 Privileged to be part of this symposium today, so my thanks to Eric and Rudy and the LC folks for the opportunity to participate. Now LC I think probably is familiar to everybody, ethical, legal, social implications. This has been my interface with the genome project initially and institute subsequently and as folks may know, one of the most innovative aspects of the genome project was the decision early on to set aside 5% of the budget for research in this domain. This was important for Congress at the inception of the initiative and continues from my perspective to be a highly important aspect of this whole enterprise. I personally would like to see this to be modeled for other NIH institutes but probably a topic for another day and an era of more flush budgets perhaps. So I'm going to talk today about newborn screening. This has been the focus of my research in recent years. I'm not going to talk so much about our research but talk a little bit more about the public policy issues and how it may be that this technology that's emerging might be applied in this particular context. So we've talked already about the enormous progress over the last 10 years or so and it's been interesting to see the evolution in the discussion of ethical, legal and social issues over that period of time. I think fair to say a lot of discussion early on was fairly speculative and fairly high level about designer babies and those sorts of issues. Much more recently I think we've gotten down into the weeds with the investigators as technologies emerge to say, okay, let's think about this specific type of testing and how it might apply to different sorts of clinical contexts. So we're moving into a much more translational era. So here's my focus. Given the power of this technology to conduct whole genome sequencing, whole exome sequencing, how can this technology be best used to benefit children? Might this technology have an application in newborn screening as the specific focus here? A little bit different way of saying this is if full genome sequencing is the new hammer, does newborn screening look like a nail? So let me talk a little bit about these programs that may not be particularly familiar to everybody here. Every baby essentially or almost every baby in the country is subject to newborn screening. So four million babies per year. All states, districts, certainly all developed countries around the world are engaged in newborn screening. So this has become a popular and effective mode of approaching disease treatment and prevention. Single largest application of genetic testing. So obviously that invites the question of how might we use these new technologies in this particular context. Now interesting, this is the 50th anniversary of newborn screenings in the United States. Massachusetts started their mandatory program back in 1963. Big celebrations going on around the country this year about newborn screening to celebrate the success of these programs. Estimates vary, but probably somewhere around 1,200 children, affected children identified on a yearly basis through newborn screening. So the key here is early identification. So these kids, for the most part, will come to clinical attention at some point. The point is to detect them shortly after birth in order to intervene either through preventive or treatment mechanisms that would reduce the morbidity and sometimes mortality for these particular conditions. What we've seen over the last 15 years and particularly over the last five, six, seven years is an enormous increase in the number of conditions that are being targeted. So 10 years ago, all but four states were screening for only six conditions. Utah at that point was only screening for two. 2013, all states screened for more than 30 conditions. And I'll talk a little bit more about why it is that we've seen this transition in recent years. And it's to a significant extent a technology-driven change, more so than a change in the basic philosophy of the programs. So here's how the system works. Blood collector from heel sticks in newborns, preferably after 12 hours of age because you want certain metabolites to build up in the baby's system based on being ex-utero rather than in utero. You want to do it before six days of age because some of these conditions, and Galactosemia being a good example, can lead to rapid decline and sometimes death in these infants. So you have to get in there quickly. You've got to get the results back quickly. Generally sent to state labs, some commercial labs involved in this, but for most part a state lab enterprise. And results return to the hospital and the physician of record for the baby. And the attempt is to try to get these results back within two weeks for the reasons that I mentioned. If the program is not highly efficient with this sort of throughput, then your risk baby is dying before that information can be brought back to bear on their clinical care. Now states report anywhere between about 98-99% uptake in this system. So enormously effective and efficient in that regard. The cost, again, varies by state and of course these are state-based programs, so hallmark is variability. But generally in the ballpark about 80 to 110 dollars, which is generally in most states charged to families. Now actually that's the charge typically that's to the hospital. The hospital will then turn around, often charge, upcharge to the family. So the charge to the family is actually greater, but that's bundled within the care bill that folks get after they've had a new baby. So you can see that the cost is remarkably low for the size of the test and the complexity of the programs that we're talking about. Now here's a fundamental issue that we'll come back to later in the talk. All states now, except Wyoming and the District of Columbia, have a mandatory newborn screening program, state mandated. So parental permission is not necessary and this was a hallmark of the program from very early on and part of the aspect of moving to a state-based foundation for the programs. There had been debate. Academy of Pediatrics was nervous about these programs back in the early 60s when they first started, was felt we didn't know enough about diseases, but advocacy organizations were effective in moving the testing into the state environment. And the perception was that the benefits of the baby were substantial enough, particularly with the first condition PKU, to warrant a state mandate for testing. Now most states, 43, do permit parents to opt out for either religious or philosophical reasons. There are no religions that have anything specific to say about newborn screening. So Utah, for example, has a religious exemption but our state is unclear about how anybody would actually claim such an exemption based on a religious tradition. Of course what happens in the real world is parents may refuse, they may give a say it's contrary to our personal beliefs and those are generally respected. Our state at least has not gone to the mat with anybody to say demonstrate that you're part of a religious tradition that has newborn screening as an element. But a central aspect of this is that the ability to opt out is not effectively communicated to parents. And this is a background fundamental problem with the larger system. The level of education of parents is typically low. And we found this certainly with our studies with population groups across the country. Folks may be aware that a heel stuck was done, the baby will come back with a band-aid on the heel, they'll know something was done. But as we talk with families about the fact that babies had 30 different tests, they're stunned. They have no idea of the extent of these programs. So a brochure goes into the bag along with lots of other stuff when you go into the hospital to deliver a baby, folks may get to read that days, weeks later or perhaps typically not. And so the ability to actually opt out is not effectively communicated to folks leading to a high level of support, tacit sport at least for the programs, 98 to 99% uptake. So what would whole genome or exome sequencing look like in this context? What would be the purpose of using this technology? And I've outlined at least a couple potential choices here. Folks here may be able to think be more creative about possible applications here, but one would be primary screening tool for all newborns. Is that what we're talking about here is potentially implementing this technology within the existing state-based program infrastructure as a new way of approaching these sorts of applications? Or we could have primary screening tools, a commercial supplement to existing state programs. Parents could choose as they have the ability to do this as a supplement. And there have been a supplemental commercial newborn screening approaches available for quite a few years and you hear stories of that being a shower gift for a pregnant woman as an expanded newborn screening panel. So here I think would be a distinction that came out of a conference here at the NIH from a couple years ago now. The difference between newborn screening, which we typically think of as a state program versus screening of newborns, which would be presumably outside the existing policy structure that we've been talking about. Or this technology could be used as secondary testing of affected infants. And as you'll see with my comments, this is where I think the biggest bang for the buck is. Identify genetic variants that impact treatment or prognosis in affected children. These conditions are not well characterized. There's quite a bit of variability. Much more needs to be known about the genetic underpinnings of those that are genetic. And this may be a powerful line of research to better understand these kids. So one basic question is can you get enough blood out of a dried blood spot on a filter paper to do whole genome sequencing? And the answer is yes. It's challenging but certainly feasible. But note that the throughput issue here is important. Eric talked a little bit earlier about and we've come to an era in which whole genome sequencing can be done in two or three days. Here I think through what the implications would be for a state-based program, and I've illustrated here numbers for two of our larger states in the country, 700 births a day, 365 days a year. California 1500 births a day, 365 days a year. So we need to think through the challenges of how this technology could be developed to the point that that level of throughput would be remotely feasible. Here's some of the ethical challenges, though, and I've got a couple slides about challenges that these developments pose for us. Multiplex platforms offer rapid expansion programs beyond the evidence-based supporting efficacy. It's a little bit of a complicated sentence. But tandem mass spectroscopy was the key technical advance that enabled a rapid expansion within the last 10 years of the newborn screening panel. And the point is what you may be looking for, one to five conditions out of mass spectroscopy, which you get results on many more. What do you do then with those results? And tandem mass is a multiplex platform. I would say whole genome sequencing is a multiplex platform. As the American College of Medical Genetics, with their enormously influential report in 2006, advocating for a uniform panel, but advocating what for many of us became a surprisingly large number of conditions. Europe typically tests for half a dozen or so conditions, or at least many countries in Europe do. I think many of us were anticipating that the evidence review process that the ACMG was undertaking would come back with a similar sort of number. Say we've got solid evidence that half a dozen conditions, or eight or maybe 10, would justify inclusion on the panel. Well they came back advocating 29 conditions and 29 conditions and 25 secondary conditions. Now we have limited data on many of these conditions, and there's wide variability in how these are treated across the clinical community. So part of this was an assumption within the ACMG process that if you have the data you have an ethical obligation to disclose it. If there's something that you as a clinician or the clinical community thinks you can do for these babies, then despite the absence of carefully controlled child to demonstrate those benefits, there's an ethical obligation to disclose. So I think this is exactly the debate we're currently having now in the sequencing domain to say once those results emerge out of your analytic process, what's your obligation to disclose those to families? And I think the fact that you're generating those through this technology offers the opportunity to significantly expand the number of things you're targeting. If indeed your conclusion is you have an ethical obligation to disclose. My claim would be that the conditions were adopted as part of that process that would not be adopted with a condition specific review. I'll talk a little bit more in a second about how the current review process works. So a couple other challenges. Research on rare conditions faces serious obstacles. And so the reason we don't have good data isn't meant to be a criticism or the research community. It's meant to be an observation about how challenging this entire domain is. When you have some of these conditions that are 1 in 20,000 kids, 1 in 50,000 kids, some of them 1 in 100,000, how do you do research on those? Any one clinician may only have a couple of kids that they see as part of their life practice. So without an opportunity to pool children across the country and have uniform approaches to research on appropriate interventions, then our knowledge and confidence in this whole domain will continue to lag. And I would draw the parallel with how leukemia care was for kids 40 years ago. The decision was made to, with children's oncology group and those sorts of systems, to pool, to put kids on protocol as a collaborative effort. And the progress with the care of those kids has been enormous because you're not dealing with individual clinicians making individual level decisions about what they think might be best. So this would be a model that I would very much encourage us to think about in this context. In addition, newborn screening programs take care of the kids until they're plugged into a subspecialist. Most do not collect data after the kids leave that initial phase of their life. And so we don't have long-term data on how the kids do. How do we know which dietary interventions are working best for kids? How do we know how well the programs in general are working without that longer-term data collection? So these are state-based programs. And the beauty of being a state-based program at this point is that it's uniformly accessible for everybody. It's designed to provide this service to every child, regardless of their geographic location, economic background, insurance status, et cetera. So that's been a pillar of this particular process. But that does provide challenges for us in terms of the costs that might be absorbed as part of that program. We'll talk a little bit more about that. So as we talked, as I mentioned earlier, newborn screening is a mandatory in most states. Therefore, we have to have, according to the, if you're following the basic justification for mandatory program that it's so beneficial to kids that you can't allow parents a choice about this sort of thing, then you have to maintain that level of benefit in order to feel comfortable in mandating tests that you may be uncertain about the long-term benefit. So the scope of conditions to be targeted is a long-term debate in this domain. Dwayne Alexander, a former head of NICHD and Peter Van Dijk, HRSA, had wrote this paper in 2006, and this has been much discussed. And what they were suggesting was that the technology changes may lead us to the point where we are thinking about what conditions we won't put on a newborn screening panel. Their suggestion here was, well, maybe we wouldn't test newborns for Huntington's disease as a relatively extreme example, but that then opens the barn doors quite wide for almost anything else that would be less contentious than something like Huntington's disease in a newborn. So here's the problem. Wilson and Younger were the crafters some 50 years or so ago of the original criteria for population screening, not newborn screening per se, but just population screening in general. And they said this general idea of early disease-detected treatment is essentially simple. However, the path to a successful achievement is far from simple, although sometimes it may appear deceptively easy. The population, and I think clinicians in general, really like the notion of screening. It makes so much sense. Why don't we just look for these things early and do something about it? In practice, though, they're much more complicated. Here's a more recent contemporary synthesis of the criteria for population screening by Anderman, and I've just highlighted a couple here that are particularly relevant to this conversation. You should have some objectives. What's your objective? What are you looking for? Do you have scientific evidence of effectiveness? Have you informed choice about the screening modality? So meeting these sorts of criteria is an enormous challenge. In addition, you have to have a philosophic foundation about what it is that justifies your enterprise. And this was a statement from the Institute of Medicine about newborn screening from back 1994. Newborn screening should not be undertaken unless there's a clear, immediate benefit to the particular infant being screened. So this has been one poll, and this is what Dwayne Alexander and Peter Van Dyke were arguing against, to say let's move away from this big benefit to the baby criterion and think about benefits to other people. Now you can see in this circumstance where the baby may not benefit at all, let's say for a particular modality. But parents are informed that they're at risk for bearing a second child. And they're informed that it entirely mattered before they get pregnant again. Or you might say, what a tremendous benefit, and there's certainly literature out there that strongly support this. These parents go through a very difficult time frequently with these rare conditions in getting to a diagnosis. The so-called diagnostic odyssey is a significant burden to families. Newborn screening eliminates a diagnostic odyssey, benefits families in that regard, even in the absence of any particular benefit to the baby. So there's where the debate is focusing. Are those benefits to other people, specifically parents, of sufficient magnitude that we don't need benefits to babies any longer? Or do we want to hold fast to this particular principle? So wide variation from state to state on what was tested for, that's changed now with the ACMG statement. And in part what's happened, in large measure what's happened, is there's a new body there, Secretary's Advisory Committee for Heritable Diseases in Newborns and Children, of which I'm a member. And this is an evidence-based process to try to tackle this question. What ought to be part of the newborn screening panel? Much more rigorous process for condition-specific decisions to get on to the so-called recommended uniform screening panel. So it's condition-specific. You show me, and this is what the Secretary's Advisory Committee will say, show us the population-based studies that demonstrate the improved outcomes with early detection for this particular condition. You'll be happy to know I won't go through this in detail, but this is the rubric for decision-making for the Secretary's Advisory Committee. They look for benefit for the baby. They look for readiness for programs to implement the testing. They look for feasibility that has to do with the quality of the test, sensitivity, specificity, cost, and able to conduct it with throughput that would support a population-wide approach. So the general point is here, we've tried to get much more formal and specific about the process by which these conditions would be approved. So the point in the next couple of slides is just that the population screening is a real challenge. And again, despite that's intuitive attractiveness, in fact, making it work is a problem. And we have many conditions for which we don't either have good data to suggest efficacy or data to suggest that they don't work as well as we might hope. So this is the US Preventive Services Task Force set of recommendations. They don't do newborn screening, but they have commented in the past, reviewed in the past, PKU congenital hyperthyroidism sickle cell disease, which get a grade A. Newborn hearing screening, though, gets grade B. And other things like iron deficiency, insufficient evidence, lead, grade D. So you can see that things that would seem to make sense in general oftentimes are challenged to meet the demands of a rigorous evidence review process. In adults, similar sort of process here. Colon cancer, pap spare, hypertension, tobacco use, these are all things that US Preventive Services Task Force at least, thought, married to the grade A for their utility. But lots of debate about these things. PSA, screening grade D, mammography is grade B. This has not been informed by some of the more recent literature about mammography that has raised more questions about its efficacy. But things like osteoporosis and even history and behavioral counseling for cardiovascular disease, they gave a grade C. And this is not to say that this is all the right answers for these things, but just to highlight for you the challenges associated with population-based screening. So what are we screening for? We could be screening for established conditions already at a newborn screening panel, but potentially move to a DNA-based platform for that. Not clear, though, in many circumstances that genetic tests or DNA-based tests are more sensitive or specific than current test modalities. A good example is cystic fibrosis. We know a lot about the genetics of CF, more than 1,000 different mutations characterized. But IRT, I mean, our reactive trypsin, is still the initial screening test on newborn screening and the definitive test for kids identified is still the sweat test. So DNA-based analysis is part of the testing sequence, but it's not the definitive diagnosis or the triggering test being used. So there's questions about whether products of metabolism may continue to be better tools for screening than DNA-based tests. Now, there are some new ones coming along. SMA, spinal muscular astrophy, for example, looks like it's got a test with good sensitivity and specificity. Folks have been talking about fragile X newborn screening for a while. So it's certainly possible that a variety of tests that may be emerging over the coming years will be DNA-based analyses, but these are probably going to be conducted as targeted test as opposed to a whole genome or whole exome sequencing approach. So there's no question that this approach would enable a large expansion of conditions targeted. The ACMG, just in the last month or two, has come out with a statement that provides some significant analysis of conditions which they consider to be actionable and that ought to be part of every sequence conducted at laboratories conducting this type of work. 57 genes and 24 different conditions are on this list. So that presumably would be part of a potential panel for newborn screening. Carrier states, of course, are ubiquitous. Cancer syndromes that are also part of the ACMG list, but just to highlight these, these may be adult or pediatric onset cancer syndromes. And then, of course, there's a whole host of things that are part of the emerging landscape out there, the 23andMe sort of testing results where folks can get putative results in what their relative risk is for a whole host of conditions like diabetes or heart disease, Alzheimer's, earwax, et cetera. So all of these things could potentially be part of this panel if this is the type of data that's being generated out of the newborn period of time. So we look specifically then at the ACMG list of 57 genes, 24 conditions as a initial target subset here. Estimates that about 1% of those individuals will have positive findings. Back at the envelope calculation, four million babies born per year. Now that's 40,000 kids with positive results. That's three to four times the current rate of true positive results within the system. And obviously that number would be substantially larger if we're talking about carrier states and other conditions. We have no infrastructure to deal with anything close to that volume. Primary care providers struggle with the newborn screening system now because they don't understand the diseases very well and they're heavily tied on the state programs and local experts to help manage these kids. Primary care providers, I can tell you, certainly aren't prepared to deal with this flood of information. And we would question whether we have the subspecialty population to genetic counselors and geneticists to be able to deal with this. Now if it's worthwhile, of course, we'll develop those resources. But at the present time, they don't exist. So here's another ethical angle here that's been part of the recent debate. This type of information would generate information on adult onset conditions in a newborn. Now the new report out of the ACMG recommends reporting these to parents for the parental benefit. So if you find a BRCA1 and a baby, that's not immediately relevant to that baby. It may well be in 30 years. But it's relevant to one of the parents who's, you're pretty sure it's BRCA1 mutation carrier. So you would report the results out on the baby because the parents might benefit from that information. This is a highly contentious element of that recent report because the tradition has mostly been not to screen babies for adult onset conditions. You screen kids for things you can do during childhood. And this is a way of promoting the autonomy of the child and concerned about uncertain psychological impacts of risk information for both the child and the parent child relationship. So this is a matter of much more debate going forward about what our ethical obligations might be for conditions of this sort that are detected in that period of time. So burdens of false positives. Most important adverse consequence population screening. We know this parental anxiety associated with this. Cost to follow up testing is one of the biggest elements of the testing programs. Finding folks, bringing them back in for confirmatory testing. The positive predictive value for most of the current tests is one to 40%, meaning the majority of positive results are false positive results. And as mentioned, we know from a long line of literature that about 10 to 20% or so of parents remain concerned about the health of their baby. The docs told them everything's okay, don't worry. Well, they do worry. And anybody who's been a parent and maybe anybody who's been a kid knows that this is a sensitive time of life. And if you're told something's wrong with your baby, that's a big deal. At least for a subset of parents who have trouble getting over that information. And ambiguous results. We also know that oftentimes folks can deal pretty well with adverse results, but they have somewhat more trouble in many circumstances dealing with ambiguous results. Hard to marshal your defense forces when you don't know what the status of your result truly is. So these ambiguous results are with us for a period of time with this type of sequencing and would have to be decided about whether you would return these results to begin with and how you would help folks deal with this information. Cost issues. Kip fee is for new board screening, as I mentioned, or about $100 or so. Charged from the birth facility. Incremental charges for new tests are often in a two to five dollar range per new board. And we have an annual discussion with our legislature about whether that is tests are warranted or not. Even though the cost is neutral from a state's perspective, they approve the Kip fee increase every year. And so if we think just on back of the envelope context, $1,000 genome will assume that cost comes down, but some of the costs associated with that won't come down. Data analysis is a human labor initiative. Family notification and follow up cost for confirmatory testing. These sorts of things, of course, will exist. So again, very simplistic sort of calculation here for thinking at a day when all of those costs might be bundled into $1,000. That's $4 billion per year for sequencing. Spending $4 billion more for kids is a great idea, but we would question whether this is the most cost effective way of spending that amount of money for the welfare of children. So couple of quick conclusions then. Current newborn screening system is highly affected for some conditions, but struggles with funding, uncertain benefits for other conditions and lack of adequate research. Population screening is notoriously complex and relatively few instances of highly effective population screening programs. So we have to have the evidence that this approach will be beneficial to babies. So whole genome exome sequencing as a primary screening tool, I think would fundamentally change the philosophy of the programs, increased cost, increased burdens of false or ambiguous results, and confer uncertain benefits without a much more robust system to conduct the research in longer term follow up. So given these additional burdens and uncertain benefits, sequencing and newborn screening would not be justified under the state mandate. So I wanna come back to that foundational aspect. These are mandated systems and you're very hard pressed to say the state's gonna mandate whole genome sequencing for a baby without clear evidence that at this point is a compelling benefit to the baby. So implementation of informed consent process would be necessary. This could be conducted as a commercial supplement without much difficulty from that perspective at least. And you could potentially get prenatal consent for this. Now we currently have a project, the NIH funded project to do prenatal education about newborn screening. I can tell you we're working hard with our OB colleagues to get 10 minutes of time to talk about this issue. And so talk about newborn screening in general, much less what might be necessary for a whole genome sequencing discussion. So Louis Thomas was a wonderful writer about biological issues and he talked about technologies at several levels. Non-technologies, counseling, hand holding, support, decisive technologies, which he characterized as his prime example there was immunization, elegant and effective. And a most creative aspect was this notion of halfway technologies, things that look sophisticated that are sophisticated at a certain level but still are founded on a basic misunderstanding or lack of understanding yet about the foundational aspects of human biology for which the technology is directed. So to a certain extent, I think the whole genome sequencing now while highly effective and beautifully designed in many respects still is founded on a basic lack of understanding about how to translate this into effective interventions for kids. So from my perspective, whole genome exome sequencing makes little sense as a primary screening tool but makes enormous sense as a research tool to better understand the genetics of a host of important complex, uncommon conditions. We need to better understand these diseases and this technology is beautifully designed to help us better understand these kids and bring more effective screening and treatment to affected children. But final point here, we need better research systems to ascertain short and long-term benefits of these screening technologies. We've got, nobody wants to stop the development of more sophisticated analytic tools but we have to understand that we need those longer-term systems that aren't quite as sexy but they're gonna provide us with the evidence about whether generating this information confers benefits on kids or families or not. And in the absence of those sorts of systems to draw those conclusions, we're gonna be stuck with a powerful technology and lots of uncertainties about the best way to use it. So, in the mecca of research here, I'm making a pitch for those systems. Okay, thank you. We have time for a quick question. Is that anybody going to a microphone? If not, the question I guess I would have for you is you made the point earlier that educating parents even about current newborn screening routines very difficult and uptake is actually very poor. Is it, if we ever moved to a world of newborn sequencing, would it even make a difference or are we so, people so low on a learning curve and uptake that it's almost a trivial difference for them? Well, and I have some hope with the innovative educational modalities to improve parental education about this. And what the public surveys consistently shows, everybody says, why don't you talk about this during pregnancy? You've got nine months. I'm very interested in anything about the babies. Let's talk about it then. So, I think there's opportunities to do that but I would have to say in this little bit more of a skeptical angle on the response. It's not been clear to me that health departments have wanted a robust education outreach because everybody signs up for this. The more they know, then they're afraid the more folks are gonna say no. I think that's not correct but that's an empirical question that we hope to answer with some of our work. Okay, we're gonna move on. We're gonna actually switch the program slightly. Our next speaker is David Williams and Dr. Williams has a lecture he has to give this afternoon back at Harvard and he has an airplane to catch so we're gonna switch it around and do the video right after his talk. And so, we're gonna now hear genomics and disparities in health and healthcare challenges and opportunities.