NIH CLINICAL CENTER GRAND ROUNDS Episode 102109
Time: 58:22
Recorded October 21, 2009

Ethics Rounds
The Ethics of Using Lotteries as Incentives for Clinical Research
Tyler Cowan, PhD
Professor of Economics, George Mason University

ANNOUNCER: Discussing Outstanding Science of the Past, Present and Future - this is NIH Clinical Center Grand Rounds.

(Music establishes, goes under VO)

ANNOUNCER: Greetings and welcome to NIH Clinical Center Grand Rounds, recorded October 21, 2009 at America's Clinical Research Hospital, the Clinical Center at the National Institutes of Health in Bethesda, Maryland, an agency of the United States Department of Health and Human Services. Today, a special Ethics Grand Rounds. Dr. Tyler Cowan, professor of economics at George Mason University, will discuss "The Ethics of Using Lotteries as Incentives for Clinical Research".

You can see a closed-captioned videocast of this lecture by logging onto http://videocast.nih.gov -- click the "Past Events" link -- or by clicking the "View Videocast" link on the podcast homepage at www.cc.nih.gov/podcast. The NIH CLINICAL CENTER GRAND ROUNDS podcast is a presentation of the NIH Clinical Center, Office of Communications, Patient Recruitment and Public Liaison. For more information about clinical research going on every day at the NIH Clinical Center, log on to http://clinicalcenter.nih.gov.

We take you to the Lipsett Ampitheater at the NIH Clinical Center in Bethesda, Maryland for today's presentation.

WENDLER: I want to welcome everybody to the first ethics grand rounds of the academic year for people who plan ahead or have their iPhones or blackberries in hand, the next one is going to be December 2nd, Wednesday, at noon, also here in Lipsett. Noon to one. And it's a very interesting case of the ethics of doing research in dying patients, so it's an intervention that had a lot of toxicity for neighbor's expected to live for a long time so the thought was what if we do it in people that are not expected to live a long time. Would that be acceptable to do given it wouldn't pose those risks to them? So that's December 2nd at noon.

I'm Dave Wendler from the department of bioethics at the Clinical Center. Today we're going to talk about money, lotteries, gambling, getting rich. Everybody can let their fantasies run wild. There's, as people probably know, a lot of debate, a lot of controversy about the ethical acceptability, the appropriateness, the extent to which investigators should pay subjects to be in research studies. While there's the debate the practice is fairly widespread. You might even say pervasive. My colleague Christine Grady had good data on this, lots of subjects getting paid to participate in lots of kinds of clinical trials.

So today we're going to talk about a particular variation on that theme which is rather than paying everybody using a lottery scheme to pay one or a few people to be in a study to use that as incentive to try to get people to enroll in studies. As people know, I give this talk on recruitment and I talk to these fellows once a year and I always say that when I first start coding this thing I thought the hard things were going to be figuring out good ideas, getting good projects and that's what's hard about doing clinical research.

Once you start doing it you realize what's really hard is getting funding though that's not as much a problem for us and getting subjects. Those tend to be the two obstacles so we're going to focus on one of the primary ones today and how to overcome it and to present us the case is Dr. Valery Gordon, who is acting chief of extramural policy at the institute for bioengineering and bioimaging which I think I roughly -- I thought I was going to butcher it, that was almost right. I was also going to make her head of a small country. We'll see how she does and which country she deserves at the end. Thanks for coming, Valery.

[Applause]

GORDON: It's a lot of fun for me to get to talk to a broader -- a larger audience about a topic that I have been thinking about for the last three or four years. And that is recruitment, specifically the use of lotteries. As Dave mentioned subject recruitment is one of the hardest things that researchers have to do. And according to this Center watch survey over 70% of clinical trials are delayed due to difficulty enrolling enough subjects. There are a number of reasons why that might be the case, we're not going to talk about that today.

So I'm going to briefly introduce this topic and then we have a fabulous discussant who will raise issues about it. So as Dave said, many trials do pay subjects, they reimburse them for time and effort and inconvenience and offering incentives to everybody is often done but it can be really, really expensive. So in clinical research it's expensive. So -- over the last several years we have noticed that -- I work on the extramural side and often work with human subjects concern identified in peer review. And lottery versus come up a few times and since I'm in the policy arena, I look at this in terms of is this going to be a topic that's going to require policy for NIH. So that raises my antenna. So what we have seen several times is the use of lotteries to encourage individuals to enroll in research. So -- and where I have seen it is in grant applications. So lotteries are used in a range of studies. These are surveys. We see this all the time. We get these consumer marketing surveys where you have to fill in a lot of questions about what kind of products you use in your household and then at the end it says your name will be entered into a drawing or something or other if you want. When I fill these out I say sure, why not, I have no particular hope of winning. But I would fill out the survey anyway. But if this was done for phase 1 studies or for observational studies or long-term treatment trials, the question becomes whether that's really appropriate. The level of payment and structure of payment the vary but this is the kind of idea the investigators enroll in a number of subjects and tell them that they're going to be -- their names will be put into a hat. At some point one name will be drawn or a few names. Those subjects will be reimbursed. Certain amount a specific amount and everybody else will get a coffee mug or the sincere thanks of the investigators or something like that. So a lottery does -- might allow investigators with a small budget to offer a reasonable amount of money to one or a few of individuals and some believe a lottery attracts more people compared to if they divided that $500, $1 each to 500 participants as opposed to having 1 person get $500 an everybody else get the coffee mug. This is what I just said. And the amount of money that specific amount is generally disclosed at the time of study enrollment. But -- question, what I wanted to say here. Right. The odds of winning any lottery depends on the number of denominator, the number of subjects who enroll. And the subjects generally don't have any idea what the denominator is but the investigators also may not have any idea as they are doing in their enrollment what the denominator will be. They can say our goal is to have this many. But the drawing will be on a particular date or something. The drawing occurs after all subjects are -- a set number of enrolled so nobody is going to know exactly -- they may know when it will occur but they might not know when it will occur and might not be able to calculate odds of winning. When I found out I was allowed to call an ethics consult, I did. Because that was a cool thing to get others to discuss the issue with me. The first thing that was raised was concerns about payment in general. Offering incentives for research enrollment can be regarded as suspect by some people that maybe that sort of destroy it is altruistic gesture that we hope for that -- the altruism that people who enroll in clinical research are supposed to feel when they are getting involved. They shouldn't be doing it just for the money. But also if one of the other concerns that lead individuals to ignore risks and that that's undue influence of some sort fail to report side effects and enroll for the wrong reasons or something like that. Those will general concerns about payment. There are concerns about lottery specifically and raises the same concerns but also additional ones. Participants may feel compelled to stay in a study if they feel that they are -- their name will be out of the running if they withdraw early. They might ignore serious risks for the opportunity of getting that what they consider a large amount of money. Investigators might ignore other reasons they should be reimbursing or compensating subjects. Certain groups are more likely to enroll, risks would be unfairly distributed, raises issue and the external validity of the trial may be in question whether you have a particularly risk taking population.

So there are questions does the effectiveness of the lottery rely on the inability to reason about their odds of winning and does the fact that there's only one or a few winners effect treatment and I'm not talking about treatment like medical treatment but how the individuals are -- how they're viewed by the investigator then there's do we view lotteries in a negative light? I would like to even broaden this question a little bit just yesterday, I won't be disclosing any confidential information. We had an application reviewed for human subjects concerns where participants were going to undergo certain number of research procedures an after each one they were going to spin a little wheel that determined what their compensation would be for participation in that procedure. A little wheel of fortune thing. It's another kind of gambling element for research. For that particular one we weren't so concerned because the range of values was very small. And -- but the investigators seem to be doing it just for fun. You could imagine if the range was large or that you could -- well, you couldn't do a negative number like you owe the investigator money. But certainly don't get the trip to Hawaii or something.

So the question is, does gambling have a place in clinical research? Then limited data suggests that lotteries in the clinical setting don't really increase enrollment. So the question is should we be using it, should we allow it? There we go. So finally, should lotteries be used or allowed if -- is there a problem with introducing having a gambling element in clinical research? How should they -- if the answer is yes how are they structured? And does the permissibility of lotteries depend on the level of risk or prospective benefit of given study? That is really -- no, I guess that's -- guess that is where I want to end. Yes. So that's the question I want you to be thinking about. And I hope that we can discuss. Thank you.

[Applause]

QUESTION: Are these legal, is it legal to use lottery?Anywhere? Is it legal to do this?

GORDON: So the question is, is it legal to include lotteries and we did have one -- I remember one application specifically where that was a concern raised about the use of a lottery and part of the response of the investigator went to the gaping commission and asked them to endorse it and the gaming commission said, "we don't get involved in research, we don't care what you do." To the general question, is it legal as far as I know with its informed consent is in place then there's no reason not to. Did you want to add to that?

QUESTION: I just want to say as I understand it, lotteries are regulated by state law. So each state has different laws about not only whether but how you can use the lottery. So at least in some states it's legal, whether in the state where the fabrication that I know about came from, state law did not prohibit it.

QUESTION: Mine is a question of definition. Gambling, the gambler has to put money on the table or buy the lottery ticket to keep that analogy going, risking that money if he or she does not win. What do you define here as what's the anti? What's the potential subject going to lose or risk by entering into this type of a proposition?

GORDON: It's the -- the question is what's the risk. The risk is the risk of participating in research. It's the risk to the research and to the research enterprise that we're talking about that, that could be affected by the lottery. But you're right. They don't have to the lottery ticket but they do -- they put themselves at risk by participating in research if there are risks associated with the research.

QUESTION: That complicates it because unless you have an experiment where there's 100% chance of each subject occurring the risk you're also anticipating that you as a potential subject are going to be the one that has harm happen to you. So you have a double… I don't quite see the analogy to gambling at all here.

GORDON: The only way I can put it, I don't know if this will be persuasive, if time is money and research takes time that's what's your --

WENDLER: Any others? Alright. Thank you, Dr. Gordon. So to help us think through, John you got a quick one for Valery? Okay. Last one.

QUESTION: So there are a series of concerns in the beginning that are definitely valid but I'm not sure which one specifically apply to lotteries as compared to just giving monetary compensation. So seems to me that what is peculiar about the lottery is the fact that one individual is going to get a lot and the others are going to get nothing. But some other concerns are general.

GORDON: Right. There are concerns associated with pain compensation in general. But the question is does the use of lotteries, the inclusion of lotteries add additional concerns or do you consider it to be the fact that one individual might benefit and everybody else might not, benefit monetarily? Does that make a difference? That is the question we're asking. Thank you.

WENDLER: So help us think through, explore, maybe try to answer a few of these questions we have with us today Tyler Cowan, Professor of economics at George Mason. Some people may be familiar with Tyler's name because he writes one of the best most entertaining informative food blogs on dining in the D.C. area so if you ever decide where to take somebody for dinner particularly if you're interested in more ethnic type meals you should definitely consult the information that he provides. Our department takes food very seriously, maybe that's the primary reason we invited him. The secondary reason why I invited him is Tyler spent a long time thinking about and writing interesting things about the use of incentives in an astonishing range of life and activities. So he had a book called Discovering your Inner Economist, I don't know if that was intended to be a frightening title but the fact that there's an economist in there somewhere strikes us as a little frightening but it offers insightful interesting sometimes challenging suggestions on how to use incentives and structure, incentive structures in such a way to help yourself do better enjoying paintings, ordering better at restaurants, having better trips and better relationships with everybody in your family. Welcome, Tyler.

COWAN: Thank you. Let me start with my conclusions and explain how it is I got there. First I think using lotteries to recruit subjects is somewhat unethical. That's my first conclusion. My second is that it's not necessarily more unethical than other things you're doing right now or what you might do in its place. So in that sense in relative terms it may look better.

I'll start by noting the irony having a group of research scientists worrying about the use of lotteries. It's a well known fact that research science itself is in many regards like a lottery, that people enter into the activity, they're not sure how many competitors they have. They don't know what the odds are. There's a lot of evidence they overestimate their chances of success. Then at the end you have some winners the people who get tenured positions, you have an awful lot of losers even among the group of tenured winners most citations go to a relatively small number of people, most people think in advance that will be them. So even if it's not truly random they make decisions on the basis of it being random. So the bottom line is science gets a lot more effort, talent and ingenuity because it's run like a lottery and everyone flocks to do it.

Adam Smith in his Wealth of Nations in 1776 put forward the general theory of professions where he noted people in general overestimate their chance of success when they undertake risky prospects. So If you look at the ethnic restaurants in Bethesda that you enjoy, you have as much choice as you do and the prices are as low as they are because those of course are run in a way like a lottery. It's not called the lottery, what gives you the choice in the low prices.

Lotteries are everywhere. They may take different form, framed in different ways, I'll get back to this point but what people are objecting to is the framing of the lottery. Much of what goes on in science is based on lottery like conditions. On one hand it's an informed consent, people know at one level they may not be the ones who succeed but at another level they don't understand the odds in terms of that reflected in their behavior. We can also draw an interesting distinction between where you might use the lottery. The case for using lotteries is that lotteries at least possibly are cheaper. So one way to use lotteries is to pay the lottery to the subjects and maybe you'll recruit them more cheaply. An alternative which is not currently legal but you could imagine it being the case, it's a good thought experiment. Let's sigh the NIH were allowed to compete with state lotteries and would issue NIH lottery tickets. And the NIH would turn a profit on this lottery as do state government and the revenue that you would take in you would use to fund experiments. With that money you might also recruit subjects paying cash. How would you feel about doing that? I'm sure the NIH could make money at this venture because state governments with monopoly, only Catholic Church Bingo is allowed to compete with them in most jurisdictions so the state offers like 40 cents on the dollar. The NIH could come along and offer 55 cents on the dollar and make really a lot of money. Most of you when you hear this you don't feel like the NIH should be running out demanding this be the way to fund experiments. You seem to not like lotteries in general, it's just not only lotteries being paid to subjects. But think about it a step deeper and look at the venture capital market and look at new start-up biotech firms and the finance literature on initial public offerings.

It happens to be the case when you have venture capital in biotech companies and initial public offerings those are risky assets. They have a lot of properties of lotteries. In fact, biotech companies can sell them for a higher price than would rationally be the case because people seem to be willing as investors to take on a lot of risk without necessarily being compensated for it. And the finance literature there are -- literature studies showing this. When it's done in a high status context you hear venture company, biotech firm, all the everyone thinks it's wonderful let's have more medical innovation, great. And there is a lottery like element is and because the lottery like element is in a way cloaked in a high status form. People seem really quite happy to support it. Even if it doesn't always give investors the best odds. But what I find striking about the normal lotteries is how low status they are from the point of view of most people. If you want to consider the selection biases from lottery there's clear evidence on this from Phillip Cook and Charles Clotfelter and who are the people that buy lottery tickets? Well they are on average they're men, minority, laborer, Catholics, the middle aged, and gamblers. Those are the people who are especially interested in lotteries. And whether fairly or not, when you put those qualities together and you have like a minority male working class Catholic middle aged gambler, at the risk -- I'm talking about societal prejudices but this is seen as something relatively low status. You look at what who runs the lotteries, its state government which is are somewhat slow status to begin with and lottery commissions are especially low status because they tend to be a bit corrupt because they have big pools of cash like the Alaska state government with the oil. When you put together state governments and big pools of cash and weak accountability you get a lot of corruption. So both on the demand side and the supply side we have low status lotteries, which people and like research science or initial public offerings we have high status lotteries where people are generally quite enthusiastic about these.

Another reason I think recruiting subjects through lotteries can be somewhat unethical but not necessarily more unethical than what is done now, has to do with informed consent. You might ask when you offer subjects the lottery, are they expressing informed consent? I don't think it's obvious that they are. One problem is you may not know the odds of a lottery but even if you know the exact odds and report them, there's a lot of evidence from economics that people just don't understand odds. Even very intelligent people, economists, sometimes economists who work in expected utility theory. When you give them concrete problems they will get odds wrong, won't understand the independence axiom, they'll make a number of other errors. So what exactly informed consent in this context means is unclear, I recall just about an hour ago driving into the NIH, if any of you parked in garage P-1, there's a sign P manufacture 1 that struck me that said something like if you park in this garage you may get bird droppings which will ruin the finish of the paint on your car and the NIH is not responsible, proceed at your own peril. They didn't tell me the odds. I was very confused by the sign when I first read I thought my goodness where are the birds? I parked my car, I was looking around for the birds. As far as I could tell the whole thing is an underground garage. So I'm still puzzled by the mention of the birds but the image I have in my mind is not that I objectively try to calculate what's the probability of the birds but rather I thought the fact they're mentioning the risk of the birds should get me worried. So I was already in this mode possibly irrational behavior. Seem it is subjects no matter what you tell them they react mostly to the framing and the fact they're being told something at all and I'm still wondering why there was the sign up about the birds when the whole garage seems to be underground, is there like a ravens on the third floor or something, still not sure. There was another part of the sign that maybe makes more sense and said there could be leaks from the roof. Leaks from the roof ruined the paint on my car, NIH is also not responsible. I think possibly the more rational conclusion is just to think that the NIH is relatively careful about informed consent. They were doing their best to give me informed consent and I appreciate that. But on the other hand the net effect perhaps was to mislead me because I'm used to the fact that when I'm warned about something that just means I start worrying about it. So did I have informed consent parking my car in P-1 garage? I guess I did but how much better should we feel about that? Seems I responded irrationally to the information I was given. It's also the case when you don't give people lotteries to do something. They still respond irrationally. That's one way even if lotteries are unethical the alternatives are also somewhat unethical. There's papers and experimental economics where simply the mere fact you give people a reason for a request it changes how they respond. If you go up to people at that time Xerox machine and say can I cut in line, end of sentence you get a lower rate of response. When you say can I cut in line because I really need to Xerox these papers you put on a cause. Doesn't matter that much because, may I cut in line, dot, dot, dot, because you have given them a reason. You haven't paid them with a lottery. Which is the tricked response? When you say can I cut in line or when you say can I cut in line because? I'm not sure, but seems to me when you offer up the because that's actually when you're tricking people that we're programmed to respond to reason even silly or non-existing reasons but when you give people reasons in part you're tricking them.

Also a lot of people do things or they partake in studies for reasons of altruism or maybe they think God will reward them for good behavior or they think virtue is its own reward. These beliefs may or may not be true but it's not obvious to me that when people respond to the lottery that they are being more tricked than they respond to simple requests based on altruism, virtue, requests interpreted in religious terms, request that simply pretend to have any reason at all why the subject should participate. If I think of my presence here which I guess in the broadest sense like as an experiment we're all undergoing, I was brought here by what's called the wage model in this literature. Not the market model, the wage model they pay you something they don't actually pay you that much and I asked myself why did I come to give this talk? I think the reason is you never are sure about your own motives. But that I really like the idea of affiliating with high status institutions and high status people. The NIH is clearly one of those. Since I had to a chance to come here and not much money but to affiliate with people in high status groups that I would do it, is that a correct response or is that a fact of me being tricked that we involve in hunter gather society to affiliate with high power individuals that was useful back then. It's not so useful back now. I could sit at home and pay off my mortgage. I don't need y'all as allies in the hobesian state of nature. Here I am maybe to some extent being tricked so what really informed consent means is a tricky question but I haven't seen an argument that lotteries are worse on that score than the other ways that subject participants are recruited. There tends to be kind of a grass is always greener fallacy when really it needs to be comparative. And again, I go back to the simple truth that lotteries as we observe them, they really are very low status and we tend to look down upon them because they're low status. This gets me to what is a possible reason to not want to use lotteries. If part of the continued existence of a hospital research team, academic career, NIH department whatever is the continuing portrait of that group as being high status, lotteries will always be viewed as low status, maybe the real reason not to do them is the reputational reason. Lotteries will be reviewed as status because they are, because they affiliate ourselves with lotteries will be viewed as low status. When you present people with an idea that possibly low status and there are morally questionable aspects of that idea. What you find people doing is coming up with reasons and objections and reasons and objections are partially true, but maybe what's driving it all is this intuition we have about lotteries being low status.

To conclude let me just say a few words about the experience of economists with lotteries because this has come up in economics. It came up some while ago, not regarded as that much of an issue. Experimental economics is a major branch of economics. And it recruits subjects but it has a very easy time doing so. Basically undergraduates are a captive audience. You can pay them small amounts. It doesn't have to be a lottery. You pay them small amounts they show up, they have fun. The chance you're doing any harm to them is quite small. Much less of an ethical issue, it's much easier to get them. In the early days of experimental economics people tried lotteries to recruit subjects, didn't seem it made that big a difference. Most people don't do it but if someone did do it I'm not sure it would be a big deal. It's kind of faded away. It's viewed -- I think having asked a few experimental economists it's viewed as a non-issue. And again most people don't need to do it.

The other thing I would stress about economic frameworks is that you need to be careful, very careful when you apply economic frameworks to biomedical ethics. Because the two frameworks normatively are in some ways die met trickily opposed. For instance it's a common economic principle in a lot of economic problems you want to do taxing the inelastic factors of production. That's a little jargony but mean it is following: Economists generally say taxes on labor are more efficient than taxes on capital. Because it's easier for capital to get away from taxes on capital therefore economists tend to favor a value-added tax on consumption which ends up being usually a tax on labor. So the economic intuition normatively speaking is to tax inelastic factors, the factors, and people, and groups that can't run away.

It's probably a good assumption for problems but it's usually a bad assumption for problems in biomedical ethics because the inelastic factors are the desperate people that do not have a lot of options and what's economically efficient according to the economists framework is basically to tax those people because they have nowhere else to go, they can grab resources from them without occasioning resource distortion. Very often this is ethically incorrect even though when an economics framework tells you that's what we should do. So I would just say when you see economists doing biomedical ethics you need to be very careful, you need to look at what are the underlying assumptions. But go back to what strikes me about the whole problem. Lotteries are everywhere. We use them all the time whether we admit it or not. Some lotteries are more fair than others. Informed consent is always tricky. We tend to worry a lot about low status lotteries. We have these worries in part because they're real problems but in part simply because they're low status. So from my point of view we have not cleared up all of the ethical questions surrounding lotteries but I think it's more important to view those on a broader comparative framework, ask our lotteries are any worse than a lot of other things we do but odds are using lottery aren't such a huge gain either. Those are my remarks. Thank you all for listening.

[Applause]

WENDLER: Does anybody have any initial questions? If you have questions if you'll go to the mics preferably, if not, give me an opportunity to repeat it. Christine.

GRADY: I want to Tyler, thank you for your remarks first of all but I want to ask how you got to your conclusions because you started lotteries on research are unethical. All the arguments that you gave the only one I can think of that got you to that conclusion was the one that maybe boils down to something like recruiting people into research is unethical because informed consent is always a sham or something like that. Is that the path to lotteries are unethical? It’s no worse than anything else we do. You said informed consent is tricky. Somehow giving people reason is tricking them and therefore I guess the question is, is participation in research ethical?

COWAN: That's a good start. Keep in mind I said somewhat unethical. The difference between somewhat unethical and unethical is maybe a big one and maybe I'm too much the economist here but as an economist I think on the margin unethicality is a matter of degree and it's a question how much one is willing to tolerate. It could be when all is said and done there's less informed consent with lotteries than other methods, it's possible but I've never seen an argument that's the case. So I would say in biomedical ethics informed consent is an important standard, less a standard in economics but for my biomedical ethics point of view, I would just say any initial unease is not dissolved by looking at how people choose lotteries which is not very rational. The question of what do you do with that? I'm not saying unethical, undecided. I'm saying somewhat unethical exactly because of informed consent like you mentioned.

QUESTION: So I had a question, I guess I was wondering about the conclusion you get that maybe because it's not as bad as other things that we do informed consent that the problem can't be dealt with separately. But I would like to give a reason why I might disagree with that. I can imagine one of the analysis problems with informed consent is therapeutic misconception. This is another irrationality, a person believe business participating in a research study they get certain medical benefits that actually don't exist or the purpose of the research is for their personal benefit which it isn't. That probably leads to bigger problems in terms of signing up for a study without proper informed consent. It's a bigger problem than lotteries I mean. I think we could all probably agree that's true. But I think then we can say why should we focus on lotteries is because it's a lot easier to fix the problem with lotteries than to fix the problem of therapeutic misconception and other problems like that. And that's to say we can prohibit lotteries from being used as compensation, therapeutic misconception if they treat the language just right, you have to make sure the effects are precisely controlled for, that's a very, very difficult. We do a lot of work trying to make that as good as possible. But it just seems that it's easier if we know the solution to fixing the somewhat unethical process of lotteries and this also might not have much recruitment given the literature, why don't we just do that?

COWAN: That's pretty close to where I come out. Namely, it's impossible to root out all the different unethical things that are done. Choose some easy targets, at least start there, maybe go further but that's a separate question. In part this needs to be done to maintain the high status of high status research organizations. And lotteries are visible and since they're so low status you get a lot of weight with people seeing you have done something to combat affiliation with a somewhat unethical, low status activity. If that's where we end up, fine. But to think about it critically that's what we're doing. And I don't want people to think that like lotteries are the problem and we somehow driven out all the unethical things we're doing. I'm here to give an honest outsiders perspective, someone that does not do any medicine at all in any guise.

QUESTION: I thought it was interesting how you suggested lotteries are everywhere. And their high status and low status rooms.

COWAN: And in this room you're all the winners.

QUESTION: I feel very lucky. But what I was confused about, what are the differences between high status lotteries and low status? It sounds like you were suggesting that part of it is being explicit about the fact that something is a lottery. The more explicit we are, the more it seems that that's the case for low status lotteries rather than high status lotteries. If that's the difference which I may not be, should the NIH sort of bow to this idea or this maybe societal problem that we are concerned about being explicit about lotteries and once they are explicit we make them low?

COWAN: I think your suggestion that the degree of explicitness matters is very important. I think also just who plays them. Doctors are high status people for other reasons. If doctors play implicit lottery, that lottery is pretty high status too. You see this on TV shows the doctor takes a radical chance, he's brilliant, turns out to save the patient. Everyone watches the TV show and they applaud at the end and laugh along the way. The people who play state lotteries for reasons I mention redirect examination less high status than are doctors so on one hand there's more lee way, but that lee way is also dangerous. There's a greater chance maybe of abuse because there's more protext. So we need to think about why are some lotteries more high status than others. Do we want to put forward a lottery which is maybe okay but the public would object to but we get it through anyway by cloaking what people actually consider is low status nature? I'm not sure if that's violating publicity or something but it's a question worth thinking about. If we could run the same lottery and make it look higher status while keeping the same structure to the pay off should that be okay? I guess I would be worried about that. But I haven't really thought it through.

QUESTION: I'm worried about the lottery undermining the science and my worry has to do with selection effect but not the fact we're selecting for low status folks. The nature of the lottery is of course you're selecting for risk takers, somebody who would do it for 1 in 500 chance, $500 that wouldn't do it for the expected value of $1 is a risk taker. People who are risk takers in money are likely to be risk takers in health behaviors as well. So you're pulling into your experiment a lot of people that are unlike in all sorts of important ways we may or may not be able to control for unlike the population we're trying to extend findings to. Two questions. One, do we have evidence on the extent to which people who are risk takers in one dimension of life, in money are going to risk takers in other dimensions as well, to what extent does it carry over to what extent is it a problem, and secondly is there any way that once we've got this bias sample we can statistically control out the effects in analyzing our data?

COWAN: Those are very good questions. It seems in the evidence if you're asking about these two particular questions is taking risk with lottery correlate with taking risks with your health, it absolutely does. But it really does run through the person being low status. Low status people take more risks with health overall, what's causal and not is tricky but in terms of correlation it's clear. And those are the people you're getting. But I don't think you're getting them because you're selecting for risk taking. Like buying a lottery ticket is not that risky. One of the safer things you can do compared to riding a motorcycle, fighting in a war or flying to Rio De Janeiro. Whether these people are the true risk lovers is really unclear but they are with their health because they're lower status and social status predicts health fairly well better than most variables sometimes better than healthcare. How well can we adjust for that, I have never seen anyone try to do it, but I don't think in principle it is hard. That is to get a sample of people who’ve bought lottery tickets, such samples exist. Some state government collected these databases given them out to researchers who will find conclusions favorable to state lotteries. The database are there. You can look at the people compare them to cohorts and see how much less healthy they are and work backwards. But when you put it that way, like is it worth the effort? My guess is no. And the gain to using lottery given you need to do this to adjust for selection bias is, you know, I wouldn't expect there to be huge gains there.

QUESTION: I'm one of the Catholic chaplains in the clinical center so by definition I don't have a problem with gambling, we built cathedrals with bingo, and that sort of thing but when you mention Bingo, there's a chuckle, and when someone mentioned earlier about the spin wheel on the lottery and see if you get the prize after you have the MRI, there was a chuckle with that. There's an unseemliness to this. Because it's unseemly doesn't make something unethical we have an obligation to protect the brand of NIH in a way we would not be associated with this kind of activity in this sense because of as you characterize it the low class or what have you.

COWAN: That's what I was saying. That's one of the main worries. If the benefits are small, and it puts you lower in the status hierarchy, given that NIH on net does a lot of good, it seems to me the net benefit is probably negative, to net cost. Why the Catholic church finds bargain worth taking is an interesting question on its own. I think it's because a lot of donors to the Catholic church which in this country, not around here but in a lot of the country, it's blue collar institution or appealing to immigrants. That affiliation is not those people seen as especially low status. It's actually bringing them closer in. Maybe that's why the Catholic church is willing to make this exchange of status for cash and maybe the NIH shouldn't.

QUESTION: One of the essential features of lotteries I think is convincing to keep them honest is convincing people that the pay-out was made. And it's usually done by publishing the name of the recipient to prove. That would of course be a problem for medical lotteries because of the confidentiality that's required. Any thoughts on that?

COWAN: My guess is that the people who show up for these things have an awful lot of trust in doctors. American versus a lot of trust in doctors to begin with. And the people who show up for experiments at all have even more trust in doctors. And they like the idea of affiliating with these high status individuals. So I'm not sure that you would need to publish something to convince them you really paid it out. But it seems also there's a way you could just like put a cashed check on the web while covering up who cashed the check and -- I mean, that could be a trick too. But you know, the moon landing could be a trick-- my guess is you can get people to believe you actually paid it out in some manner or another without violating confidentiality. I wouldn't be too worried about that.

QUESTION: Thank you.

QUESTION: I think this is has been alluded to earlie but do you think one issue might be if people are more likely to play lower lotteries of status and in many areas of their life as more vulnerable that we would be perceived as exploitative to use this mechanism -- exploitative to enroll people?

COWAN: It could be perceived that way but again it gets back to economics and bioethics. In economics the tendency is to say it's not deep informed consent but they did it and people are faced with lotteries all the time. It's not a unique processing problem that they have never seen before. They did it, what the heck, let them do it. Bioethics I know less well but it's more problematic under the standard of biomedical ethics. What should rule here, standards of economics of biomedical ethics, you probably think biomedical ethics maybe I am a bit more favorably inclined toward economics but again if public perception is the main issue you don't need to necessarily resolve that dispute. But if you take poor people who would play the lottery in anyway, you sell them extra lottery tickets, that doesn't personally bother me so much. Even though I don't think it's exactly deep informed consent either. One interesting question you might have wondered about in the literature, if a person is sold extra lottery tickets does he or she cut back buying lottery tickets elsewhere? According to Phillip cook the only paper I have seen, Phillip cook says no. At the margin you buy more lottery tickets you buy more lottery tickets so you might think you're displacing state revenue but as far as I can tell from this one paper you're not. Like at the margin extra risk people end up accepting.

QUESTION: I would like to move back from the perception field to the maybe more substandard field, not that perception is not important. In the -- let's start from areas of agreement and maybe moving areas where we don't agreement there is a general principle I think and everybody would agree in this room, that people should not be -- the amount of reward should not be high enough to be coercive. We don't want to force people by giving them -- if I submit a protocol where I'm giving a thousand dollars for a blood draw the IRB says it's too much. So the principle is people are entitled to be rewarded either monetarily or in other ways for what they do. So does the lottery -- I guess the answer should be case by case but does the lottery infringe any bun of those two principles for them, a lack of a word for others?

COWAN: Two responses. One this brings us back to the difference between economics and biomedical ethics perspective. The economists high rewards great. No problem, no coerce -- no coercion. People chose it. From the point of view of biomedical itself ethics maybe not. If you look at how it's run it's a common pattern. Well run lotteries tend to have extremely high prizes for the number one winner and then a big drop off for all the prize beneath that. And a larger number of prizes. That suggests to me that when people hear the absolute size of the prize their primed in a certain way, in terms of order of magnitude. They don't really adjust properly for the odds of winning. They're in this way tricked. They hear the prize is $200 million they don't think it's spread out over 30 years or chance of winning it is so tiny, whatever, they think in terms of $200 million and they apply the category small chance to $200 million. If that's the way people behave the greater the extent that you have very narrow odds and really huge prizes, that's when you're tricking them. Given how many people partake in these experiments relative to lotteries it's a small number that you have. So people are getting small odds, so they're not tricked into thinking too highly of the offer, they're in a way tricked to thinking too lowly of the offer given the other evidence we have on lotteries so in that sense I'm not so worried about the coercive aspect if it's a 1 in 40 chance it's quite possibly they're undervaluing in what it's worth. It's the 1 in 200 million chance they overvalue.

QUESTION: But you don't know at the beginning if it's 40 or 200,000.

COWAN: Well if you're participating in a medical experiment it's unlikely you'll think there will be 200 million other people in it. You may not know if it's 4400 or 200 or 157 but odds are it's a pretty small numbers problem and you'll be underestimating what your chance if the reward is like. That's my guess.

QUESTION: But there's genetic studies to where you can enroll a very large number of parents.

GORDON: We talked when I had the consult at the bioethics department we talked a little bit about the utilitarian perspective, the happiness of the one winner counter act the relative unhappiness of everybody else who doesn't win and is there a real justice issue. The consensus seemed to be although I don't know whether it's completely comfortable with it but the consensus seemed to be that since everybody knew up front there was going to be one winner and what the amount would be, they are all equally informed equally informed to begin with. That negated that justice question whether it was undue inducement or however you're talking about it. In terms of that one person who wins the large amount and then the rest of the people who don't win the large amount.

WENDLER: So there's equal treatment in one sense that everybody basically has the same exanti odds of winning there's obviously not equal treatment in the results that one gets it and everybody else doesn't. Did you want to follow up on that?

QUESTION: I wonder whether for that specific individual that wins the money has providing he understands correctly the odds and he wins and we're giving him a thousand dollars for a blood draw. Is that ethical? In that specific case?

GORDON: Well, if it was given before participation, I would say that would be a bigger problem. But after everybody has participated and everybody has had that blood draw and everybody -- he may be the only person who understands the odds but everybody has had equal probability of being the winner. Yes, he is compensated extra. But everybody knew that one person was going to be compensated extra to begin with.

QUESTION: But did we do any wrong to that specific individual? That's the question ethically.

COWAN: To winner? How can you do wrong to the winner, you won the research lottery, you work at NIH and you're successful in the medical field. Did we do anything wrong to you?

QUESTION: Let's say it involved some risk and the person participated because he wanted to win and does win and does get the risk. The person could tell you had I known there was no possibility of winning a thousand dollars I would have not participated. Now I have hepatitis as a result of this.

COWAN: But that's not a problem unique to lotteries, that's unique to whatever you do to get people to participate.

QUESTION: This was my initial point.

COWAN: That applies equally to all methods of subject participation.

WENDLER: Last question.

QUESTION: Lotteries if they're accepted and become adopted widespread, I don't see -- I see a lot -- from an economic perspective I see it to be highly problematic. If you had five trials for Lupus for example, and they were all lotteries, then the investigators would try to essentially coerce people to come to their trial and you would -- it would be a competition for the prize money. Seems like there would be just a increase in the pay-out. Then policy would kick in and there would be problems there. So this seems to be, maybe an issue.

COWAN: Sure. This is exactly true. The practical benefits of lotteries that you can get people more cheaply, if only one someplace doing it they have a quasi monopoly but the more places that do it that advantage erodes and ex-post the competitive outcome is easy to get people with the lottery if not and the lottery doesn't have an advantage. It evaporates over time. And this gets back, to not like a practical advantage and there are possibly some other problems maybe just he's off a bit. That's exactly right.

WENLER: So join me in thanking Tyler and Valery.

[Applause]

WENDER: We'll see you December 2nd.