got to decide that even as the literature was very weak on the subject, and then the debate over that got censored.
Updating conclusions as evidence evolved — that turned out to be quite difficult. A huge amount of evidence came out very early that closing schools wasn't particularly necessary and yet schools stayed closed in my native California for years. My kids didn't see the inside of a classroom for a year and a half. And then finally, maintaining public trust — the collapse of public trust in science and medicine has been catastrophic to watch. Especially in medicine. I think in the United States there are surveys published in places like the Journal of the American Medical Association suggesting that 60% or more of people do not trust their own doctor. And in science, in 2024 there was a Pew study that found that one in four Americans do not believe that scientists are working for the public interest.
Science can't survive that. Science needs the public's trust in order that the public continues to support science. Okay, let me give you one other measure of this, and that is life expectancy. The mission of the National Institute of Health is to support research that produces health and extends life. If you look at the trend in life expectancy — now, your eyes are automatically drawn to the right of the graph and you're looking at the collapse in life expectancy in the US versus Sweden. I am not here to talk about that. I want you to focus instead on the flatlining since 2010 in the United States in life expectancy.
Why is that? How is it that despite tremendous technological and scientific advances — and legitimately scientific advances — that happened since 2010, those advances have not translated over to better health and a longer life for the people of the United States? And even now we're only back up to 2019 levels. That should give us pause. The trust that people have in scientists — the US budget for the National Institutes of Health is almost $50 billion, that's with a B, that's a lot of money. How has that translated over to better living for people? Because that's what they've task'd us scientists with. What you have instead is a flatlining of life expectancy.
Now, when I present this graph, I often get the pushback to say, look, life expectancy is a complicated outcome. It is the outcome of not just medicine but of a tremendous number of other demographic factors and a whole host of social factors, which is absolutely true. I have a demographer hat that I could wear. I did once upon a time wear. And it's true. But if medicines and the investments in science that we have are not helping address the life expectancy increase in the context of the societies as we have them, then what is their purpose?
Now, I put this graph up as a challenge, but I want to emphasize I do fundamentally believe in the power of science to improve lives, and I think everyone in the room believes that. The question is: how can we do that? How can we actually have those advances improve people's health and lives? There are a few lessons I want to emphasize in this talk that I think are key to why we have not seen the translation of scientific advances to improve and extend life. First, we have to understand baseline health and we have to take baseline health seriously. One of the reasons why the US did so much worse than Sweden is that the United States is among the least metabolically healthy countries in the world. If you have a population that is metabolically unhealthy, when faced with infectious disease threats, especially new ones, you're going to do poorly because there's a connection between broad-based health and infectious disease. Infectious disease doesn't just have one infection fatality rate. It's an infection fatality rate that depends on who you are, what your circumstances are, how old you are, and a whole host of other things.
Second, you have to earn trust before a crisis. The trust collapses during a crisis because the evidence standards to which the scientific community holds itself before a crisis are not high enough. This is why I think the work of the Center for Evidence-Based Medicine is so important. Third, dissent improves science. Dissent improves science. It became very difficult to voice ideas that were contrary to the narrative that was promulgated during the COVID pandemic.
So I'll just give you one example. I mentioned masking toddlers. At one point during the pandemic, I was on a public roundtable with the governor of Florida who asked me what evidence there was about masking toddlers. And I told him that there's — you can laugh, this actually happened — I told him, this is in 2021, with my voice shaking, sir, there is no good evidence. There's no evidence at all. There's not a single study I found about the effect of masking toddlers on preventing the spread of COVID, which is true. I think still to this day there's zero evidence, or at least no evidence anywhere near the top of the pyramid. And that comment then resulted in the video of me saying that being removed from YouTube and censored at scale.
How does that help science and trust in science to not allow people to say no? You have to have the ability, because very often in science we're just wrong. The typical experience I have of being a scientist is that my first idea is always wrong. I get my students telling me I'm wrong all the time, which is an absolutely annoying thing, but also very wonderful at the same time. And it's by allowing a culture of dissent, where instead of destroying the careers and reputations of people who say no — especially on the most important issues — that you allow science to advance. That was the heart and soul of the first scientific revolution: that you allow dissent when there are people who say no in the face of high scientific authority.
And so the problem here was made quite clear during the pandemic. Fourth, transparency strengthens institutions. I've seen firsthand the incentives and the instinct to hide data and conclusions and discussions when you are in the halls of power. And yet, if you do that, what you'll have is a public that doesn't trust the conclusions that come out of those institutions. You have to be willing to put forward data and ideas, allow the conversations to happen, especially if you're in power, that run contrary to the ideas of the institutions that you lead. And those institutions must be willing to change course in the context of those discussions if it turns out that we're wrong. Really very difficult to do.
Finally, we have to be prepared in a way that relies on strong foundations of evidence, and this is why the CEBM is so important. To maintain the highest standards, because it's very tempting — a tremendous amount of money depends on not maintaining the highest standards of evidence. A tremendous amount of power depends on not maintaining the highest standards of evidence. And it's very difficult, because when you stand against history saying, no, I need evidence from the top of the pyramid, you're going to be subject to tremendous pressure to recant. So you have to have structures in place that say, look, there's lots and lots of data out there, but if they're not the highest quality, if it's not the highest standards for scientific reality, then I'm not going to believe it. Even if you maintain the highest standards, you're going to have debates, tremendous disagreements. But you have to maintain the highest standards before you decide that you're going to recommend something at scale to the population as a whole.
So what are we doing at the NIH to meet these challenges? First, you've heard all about the Make America Healthy Again movement. It's the most obvious thing. We have populations that are flatlining in life expectancy, have chronic disease problems like type two diabetes, heart disease still killing at scale, cancer, Alzheimer's disease — all of these are essentially chronic conditions. And so we are focusing on the chronic disease epidemic, which I think is quite even more scary than the COVID pandemic was, if you think about the actual threats to human health. We absolutely have to make America healthy again, and by doing so, hopefully make the whole world healthy again.
Second, we're supporting standards of gold-standard science. A lot of that — if you've seen me, God hope you haven't seen me too much, but every talk I give, I talk about replication as a standard. The reason why I talk about replication as a standard of truth is again this epistemological fight. It's very important if we're going to believe some scientific idea that that scientific idea be seconded and thirded — is that fourthed? Is that a word? — by other independent scientists who look at the same topic, maybe come at it from a different angle, and tend to find the same thing. And yet it is very difficult in science to make a living based on doing replication work. Even if you do high-quality replication work for a living, you'll find it difficult to get your work published in top journals. You'll find it difficult to get large scientific grants. And many universities will look at your work and say, well, it's not original, and so you won't get tenure.
That by itself is an indicator that we have a scientific culture that doesn't value the right things. Because if you're going to work on ideas that are at the bleeding edge, which we should as scientists, at the same time you should have a scientific culture that values replication as a really important, fundamental part of scientific culture — that rewards it, that gives it the same kinds of honor that it gives to new discovery. Because new discovery without replication is a recipe for disaster. A large portion of the new ideas that we have as scientists are wrong. It's only in the collaboration with the scientific community at large, looking and checking our ideas in a collaborative act, that we have the chance of identifying which of those ideas are true and which of those ideas are useful.
So that's what gold-standard science is about. Third, we have to protect free speech and scientific dissent. I've gone on about that for a while, but we can talk about specifically how that should happen. I think the key thing is to democratize the centers of power. You cannot have a few journals deciding what's true or false for the entirety of the scientific community. You have to have the scientific community together deciding what's true or false. We'll talk about exactly how we're thinking about implementing those ideas. You also have to not go to social media companies and give them blacklists and tell them to censor people unless they go out of business. You have to stop doing that. The US government has stopped doing that, although it did that during the pandemic.
Fourth, you have to enable cutting-edge, high-risk, high-reward research. The portfolios that the NIH has had have been actually in many ways too conservative, focused in recent years on safe ideas that are certain to produce papers in top journals, but not necessarily at the bleeding edge. And I'll show you some evidence that it wasn't always this way. And finally, you have to restore public trust through transparency. You have to be absolutely clear about when you have negative results, for instance — that the hypothesis that you had, that you thought was true, is not true. And you have to encourage scientists and reward scientists to do that.
So chronic disease — I've talked so much about this already, so I won't talk so much more about it. But you have to support prevention, treatment, and implementation science. Let me just, as an example of this, talk about the HIV epidemic. The HIV epidemic has been going on now for 50-some years, 40, 50 years. But we are now finally at a point where we can potentially end it. We have in advance a drug and set of drugs and strategies that, with a single injection, you're not going to get HIV for — even if you're exposed. We have a suite of drugs that, if you do have HIV, can reduce your viral load to near zero. Those two together are enough to end the HIV epidemic if we just make a concerted effort to understand where we need to intervene.
Even without an HIV vaccine, we can end the HIV epidemic. And so that's something we're focusing on. Implementation science is not the traditional way we invest in HIV, but yet it's the right thing to do given the technological opportunity. Let's take the technologies we have and use them to improve population health. We're focused on repurposed drugs. There's the potential for using existing products that are potentially off-patent to prevent or treat other illnesses that it is not currently used for. You need still very high evidence standards in those cases, but you also have to decide to invest in them. Typically drug companies have very little incentive to invest in that kind of research because you're talking about drugs that are off-patent. So at the NIH, we're doubling down on our investments in relatively inexpensive products to see if they can address some of the key chronic health problems — and not just drugs, also lifestyle interventions and a whole host of other interventions, an all-of-the-above approach to deal with the chronic disease crisis.
It's not going to be a single solution. This is a complicated problem that has a lot to do not just with the medications that we have, but how our societies are organized, the incentives that we have for eating the wrong things. All of that is really important. Some of the things that you've seen coming out of the United States may seem like small things — food dyes in M&Ms, right? That seems like a small thing. But the thing about that is that it's in the accumulation of small things that you change society. If you have high standards for everything, you can finally address large problems that are complicated and multi-dimensional.
Here's our favorite pyramid. I wanted to emphasize in this pyramid the reproducible science. Meta-analysis is about reproducible science for evaluating whether science is reproducible, right? That's why it belongs at the top of that pyramid. And yet we at the NIH I think underinvest in replication work, in reproducibility work. That's changed under my leadership. We're going to have large grants available for scientists who do replication work. I'll tell you the key problem I'm trying to address. The key problem here is that you have a culture that views replication as a threat. If someone comes to you and says, please, Jay, can you share the data that you have with me? I want to see if I find the same thing as you — I'm going to view that as a threat to my career.
Right? If Carl Heneghan is looking at your work, you're going to go, oh my god, what if he disagrees with me? The fact that we view replication as a threat when our work is subject to replication indicates a sick scientific culture. Because in reality, if somebody comes to you — a scientist comes to you and says, I want to see if your ideas are right, I want to check them — well, that's a high honor for you as a scientist. That means your ideas are worth the attention of other scientists. When Carl Heneghan says to you, I want to see if your work replicates, you should be very happy because that means your work is important enough to attract the attention of Carl. We have to flip the way that we evaluate what's valuable in science. It shouldn't be: I get a paper published in Science or Nature and therefore I'm a good scientist. What it should be is: I have a really important idea, other scientists find the same thing as me, and it's valuable to people.
The key thing then is we have to change the culture of science so that we value replication. You can have scientific careers where you advance, you're appointed tenure because you are good at doing replication work. It'd be amazing to have a thousand John Ionnidis walking around, a thousand Carl Heneghans and Tom Jeffersons walking around, where people want the attention of those scientists to see that their ideas are actually correct. And when we have an idea that turns out to be incorrect, we punish our scientists too much. We need to give people second chances. Kind of like they do in Silicon Valley, where if you found a company and it doesn't work out, it's a constructive failure — you may get a second chance.
So what we're doing — I'll talk more about this in the Q&A — but we're going to fund research in replication. We're establishing a journal at the NIH where you can publish your replication work. The idea of that journal is, unlike many other journals, it's going to be minimally gatekept. You're not going to be fighting with Referee 3 in order to get the paper published, but it will be peer-reviewed in a transparent way. So every paper in that journal will have peer reviewers, including the person being replicated, with a chance to do a peer review of the replication work, and then all of the peer reviews published alongside the paper.
And then third, what we're going to do is, when you search in PubMed — where you're looking for papers, everyone in the room knows what PubMed is in this audience. It's the way that we organize all of the published biomedical literature. We'll have a replication button. You click the replication button and there will be a knowledge graph that will pop up that will give you all of the related papers for every single idea that's in the paper. And there'll be a little AI summary for each of those ideas — what the rest of the literature says — and then you can click on the papers that are cited by the AI summary to decide for yourself whether you agree with the AI. And then we will take feedback to improve the system based on feedback from scientists in detail.
The idea is to elevate the reading of the papers and deciding if you agree over the abstract and where the paper is published for deciding whether it's true or not. By having the replication button, you'll be looking not so much at is it published in the New England Journal of Medicine to decide whether it's true, but what does the rest of the literature say, whether it's published in the New England Journal or not. And the goal is to make the data more easily available for researchers at scale. Instead of going to researchers and begging for the underlying data, you will give the opportunity for researchers to have the underlying data published alongside the papers so people can check for themselves.
Finally, we're going to be investing in a set series of metrics. Does everyone know what an H-index is? Okay, so think about an H-index — like what does it actually incentivize? It's the basis for getting tenure, the basis for promotions, the basis for so much in science. But what it elevates is it elevates your influence in scientific society, measured by citations, and it elevates your volume — how many papers you've published. Are those really the only metrics by which you decide whether a scientist is a good scientist? Absolutely not. Right? And yet that's the metric we use for deciding so much in science. What we have is a paucity of metrics for deciding what a good scientist is. It should include among the things that make you a good scientist — maybe how often do you try new ideas out in your work. How often do you work on replication studies? Do you share your data? Do you share your code? Do you share your tissue samples? How often is your work subject to replication? That should be a positive measure.
All of these are things that are measurable and we'll start to measure them at scale. The idea is to change the culture of science around the evidence standards that we want for scientists and make scientists want to do the things that make the scientific culture healthy rather than the unhealthy scientific culture we currently have. We have to protect scientific dissent. I will never participate in a devastating takedown of another scientist. We do have to maintain standards to make sure that there isn't scientific fraud in our funding decisions, but we have to not attack scientists simply because we disagree with their ideas.
I think that's going to be a big change in the scientific culture that happened during COVID, because I'll tell you firsthand that if you stuck your head up and said, look, no, I don't agree with some idea or other, it was very easy to get it chopped off. Okay. Fourth, we have to support high-risk, high-reward research. Let me show you a graph from a paper I wrote a few years ago. Let me describe this paper before you look at the graph. So here's what we did. The question was: how new are the ideas in NIH-funded research? How new were the ideas? So what we did is we looked at every single paper published in 1940 in PubMed, took all of the words and word phrases in them and collected them. Then we did the same thing in 1941 and subtracted off the words from 1940. And what you're left with is the new words that were introduced in 1941. But those new words represent the new ideas that were introduced in 1941. You do the same thing for 42, 43, 44, 45, and so on. And you take out all the synonyms and what you're left with is a history of medicine in the words that are in the published biomedical literature.
So PCR — polymerase chain reaction — shows up in 1982. I think small interfering RNA shows up in 2001. You have a history of biomedicine embedded in the words in the published biomedical literature. And you go back to the papers themselves and ask: how old, how new are the newest ideas in every single paper? So a paper this year that's published that has small interfering RNA as its newest idea is relying on old ideas — 25-year-old ideas. Then you go back and ask: what fraction of the papers that have zero-year-old ideas as the newest ideas, one-year-old ideas, two-year-old ideas — that's the x-axis — what fraction of them are NIH funded? The green dots are the papers that were published between 1990 and 1999 in PubMed. And what you can see is that more than about 55% of the papers on the left side of the graph that had zero-year-old ideas in them had NIH funding. One-year-old ideas — a paper that had a one-year-old idea — about more than half of those were NIH funded, and so on. Correctly, appropriately, the NIH should be funding papers at the bleeding edge. So you see a monotonic decline as the age of the newest idea in every paper goes to the right — the fraction of NIH funding declines.
If you look at papers that were published in 2000 to 2009, on the left side of the graph what you see is a collapse or a decrease in support for the newest ideas. This correlates with the graying and aging of the scientific workforce, but also with a culture of science that essentially punished being wrong with the newest ideas. We're working to change this. And the key idea is something called the unified funding strategy. The idea there is instead of — so we still have peer review as the core of... I'll just tell you my main job. My main job is I have 100,000 applications a year for money, and I got to pick 10,000 of them. And I have a staff of 20,000 people, many of whom are brilliant scientists, to help me pick the 10,000 out of the 100,000. Because it turns out it's not possible — unless you're Tom Jefferson — to read 100,000 papers. So Tom, if you want to come work for me, the offer is still open. So you have 100,000 applications. So the way that many of our institutes — what would happen is we have peer review where three reviewers read...
every single application. Three different reviewers read every single application. They give them a score. In fact they give them a suite of scores. You give them a score based on methods, a score based on how innovative the idea is, and then an overall score. It turns out that the overall score correlates very strongly with the method score, but much less strongly with the innovation score. Many innovative ideas are written by early career scientists who don't have the grantsmanship skills, but they're really neat ideas. And they end up with worse overall scores even though the idea is quite innovative. One of the methods used to decide which 10,000 to fund in some of the institutes has been a payline where you take a look at the money that you have, you rank all the applications, you do a cutoff and say we're going to fund everything above a certain payline. But you can see what happens is that you get proposals that are certain to produce really high impact papers but you leave a lot of really innovative ideas down below the payline and don't fund them. The unified funding strategy addresses that. It allows the institute directors to have the leeway to fund a portfolio of projects where success is measured not by has every single project produced papers in Science and Nature but rather by does the portfolio as a whole advance science? Does the portfolio as a whole advance human health? And so that means that if I fund a portfolio of 50 projects and 40 of them fail or 49 of them fail and the 50th cures type two diabetes, well that's a successful portfolio. It's a different way of measuring success of a portfolio rather than you published a paper in a top journal. It's did you advance human health? Do you fundamentally transform your field and knowledge of your field? And the failed projects will be encouraged to report what went right, what went wrong. There'll be constructive failures and people will get credit for that. So we punish failure less. It's a little more of a Silicon Valley approach to deciding what the portfolio is going to look like.
Okay, finally, restore public trust. We're going to focus on transparency about methods, transparency about findings, and transparency about uncertainty. It's very difficult, I'll tell you. If you are in a position like I am to say I don't know — and you saw this during the pandemic and I'm doing my level best to do that. I will do that, I'm sure, in response to many questions you have to me during the Q&A session. But it's, I think, absolutely vital for public health leaders to maintain that epistemic humility because there's a lot we don't know about in science with any degree of certainty. And the epistemological challenge: if we have high standards of evidence required for us to say look I know the truth about a topic, that means many many times we have to say I don't know. When it comes to fundamental questions about should I take this protein product, should I take this vitamin, should I take this drug — often the answer is I don't know.
Okay. I'm going to just end with one case study and that has to do with pandemic preparedness. I started with COVID. There's an ecosystem whereby we have created a situation where a pandemic happens, we're going to react with lowering our evidence standard. And the ecosystem essentially has this structure: the idea is that we can prevent pandemics by going out into the wild places, finding all the pathogens that are in bats or whatever, bringing them into labs — many of which don't actually currently infect humans at all and will never infect any humans — altering them or investigating them in ways that potentially are quite dangerous. And then depending on that investigation, you identify, you triage those pathogens and say, okay, here are some pathogens that have never infected a human that are more likely to infect humans if they evolve in certain ways in the future. And then develop countermeasures to those pathogens — countermeasures that cannot in principle be tested in any human because that pathogen has never infected a human. And then give contracts to stockpile those countermeasures at scale. This is the structure that I think of the idea of the last 25 years of how we prepare for new pathogens, disease X and so on. The idea is that we go find in the wild these potential pathogens that might make the leap into humans and then develop at scale a huge number of countermeasures that cannot in principle be tested with anywhere near the bottom of that evidence pyramid because they've never been used — because the pathogens have never infected a human.
Then just take one step further. If and when the pathogen does make the leap into human populations — god forbid — that pathogen will very likely, given the vagaries of evolution, not look anything like the thing that we've developed countermeasures against. And yet, when it does make the leap, there'll be tremendous pressure to use those countermeasures. Tremendous pressure to use the vaccines or the antibodies or small molecules that we developed that we believe would be countermeasures but would never have really been tested in human populations. We saw this during the COVID pandemic. The COVID vaccines — there was tremendous pressure to use them even in populations where it was really clear that the marginal benefit probably didn't outweigh the marginal harm. Instead, the right way to deal with pathogens is first study pathogens that actually infect humans, hold the newly developed countermeasures to very very high evidence standards, and work on the chronic disease crisis. Because by addressing the chronic disease crisis you make a population more resilient when there's new pathogens that do emerge. That's the new idea for pandemic preparedness at the NIH.
The future of evidence-based medicine I think is very bright and I'll tell you why. People really want good answers. They're tired of ideas or products that are sold to them that don't work, that produce no real value but cost a lot of money. And so the future of evidence-based medicine: these standards are transparent where evidence can be examined, challenged, and verified. Conflict of interests are disclosed. Reproducible where reliability becomes the foundation of scientific progress. Open where there's scientific debate at the heart of how we decide what's true or false. And those debates, even if they're vicious, don't result in devastating takedowns that lead to people being unable to participate in future debates. Patient-focused where research priorities reflect the greatest health burdens. So the money should go where the disease burden is, but also to places where they're scientifically important. Rare diseases often produce advances in biological knowledge that benefit everybody. So we should be thinking very broadly about this. Adaptive where the institutions change, are much more humble when the evidence changes. And then finally trusted, where trust is earned through honesty, transparency, and accountability. I will end my talk there and I thought I'd go 20 minutes but I can't help myself. Thank you everybody.
