Friday, February 3, 2006

Session 6: Newborn Screening for Genetic Disorders/Diseases

Jeffrey Botkin, M.D., Associate Vice President for Research Integrity and Professor of Pediatrics, University of Utah, Salt Lake City, Utah

Michael Watson, Ph.D., Executive Director, American College of Medical Genetics, Rockville, Maryland

 

CHAIRMAN PELLEGRINO:  Let's begin the session.

Members of the Council, please be seated, and speakers as well.  Members of the Council, please be seated.  We need to send our border collie out and nip at the heels.

This is our last session for this particular meeting, and we'll be turning to the subject of newborn screening for genetic disorders.  Our first speaker will be Dr. Jeffrey Botkin, Associate Vice President for Research Integrity and Professor of Pediatrics at the University of Utah in Salt Lake City.

The speakers have promised to be clear and brief, and I hope that will be a sample of what we would expect also from the interrogators.

Thank you very much.

Dr. Botkin, if you don't mind beginning, Id' very much appreciate it.

DR. BOTKIN:  All right.  Thank you, Dr. Pellegrino.

It's certainly an honor to be here for the Council today.  This is a smaller core set of issues, I suspect, from what you've been discussing at least over the last two days, although I'm aware that Dr. Fost was here a month or two ago and did have some information to provide you about newborn screening.

Nevertheless, my plan was to go through a number of talking points.  What I'd like to do is touch on a brief description of the program, talk a little bit about what the contemporary controversies are with some specific focus on some of the ethical issues, and then end with a couple of brief points about at least where  I would like to see these programs go over the next 20 or 30 years.

These are issues that I think have been or were sleeper issues for a number of years.  As a practicing pediatrician in my younger days, we were aware of the programs, but they weren't particularly controversial.  They were really occurring below the radar.

I think to a large extent that's no longer true, and that's true both because of some technical innovations that have changed the nature of the programs, and I think some social dynamics that have been brought to bear that have changed how people, parents look at these programs.

Certainly the largest application of genetic testing in the U.S., four million infants per year being screened with these technologies, and it is most frequently described these days as a system.  So I want to emphasize that concept.  I think too often it's thought of more strictly in terms of a test, but should be thought more broadly in terms of a system, because I think it's the system issues that raise a number of the difficulties.

Parent education may or may not be a component of the system.  Theoretically it is, but oftentimes quite limited, and this would be education that occurs in the perinatal period, most typically after the birth of the child.

A test is done, and what I've done is put on the table here what's known as the Guthrie cards.  Heel sticks are done in the infants; blood spots are provided or put on these circles and then sent to the state lab.  And in fact, this is an example of a relatively simple technology, but a technology nonetheless, and it's these cards that I think were a significant invention of Dr. Guthrie to enable mass screening to be conducted in an efficient manner with babies.  If you had to send two or three cc's to the lab on every baby, it wouldn't happen.

So the system involves in the blood test or the blood spot that then goes to the laboratory.  The results typically go back to the primary care physician.  The primary care physician, if there's a difficulty that needs to be followed up with the family, contacts the family.  There is an established mechanism by which the diagnosis is confirmed.  There's short-term intervention for kids who need to have interventions and then long-term interventions to maintain the treatment for the affected child.

Now, any one of these components, of course, can be limited or weak and, therefore, undermine the efficacy of the system in general, and I would say that the test itself -- well, of course, highly technical is the easy part.  The tougher parts, the more expensive parts are those other  system components that involve the parents, the primary care providers, and the subspecialists who are all part of the larger system that helps support these kids.

Now, one of the aspects of the program that is strongly historically based is that they're state based.  It's state health departments that take responsibility for these, and I think probably one of the few large health programs that are narrowly focused within state control, and as a result there's some substantial variability from state to state, and that has become a focus of significant dialogue in recent years.

And this has to do with the tests that are performed, as well as the nature of the tests, the specific technologies that might be employed, and other program components, like who it is that follows up on the test, et cetera.

All but three of the programs are mandated [meaning the informed permission of the parents is not required to conduct screening]. Maryland, Wyoming, and the District of Columbia are the exceptions, meaning the informed permission of the parents is not required to conduct screening.  And I would have to say that this, within the nursery environment, this certainly is consistent with my experience.  This is just not a high priority item for care providers.

Frequently, in this day and age, babies go home less than 24 hours of age.  Particularly for new parents, there's just an enormous amount of material that needs to be covered.  People need to rest.  People need basic education about baby care, and they want screening as simply not a high priority item, given the a priori risk for many of these infants is quite low.

So it is an issue that does not end up having a significant amount of information related to parents about the nature of these programs.

Now, all but two states permit parental refusal for religious or philosophic reasons.  However, parents are not routinely informed of their option to refuse, and in those states that do have an informed consent process, like Maryland, the number of parents who actually refuse screening is literally a handful per year.

Now, residual blood spots that I won't at least in my comments stress at all is another interesting domain here.  Virtually all cards will have residual blood left after the mandatory screening programs have been conducted.  So states will store these for various periods of time, various lengths of time.  Utah, for three months and then destroys them.  Some states at this point for 21 years or indefinitely, and an interesting potential source of DNA on the population.

Somewhat unclear how long the analytes for a variety of the other tests stayed good on those cards, but the DNA appears to stay good for quite a long period of time.  So if many states were to retain these resources for decades, you would essentially have a DNA sample on a large segment of the population that would be potentially quite useful for a variety of applications.

Now, I will say that these programs really have been remarkably successful.  To the extent that I've been critical of some issues, I don't think there's any question that the programs in general really have been extraordinarily beneficial, and particularly for conditions like PKU and hypothyroidism.  These are conditions that will cause devastation, neurologic impairment in children by the time that they are detected clinically.  So a screening approach is essential to pick up these kids prior to that damage, and the treatments are relatively straightforward.  Although sometimes the treatment for PKU is underestimated in terms of the burden that that can provide to the child and the family, nevertheless, these are paradigm conditions for this approach, and I think this has worked remarkably well.

Part of the question is whether other conditions mirror the characteristics of PKU and hyperthyroidism and our ability to adequately intervene and protect children, as has been the case with these conditions.

All right.  So what are the contemporary issues?  Wide variability from state to state without clear justification for why that is.  Local political pressures, local lay pressures, local expertise, different looks at the data at the state level.  There's a wide variability, and I think there has been a concerted effort in recent years to say that this is not a rational system.  We ought to have more uniformity across states with the types of tests that are being offered, and perhaps with the technology that is being used for the test.

Rapid increase in the number of conditions on the screening panels, and that's the majority of my slides here to illustrate for you.  These are from Brad Therell, 2000.  I mostly just want to point your attention to the purple, more than eight disorders, again, 2000, largely East Coast, north East Coast that had more than eight; 2001, 2003, 2004, October 2005.  My home State of Utah is now purple having implemented tandem mass this past month.

So we now have quite a few states.  Thirty-six states, I believe, have more than eight disorders on their panels now, whereas in 2000 there were eight states with more than eight disorders, and this expansion is really due to new technology, and specifically tandem mass spectroscopy and do a test simultaneously on 30 or more conditions from a single blood spot, and I think this is central to the ethical issues that are part of the newborn screening debate at this point.

And I think folks are anticipating the prospects of DNA based testing in the not too distant future, and so the number of tests that conceivably could be done on newborn screening spots is really quite large in number.

So part of the challenge that I'll mention here in a few minutes really is that.  How do we draw appropriate lines around the types of test that ought to be implemented with this type of system.

I think this expansion is due to a variety of things, the technology, of course, and also lay advocacy in an emerging commercial market for newborn screening.  Here's a screen shot of the Pediatrix Web site, a company that provides commercial newborn screening.  Some states contract with this company to provide their screening.  I believe they're up to about 50 conditions now on their panel, and these are marketed through women's magazines, et cetera, to parents directly as a way of providing care for their children.

So I think this commercial market has helped foster and encourage states to look pretty hard at a competitively large panel.  I'd be interested in Mike's comments on that.

From my perspective, I think data on the efficacy of population screening for many of the conditions that are now part of these large panels is limited.  It's very difficult to do research on rare conditions.  Some 12 of the 29 conditions that are on the ACMG panel have an incidence of less than about one in 100,000.  Much of the information to support screening for these rare disorders comes from small, uncontrolled observational studies, and given the rare nature of these conditions, if you're a clinician who has seen three or four kids with any one condition, then you are the world expert.

So oftentimes it's historical controls, I think, that are used to try to determine whether newer interventions are effective or not.

Again, from my perspective I think population screening has proven to be rarely an effective intervention for very many things in medicine.  We have enthusiasm over the years for a variety of different approaches, and once the data is collected, we're disappointed.

I'm old enough to remember when a chest X-ray was done on every kid, every person admitted to the hospital.  It seemed like a great idea.  Once people collected the data on it, it's not that it never worked.  You know, the argument is that you never found anything of value on those films.  The argument is the benefits, the occasional good information that was provided was swamped by the false positives, the expense and the complexity of the program.

And I think we see that today.  Mammography is recommended for women 50 and older.  It's not recommended for women under 40.  Is that because breast cancer doesn't occur at that age or that mammography is not useful at that age?  You know, no.  You could pick up cases at the younger age.  It's just that the value of the screening program, given the low incidence and the high number of false positives overwhelms the benefits that are achieved by those detections.

So I think the questions are still out there with respect to many of these conditions about whether that will be the case for this population screening.

Now, of course, these are conditions that are not curable in any sort of quick infectious disease type sense.  They're manageable or partially manageable through oftentimes fairly complicated and burdensome interventions over time, and the question then is is that part of the system.  How well does our system support kids and families who have to maintain these complicated support services for kids over time, and if they're unable to do so, then the value of the screening program is substantially reduced.

Now, negative impacts of screening.  I alluded to them a little bit, but I'll enumerate them a little bit more explicitly here.   The positive predictive value from any newborn screening test is less than two percent.  Positive predictive value means the percentage of positive tests that are true positives.  Okay.  So a two percent positive predictive value means of 100 people who end up positive on your test, two of those are affected with the condition.  Ninety-eight are false positives.  Many are less than one percent.  For a tandem mass [spectrometry test] the figure is relatively good, approximately ten percent, meaning of the initial positives ten are true positives and 90 will be false positives.

Now, that's an enormous cost to the program because much of the system is designed around trying to track down these parents, find out whether the kids are true positives or false positives, and those personnel issues end up being problematic.  Simply finding people in this day and age has become enormously problematic.

But probably more importantly is that the research consistently shows that some parents with false positive kids continue to believe there's something wrong with their child.  Somewhere between five to ten, some studies up to 20 percent.  Interview these folks later.  They're still anxious about it, or in fact, they may say, you know, "I'm not sure.  You know, they told me he might have CF and then they said it was okay, but I'm not convinced."

From the clinician's perspective, you give them the good news.  You tell them, you know, it was all false positive; don't worry about it.  Well, a percentage of folks do worry about it, and that's the burden of the programs.

The nature of many of these rare conditions in particular is not well defined.  Screening detects individuals who have abnormal biochemical profiles, but no apparent clinical disease.  So these individuals can be harmed if unnecessary treatments are implemented when they're not indicated.

So if you look across the spectrum of severity for the condition, you have folks, kids who may die in the neonatal period no matter what you do, and you have folks on the other end of the spectrum who may never have manifestations during the course of their life.  You can't tell those apart necessarily in the newborn period.

So if you've got a restrictive diet, if you've got an intervention that you're going to use for those kids, it may be applied to that segment of the population that doesn't need it because they've got a more benign form of the condition, and this was fully illustrated with the PKU experience early on.

I know Norman Fost talked a little bit about this. Hyperphenylalaninemia basically is a benign condition, looks like PKU.  Some of those kids were put on restrictive diets.  Some died; some had malnutrition as a result.  I think Norm has larger estimates of the number of kids out there than many folks do -- I don't know -- that have seen data on that, but it's clearly a problem.

I'll illustrate that quickly just with some more recent figures.  MCAD is, I think without question, the one condition or a condition on the tandem mass panel that most folks are quite convinced is a worthwhile modality.  So a large study done in Australia and a couple of territories there by Bridge Wilcken, they found 16 affected children per 100,000 kids screened, and what they did was look retrospectively at how many kids had been diagnosed clinically within these populations during four-year periods, and in general two to seven kids were diagnosed clinically in any four-year period before tandem mass was implemented.

Once tandem mass was implemented, during that four-year period they found 17 children.  So what's the explanation?  I mean, it might have been that some of the kids died and were not adequately identified as having MCAD previously during the clinical phase, but it's also quite possible that they were diagnosing, finding folks with MCAD who never would show up with disease, and maybe they're symptomatic and struggling with different aspects of their health, and this would provide benefit, but I think we're uncertain about that.

All right.  What are the current controversies?  What criteria should be used to include the test on a newborn screening panel?

The next few slides here, and I think your handouts show what have always been the classic criteria.  Wilson and Jungner from 1968, World Health Organization publication, and these have always been public health criteria.  These were not developed for newborn screening per se, but for public health screening more broadly.

And I'm not going to go through these individually, but part of the point here is that they are quite front loaded for benefit to the individual involved and quite a bit of knowledge about condition and systems in place to support affected individuals.

Institute of Medicine, with their report in 1994, quite a bit more terse.  They had three conditions specifically designed for newborn screening, number one being for conditions for which there are indications of clear benefit to the newborn.

So clear benefit to the newborn has been traditionally a central justification.  I think part of the dialogue, a significant part of the dialogue in recent years has been questioning that.  Is it appropriate to consider tests for which there may be marginal or even no benefits to the newborn if there are benefits to family members?

And there's a couple of family member benefits that are really quite real:  elimination of the so-called diagnostic Odyssey.  Many of these conditions will take families through the clinic multiple times, enormous levels of frustration trying to figure out what's wrong with my baby, and ultimately the diagnosis is made.

Screening may eliminate the diagnostic Odyssey in many circumstances.

Secondly, information for reproductive planning.  If it takes time to make the clinical diagnosis for a child, say a, year and a half, two years, what's happened in the meantime not infrequently is the second affected child has been born.

So even if there's no benefit to the child from screening, the argument goes families are forewarned of their reproductive risks and can take measures either to have no babies, do prenatal diagnosis, whatever they might choose to do for future pregnancies.

Now, there have been concerns about whether this early diagnosis of kids prior to the time that they would present clinically eliminates sort of this window of -- I don't know the right term for it .- sort of blissful ignorance for lack of a better term that parents might have with their child, and folks have looked at this in the context of Duchenne muscular dystrophy.  This is a condition that there was pilot screening for in Pennsylvania and Wales.  Nothing can be done for kids with muscular dystrophy by way of treatment that will eliminate the mortality track.  Supportive treatment, but this would not be a condition that you would screen for because you can do early treatment.

So they talked to parents about whether it be neonatal diagnosis of these kids rather than a diagnosis more typically at two to three years of age, in any way impaired their relationship with the child, and the research did not show that that was an issue.

So I think ethicists remain concerned about that issue, but the limited data we have available right now suggests that that may not be an actual concern.

One paper recently illustrated another potential approach to newborn screening outside baby benefit, and that's what they called reverse cascade.  They were looking at hemochromatosis, a hereditary condition of iron metabolism, problematic for population screening, and so what they did is they screened babies.  This was in France, identified carriers and affected infants, and then went back to their parents and grandparents to screen for affected individuals, and they suggest that's an appropriate screening strategy primarily to benefit the parents and grandparents.

All right.  The second or third concern here, what are the ethical responsibilities to report test results for conditions that are not specifically targeted by the program?  Tandem mass is a so-called multiplex technology, meaning it analyzes the chemistry of these specimens rather exhaustively.  You may be only looking for one or two conditions on that output, but you've got the output for all 30 or 40. 

What do you do with that information?  Is it ethically appropriate to ignore that, knowing that there may be implications as a result to the kids and families from not having disclosed that information or does the fact that you get that information provide an obligation to relay that to folks even if it doesn't fit what would be the standard criteria that have been in place for a while?  I think this is a tough question.

What's the role of parental permission with newborn screening?  A separate issue, long debated.  I would say programs and providers do not support an informed permission process.  A number of professional organizations, including Institute of Medicine, however, have advocated such a process.  The few studies that have asked parents about this, parents don't actually expect an informed permission process.  They're willing to go with a system in which they recognize that there's a unique environment within the newborn nursery.  They don't necessarily feel like they have to talk to folks in detail and make these sorts of decisions, the majority at least.  Obviously there's a spectrum of opinion on that.

So I would say from the program perspective, there's probably little emerging pressure to have a permission process for things like PKU and hyperthyroidism.  Where it becomes problematic is when you have other conditions for which you really don't have good data of the efficacy of the program.

If you've justified your mandatory program on clear benefits to the baby and you don't have those clear benefits, how is it that you're justifying the mandatory aspect of those programs?

And then lastly, what's the appropriate approach to establish efficacy for screening programs?  I think there are multiple challenges here, rare conditions, how you do research when conditions are less than one in 100,000 kids.

The state-based organizations, I think, are a barrier to conduct research.  Health departments are there to provide services.  They're not academic institutions where research is a significant component of their daily set of activities.  I think hopefully additional research funds can be funnelled through health departments, but at this point I think it's simply not the same kind of environment that academic centers provide for research.

Limited funding has always been a problem, and then ethical concerns over randomized controlled trials.

All right.  So that's my personal vision, and I would say that these are issues that are not going to strike anybody as particularly unique.  I think Institute of Medicine articulated many of these a decade ago now when they looked at these issues, and I think Mike and I are in agreement about many of them as well, if not all of them.  I think there has to be an effort at professional consensus on some of the key ethical issues.

How this would be done I'm not sure.  Perhaps another Institute of Medicine type of approach on some of these issues would be beneficial, although I think that there's always a challenge in translating professional statements into actual policy that's a matter of some frustration.

Nevertheless, the issues that I think need further discussion are questions around multiplex technology.  I think we're going to be increasingly seeing this with DNA technology.  We see it with tandem mass.  What do you do with the full spectrum of results when you're only looking for a smaller subset that meets traditional criteria.

Secondly, informed permission from parents, I think, needs additional discussion particularly in light of the experimental nature that everybody is going to be advocating for better addressing these issues.  We need to think more clearly about what the role of parental permission in education is.

Third, we need to think more clearly as well about the criteria for newborn screening.  If it's significant benefit to that child, then I think there will not be significant difficulties that arise down the road.  I think if we're willing to say, well, this is an opportunity to provide benefits to other people because we have a captive audience within the nursery that's a wedge into the larger family, then there's really no logical line to draw around that.  We could be doing an enormous number of tests on babies in the not too distant future, particularly if the $100 full sequence genome becomes a reality in the next 20 years or so.

Research ethics issues I think are still with us, how to do research in this context, particularly where it's problematic for control trials that I tried to address in the paper that I think was in your folder.  Adequate funding is certainly a question.

Longer term, how should the system look?  Regional and national pilot programs to assess efficacy of screening prior to full implementation, but that would just be expected that one moves from evidence of benefit in small trials to pilot studies with consistent data collection across the country.  If these tests are important for child health, then, of course, they're important to  do in the right way with the right data collection.

There ought to be a national registry of affected children, and along with that longitudinal surveillance of the affected kids for health status, developmental status, psychosocial, and cost impacts as well.

And one way to conduct this research actually is something that Mike may speak more about, but it's setting up something like the Children's Oncology Group.  Cancer also fortunately is a rare phenomenon in kids, and what's been done is to set up several centers around the country that are conducting research on cancer, and a large percentage of kids with cancer in general are enrolled in clinical studies.  That's probably the single greatest reason why there's been so much progress made in childhood cancer compared to adult cancer, where a very small percentage of adults with cancer are enrolled in clinical trials.

So you can feed kids who are affected with these conditions in the central centers.  They can use protocols to compare established treatments with innovative treatments and try to draw some conclusions within a reasonable time about whether these things work or not.

Under the current system we will be arguing 20 years from now whether these tests benefit kids or harm kids, and if we're going to move forward here, I think we need to do so in the context of a comprehensive research enterprise.

CHAIRMAN PELLEGRINO:  Thank you very much, Dr. Botkin, for your conciseness and especially for touching on some of the ethical issues and some recommendations for us.  I appreciate it.

We're going to turn now to Dr. Michael Watson, who is the Executive Director of the American College of Medical Genetics.  He will make his presentation, and then we will have interrogation.

Thank you.

DR. WATSON:  Interrogation.  All right.

(Laughter.)

DR. WATSON:  Well, thanks for -- that's an interesting use of the word.  We'll see how it goes -- I appreciate your inviting us here to talk.  Jeff and I actually were at a meeting together just last week and recognize that at this level, I think, of entry into this topic, that it was probably better for us to just both speak and then take questions sort of together because I think, as he alluded to, we probably agree on the great majority of the system and the issues.

However, the difference, I think, in my perspective was driven by the fact that we were asked as an organization to address specifically the kinds of conditions that ought to be included in a uniform panel of screening, which placed us where the rubber meets the road, and it's nice to talk about all of the past policy decisions that have been made.  To operationalize that is another problem entirely.

So what I want to do is not surprisingly Jeff and I, our slides may almost be identical.  So I'll try to avoid saying the same things he said unless I disagree with what he said, which there's very little of that, in fact.  And what I will do though is try to drill down a bit more into some of these issues he alluded to because they are the issues that we specifically faced in trying to operationalize the decision making around newborn screening.

And I'll do it actually through a few different mechanisms.  I think thinking about the development of phenylketonuria screening is a very good example of some of the problems that we had that actually none of the systems have ever addressed.  The systems are quite difficult in that they are a mix of quantitative and qualitative measures that can become very difficult in trying to make decisions around the difference between objective criteria, like the incidence of a disease, versus the subjective, about what is perceived as a benefit to either the child, the family or society, a very difficult process to work your way through on a condition-by-condition basis, and realizing that when we did our analysis, we looked at 84 -- well, 79 conditions -- within a couple of clusters of conditions, which left us with a feeling that it was different in the past.

In the past you looked at a condition and said, "Okay.  It seems to meet these criteria in general."  You may prioritize one criteria over another, but you look at a condition on an individual basis, and we're looking at 79 to 80 conditions and now are trying to rank them against one another really as to what is more appropriate because I think as Jeff suggested, the potential for screening for enormous numbers of things based on the technologies that are becoming available is clearly moving this field very rapidly.  But we don't have a good approach to sorting out the issues that really are prioritized.

I think David Atkins captured it well recently.  He's with the AHRQ, their evidence based medicine approach, where he, when looking at newborn screening and many of our recommendations, suggested that somewhere in the neighborhood of 90 percent of the differences in perception were really ones of your perception of the data and your value system, and about ten percent of the difference was actually in the evidence itself.  So we had a complex problem to deal with.

Excuse me just a minute.  My arms aren't quite this long.

All right.  So very briefly as Jeff indicated, newborn screening today is a public health program.  It's mandated at individual state levels.  That does bring significant difficulties to the process because a test in one state is not the same test in another state.  So just aggregating data becomes an enormous difficulty in trying to compare performance and to bring all of the data that has been collected across the country together to make decisions.

The fundamental basis of the programs historically has been aimed at the identification of conditions for which early intervention can prevent mortality, morbidity, and disability.  And I would say in the earliest days, it was very much focused on conditions in which one made a very large difference in the outcome of the individual, and a lot of what we're beginning to see is incremental improvements and recommendations from what is now a very strong consumerism movement in the United States to recognize incremental benefits as being useful, not just the very large benefits between a severely affected child and a near normal outcome.

For the most part, newborn screening is based on diagnostic markers identified in blood spots, but there is another trend.  Jeff showed you the blood spot card where the blood is collected.  However, there is another whole set of newborn screening that arising in the nursery environment.  Jeff alluded to hearing loss in the one child he showed.  That is not a test that is done through bioanalytical markers.  It's done in the nursery.  It screens the child for hearing loss and becomes a very different system, and there are an array of conditions now that are actually being  pilot tested within the nursery:  hyperbilirubinemia as a screening study since we've had an up tick in the occurrence of Kernicterus in the country.  There's quite a number of conditions that could potentially be done in the nursery, and then tied to the public health reporting system so that the state's responsibility would really be the oversight of the program and ensuring that everything happened in a public health setting for the screening of that condition.

But that's an interesting evolution of the programs that has occurred just over the last couple of years. 

And the extent of the program and the analytical platforms that are chosen for screening are decided on a state basis independently, and there's enormous differences in the tests chosen for a particular analyte to be screened that, again, makes it increasingly difficult to have that comparative data, which is critical for rare diseases because one state, for instance, the condition homocystinuria, most states won't see one of those in a year, which has implications for quality of testing and lots of downstream implications, and then to have them identified through different technical approaches makes it increasingly difficult to compare really what you started with in those different states.

So I'll come back to this in just a moment and say that just a little bit more around what is the newborn screening program and system because, as Jeff suggested, it is a system.  Screening does not operate in and of itself, though it does in the private sector.

Clearly, there are two competing kinds of screening involving the state mandated screening programs and another component that's referred to as supplemental screening, and in fact, some states have actually mandated that mothers in the nursery be made aware of the availability of supplemental screening, and that's particularly in the context of the state feeling that other states have progressed further than they have, and they have taken on a responsibility feeling that they need to tell the parents that there is more available out there that can be purchased privately.

So screening is central to the system, but there's a short-term responsibility of tracking those children down that the state program takes on.  That's greatly complicated by the fact that the hospital system has not assumed responsibility for the tracking of the children that come through the system for newborn screening.

The newborn screening program has been shifted into the laboratory component of hospitals where it's a reference lab send out type of system, and the hospital does not necessarily take the same responsibility for those specimens as it does for an in-patient within the hospital itself.

Once identified and tracked down as having been screened positive, the diagnostic component of the system kicks in, and that's largely through -- in some states, in fact, the referral is made directly to specialists for the condition related to the marker that was found to be elevated.

In other states the primary care physician is the front lines, and for some conditions like sickle cell, there's a feeling that one needs to tell the family and the primary care physician and the specialist simultaneously, with the feeling that that family may not have the same kind of access into the health care system, and therefore, the family becomes much more directly in the loop.

In most metabolic screening, the primary care physician is given the information and that is then communicated to the family.

Once the diagnostic system has kicked in and you move into the management and treatment side, which is presumed to be available because that's one of the really critical criteria by which one assesses a condition, that there is a management or treatment intervention that can make a significant difference in the outcome of that child, and then two critical system components, one being evaluation of the program.  Is it working or not?  And that was stunningly difficult for us because of the great differences between the states and their approaches to screening their technologies, their cutoffs for analytes made it particularly difficult.  But one assesses the entire system right from identification in the screening process to did you realize the outcome that was expected based on having identified and intervened.

And then education.  Education is critical to the development of these programs, and it's not just the education of the family who is participating in newborn screening.  Clearly, this has been a program that has occurred in the nursery.  I can't think of a worse time to try to deliver education than in the nursery at the time of delivery, and there are significant efforts taking place to shift that up into the prenatal period and shift the burden from pediatricians in the nursery down to the obstetrician-gynecology community who have a much longer, ongoing relationship with a mother who is pregnant and has a much better opportunity for education.  So, I mean, clearly a problem, but one that we have begun to address as at least being fixable to some extent.

Now, Jeff showed you this same thing.  I'll zip through this since he gave you the perspective already.  It is expanding fast.  This is actually from the perspective of the percentage of babies born in the United States as opposed to the number of states.  Wyoming is not going to make a significant difference in the total number of babies born in the United States.  California, Texas and Florida comprise nearly a third of the babies born in the United States.

So in 2002, when we were asked to do a scientific analysis of the conditions under consideration, about four percent of the babies born in the United States were screened for more than 20 conditions.

I'm going to blast through this because you saw this already.  I won't even dwell on it.

In 2004, while our study was still ongoing, it may have had some impact on some of the decision making in the states because we were quite open about our process and where things were going.  At that point in time about 37 percent of the babies born in the United States were being screened for more than 20 conditions, and you can see that there's about six states where only four conditions are being screened.

And now, as of last week, 66 percent of the babies born in the United States are being screened for more than 20 conditions, and you can see that the minimum number is now up to eight, and that's a bit of an aberration.

California recently announced that we're going to screen for 77 conditions.  The states don't even count these things the same.  So it becomes a rather murky kind of topic to address.  California got 77 because it said we're going to screen for sickle cell anemia, but we're using a technology that is multiplex in its nature, isoelectric focusing and HPLC will show you all of the globin variance, and what they said was, okay, there's 25 clinically significant combinations of variants, each of which is a condition.  So there's this fine line between marketing and counting that gets very complex, and clearly the private sector has driven that to some extent in claiming that we're screening for increasing numbers of conditions.

I think if you take PKU as an example there are sort of two competing problems when you look at how it evolved.  We thought we were screening for PKU.  We were screening for too much phenylalanine being present.   Jeff alluded to the fact that there is a benign hyperphenylanemia usually characterized by intermediate levels of phenylalanine, but over the years what we've learned is that there are other conditions with increased amounts of phenylalanine.  There are biopterin regeneration defects and biopterin defects of generation of the material.

Each one of those is two separate genes.  So for PKU now we have at least five different genes that we are considering when we think about it.  Four of those are so rare that you couldn't make a decision based on them independently.  However, we're screening for the marker of too much phenylalanine, and I think that's an important consideration.

Jeff alluded to it in talking about all of the extra information we get and whether we should tell people or not.  The reality is that when we screen for too much phenylalanine, we're really looking for PKU, but we acknowledge that during the differential diagnosis of having too much phenylalanine, it's quite possible that you're going to find that the patient actually has one of these other conditions.

But that is not something that the system is revealing.  That's actually a physician-patient relationship and sorting out what that baby's reason for having too much phenylalanine is and is not subject to, I think, decision making about whether or not I tell the patient that.

And that actually accounts for probably 90 percent of the conditions in our panel that are called secondary targets.  The condition is part of the differential diagnosis based on the analyte used to identify a core condition in the panel.

So as you see, we have great disparity across the country in what's available.  We have a very mobile society now.  So families, in fact, have been the driving force in newborn screening since day one.  The Guthrie card was Robert Guthrie who had a child with phenylkonuria, drove the system, organized families, and lobbied for the development of newborn screening.

That and technology have probably been the two main driving forces, and our organization and approach to sorting out what we should and shouldn't do and what are the criteria by which we make those decisions has lagged tremendously behind.

As you look, I mean, newborn screening is often said to be a universal screening program.  That's why we embed it in public health programs, because in an environment in our health care system where access is increasingly difficult to large percentages of the population for various resource or other reasons, the only place where we could assure access to all newborns was in this public health system.

So it's imbedded there, and we acknowledge that that was an important place to keep it in order to provide that broad access.

There are now, as I said, there are more than just those three conditions.  Sickle cell is now screened in all states, and at least three other combinations of a sickle allele with anther hemoglobin chain are also identified.

It's not perceived as fair by the public.  Newborn screening programs have gone through episodes where they try to target specific populations.  This was done in the U.K. for some conditions thought to be most common in the African American population, failed miserably. 

In the United States we've had similar difficulties because in reality selecting a population by some ethnic criteria is not a surrogate for long distant ancestry of origin of that particular population that is at true risk, and those are very different concepts, and selected population screening has not worked well.

So do limited pilot programs, tend to be some of the larger states.  There's probably only two states that have a significant research component to their newborn screening program because they are much more service oriented.  So it has been very difficult to do the pilot studies and develop the kind of information that's needed.

The NIH is currently looking at developing a group of states within which they might be able to pilot studies much more effectively than we have in the past, and it has been left to consumer initiative.  As I said, some states have said, okay, we're just going to tell them that all of this other stuff is available in the private sector, and it's not uniform as I think you've seen.

So getting down to that, sort of drilling down into the detail of what you have to do, where the rubber meets the road and you're going to be trying to make decisions about conditions, you have to have some criterion, and as I said, Wilson and Jungner criteria are not particularly quantifiable.  There are general considerations.  More harm than good, you know, is a reason for not screening, but that's a unitary sort of value and very difficult to quantify because you're aggregating a lot of different criteria and values together in making your ultimate decision.

There has been prior policy guidance.  I gave you a paper by Brad Therelle that touches on a series of those, the work by WHO in 1968 that really resulted in the Wilson-Jungner criteria, criteria that were actually developed around, I believe, chronic disease screening in adults.  So it was not ideal for newborn screening, but conceptually had a lot of useful information.

1975, the National Academy of Science expanded on that, and what they did was take those criteria for the most part and develop some procedural guidance around how one might use them, but still considerable qualitative kinds of information being considered in deciding.

In 1994, the Task Force on Genetic Testing that I co-chaired looked at newborn screening to a limited extent and tried to lay out some more specificity about the nature of the laboratory testing and some of the quality assurance issues needed to be able to really understand whether or not something was going to work or not.

And then in 1999, probably the most recent very useful study was the American Academy of Pediatrics Task Force on Newborn Screening that looked at the full breadth of the newborn screening programs and made general recommendations across the full spectrum about how we as a nation can improve newborn screening.

One of the recommendations they made is that there should be an organized activity towards determining what is appropriate for screening at this point in time.  It was clear that many states had already expanded newborn screening with most of the expansion really being tandem mass spectrometry and then counting, in the way you count up how many times you're screening for.

The condition itself is a mix.  There are some very objective criteria there.  The incidence of the disease, if you have good population data, you can say this is the incidence.  But now what we realize is that technology has changed our perception of incidence.  Whereas you might not screen for a condition that's found in one in 100,000 or one in 200,000 in the population, the technology may actually through its multiplex capacity allow you to screen for that and a whole lot of other things.

So if you look at tandem mass spectrometry, there's about 43 total conditions that you can diagnose through tandem mass spectrometry.  Many of them we didn't think appropriate for newborn screening, and, in fact, there's a couple of different tests because it's one scan of the system gets you a group of acylcarnitines for the fatty acid oxidation defects and organic acidurias.  Another scan gets you the amino acids.

So there's a couple of different tests mixed into it, but among the amino acidurias there's a half a dozen.  So when one thinks about the incidence of any one of them, the test is actually doing them all at once.  So that criteria gets quite muddied.

Other aspects of the condition is clearly the burden of the condition.  If you think about something like cystic fibrosis where if you did it by DNA, you could be identifying individuals at risk for sinusitis through mutations in the CFTR gene.  You want to focus on those things that really are related to severe disease.  so you begin to narrow down.  If you were to do it mutationally, you'd want those associated with severe disease presumably for screening, but the preference is for functional assays that tell you that whatever mutation might be present in the gene, that you're really looking at the expression of that through the elevation of product of a particular pathway.

And those get you at least closer to the endpoint of a functional identification of an at risk individual as opposed to the DNA level where the variable expressivity of the conditions is quite a problem that will be faced in the future.

The screening test itself, obviously without that you don't even think about screening for a particular condition, and when we analyze these 70, 80, 90 conditions, there were 27 where we just said there isn't even a screening test available both through a cost perspective.  There are tests available, but certainly that don't meet the screening criteria.

We actually considered the fact that a test had a multiplex capability as something of value in a public health setting.  When you reach the diagnosis level, there's a number of considerations that come into play.  Diagnosis can be done by a primary care physician, for instance, in hyperthyroidism, a relatively common condition in newborn screening.

For many of the metabolic diseases, diagnosis takes place at the level of a group of specialists who are in incredibly small numbers in the United States, and that's a problem in and of itself.  And we ascribed greater value to a diagnostic that could be done more broadly in the physician community than we did to one that was restricted to the smaller group of experts and have gone the next step now of really developing materials that facilitate people's understanding of when they need the experts and how to get them and loop them into the system so that as a collaborative management approach to many of these rare diseases.

Management and treatment needs to be available.  Many of the same issues, complicated by the fact that with many of these conditions, given that there is mutation variability in the gene that has implications for severity, that management and treatment is not a single entity.  Some will present earlier in life as Jeff alluded to and perhaps before the newborn screening result comes back.

Others will be milder.  That's the reality of genetics in the current day.

Potential harms?  We approached the potential harms really from a much more clinical perspective.  We did our work through a contract from the Health Resources and Services Administration, the Material and Child Health Bureau and were specifically asked to address the scientific issues and evaluating conditions.  So we did not spend a lot of our time thinking about the consent issues.  There were parallel projects occurring in the country that we're addressing, many of those kinds of ELSI and consent issues and other aspects of newborn screening.

But the hard part is putting it all together at the end.  You've got all of these criteria.  Some people think that the ability  to treat is the major one.  If you can screen and you can treat, then that, you know -- those are really the trump cards.  So there's a lot of difference of opinion, and that's where the value system begins to come into play as to where are you going to place your priorities across these various criteria, some of which are clearly quite subjective. 

What is the benefit?  Jeff has alluded to many of those.  There are families in Fragile X Syndrome, for instance.  It's an interesting condition; didn't meet our criteria for newborn screening, but as you think about it, it's relatively common.  It's more common than many of the things in our panel.

But the improvement is one of early intervention programs for mental retardation and developmental disability.  Very much more incremental and significant questions as to whether or not in the mainstream school system you realize the kind of benefits from early intervention that you might in the best center in the country, where very few of these individuals are going to be seen.

So that's a problem.  It's a condition though where Jeff suggested about two years was the diagnostic Odyssey length.  In Fragile X Syndrome, it's closer to four years.  Most families have completed their family planning by the time the diagnosis is finally established and argue strongly that one should broaden that benefit consideration out to those things that might have implications for reproductive decision making.

I'll keep moving quickly now.

CHAIRMAN PELLEGRINO:  Thank you very much, Dr. Watson.  Oh, I'm sorry.  I thought you --

DR. WATSON:  Well, this will go actually quite quickly.

I won't spend a lot of time on many of the issues that Jeff has already alluded to.  Things that complicated our problem were the clinical, biochemical, and molecular complexity of the conditions that we considered and this interrelatedness between the markers and their involvement in multiple diseases that are part of  really a continuity in a biochemical pathway that make many of the conditions quite similar at the analytical level.

The impact of these multiplex platforms is significant, and their ability to identify numerous biochemical markers.  There were tradeoffs here.  Clearly one can adjust tandem mass spectrometry so that you can focus it on certain conditions, but there are tradeoffs in the quality of the testing and it's important to realize that most newborn screening laboratories in the country have been run by clinical chemists who have been in those systems for a very long time.

Many of the new genomic technologies are brand new.  So they had a long learning curve in developing that technology and bringing it in.  Jeff suggested a positive predictive value was a significant problem.  We've already seen in tandem mass spectrometry in the country that in some states the positive predictive value of screening by tandem mass spec is two percent.  In others it's 50 percent.  They have maximized the performance of their testing systems, and it had implications throughout the way they approach the patient. 

If you're 50-50 when you go out for diagnostics, you don't just begin confirming the condition.  You start looking for hyperanenemia and associated things because you want to be ready to treat, and it's not going to cost you as much to take that step.  If it's one in 50, you're probably not going to make some of the investments in some of the residual phenotypes of these conditions, but really very much more focus on confirming the fact that it really is before you start down that pathway.

The process by which we understand genetic conditions, we really don't understand genetic conditions until we have done them in newborn screening.  We go through a long series of biases.  We start with the most severe versions of the disease.  We move into families that we see less severe versions within the family.

We move out into phenotypes and look at that group, and it gets less severe again, and it's not until we do the general population that we actually understand the implications of this gene for the broader population.

There's a lot of gaps in the level of clinical knowledge among the stakeholders.  It think it's well accepted that there are many stakeholders, families, organizations like yours, and there's a wide range of expertise and understanding of genetics in these conditions that come to play.

And then the whole consent issue, which I won't touch on.

And then lastly, things on the horizon.  Lots and lots of new technologies and treatments are already in the pipeline.  The field is moving at a remarkable pace.  I've mentioned a few, and we can perhaps discuss others during the discussion section.

It's recognized though that this has become very much a national issue, and there have been moved made already.  There is a national advisory committee, the Advisory Committee on Inheritable Disorders and Genetic Diseases of Newborns and Children, advisory to the Secretary of HHS is developing strategies for a national level look at conditions to consider their appropriateness for newborn screening.  They're looking at models like the immunization practices models that have continuously upgraded the vaccination programs in the country as an approach to bringing this scientific information together with the other interest.

Lots of new indications.  We've talked about conditions with less dramatic intervention or improvements from interventions, reproductive decision making.  Lots and lots of talk about late onset disorders where to this day many of the interventions are much less clear, but we're already seeing a group of conditions like the lysosmal  storage diseases where there are versions of those that will present in the newborn period and can be intervened, treatments that have gone through FDA and are approved, but there are subsets of those patients who don't get the disease until they're 40 or 50, and they all show up on newborn screening, so that we increasingly need to have an organized system to move people into expertise that will deal with their particular diagnostic issues as best as can be done.  And then expansion, as I said, into the nursery. 

So the big things we need.  Data collection is huge.  The Children's Oncology Group model is a nice one, but genetics is going to be hard because it's everywhere.  Oncology at least, I was in the Children's Oncology Group for 20 years, and it was a spectacular program.  It allowed us to raise the standards for the laboratories because we were doing clinical investigation to a large degree.

Raising the standards for those labs so that the best information was coming into the system on which decisions would be made about what to do and what works was critical, and we don't have that in the current laissez faire kind of marketplace for how we work through these patient populations.

Education is an enormous problem.  Genetics and knowledge and genetics has expanded enormously over ten years.  In 1990, 50 percent of the medical schools in the country didn't even teach it, outside of Mendel's peas and that kind of perspective of genetics.

And the quality of testing is becoming an interesting problem as well where we have programs to improve performance across the country, all of which are going to be necessary to get to that kind of uniform data that we need to ultimately make the kind of decisions that have to be made about what's appropriate and what's not.

So I'll stop there and be interrogated.

CHAIRMAN PELLEGRINO:  Thank you very much, Dr. Watson.

(Applause.)

CHAIRMAN PELLEGRINO:  I want to remind the Council members we have 30 minutes for questions.  You can direct them to either of the speakers or simultaneously.  Dave will keep you in order, the order of speaking, that is, not otherwise.  But I would invite the speakers to respond and also to quiz each other if they wish to do so on the same question.

I have first Dr. Gómez-Lobo, followed by Dr. Hurlbut and then others as you show and manifest your sign that you wish to speak.

DR. GÓMEZ-LOBO:  I wanted to ask an information question because I'm a bit worried from an ethical point of view about the notion of, say, doing something for which the vast majority of people would not be giving informed consent, if I understood correctly.  Many parents would not be even asked apparently, and those who were asked apparently comply very, very quickly.

I'm among those parents, too , who rapidly comply to whatever they tell me, but the problem I see is this.  Do you see any pressure to go from newborn screening to preimplantation diagnosis?  In other words, isn't there here a monster that we might suddenly find ourselves with a program that screens embryos before they're implanted?  Well, we can imagine what would happen with that.

CHAIRMAN PELLEGRINO:  I presume that's directed to either of the speakers.

DR. BOTKIN:  Well, that's a great question, and I think there has not been a whole lot of attention to the interface between newborn screening and prenatal diagnosis, and they're clearly avenues of information that are important there.

Oftentimes newborn screening will pick up carrier infants, meaning the parents are carriers, and that can provide some reproductive information for them.  It may then indicate that the child will be at risk for adverse pregnancy outcome 20, 30 years later when he or she has a baby, and so there's those sorts of issues.

I have not seen anybody explicitly talk about tests that might be implemented in a newborn screening program, particularly if they become DNA based, translate that back to pre-implantation genetic diagnosis platform, although I don't see any reason why from a technical perspective you couldn't do that.

Presumably you'd have to have a different set of criteria for what tests you might apply in those contexts, but you'd have to -- from a technical perspective it seems quite feasible.

CHAIRMAN PELLEGRINO:  Dr. Watson.

DR. WATSON:  Yes, I think it's feasible, but I haven't seen much evidence.  I directed prenatal diagnostics at Washington University for 17 years before I moved to the job I'm currently in, and from my own experience in that, as treatments become available, people don't tend to move back to reproductive decision making.  When there is an expected good outcome from having been identified at birth, they tend more toward the side of going down newborn screening and appropriate management and treatment.

One thing that will become apparent though, I think, in broad analyses of conditions as to whether or not they are appropriate for newborn screening is if they're not and it's because there is no treatment in the outcome and burden are huge that will actually be defining the conditions that are probably -- that many would consider appropriate in a prenatal diagnostic environment, in which everything from CVS to amniocentesis to PGD is an option.

CHAIRMAN PELLEGRINO:  Dr. Hurlbut.

PROF. HURLBUT:  First a comment and then a question.  I have a vague apprehension about this whole thing of screening because as a physician it feels as though the emphasis is increasingly likely to be put on detecting things rather than  curing things.  Now, I know that it's a very good thing to make your diagnosis early in a lot of conditions, but you could easily get the situation where you're spending an awful lot of the health care dollar screening for things rather than working toward cures, and I think what was implicit in Alfonso's question was the question of whether this will translate into what Leon Kass has called curing the disease by preventing the patient.

I'd just like you to reflect backwards on the prenatal screening a little bit, but my comment really spreads a little wider because I was talking with a woman recently who was prenatally screened, and she said that the prenatal screening put her under the pressure of a huge amount of anxiety while she was pregnant, and even after the tests came back okay, as you said about those with false positives, there's something funny in this.

As a physician I feel it.  Well, that's a broad question, I realize, and a very vague question, but my specific question is you mentioned the potential with DNA arrays and so forth to screen the entire genome.  I was talking to a venture capitalist the other day who said that now there seems to be a technology for $1,000 to screen all six billion base pairs.

And we're going to be looking at commercialization of that and all of the dangers that come with people being encouraged to do things that they don't understand, and my big question is are you thinking about making negative recommendations in the broad front on that.

But before you get to that, would you comment a little bit on the statistical nature of these findings?  A lot of the concordance between identical twins, for example, is surprisingly low in a lot of gene expression patterns, and we have an overblown notion of genetic determinism in our society, and you can see that as an equation for an awful lot of social disruption that isn't just anxiety provoking, but would influence therapeutic decisions, educational decisions and even just attitudes toward your children.

DR. WATSON:  Well, that's an easy question.

(Laughter.)

DR. WATSON:  I'll start with the first one.  I don't think screening is actually at the expense of finding the cure.  In fact, to consider screening, at least newborn screening, one has the cure in place or the treatment in place that makes that significant difference at least in our current, you know, value system about the criteria and the magnitude of improvement.

PROF. HURLBUT:  I guess I was really thinking more of prenatal screening because there was a recommendation from some scientists at UC-SF about a year ago saying that all women, not just after age 35, ought to have prenatal screening.  Think of the amount of money that will cost and increasingly even the March of Dimes is speaking in terms of preventing disease by prenatal screening and the implicit endorsement of abortion.

Something seems a little askew in all of that to me.

DR. WATSON:  You know, I think certainly in genetics, I think genetics has made a relatively small proportional contribution to terminations in the United States.  And I think that certainly it has arisen around those -- I mean, it's sort of the way you find out about genetics in a family.  There is prenatal screening of the type that's done for Down's Syndrome, where people's decisions are changing, you know, over time as to how they think about having a child with Down's syndrome, but the vast majority has been very severe disease, commonly found in an individual in the family, often in the pediatric setting, often lethal, and much of prenatal has been those kinds of conditions where families are making decisions based on what has happened in a prior pregnancy.

That works, not prenatal in and of itself, but that kind of detection works best the rarer that the disease is because the rarer it is, the more likely that the carriers are within that family and not everywhere else.  So it's an effective way of identifying people and then clearly you provide information to people about what you know about the range of expressions of that condition.

Your question on arrays is actually not — there's some underlying issues around a raise, I think — that make different kinds of arrays different.  A sequencing array would be the nightmare of all time for prenatal screening, newborn screening, medicine in the hands of people who don't really understand them.

If you look at the cystic fibrosis gene, roughly 90 percent of patients with a mutation have one or the more common ones that is reasonably well understood.  Ten percent of patients with cystic fibrosis have a mutation that's private or rare in their family.  Initially on identification it was was unknown sequence variation, and a lot of other analysis was done to say, "Okay.  This is probably pathological because of its distribution in the family and other evidence about the domain of the gene in which the mutation may have occurred."

You let unknown sequence variation loose on people and it's going to be a nightmare.  But arrays can also be highly targeted at specific mutations, and in that sense, you know, they do begin to move their way into this discussion, and I think it is very much complicated by what we've alluded to as that variable expression of many of the mutations.  They operate on a background, and those genetic backgrounds are not identical in the population, and that contributes a lot to the variable expressivity of the conditions.

But New York State is currently beginning to test Affymetrix arrays for hearing loss.  If you take hearing loss, it's about 117 genes with a lot of different mutations in many of those genes.  If you take the panel and the secondary conditions that we recommended, that's about 155 genes' worth of conditions, many with a wide distribution of mutations across the genes.

So, you know, at least it's a more tractable issue there where you're targeting something.  You have a lot of the clinical questions about the targets, but as long as we have more functional kinds of assays.  In hearing loss, we test to see if they can hear.  We don't test them for that wide range of mutations.

So the further we step back from the functional expression of the gene to the gene itself, I think more and more issues get sort of overlaid on the problem.

DR. BOTKIN:  A couple of quick responses.  I do think that there's competition between money that's going into this screening enterprise from the care enterprise, and you know, looking at the larger picture of newborn screening I personally would have been much happier had all states been screening for six things, but yet there were additional resources going into the other components of the system

Utah is an excellent example, and I think many states.  For many years, the insurance regulations within the states did not require insurers to cover foods.  Insurers considered PKU diets and formulas to be food.  So many families struggled terribly to try to maintain their kids on the appropriate diet.

Well, we passed legislation, the state did, a couple of years ago to require that.  It's now back at the legislature this year with the insurance lobby saying we don't think that we ought to be required to provide these foods.  Well, that's the nature of the intervention for many of these conditions, is dietary manipulation, and if those are off the table for support, then the whole system is going to collapse.  It doesn't matter how good the test is.  It doesn't matter what the positive predictive value is.  If families can't afford the food, then the system doesn't work.

So I think more attention to those larger aspects of the system, I think, may be at the expense now of significantly expanding the number of tests rather than the quality of the service, and I just would comment, too, that I think that what we're seeing in a lot of different domains is a large volume of genetic information that is coming into some very narrow funnels in terms of the clinicians being able to manage and handle and think through that kind of information, and people have to develop criteria in this context and other contexts about what's the minimal criterion for allowing information to flow through, and if it's satisfaction, people want to know, then that's pretty low level, but folks will be happy to provide that kind of information and vendors will be happy to sell that kind of information and so it becomes problematic then to step back and say, "Okay.  We're going to have some clear criteria, and even though we've got information, we're not going to allow it through that funnel."

That's a significant social challenge that's occurring, I think, across the range of genetic testing right now.

CHAIRMAN PELLEGRINO:  Dr. Rowley.

DR. ROWLEY:  Well, I want to make one comment and then ask two questions.  The comment is that use of the Guthrie cards at least in Europe are saved for a long period of time, have been very important in understanding that for infants and children with leukemia, the leukemia has often started in utero, and you can find the same genetic abnormality that you find in the leukemia cells in a two, three or five year old in DNA in the Guthrie cards.

So unfortunately, many states, Illinois being one of them, throws them away after two or three months so that you can't do those kind of studies.  So they have a scientific benefit that nobody expected.

So the questions I have, one of which is related to the comments that both of you made about the insensitivity or the high rate of false positives in some tests, and then the variability among states, and, Mike, you indicated that at least that latter variability might at least be ameliorated somewhat by better tests or better national standards, and I think that that's good, but if 98 percent of the tests are false in one sense, that's a very grave concern.

The second is a different issue, and it more relates to the fact that as you may be aware, the Council has looked on many different issues related to children, child health, and the bioethical issues in the different arenas, and I would be curious. 

If you look at all of the aspects that the Council has considered, is a concern over the problem of genetic testing of sufficient national ethical proportions that the Council should take this on as a major issue, or is this something that in the broader perspective of child health issues is something that it's important that we be aware of, but not that we focus on in a major way?

DR. WATSON:  That's interrogation.

Well, on the latter point about genetics broadly, I would say that it ain't all the same thing.  Diagnostic applications of genetics I don't think have many of the issues that you allude to, but as you move out of dealing with patients and families, you do begin to collide with the kinds of issues that I think are of interest to you.

Carrier screening is an area that is expanding rapidly in the United States.  This is broadly offered to individuals or the entire population.  Cystic fibrosis carrier screening is being offered to everyone during pregnancy now.  If you're from an Ashkenazum background, you're being offered, you  know, probably 11 mutations from six or seven genes in a screening test where your risk is anywhere from one in 29 of being a carrier to one in maybe 80 or so for many of those.

And you know, it's a lot like newborn screening.  You start at a certain place, and then you begin to wonder about the cost benefit.  If that sets your threshold, you know, there's a reason neither Jeff nor I talked about cost benefit studies in newborn screening.  Well, it's because they're not.  They're extremely limited, and those I've seen have almost never acknowledged the genetic aspect of it, which is not just the individual you identify.  It's also that whole family now, and it's possible considerations of prenatal diagnostics, almost never factored into any cost analysis of newborn screening.  I think it needs to be acknowledged in those kinds of things, but it hasn't been.

You think about the quality adjusted life year measure as sort of a standard of when you should do something or not in a screening environment.  Well, $50,000 in a ECHO in the hospital  is sort of the line, and most of these beat the heck out of that.  So it becomes very difficult, you know, especially when you start with something like CF in carrier screening.

I can't think of a worst place to have started in carrier screening, but it was imposed by an NIH consensus conference that said, "Let's go do this."  So we tried to make sure it would at least be done appropriately.

I don't know if that got at all of yours.  Did you start with -- oh, the cards themselves?  They're tremendous resources.

DR. ROWLEY:  No, but the other was the variability in tests and what could we do to make sure that we really have the best possible test, which is a moving field so that those kind of regulations also should move.

DR. WATSON:  Yeah, my feeling about the tests themselves is that the best way to do it is to shine a light on it.  If there is that disparity and at this point in time we have nothing that allows us really to say State A is doing better than State B, and you know, at the end of the day, who really is impacted and will be influenced to make the change?

Well, a lot of false positives obviously impact physicians because they have to manage this, and they impact the families, but I think the physicians may be in the best place.  Once they begin to appreciate a standard, and this is what we did in the oncology groups, as you know, we raised the standard.  We said, "We expect you to be detecting 70 percent of patients with ALL, for instance, in a cytogenetic test."  That's not out there in newborn screening yet where they're told that we expect this to operate at this level.

The states are very independent in the tests they choose to use and the standard they establish for that test.  There's programs developing through the National Newborn Screening and Genetic Resource Center to collect this data and allow some comparison so that presumably evidence of great disparity would drive many of those to improve their programs.

CHAIRMAN PELLEGRINO:  Dr. Botkin.

DR. BOTKIN:  The saving of samples, you know, this is just such an enormously valuable potential resource, and I know Utah is one of the states that saves them for the least amount of time, which is three months, and it's a liability issue.  They say that they save the sample long enough to confirm the diagnosis.  If there's a sample that needs to be retested, fine, but beyond that, they think that liability for retaining those samples longer.

We're trying to convince them otherwise, but the other hobbling factor with that resource is the fact that they were obtained without consent.  So that may mean then that according to at least a number of national recommendations they shouldn't be used without individual permission.  Well, that's a pretty serious limiting factor.

So I think additional thought needs to be gone into how these can be used in a creative research enterprise without the constraints of an individual consent process, and I think that's possible.

And then I think your question about whether this is something that the Council itself would -- well, whether all of us would benefit from the Council taking a look at it perhaps is the right way to look at it, and I guess my short answer would be I don't think newborn screening per se, but I think the larger issues that are coming down the pike in terms of the amount of genetic information, the information that's being generated both through sort of public health as well as commercial sources and how to manage that information in an appropriate way so that there's a fair amount of confidence that that's likely to benefit people at the other end of that pipeline seems to me could be a very beneficial enterprise.

CHAIRMAN PELLEGRINO:  Dr. Meilaender.

PROF. MEILAENDER:  Two just sort of focused questions.  I'd be interested to know to what degree the two of you agree or disagree in your analysis.  One has to do with this rapid increase in that both of your presentations, though configured slightly differently, showed an increase in newborn screening.  Just sort of what your view is and what's driving that.

I mean, it seemed in part to be or somebody talked about parents as a kind of consumer movement here, that sort of desire.  There's also just the technology in the sense that you become capable of doing something and you want to do it.   Maybe there's some state envy at work.  I don't know.

But if it's technology and parent pressure, which of these seems more important?  I mean if there really is this parent pressure, it's kind of surprising that there doesn't seem to have been the same pressure for parents to give consent or permission.

So just say a little more about your own view about what's driving this increase, and then the second, I mean, I don't know if this can be easily answered, but would universal and uniform newborn screening that eliminated these differences among states be a qualified good?

I mean, I can see certain data collection and so forth that I can understand, but are there benefits to be gained from this sort of somewhat more hodge-podge kind of way that we have it, or would it just be better if that didn't exist any longer?

DR. BOTKIN:  Well, great questions, and I think what's driving the increase in the number of tests, and I think you probably mentioned all of the things that I would identify as etiologies.  I think there's a strong social attitude that screening is a good thing, and I see it in the paper every morning with the body scanners.  You know, spend 600 bucks.  Detect disease early and save your life.

Well, there's no data to support any of that, but it's part of the social consciousness now, and I think how that's translated into newborn screening is the strong sense that if you've got five tests, that's good.  If you've got 20 tests, that's really terrific, and any self-respecting state, you know, should not have less than 40 tests on its panel.

In fact, there was a senior pediatrician in Utah who literally at a public meeting said, "You know, I'm concerned that businesses will not relocate to Utah because we don't have enough tests on our newborn screening panel."

(Laughter.)

DR. BOTKIN:  Well, that's how to get the legislature's attention on this kind of thing.  So I think it's that aspect and sort of the social receptiveness to the whole idea that screening works when we know in most circumstances it doesn't.

There's peer pressure, the technology, and then I think some of the lay pressures that are emerging also with lawsuits about this kind of thing are sort of all part of that general picture that, again, any self-respecting state ought to have a whole bunch of tests to be done.

The variation issue I think is a very interesting question that, you know, I've sort of struggled with, too.  I see it as a good in terms of a natural experiment in which if there is variation then you can actually compare the outcome of one state that screens to another state that doesn't screen.

Now, you've got cohort comparison problems there that would need to be dealt with, but at least it's a way of getting around the randomized control problem that bedevils a lot of the basic research in this particular area.

So it's good in that respect.  Whether it's sort of good with respect to ethnic make-ups, racial make-ups within states, I think that's a harder question.  Utah did not screen for hemoglobinopathies for many years because we have less than one percent of the population that's African American.  Well, is that justified or not?

I think a lot of folks eventually decided it wasn't justified in particular because lots of other folks, too, have hemoglobinopathy problems, but I mean, you can push that sort of question.  If you've got a genetic condition that is almost exclusively within, say, the Hmoung population, and Minnesota has a Hmoung population, but Utah has none, what does that mean in terms of Utah's obligation to provide population screening for that condition?

Well, that doesn't seem to make sense.  So I think there's some boundaries that have to do with social justice issues as well as prevalence that are important.

CHAIRMAN PELLEGRINO:  Thank you very much, Dr. Botkin.