Apr 8, 2016


Dr. Schiffman: Genomics offer extraordinary opportunities but we need to use it wisely. More about this coming up next on The Scope.

Announcer: These are the conversations happening inside health care that are going to transform health care. The Health Care Insider is on The Scope.

Dr. Schiffman: My name is Joshua Schiffman, a pediatric oncologist here at the University of Utah and I'm here with our special guest, Dr. Wylie Burke, Professor in the Department of Bioethics and Humanities at the University of Washington. Phronesis, can you tell us a little bit about what is phronesis and how should this be used to guide genomic testing?

Dr. Burke: So phronesis is a concept that Aristotle articulated and his notion was that there is a kind of expertise that is derived from experience, from the right kind of experience, experience that occurs under a morally informed mentorship. And it is this kind of experience that leads to knowledge about appropriate course of action in a variety of different settings. And I was interested in this concept the phronesis because it seems very applicable to clinical medicine.

What we try to do in clinical medicine is to train medical students, residents, and fellows with the right kind of mentorship, the right kind of guided, morally informed mentorship that helps that helps them to derive experience that leads to good practice and, to me, this is a good example of phronesis.

Dr. Schiffman: Can you give us a specific example from the world of clinical medicine that you've encountered, how clinical experience and wisdom should inform genomic testing?

Dr. Burke: I think the most important and relevant example from clinical medicine is an acquired experience and wisdom around the issue of screening. Screening sounds like a really good idea. The concept of screening is to go look for a problem before it happens so that we can either treat it early or maybe even prevent it. Colon cancer screening is an example of that.

We can use colonoscopy to find polyps, remove them and prevent colon cancer from developing. It's a simple concept, it sounds great, but it doesn't always work well. And it doesn't always work well because when you go looking with things in healthy people, sometimes what you find is misleading information, false positive information, information that causes you to act in ways that harm people when they would have been just as well, maybe better off, left alone.

What we've learned from our experience with screening is that we need to be very careful to understand how well a particular test actually predicts disease. We need to figure out how to avoid false positives. We need to figure out what actions are prudent following the information how people should be followed up and in particular because every screening test has the potential for false positives. We need to figure out how to sort out who has a false positive from a true positive before we take action. All of those lessons are relevant to genomics.

Dr. Schiffman: Here at the University of Utah we spend a lot of time thinking about newborn blood spots and certainly through our Department of Health, we clinically, like all of the other states in the country, do newborn screening. What are some of the issues that are relative and specific for newborn blood spots in testing genomics and so forth in minors compared to adults?

Dr. Burke: I think the most important principle to bring to bear when we test newborn or any child is that we wanted to medical testing that will enable us to take action now, rather than thinking about testing for things that might happen later in life. We really don't want to test a newborn or any child for an adult-onset condition when there's no action that we would take before adulthood. Rather. we would like to give them the chance to wait until they're adults and make decisions about testing.

But with newborn screening, there are two particular issues. The first is that we really only can justify screening if we are going to take action immediately. So what we want to screen for are conditions that are urgent. Phenylketonuria, PKU, is the classic example. You need to find kids who are predestined to have PKU as soon as possible after birth in order to prevent the brain damage that will occur if they're not treated.

So that's our really important principle: time-urgent, action. The other principle that's important to bear in mind is the general screening principle, and that is that we are screening healthy newborns, people from the general public, people who we have no reason to think are sick and we want to make sure that our test really works. We want to make sure that this screening process is really going to result in net benefit to the population of infants that we're screening. So that means we have to have a lot of evidence. We have to have good quality evidence that says we understand the test we're using and we understand what actions we should take after the test.

Dr. Schiffman: How do you suggest we handle this situation where we'll actually have the data in hand? Should we look at it? Should we analyze it? Let's take an example you discussed earlier, their BRCA1 or BRCA2 genes. So breast and ovarian cancer syndrome, it's only going to affect women when they're older. Certainly nothing that we would intervene with during childhood. But now, if the technology is moving so rapidly, we'll have that information in front of us.

Should we look at it? Should we blind it and wait till they turn the age when will affect them and then look at the results? What would the implications be for the parents and so forth? So tell us a little bit about your thoughts on now the technology is so good, we'll be getting information that we may not necessarily want to use and how do we deal with it if we have it?

Dr. Burke: So I want to challenge the assumption that we will be getting the information because my first reaction to that scenario is to say, "Let's acknowledge that the low cost of exomes and genomes is a temptation and it's a temptation that we should resist." I don't think it's useful to get an exome on a newborn, but suppose somebody comes back and says, "Wait a sec, doing an exome is really a good way to create the information platform to pull out certain clinical diagnoses that do have time-urgent treatment in a newborn period."

If we do the exome, we can find x, y, and z additional genetic diseases that we can treat in a newborn period. The exomes, the only way to find them we really need to do the exomes. Okay. If that's the scenario, then what we have to do is think very carefully about how information is generated from a DNA sequence. The DNA sequence is not the information, the DNA sequence is a platform from which we gain the information and there are several analytics steps. And we can make decisions about what portion of the genome we subject to those analytics steps. And I think it's an important policy issue to determine how to do that.

So I would say we get the DNA sequence or the DNA sequence of protein-coding genes, which is what the exome is. And then in the newborn period, we make very deliberate decisions about screening . . . I don't know how many genes, it will be 10 genes, 12 genes, maybe 50 genes that will give us information that's clinically important in the newborn period and that's all we do. So we're making a deliberate choice not to generate information that will not contribute to the health care of the child at that time.

Dr. Schiffman: How do you respond, though, with the parent's information which is obviously, in some cases, contained in the child's genome? So if we don't want to look at the information that's not going to affect the child but if we have that information and it could be important for the parent's health. Is it better just to sequence the parents and say, "This genome belongs to the child and separate to two."

Dr. Burke: Yes, I want to really distinguish between having the information or potentially taking action to get the information. If we take what I would call a disciplined approach, then what we're doing is we're creating a particular DNA sequencing platform that we're then analyzing for . . . again, 50 genes perhaps that are important to the child and we're not doing anything more with that sequence, that's our protocol, that's our screening protocol. To say that we "have the information" on other genes is to assume that once you have the DNA sequence, you have to go look at all those other genes.

And we're not going to. There are a lot of genes we're not going to look at. For example, no one would suggest in that circumstance that we should look to see if there's a Huntington's disease mutation because no one is proposing at this point in time that that's useful information. Even though if you found it, you'd know something about the child's future and something about the child's parents. I think we have to think very carefully what information we want to pull from the sequence and what information we don't want to pull.

We can justify the approach that I'm outlining in two ways above and beyond focusing on what's good for the child. One of them is it takes extra work, time, labor, resources to do every bit of analysis that you do on the sequence; it doesn't automatically appear. At this point in time, it takes a fair amount of manual labor in terms of interpreting gene variation and trying to figure out what's clinically significant and it's very hard to justify doing that on genes that aren't relevant to the child's health. And sequencing technology' going to get better.

I'd say get the sequence, pull the information you need, discard the sample. We'll have plenty of opportunity to get additional sequence information that's clinically relevant down the line. The deceptive appeal of personal genomics is this idea that I'm going to find out everything about me and it's going to empower prevention, it's going to make me healthier. And in fact, genomic information doesn't have much utility when we're talking about most of the disease burdens that most people deal with: risk of diabetes, risk of cardiac disease, stroke, etc.

Yes, genomics contributes. We all have some degree of genetics acceptability or up or down to those conditions, but the actions we need to take really have nothing to do with genomics. And what genomics will tell us, most of the time, is, "Yeah, you're a little bit above average or a little bit below average." Good luck with that. If fact, what we need to do is learn how to help people to live healthier lives. We need to learn how to help people to quit smoking, we need to learn how to help people to eat healthy and be active. Really, it doesn't matter what your genome is, you still need to do those things.

Announcer: TheScopeRadio.com is University of Utah Health Sciences Radio. If you like what you heard, be sure to get our latest content by following us on Facebook. Just click on the Facebook icon at TheScopeRadio.com.

Sign Up for Weekly Health Updates

Weekly emails of the latest news from The Scope Radio.

For Patients