PSYC 123: The Psychology, Biology and Politics of Food
|Transcript||Audio||Low Bandwidth Video||High Bandwidth Video|
The Psychology, Biology and Politics of Food
PSYC 123 - Lecture 12 - Public Health vs. Medical Models in Nutrition Change: Saving Lives One or a Million at a Time
Chapter 1. Antibiotics in Meat; Differing Approaches towards Food [00:00:00]
Professor Kelly Brownell: Okay, good morning. I’m assuming you guys are all preparing very hard for the mid-term on Wednesday, are totally on top of all the material. Just a few logistics of about that, as you know, the exam will be in this room and it will be during the allotted time for the class from 9:00 until 10:15 and that should be plenty of time to finish up the exam. You’ll notice — and you probably have noticed this already — if you look around the room there some spots in the room that are lighter and others that are darker just because of the way the lighting happens to go, so when you come in for the exam you may want to try to find a seat in an area that’s one of the better lit areas. If you have any questions about the nature of the exam, please see me after class or talk to one of the teaching fellows.
I wanted to follow up on some things I meant to get to in the last lecture but we ran out of time, and just go over several of those very quickly. Then we’re going to launch into the discussion today of public health. The two things that I didn’t get to in the last lecture were two. You won’t be tested on these for the exam, so that’ll give you a little bit of a break, because I didn’t have a chance to go over them in detail in the lecture.
One of the concerns, one of the issues pertains to the use of antibiotics — heavy use of antibiotics in farm animals, especially animals like chickens and things like hormones that go into farm animals like cattle. The concern with the antibiotics is that so many — such heavy use gets made of these, especially in animals that are in close quarters, which as you saw from the discussion of industrial farming, is happening more and more. These animals, clustered together, can get sick and spread disease in a very short period of time.
The farmers found that you could give antibiotics to the animal’s prophylactically, that is in anticipation of getting a disease rather then waiting for a disease outbreak to occur, and then found almost as an aside, that the animals grew better with the antibiotics. The heavy use of antibiotics allows the chickens to thrive in more of a disease-free environment, promotes growth in the animals, but the environmental concerns are that the heavy use of the antibiotics gets — the antibiotics gets into the groundwater, goes from place to place via the water tables and things like this. Therefore resistant strains of bacteria develop.
There are some studies of people who are around, who live in the vicinity of some of the places where there’s a heavy concentration let’s say of poultry farming. So for example, one of the places where this happens in greatest concentration, Arkansas, is a big chicken raising state. Also there’s a part of Maryland called the Delmarva Peninsula, which is the Delaware, Maryland, and the Virginia confluence, and Purdue chickens is in that area, but there are others as well where researchers from Johns Hopkins have studied this resistant strain of bacteria in that area, and have found what they believe is very alarming findings. So of all the different concerns with modern agriculture and the modern raising of animals, the heavy use of antibiotics and the resistance that people can develop to these antibiotics and — I mean resistant strains of bacteria that develop in response to this have become a very real issue.
Then the other thing that I mentioned at the end of the class is that — and we’ll come back to this later — people are interested in food from a lot of different points of view. So if you look at the people that are interested just in the issue of sustainability; you have people that are concerned about the environment; you have people that are concerned about biodiversity; people concerned about animal welfare; and then of course you got the whole public health overlay with people concerned about the effects of food on physical and mental well being in people.
All these different philosophies or conceptual approaches are passions that people have get played out in different ways people organize, so there are different groups that represent each of these different areas. One thing that hasn’t happened is that these groups come together around a common goal: the common goal would be better food. We’ve talked at The Rudd Center about possibly convening leaders from these various groups to come together and talk about how they each could band together and form a stronger, more united voice if a unifying theme could be found, and that is people need to eat better food. Later in the class, we’ll talk about how those groups might come together, and what they might do.
Chapter 2. Defining Public Health and Epidemiology [00:05:05]
Okay, some of you may have seen yesterday’s New York Times Sunday Magazine. That just so happened, it contained a number of excellent articles on issues pertinent to the class. Michael Pollan had an article, Mark Bitman, a well known food writer had another one and you see the — there are more articles in the magazine than just this, but these are the ones that I thought were most pertinent to the class. If you haven’t had a chance, I urge you to get a copy of The Times Magazine, you can read it right online if you wish, you don’t have to go find an actual copy. Some of the articles are really interesting and have up-to-date information on some of the issues like the potential of a Green Revolution in Africa.
I thought I’d start off today with something fun. Some of you guys will have listened to one of the NPR shows on the weekend where they had people read crazy tales, and the contestants on the show have to guess which one is true. Well I’d like to do a little thing like that today, so I’m going to show you four things and one of these is true, and I’d just like to see a raise of hands to see if you can guess.
The dairy industry did something in 2006 in San Francisco. Option one is that they put a device in bus shelters that released the smell of fresh-baked cookies, in order to encourage people to drink milk that might be associated with cookies; number two, they tested a campaign where children were given free milk bottle costumes and milk mustaches for Halloween; (c) they tested adding subtle milk flavor to sweet non-milk products like soft drinks, in an attempt to increase desire for milk; or (d) they tested the use of electronic signals that stimulate the part of the brain that gets activated during the act of breastfeeding to increase desire for milk. All right, so how many of you believe (a) is the right answer? All right, how many (b)? All right, (c) and (d)? Okay it looks like (a) and (d) got the most votes in this case.
The answer is (a). They tested a device that released the smell of fresh-baked cookies into bus shelters to see if this would increase desire for milk. There was immediate outrage about this because people were feeling there’s nowhere you’re safe from marketing, but that this was sort of an obtrusive and sort of guerilla way of marketing that we’ll come back and talk about in a subsequent class. The industry didn’t actually do much of this, because the public outcry was so great, and the press got after them, but it was interesting, the different ways that industry is using to sell food.
Let’s talk about public health. Now why am I going to talk about a profession? Everything else in the class is — or a concept, if you will — everything else in the class is about substantive material, and this is more about a conceptual approach in a profession, namely, public health, and what it means. Well it’s so important here because public health — in the views that public health promotes, in opposition to the traditional ways we look at things, are very important in the context of changing diet, and changing the health of the population.
If you look back in history several hundred years, you get some very startling statistics like this one. Then, in this case in France, the median age of marriage was older then the median age of death, and the only way that could occur of course is if many, many people are dying young. This has been turned around by public health. This is obviously no longer the case in any part of the world, and the question is, what’s happened, and what are the health victories been that have encouraged this sort of thing?
What changed? Things like sanitation changed. There are many other things that changed as well, but sanitation would be an example of a major public health victory that has really saved many, many, many lives. Here’s the definition of public health:
“The science and practice of protecting and improving the health of a community, as by preventive medicine, health education, control of communicable diseases, application of sanitary measures, and monitoring of environmental hazards.”
Now it’s interesting here, because here the unit to be protected isn’t the individual but it’s the whole community. The community can be construed broadly as a state, a country, or even the whole world. Instead of working with individuals one at a time, the focus is more on large groups, as in a community. The protection of people from environmental hazards becomes an important part of that, so protection of people from water pollution, or from air pollution, would be an example.
One question one could reasonably ask is, is the food environment toxic? And is that an environmental hazard? Not as — like a poison that’s in the water supply; but is the environment sufficiently toxic to produce enough disease that people deserve protection from it? And should government get involved in that process? Again, everybody will come down in a different place with that, but to the extent that holds true, public health becomes very important.
The classic start of modern health has a very interesting history. Some of you may have heard this and I know some of you have a background in public health already, so you all have heard this, but it’s a very interesting start. Certainly, the concept of public health goes back well before this particular event that I’m going to talk about, but this event was pretty noteworthy, and it led to the classic start of modern epidemiology.
In August of 1854 there was a tragic outbreak of cholera in London, and many people were getting sick and losing their lives because of it. The prevailing theory out there was called the Miasma Theory, and that had to do with spontaneous generation. The idea was that disease came about from spontaneous life forms, from things like swamps and putrid matter. This led mainly to treating the disease when people had it, but there weren’t effective treatments and some theorizing about where it might come from, but not much else.
There was an alternative that was embraced by a small number of people called The Germ Theory at the time and the idea here was that one didn’t just get the disease as its own entity, but there were specific things that went into the body, microorganisms (although they weren’t called that at the time) that invaded the body and made the person sick. If there was a way to find the source of those, one might do something about it; but this was very much the minority view at the time.
Enter John Snow, a physician and considered now the father of modern epidemiology, but he was a traditionally trained physician. He was an anesthesiologist; well known in England, in fact, ministered to the Queen and developed something called the Chloroform Pump that was used for anesthesiology, so this is what he was known for.
During the Cholera break he got interested in what was going on, so what Snow did was he rejected the Miasma Theory and believed there was some toxic agent that was invading the body. He wanted to find out where it came from, and so he did things that today we would consider pretty routine, but were really unknown at the time: he tried to trace the spread of it.
There was one notable family that gets discussed in this history called Barnes, where there were a lot of cases of this within the same family. He did more than that. He suspected that the transfer was through the water supply and so he did geographic mapping, much like pins on a map. He tried to find out where the cases were occurring, and where they were clustering. He wanted to see whether there was something common to the geographic area that might explain why the disease was occurring in greater concentrations some places than others. He found that a lot of the deaths were occurring nearing a particular pump called the Broad Street Pump. It took him a lot of effort, and a considerable amount of time, to persuade authorities that this might be a water-borne disease and that this particular pump might be one of the main sources of the disease.
In September of 1854, he convinced the community leaders to remove the handle of the Broad Street Pump and the Cholera problem stopped spreading.
This was a remarkable conceptual breakthrough, nothing remarkable by today’s standards, but certainly at the time was a complete rejection of traditional thinking and going about things in a different way, and it led to a really impressive public health victory. In fact, this is — this little cartoon that on the bottom says, Death’s dispensary, was an example of how the recognition that The Broad Street Pump was the source — one of the sources of disease and you see the skeleton up there pumping the water, got played out in cartoons. There was some developments later, more traditional medical developments, like more work by Pasteur on the germ theory. Then there was the isolation of the particular microbe; this happened 26 years after Snow died. You can see here that the public health intervention which was to prevent the cases from beginning in the first place, rather then trying to treat them, was very effective in this particular case.
If you travel to London there’s a John Snow Pub and there’s the reconstruction of the Broad Street Pump without the handle, in order to commemorate this particular public health advance. There’s very interesting background on John Snow. In this particular case in this book, this booklet created by a professor at UCLA, and I’ve offered up the website that you can connect to to download this information on John Snow. The book is published by Oxford University Press.
It’s a very interesting history, but that particular moment in time really helped change the way people approached disease, and later on this new conceptual framework to use for tracking disease and trying to prevent it onto the traditional medical kind of an approach.
Chapter 3. Giving Public Health the Due Credit [00:16:03]
When we think about the major public health victories, there are many, many diseases that used to ravage the world that just aren’t present at all anymore, or present in very low numbers.
Many people believe that medical advances are responsible for those phenomenon, and to some extent they are, the use of antibiotics for example was a major breakthrough and there are countless medical breakthroughs that help explain the reduction in disease. But many people believe that the biggest reductions have occurred because of public health interventions, and have focused primarily on prevention.
Here would be an example of this, where a particular public health intervention is available. This shows rates of smallpox and this line — and this is dated from the UK — so here are numbers by date of the number of cases of smallpox and you see it really going down to almost zero. The black line up here represents the number vaccinations that were done. Here a vaccine was discovered, many people were immunized, and smallpox was virtually eliminated. This is a pretty clear public health victory.
Now there are other cases like with tuberculosis where the rates went down — excuse me — an awful lot before specific treatments got developed, and so this kind of reduction is due to public health interventions, and to some extent, medical interventions before real treatment came in. There were many examples of this kind of thing with public health.
Now an area that’s a little more pertinent to today’s picture would be changing rates of diseases related to the heart. If you look at total cardiovascular disease; the blue line on the top, other diseases of the heart; the green line, coronary heart disease — and we don’t need to distinguish these from one another so much — and then stroke on the bottom line, you see that beginning in the 1960s and then proceeding through the current time there have been significant reductions in rates of these diseases and death from them. What are some of the things you guys think these — this might be due to?
Professor Kelly Brownell: Okay, more awareness, so people may get help earlier if they experience symptoms. Any other hands? There are a lot of possibilities here. Yes.
Professor Kelly Brownell: Okay, the surgery is better, so people are more likely to survive if they’ve had a heart attack, let’s say.
Professor Kelly Brownell: Okay, fewer people smoking absolutely, one of the major contributors. Other things you guys might guess at? Yes.
Professor Kelly Brownell: Well could be — other diseases are competing with it like potentially killing people earlier; probably the opposite is true because the diseases — as we discussed earlier in class, the sort of diseases that would kill people early in life have — some of those have been eliminated or curtailed a lot, and so people then get to the ages in life where they experience things like heart disease and cancer. But your point is a good one, if there is some other disease like cancer that might have stepped in and been explaining a lot of the deaths, then fewer people would be able to die from heart disease. That’s not necessarily true but — yes.
Professor Kelly Brownell: Okay, better screening and earlier diagnoses. Yes.
Professor Kelly Brownell: Okay, I mean you’re — so more medicines, you’re absolutely right. Think of all the people that are on statin drugs or high — for their cholesterol or their lipids, people on high blood pressure drugs and things like this. So those are medical advances. Better emergency response is another one: we have a much better emergency response system then we used to have. People doing prophylactic things like taking aspirin as a preventive measure; a lot of things are happening to explain this picture.
What you see is a combination of things that are medical interventions, better emergency response let’s say, better surgeries, better treatments, better medicines, combined with public health things like less smoking. Potentially better diet, some physical activity gets mixed in, a lot of things like that that might be active here. This is a partial victory, you’d like to see the numbers look much better than this, but it’s progress nonetheless. This can be explained not totally by medicine, not totally by public health, but by their combination.
It’s very interesting to look at tobacco in this regard. The world’s smoking rates you can see vary a lot and it’s — the type is a little bit small but you can see the United States down here at 24%. Now it wasn’t long — that many years ago that the United States was up in this sort of territory, where about half the people in the country are smoking. There are several things that are hidden — I mean that are embedded in this slide that are interesting. One is we have to ask how the United States got better from half the people smoking to a quarter? The other thing is, why do we have such high rates of smoking in other parts of the world?
If you look at the countries at the top of that list, a lot of them are the developing countries. There is a long and sad history of the exploitation of developing countries by the tobacco industry. Some of it are the big multi-national tobacco players that are headquartered outside these countries, but in some cases like China, it’s state-owned tobacco businesses that are driving this. The numbers in these countries are very high, and this idea of multi-national companies using the developing world as an emerging market for things that may really hurt people, is a topic that we’ll layer onto food as we get further into the class. The parallels with tobacco aren’t perfect by any means, but they’re interesting.
What happened? Well, America went down about half in smoking and a lot of interesting things happened, so these are data from New York City beginning in 1993. Out here you have a 22% smoking rate occur, but that’s about half of what it was before, so a lot of advances occurred. If you could plot this line before 1993, it would be up here somewhere.
This is very interesting, and what they point out is that there were specific public health regulatory legislative-type interventions that were made that help explain the decline from the 22% down to 17%; and while that doesn’t seem like a big number, in a city the size of New York that’s actually a lot of people affected. If you go back here before this particular time, there are things like heavy taxes because some of the heavy taxes kicked in around here, but there was certainly big tax increases in cigarettes before surgeon general’s reports, concern about the health consequences of smoking, lots of information on secondhand smoke, etc.
This is a real public health victory brought about really almost not at all by medical advances, because the treatment of diseases that smokers get — with heart disease being to some extent an exception but — like lung cancer are very difficult to treat, and so the big advances have come through public health interventions.
Public health is very often important where medicine is powerless. If you don’t have a specific disease that can be treated with a specific medical intervention, then public health tends to be the primary way to go. Cigarette smoking is an example of that. 90% of lung cancer cases are due to smoking; 90% of those who are diagnosed with lung cancer die regardless of the treatment they receive; so this is not a very treatable problem that yields to medical interventions, but is one that does yield nicely to public health interventions.
The differences between the traditional medical model and the public health model are quite profound. Now by the medical model, what we mean is that you deal with people who have a disease, and the hope is that you apply some treatment to the disease, the disease goes away, and everybody is happy in the process.
There are lots of examples of this. Somebody — a child gets an ear infection, the parents take the child to the doctor, the doctor treats it with an antibiotic, the disease goes away, so that would be wait for the disease to occur, intervene with it, hopefully have an effective intervention and you have some success. Hopefully that intervention can be used on a broad scale in a cost effective way, so that all parts of the population can get exposure to the helpful treatment.
The public health model is different. Here’s how they differ fundamentally from each other: the medical model really focuses on the individual. What’s causing the disease in a person — not in the population, but in the person? What are the consequences for the individual, and what are the remedies that will help the individual?
The public health model looks at the population, in contrast. What are the causes of the disease in the population? What are the consequences of the disease to the population? How many people have the disease? How serious is it? What’s the overall health care burden of this? The remedies are delivered to populations, not necessarily the individuals. As you can imagine, what these are leading toward, is a focus on treatment in one case, and prevention in another.
Now there are many, many cases where treatment is absolutely appropriate and life saving. There are other cases where the treatments aren’t so effective or can’t be applied on a broad scale, but there is prevention.
The issue is, if you want to change the world’s diet, and lead to a healthier food environment around the world, which of these two models applies? They both apply to some extent. For example, you can have the treatment over here on the left apply when people have been eating a terrible diet, they get heart disease, they need quadruple bypass, you intervene with the expensive surgery. That would be a case where treatment could be life saving and potentially helpful, but you can imagine the cost and you’re only helping one person.
You could take a public health point of view and focus on prevention, and potentially with this same amount of money that gets spent to help one person with a medical intervention, you could intervene with many more people at a preventive level, and potentially have greater impact. We have to ask ourselves, what’s the potential for the traditional medical model versus the public health model with diet?
Now the one metaphor that gets used a lot in public health that is memorable and meaningful in this case is the upstream/downstream metaphor. Let’s say you start with a nice clean mountain stream like this, but then it goes through farmland, it goes through cities, it goes through a number of polluting opportunities, and you end up with a polluted stream that looks like this. Medicine really works here, you wait until there’s disease that exists in an individual or in a body, and then you try to clean it up.
The public health works over here, where you try to find out what’s happening between the upstream and the downstream and then intervene, so you never get what you see on the right hand side here.
Chapter 4. Prevention, The Epidemiologic Triad and Key Steps [00:28:23]
The focus on prevention in public health leads to three different terms that get described — that get used to describe prevention. One is primary prevention, and the object here is to avoid the development of disease in the first place. The second is secondary prevention, which focuses on early detection of the disease. So you’re not preventing the disease, but you’re catching people who have it and then you try to prevent its progression; and then tertiary prevention is try to reduce the impact and the complications of disease.
Now as much possible, you’d like to see primary prevention occurring. That’s not really always possible. But one thing you could do for example with traditional infectious diseases, is you can take people who have the disease and then quarantine them, or give them medicines that stop the spread and that would be not primary prevention, but more secondary prevention. All these kinds of prevention — types of prevention have their place.
One thing that the folks in public health talk a lot about is the epidemiologic triad. What they say is that the disease — when people or populations get disease, it’s a combination of an interaction of three factors: there is the agent, the host, and the environment. The agent is what is it that’s causing the disease. What’s toxic in the environment that causes the disease in the first place? The host is the — whoever gets the disease or the population that gets disease and they would have their own set of vulnerabilities. Then in the environment of course, those are the location, or the sources of the potentially toxic agents.
The philosophy in public health is you really need to understand all these things in order to get a full grip on what’s going on with disease and finding a way to prevent it.
Just as an example of this, the agent, you could have a specific toxic agent like bacteria, a virus, whatever it happens to be that’s causing the disease. The host, some groups are more vulnerable to the disease than others.
For example, earlier in class we talked about how people of Asian descent have much greater — have many greater — more severe metabolic consequences of increasing weight, than people of other nationalities. So at a given level of overweight, Asian people are more likely to get some of the metabolic consequences. That particular group would form a vulnerability group, and they would be a host that would be particularly susceptible to these kinds of food-related problems. You could find groups that are vulnerable by virtue of living in poverty; or there are groups that are especially exploited I guess you could call, are targeted by industry that might be selling some harmful product that could fall into a host vulnerability kind of group.
Of course we’ve been talking a lot, and we’ll talk even more about environmental issues pertaining to food. I don’t just mean the food environment, pesticides, and herbicides, and pollution, and things like that we’ve been talking about, but a lot of the factors that we’ll be discussing later in class. What about portion sizes? What about heavy marketing of unhealthy food to children? These are all factors that — and we’ll talk about how many calories people are consuming in liquid form today, compared to what was occurring before. We’ll talk about fast food; we’ll talk about economics; we’ll talk about a lot of these factors that focus on the bottom right hand side of that triangle. What are the environmental factors that are contributing? The problems of what might be done about them.
This little chart from The World Health Organization will give you a sense of how public health might proceed with dealing with different problems. There are these four boxes where you end up with relationships that look like this. The first box is surveillance. You need to find out where the problem is occurring, what is the problem, and some definition, some description of its severity, its prevalence, etc. Just to layer in some examples with diet, you might assess diseases related to diet in the population. How much cancer is happening, how much heart disease, how much of this is related to diet? That would be an example of surveillance where you’re just tracking something.
The next step is to identify causes of whatever the diseases are, and to identify both risk and protective factors. Again, you can see how risk and protective factors might be different for the population than it might be for specific individuals. Here you’d look at what parts of the diet raise risk, who’s vulnerable, and why they’re vulnerable, and are there protective factors that might be relevant?
The next is to develop and evaluate interventions, try to find out what works and what doesn’t, so you can intervene with a problem. Here you could create programs, do small clinical trials with individuals or larger systems, and test cost effectiveness so you know what might get rolled out to the broader world. Then the last step, as you might imagine, is scaling up with policies and programs, so you could apply effective interventions to many, many people. Here you’d make social, economic, policy, public opinion changes in order to maximize the benefit to cost ratio. I’ll give you some examples of really stunning public health victories where this sort of thing has been played out with specific problems.
Now, if you look at the way medicine, or even my native profession of psychology deals with problems, they tend not to look at the world this way. For example, instead of looking at what are risk factors and vulnerability factors for the population, they’re looking at what constitutes risk and vulnerability for individuals. Now, here’s an example of where this might play out, and I think I mentioned this example before. There are many people out there in the health professions who are looking for the gene — for obesity genes. What they want to find out is when obesity occurs; are there genetic reasons for it occurring; and who might be most vulnerable; and what are the conditions that create this vulnerability?
Now there’s great excitement about this, a lot of funding for it, and there’s some really quite exciting science going on in this context. One point of view that would say that it’s — that’s really not going to be very helpful, because why do we think that genes are going to explain this problem? I mean you have rampant obesity in the U.S., Mexico, other countries, and you have almost none — I mean not none, but you have very little in countries like Somalia and Ethiopia. Is that — can you explain that by genetics? Are we genetically different from people in those countries that explain that?
Well no, absolutely not. How do you know? Because studies show that if people move from those kinds of countries to the U.S. they gain weight; people from the U.S. move to those kinds of countries, they lose weight. If you take the sort of typical American diet, feed it to laboratory animals you get obesity. We’ve shown you examples of that. There are many reasons to believe that this is an environmental problem, and that you’re not going to learn an awful lot through the genetic discoveries. Now maybe, maybe the genetic discoveries will led to some drug that can override the toxic influence of the environment. There’s nothing on the horizon that looks very promising in that respect, but you never know. If something like that came about it would be a great advance.
The traditional medical approach would be to look for the gene because you really want to know why individual A has a problem. From a public health point of view, you’re not paying attention so much to whether individual A has the problem, but how many individuals in the population fall victim to the problem; and then that leads you to a sense of what the causes are. You can see with this process here how you’d start with surveillance; end up with broad scale interventions.
Now the other thing that very often happens in my field, and in medicine, is that people work up in this area and they might find out for example, that people with high cholesterol are at risk for heart disease; or people that eat too much saturated fat are at risk for heart disease, and those would be vulnerability factors. So they published their papers and that’s their job, that’s their part of the world and they do a good job at that, and then they hope that there’s uptake of that information by other people in ways that will make a social difference.
Some people work over here, and develop programs; so for example, there are plenty of people in my field who do intervention programs in the schools, for example. You go in the schools, you try to get rid of the soft drinks, you teach nutrition education, you put in a physical activity program, etc., and so they publish their work in the journals and then they hope that there’s uptake out here. But very often there’s not because the people that are out — responsible for doing this very often don’t know about this kind of work, and the people who are doing this don’t communicate very well with this group. So there’s a real breakdown between various segments of the scientific world.
The hope in public health is that all these things come together, and you’ve got somebody watching the whole ship. Now that doesn’t always happen because there are these segmented parts of public health as well, but there’s more hope for it in that kind of regard. The job isn’t done if you accomplish something here or here, but only when these things happen, is the job really done. Again, I’ll give you some very interesting examples of that.
Chapter 5. The Importance of Science and the Problem of Self-Interest [00:38:56]
Now, we’ve talked a lot in this class about science. I’ve shown you slide after slide, after slide of scientific data. Why is that? I mean we could — this could be a course on focusing on the history of food and food systems without respect to scientific information. We could be talking about the anthropology of food, we could be talking about food and literature, we could be doing a lot of things that don’t really have science as the background.
Why do we focus on science? It’s terribly important in this arena, and here’s why. First of all, you have to have agreed on standards for truth. I’ll give you an example of — in a later class of a study that my colleagues and I did looking at the impact — in fact I think it’s in your readings — looking at the impact of soft drink consumption on health outcomes. As you might imagine, this study found that the more soft drinks, sugared beverages people consume, the higher their likelihood is of eating a poor diet in general, developing obesity and diabetes; and the stronger the methods get within this literature the more likely that finding is to be the case. We published that in the scientific literature.
The food industry, the trade association for the soft drink industry, then funds a scientist who has a very well-known reputation for publishing things friendly to the industry, goes out and finds the studies, the university of studies that we reviewed, reviews them himself — there was a group of people involved in this — and then publishes a paper that says the opposite of what we did. No, that sugar beverages really aren’t bad after all. Well, what’s the truth here? What should people pay attention to? Is ours more credible or less credible then the other one? Well, we’ll sort that through and we’ll talk about this interesting issue of conflicts of interest, and how science gets used in many interesting ways.
We need some standard of truth. If you’re going to change your diet, take some supplement, start exercising because you think it’s good for your health, you darn well want to make sure that it’s reliable information that you’re making those decisions on. What if you decide to take fish oil supplements, for example, because you hear in some place that fish oil supplements are good? What if they’re not good? What if that was crummy science? What if it was industry funded stuff that would suggest that fish oil is good when it’s not? That would be an example of you making the decision that could really affect your health and well being, and then there are lots of you out there so it affects population, health, and wellbeing when the standard for truth is being undermined. Now in — I don’t want to give the wrong impression about fish oil, in fact, fish oil is good and there’s a lot of research on that kind of stuff, but that would be an example, just a hypothetical example of how you might find something like that.
You need to have some commonly agreed upon rules about what constitutes proof. How many studies constitute reliable evidence that something is related to something else? How good do those studies have to be? How many subjects do they have to have, etc.? The science is one of the ways, although it gets undermined by the kind of thing I talked about a moment ago with the soft drink papers. It erects barriers to lying, and you hope that it helps erase the self interest that some parties have in things.
As another example, the cereal companies have been funding a lot of research by various investigators. The money tends to go to certain investigators that reliably produce results favorable to industry showing that people that eat ready-to-eat cereals for breakfast do better, they’re healthier, they do better on cognitive tasks and things like that, than people that don’t eat any breakfast at all. What they — in that — what they conclude from that is that any breakfast cereal is good for you and is going to help.
Of course, the comparison here shouldn’t be breakfast cereals as a group compared to not eating breakfast at all, because we already know that it’s a good idea to eat breakfast. The most relevant comparison would be crummy breakfast cereals and good breakfast cereals compared to one another. By crummy I mean high in sugar or high in fat, or even high in sodium. The industry sets up the research where they’re going to get a positive outcome. They know in advance which scientists are going to give them the positive outcome, and as a consequence, you’ve got self-interest in there and you’ve got distorted biased science. The fact that you use science helps control for that to some extent, but doesn’t eliminate it entirely.
This was interesting quote from a very well known marine biology scientist that says science is nothing more then a system of rules to keep us from lying to each other. Well, why would we lie to each other? Well, you’ve got a lot of interested parties with a big stake in the game, and so it’s easy to see why lying would occur in science, to some extent, as protection against that.
We’re going to talk about science, and the question really becomes who’s the referee here when you got self interest involved, when you’ve got the soft drink industry saying, and science supporting it by their own funded scientists saying, well soft drinks really aren’t bad, it doesn’t hurt anybody to drink sugared beverages, and then you have other groups of scientists saying, well hold on there’s an awful lot of science on this that suggests the opposite. Who’s the referee? Who’s going to decide?
Well, that’s where you need an informed public and you need people to understand as much as possible about what’s going on in science, and that’s what I’m going to try to help foster today. There are a number of very interesting issues pertaining to the science of diet and health. One of the kinds of studies that get done, so for example, observing large populations versus doing studies in the lab.
I’ll talk about correlation and causation. It’s very important in studies to control for confounding factors. We’ll talk about cross-sectional versus longitudinal studies.
There are three common methods that get used in public health to test the relationship between certain things and diseases, and then to test whether you can turn the picture around. There are cross-sectional studies, and one of the types of those is called a case control study. I won’t talk about that in particular, but I will about cross-sectional studies. There are longitudinal studies, and experimental studies, so let’s break those down.
Chapter 6. Cross Sectional Studies: A Snapshot in Time [00:45:55]
Cross-sectional studies might look like this, where you’re taking a snapshot at one point in time of a group of individuals and you might, if you have enough subjects, be able to compare groups of individuals. Let’s say this is a hypothetical example that looks at cholesterol levels, smoking and heart disease in a group of males and females. You do this one shot in time assessment, so let’s say you get 5,000 males, 5,000 females. You do a big study on this and you get collect data where each of these X’s appears.
What can you do with that information? Well, you can look to see how, let’s say cholesterol and smoking are related to risk for disease, you could compare what happens with males and females, so a study like this will take you partway down the road to establishing scientific certainty, but leaves a lot of things unanswered. One of the problems here, is that let’s say you find the relationship in the females between smoking and heart disease but not in the males, and you wouldn’t — that’s not what you would find but let’s just say you did — then does that prove that smoking is causing heart disease in females but not in males?
Well no, it doesn’t really prove that because you’ve just got this one snapshot in time, and so it could be that there are things related to smoking and risk for heart disease that you haven’t measured, so let’s just take stress. Let’s say women are under higher stress then men, women smoke more because of the stress, and they’re dying of heart disease because of the stress not the smoking, the smoking is just in there as an incidental variable. Now that’s not the case: smoking is a very strong predictor of heart disease; but you can get a sense of the weakness of cross-sectional designs. These are the quickest way, and the most inexpensive way (although not inexpensive in an absolute way), to start to get a sense of what’s going on with risk factors and disease.
You can also do cross-sections over time, so it might look like this. Let’s say that a series of scientists collect data in 1978, 1988, 1998, and 2008, and that they do have data on females and males. In 1978, they get a group of females, they also get a group of males, but of course they’re different from each other. Then in 1988 you get yet another group of females, and another group of males, now this is cross-sectional over time because the females let’s say up here, this group of females is different from that group of females.
If they were the same, and you were following the same group of people over time, then you’d have a longitudinal or a cohort study that I’ll get to in a minute. If you’re just taking this snapshot in time in 1978, then you get another group, let’s say a random group of people from the population then you’ve got another group of females, another group of males, but they’re different from the original groups. Then you could fill in these other groups over here, and then you could look at trends over time and you could see what’s happening with say rates of disease or things, and you could put in data points, so up at Group A you’d get a data point there, with Group B you’d get a data point there, and then you get all the other data points and then you can compare these data points in various ways.
For example, you could compare the males and the females at one slice in time, that’s what we talked about before, but you could do sequential cross-sectional studies and so some kind of tracking over time. But there are some weaknesses in this method as well, but it’s better then just having one snapshot in time of course. This would be an example of a cross-sectional study, but you’re just taking different cross-sections over time.
Here would be an example of a cross-sectional study, and I’ve showed you this graph before. This is the graph that shows body mass index, level of weight increasing on this axis, and then mortality rate on this axis, and you get a curve that looks like this. As weight goes up, risk goes way up as well, and we talked in this about this part of the graph — let’s see if we can get it up here.
First, this is an association, it’s a correlation, but it doesn’t really prove causation. So as weight goes up risk goes up, but you don’t know if it’s the weight that’s causing it. Let’s say heavier people are more likely to be in poverty which is the case, and it could be that the poverty is causing the mortality rather than the weight itself, weight is just a marker of it or a correlate of it. That would be an example of a correlation and one of the weaknesses in cross-sectional studies. I also mentioned this up curve over here that we — as we discussed in class might be attributed, in fact, is from the way the studies go to cigarette smoking.
Now, if all you were looking at is correlations and you didn’t have data on whether people were smoking, you would assume that there was something bad about being underweight or being in the low weight range that was leading to increased risk. In fact, it’s not the weight itself, it’s just that the lower weights are associated with smoking, and smoking is what’s driving up the mortality. That’s why you need — in these studies you need to look at causation rather than just correlation. The solution to that particular problem was to compare smokers to non-smokers in a sample like this, or you can take statistics and control from the impact of smoking and find out how much it’s contributing to that little J part of the curve.
Chapter 7. Longitudinal Studies: Cause and Effect over Time [00:51:47]
This is interesting, and we can learn things from cross-sectional studies. Longitudinal studies are the strongest way. They’re more expensive, more difficult to pull of for reasons I’ll explain, but they’re the strongest way to nail down cause and effect relationships.
What we would have here would be a group of people studied over time, and each of the dots would represent a data point. Let’s say Time 1, Time 2, Time 3, Time 4 separated by five years, for example, and instead of studying different people at different points in time, you’re taking a single group of individuals and tracking them over time. They are called a cohort, and the study is a longitudinal one.
In this case, you’ve got a group of females, so a group of females is identified out here, those same people are studied at this time point, that time point, and that time point and the same might be true of males. Now the nice thing about a study like this — well first of all it’s difficult doing studies like this because one thing you need if you’re doing a scientific study is to keep your subject pool intact. If you start off with 100% of subjects and you end up with 20% you’ve got a study that’s not terribly valid because you don’t know what happened to all those 80% of people.
It’s pretty hard to keep a cohort intact like this over a sufficient period of time, but some investigators have done it. You remember I talked about the Framingham Heart Study. That would be a perfect example of a longitudinal cohort study where a group of people in Framingham, Mass were followed over a period of time for risk factors related to heart disease.
Now the nice thing about these particular studies is that you’re controlling, because people are their own controls here. You’re controlling for the factors that might otherwise complicate interpretations that you have of associating certain risk factors with disease. Let’s just say family history is one predictor of disease. Well, if you’ve got a bunch of people in a cross-sectional study studied one point in time; they’re all going to have different family histories and you’ve got to measure that, control for it, or whatever. If you’re making — using people as their own controls over time. The individual’s family history isn’t changing as they go from Time one, to two, to three, to four but it remains constant, and it provides a stronger method for looking at the association of certain risk factors to disease.
A great example of this would be let’s say this is a dietary intervention study, and you get baseline data out here at Time One, at Time Two you intervene with some diet intervention, let’s say have people lose weight, eat more fruits and vegetables, take supplements, whatever it is. Then you follow people — that same group of people over time and you can get a pretty strong cause and effect inference out of that.
The reason you can draw a stronger inference is in part is because the people, as I said, are their own controls, and so variability that gets introduced into studies because of differences between people don’t exist using this method, because the same group of people are followed over time. You’ll see longitudinal studies, cross-sectional studies mixed in to the sort of papers that I’m having you read. It’s interesting for you to think about what you — what it means when it — in the abstract of a paper it says, this is a longitudinal cohort study versus a cross-sectional study.
Here’s an example. Let’s take a study and interpret this. 2006 there was a paper published in The Journal of the American Medical Association looking at green tea consumption, the mortality due to cardiovascular disease in a Japanese sample of individuals. I’m going to talk about certain things from the abstract of this particular paper to point out some of these issues.
First, the context of this is there’s some research suggesting that green tea may have some protective benefit. What they did was they did a population based, which means that it’s more or less a random sample of the population, and a prospective cohort study. Prospective and longitudinal mean the same thing: you’re taking the same people and following them through time, the word cohort means a group of people were identified and followed over multiple points. They looked at all cause mortality here, so that would be death from any reason.
Now here are the results, and this is going to be very difficult to sort through but let me try to help guide you through it. What they do in studies like this is they’ll say, look at people at five levels of green tea consumption. They’ll take the bottom fifth, a second fifth, on up until the people that are consuming the most green tea. They’ll take the group that — at one of the extremes, usually the people with the lowest level in this case and say okay whatever risks those people have we’re going to call it 1, we’re going to establish that as 1. They’ve taken the people with the lowest green tea consumption and said their risk is 1. Let’s see what happens as green tea consumption goes up in those other four groups (because they broke them into quintiles, or five groups), and see what happens to risk as a function of 1 for the referenced group, and that’s called relative risk. You see the terms relative risk used a lot in these epidemiology studies.
I’m going to point out this particular part here. Now what this says, in women, the hazard ratios — we don’t need to talk about what that is so much, of cardiovascular disease, mortality across increasing green tea consumption categories were 1. Okay, so that’s the lowest group of green tea consumption. Then it goes to .69; .69 for the next group. Now let’s see where we are. Okay. .69 is the highest group; 1.0 is the lowest green tea consumption. It looks like 1.84, and then it goes down to the .69. Okay. That would say that if you are in the highest level of green tea consumption you have 69% of the risk that people have if they’re at the lowest level of green tea consumption, so this would represent a 31% reduction in risk. That’s what relative risk means. This study sounds pretty good, high green tea consumption, 31% reduction in cardiovascular disease. Now let’s skip that.
Now, here are the numbers of how it works out in a study like this. I mean who wouldn’t do something like drink green tea if you’re going to get a 31% reduction in risk for something as serious as heart disease? Even the big studies become interesting in this regard. In this particular study, there were 40,000 subjects that started initially; 86% of those completed, so 34,000 completed the study.
The number of people who died of heart disease in that particular study — cardiovascular disease is the CVD — is 892 of that 34,000; the 31% reduction of that would mean that 615 people would die of heart disease out of that 34,000. If you subtract the 615 from the 892 there are 277 people who would have benefited out of the 34,000. What does that mean in terms of your own personal behavior and your own choices? Well, if you’re one of those 277 you’re golden, and you’ve got benefit from being in that particular high tea consumption group. But the chances are that you’re not going to get heart disease if you’re living in Japan and in this particular sample, at least during the time of the study, and even if you get heart disease, there’s not a big chance that you’re one of the ones who would benefit from the green tea. Do you take the green tea? Well personal choice of course; but that’s how these numbers were work out in these big studies.
Here would be an extreme hypothetical example of that. Let’s say that somebody has proposed that there is a 33% reduction in the — 33% drop in some disease if you eat some incredibly icky food. Remember that little NPR clip we showed about that fruit — I forget what it was called — but people were making all kinds of faces and eating the fruit, so that would be an example. Let’s say this is fruit that you just can’t tolerate very well, but it leads to a 33% reduction of disease; that sounds pretty good.
Let’s say somebody did an enormous study, nothing like this has ever been done, but did a study with ten million people and of the people who don’t eat that, three get the disease. Of the people who do eat it two get the disease, and of course that’s your 33% reduction. The chances that you’re going to be helped by eating that bad tasting fruit isn’t very high. So that’s why the distilling science into terms that the public can make sense of, is very important.
Chapter 8. Experimental Studies: Manipulating Variables [01:01:28]
Now there are experimental studies that are interesting in this regard. Experimental studies are when you manipulate some variable. You’re actually doing an experiment where you’re manipulating some variables. Randomized control trials, RCTs, are the gold standard for doing this. Where you randomly assign people to groups getting an intervention or something that’s not the intervention, and then you typically have control groups. In these, you want to control for as many confounds as possible. I found this funny little figure in a psychology journal once that talked about the need for control groups in these studies, but of course the need for out of control groups as well.
Here would be an example of an experimental study. Let’s say you get access to schools, and you hypothesize that a nutrition intervention would improve what kids eat in schools and at home, and ultimately their health. Let’s say you have four schools available to do this kind of thing. How would you design the study? Trying to be as scientifically sound as possible, you have an intervention, you hypothesize that it’s going to help kids, and you want to find out whether it really does, and you’re lucky enough to get four schools to work with: how would you design a study like this? Anybody take a shot at it? There’s a pretty straightforward initial answer, yes.
Professor Kelly Brownell: Okay, so that would be the typical approach, is you take two schools and give them the intervention, and you take two schools and not give them any intervention. If you were being — if you were going a good job at controlling things, you would randomly decide which schools got the intervention and which schools didn’t. That leaves out at least one source of bias.
Let’s just say you’re the investigator. You really want to find that your intervention works, and you know that the staff in one school is particularly motivated to carry this out. If you put the intervention in that school and don’t randomize then you’re likely to get bias in your results, so randomization helps with that.
The unit of randomization becomes interesting. You could randomize by school and that would be the obvious way to start doing that. You’d have school 1, 2, 3, and 4 and just say by random assignment you get something like this, you get your intervention and control, two in each condition.
Now the problem with an approach like this, is that you’ve only got — you basically got four subjects in this study. You’ve got a lot of kids within each school, but only four subjects when it comes down to doing your statistics. If there’s something that makes some of these schools different from others, it’s going to be very hard to figure out whether any results you get are because of the intervention, or because of these predisposing factors that might have led one set of schools to do better than the others.
Another thing — another way to randomize would be to do it within classes across schools. In School 1, let’s just say you’re going to take the sixth graders in School 1 and there are two sixth grade classes in that school. You do a random assignment to intervention or control. Then you go to seventh grade classes in that school and do random assignment for Class 1 and 2. Then you go to the eighth grade and do the same thing; and then you’d repeat that with School 2, 3, and 4 and then instead of four subjects you have many more subjects because your unit of randomization is classes within schools. This would be preferable because then if there are any things going on that are specific to a school — like let’s just say School 1 is one of the school’s with highly motivated staff, that’s going to occur evenly across all your conditions. So there are equal numbers of kids in the highly motivated school who got the control intervention, and ones who got the control versus the intervention. That helps rule out one potential source of bias.
This would be an example of an intervention trial, but there are lots and lots of intervention trials that we’ll discuss some of those later in the class. Ultimately, if you want to find out that some change in something leads to benefits, you’re going to be doing some kind of a control trial.
Let’s try to pull all this together into some overriding conceptual scheme. If we look at public health, how does it approach the world in a unique way compared to the way traditional medicine does? What does that lead us to in terms of trying to help people? The traditional approach, as I mentioned with medicine, is to focus on the individual. You hope that by focusing on the individual, you get a good outcome: improved health and well being. In order to do this you try to motivate the person to make a change, or you give them knowledge, or in the case of medicine of course you can intervene with drugs and things like that.
The idea here is that you apply something to the individual, the individual gets better and their health improves as a consequence. Here you educate, you implore, and you hope that the world works this way, so all these things fit together into this neat little picture. Cigarette smoking, for example, you tell people smoking is bad for them, you tell kids that sugared cereals aren’t good for them, that soft drinks aren’t good for them, that fast food’s not good for them, and they should eat these things in small amounts — that would be educating and imploring people and you hope you get a good outcome.
Now in the case — there are advantages to this kind of thing. Government doesn’t have to get so involved necessarily with changing policies. There are disadvantages of this as well. One is that it tends not to work very well. But we have to ask when this will work, or are there enough resources to make it work? You’ll see some examples of that in just a minute.
Another approach, a different conceptual approach is to change conditions that affect the individual. You’d back these things out, and instead focus on things that occur before the individual gets involved. You’re changing the conditions that affect the individual. So you might change economic conditions, so maybe the fundamental cost of food would be an example using taxes. You’d use legislation to intervene here, maybe legislation that would prohibit marketing of unhealthy foods to children.
The environment could get involved by making healthier conditions, and there are a million examples of that; and government can use its regulatory authority, for example, banning Trans fats in restaurants. To create what economists call optimal defaults.
This is a key concept, optimal defaults. The idea here is that you want to set up environmental conditions where the optimal behavior becomes the default rather then suboptimal behavior becoming the default. Then that, in turn, affects the individual and then you get the increased health and well being.
Chapter 9. The Sagacity of Optimal Default [01:08:40]
Here’s an example drawn from economics. There’s a group of economists who have studied enrollment in pension plans. Some employers automatically enroll people in pension plans, which are a good idea by the way, because then if you’re investing money throughout your work life, you’re less likely to be dependent on the state later in life. Some employers make you — enroll you by default others — but you’re allowed to opt out of it. Others don’t enroll you but you’re allowed to opt in, so the same choices but whether the default is in or out varies across employers. The number of people who join pension plans varies a lot depending on the default. Same choices but just different defaults.
Here’s another stunning example, people who become organ donors. These are data from European countries that break down about 60%/40% into ones that use the U.S. model where you’re not an organ donor by default: you can choose to be an organ donor, but you’d have to take active steps to become one. Other countries you’re an organ donor by default, but you can opt not to be one if you wish. Here are the rates. Using the U.S. Model, Denmark, The Netherlands, the UK, and Germany have this percentage of people who are organ donors. The other countries are like this.
Now that is breathtaking, that difference. Imagine if you wanted to produce this effect from this baseline with education, motivation, imploring people and the like. You could never get an effect like this and it would cost an absolute fortune to do it, or you can just change the default. Just changing the default becomes an important theme that will play through all the rest of the class. We hope to create optimal defaults in ways that change public health. There are some classic examples of this in the environmental area like chlorinating water, using immunizations, and you see some of the numbers here. These are classic examples. I’m going to skip over this because I want to make a point here.
You remember when we talked about malnutrition, and some of the diseases and maladies of the body that fall from that. One in particular was Vitamin A deficiency. Scientists discovered that Vitamin A deficiency provoked by malnutrition leads to a number of bad things, and you see them listed here. Here’s a case study of public health in action, a very successful case study.
There’s a particular scientist at Johns Hopkins in the School of Public Health named Alfred Sommer. He was an ophthalmologist by training, but also a public health expert. He was one of the ones who initially did some of the documentation that Vitamin A deficiency was linked to a number of diseases that were killing many millions of children. He then conducted small trials. Remember the WHO slide with the surveillance becoming the first step, and then you establish the risk factors, and then you do small studies to see if you can correct what the risk factor is? Here’s all that happening by one person. Here’s the risk factor Vitamin A deficiency; leads to these; then he conducted small trials of supplementing by doing Vitamin A supplementation for children in poor countries. Here’s a picture of Sommer doing those kind of studies.
What’s most impressive here is that he connected science with public policy in a very impressive way. He now, throughout the world, there is wide scale Vitamin A supplementation due mainly to the scientific and the public health efforts of Al Sommer. You can see from the results here that huge, huge changes have come about and you can see that it’s very cost effective to do this kind of thing.
This would be an example of a startling public health victory where the science gets connected to the public policy in a way that really affects the world. There are many examples of this, but this is one of the most startling involving the nutrition arena. If you’re interested in reading more about public health I’ve listed a couple of books here. Those of you who might be interested in pursuing a career in public health, the website down here, The American Public Health Association, is a good one to go to for information on different degrees one can get. That’s all listed there, and of course this will on the course outline. The creed here is that instead of saving lives one at a time you try to save lives millions at a time.
[end of transcript]Back to Top
|mp3||mov [100MB]||mov [500MB]|