Extended Interview with
Stewart Agras, MD, FRCP(C)

Describe the study you are doing.

We are looking at two different ways of preventing obesity in children between two and four years of age. Basically there are two big ways of doing it, first, altering diet and activity, or second, changing parental behaviors towards their child when they are eating.

What prompted this study?

What prompted it was another study that we began about 15 years ago which followed up infants all the way from birth to eleven and a half years of age. We were looking at early risk factors for obesity. The first thing we found was that parental obesity was the big risk factor, but that was modified by the temperament of the child. The more difficult the temperament was - irritable, crying, weepy kind of kids - that doubled the risk for the child, in terms of becoming overweight. We then looked at the data and found that it was probably being modified in some way by parental behavior. That is, parents overreacted to these kids, and over-controlled them, and probably over controlled their feeding. If parents have a difficult kid, they are probably going to either ignore them or over-control them. Probably both are bad. They probably will not only over-control them in terms of when they have a temper tantrum or when they’re difficult, but in terms of their feeding, because the same behaviors occur at the table as occur when they are playing, when they fall down, whatever. So the over-control will impact feeding, and over-controlling children, as we hypothesize, means that they won’t learn self control. And if they don’t learn self control when they go out in the world, it’s going to be a bad thing for them given the huge food environment that we have out there.

What’s the difference between control/over-control and structure (the latter being presumably the desirable?)

As far as family structure goes or parental structuring, one has to really have the optimal structure; not too little of it, not too much. And as most parents know that’s jolly difficult to do. But that’s basically what we want to do: trying not to overtake the child’s natural feeding ability, and push your own feeding rules onto the child.

What’s an example of where that line is crossed?

Well basically when the parent takes control of the child’s feeding, so the child doesn’t want to eat. Now the parent pushes them to eat. Or the child today eats what the parent thinks is too much, and [the parents] try and stop that. So, that’s the kind of thing we are talking about.

What about indirect methods such as controlling what food is put out, when it is put out, and maybe even more subtle methods such as just cutting back on this, that and the other thing?

Feeding it a very complicated behavior, so there are many things that impinge on it, right now we are just looking at the family impingement. So, parents, for example, might elect not to have regular meal times, or regular snack times, and that is not very conducive to the child learning about their appetite, about hunger, and soon and so forth. So it’s much better to have a regular, three meals, and a couple of snacks, for young children.

What is too much control?

Any interference with the child’s eating is over-control. On should really allow the child to eat what they want.

But if you have regular snacks, and regular meals, you’re interfering. If I’m hungry between snacks-- Right but then the child really needs to learn about hunger. For example, if a child refuses to eat breakfast, and the parents are trying to push him to eat something and he refuses, doesn’t eat a thing. Now they’re going to get hungry before lunch or before their snack, and now what does the parent do when they want food? If they don’t wait for the next snack, and feed them in between times, the child is not learning if they miss breakfast they’re going to be hungry, and want some food before the snack time is due, so the whole feeding sequence gets disrupted.

Although if that child is genuinely not hungry at breakfast, why should he eat just to fit in with the program? If he’s not hungry, the parents should ignore it, that’s fine. But then the child learns that they’re going to be hungry before their next meal, and that’s important. If that’s interfered with and the parents push the child to eat the breakfast, maybe a little bit, now there’s a snack, coming up in two or three hours time, and they abort the situation by giving them some food an hour later. So the child doesn’t learn that they have to wait for a meal.

But why not do on-demand feeding as you do with breast feeding? Breast feeding is on demand, less so if it’s on bottle feeding, but parents have to make that switch between them feeding their baby to the baby beginning to feed its self, and then moving towards a regular sequence of meals. And that’s a difficult transition for some parents. (Of course they don’t really realize that it’s the baby who’s in control of the feeding, not them.)

Why shouldn’t we all eat on demand? If we’re truly in touch with our hunger cues, why should we be constrained by social norms? That’s just the way life is arranged isn’t it? We all run to schedules. Perhaps if we were in the wild, we would in fact eat on demand, in fact it’s very likely that primitive people did eat on demand. They wouldn’t have food available all the time, they had to go out and get it.

So that’s what it comes down to, that’s a pity…

What are you comparing and what are you looking for precisely? We are comparing basically Ellyn Satter’s model, to the recommendations of the National Institute of Health, which is more activity, and moving towards a more vegetarian based diet: vegetables, whole grains, and that kind of thing. And what we’re interested in is not only the effects of those programs on weight at all; in fact we’re not even measuring weight in this study, not the child’s weight. We’re not even dealing with the children; we’re dealing with the parents, because it’s parents who institute the feeding practices. So what we are interested in is what parental behaviors change with each one of these programs. Are there differences? Or do the same behaviors change in both programs?

And what are the different behaviors you’re looking at?

Basically, what we are looking at is what the parents feed the child. Do they in fact change the child’s feeding towards more vegetables and so on and so forth, and whether there is any change in parental control of the child’s feeding, and we have a number of measures of that.

Besides excessive control, is there anything that has been measured in parental behavior that contributes to obesity that you know if?

There are many other things that parents do that contribute to obesity, we’re just looking at a little tiny slice of the problem. For example, the entire family structure may predispose towards obesity. There was a study done many years ago of the pets of obese and thin families, and of course in the obese families the pets were obese; in the thin families they weren’t. So it shows how the entire family environment really can change things. So we are looking at the interaction of the parent and the child, basically how that changes with two different widely used programs, one focusing on altering diet composition and activity levels, the other focusing on parental child behavior. Describe in summary form if you can, program A and program B and what they’re based on. In summary form, the National Institute of Health program basically has the two aims, reducing caloric intake and increasing activity levels. Obviously that at a crude level what obesity is about: intake, and output. The other program focuses quite differently using Ellyn Satter’s model, on the way in which parents and children interact over feeding situations.


Well, for example, basically it’s the division of responsibilities. That is, the child’s responsibility is to eat, and the parent’s responsibility is to provide the food. That’s the simple way of putting it.

How are those two not compatible; in what ways do they differ?

Well, for example, I think parents find it very difficult to stand [it]if a child doesn’t eat at a meal, or if the child has a temper tantrum over something that the parent [has done]. Parents I think find that very difficult, so they tend to interfere with the child’s eating, that is, they take over the child’s responsibility. On the other hand, the child can take over some of the parent’s responsibility by being very stubborn, and the parents then provide special food for the child.

To what extent are the red light, green light, yellow light foods in the National Institute of Health program restrictive in a way the other program isn’t?

I don’t think the NIH food program need be restrictive because parents, after all, are serving food to their children, it just means they’re going to change some of the content, and do it very gradually, they’re not going to rush in and change everything at the same time, so it needn’t be restrictive in that sense. It’s a little more prescriptive then the other program, but it could result in very similar changes as the other program. Parents start to think about the whole business of feeding and then change certain behaviors. That’s what we would look at--and delineate what behaviors of the parent actually change.

Ellyn Satter believes that we all have internal regulators and that Life gets in the way of them. To what extent do you think most people in this world are able to return to those?

Well internal regulation is what one should learn during childhood, adolescence, and so on, and we should be able to theoretically regulate our intake and regulate our activity levels. There are enormous barriers to that. For example, the food industry provides nine hundred or a thousand calories more, per person, than we can eat. So, we are awash with food in this environment. And I think that’s one of the things that’s changed over the past thirty years to provoke this increase in overweight and obesity now both in adults and in children. For example, if you simply look at portion size, I remember hamburgers were really quite small things, and now they’re huge! So that’s one thing that the NIH program teaches: how to reduce portion size by perhaps reducing the size of the plate, that kind of thing.

Rattle off some of the big culprits if you will.

The big culprits that may underlie obesity: Well of course, we can neglect genetics, because certainly genetics haven’t changed in the last thirty years, so it must all be the environment. The other big thing that has changed is activity level. People are able to do much less activity than they were thirty, forty, fifty years ago. First of all jobs are much more sedentary, certainly mine is, I think most people’s jobs are far more sedentary than they used to be. We are losing a lot of factory jobs that used to be very energetic: energy-spending jobs. And we used to be able to walk safely through neighborhoods, now there are many neighborhoods that people just don’t want to walk through because they’re not very safe. So activity levels have dropped down from that point of view, we use automobiles instead of walking to things. Instead of walking to the bus, and catching the bus, we go by automobile. So there have been massive changes, in terms of energy output. Then you say, ‘Why don’t people exercise in discrete ways?’ Many people do, if they have the resources to do so. But many people don’t have the resources, they don’t live in neighborhoods that can support activity levels, and they don’t have the time or the money to do it.

What do parents not know that you would like them to know?

I think many parents don’t know the details of a good diet, or how to implement those details. And to get them to implement those details one has to examine to some extent the barriers presented to each individual that may be different and then figure out ways to overcome those barriers.

Give me examples of barriers and how to overcome them.

If you don’t have any fresh food in a neighborhood, for example a farmer’s market and only small grocery stores that don’t really sell much in the way of fresh foods, where are you going to get the NIH diet from? That’s a big barrier for many people.

And what’s the answer?

This is where obesity gets complicated. Our answer is really at a much higher level, it’s at a societal level. How can a neighborhood get itself access to fresh vegetables? And there are many studies now going on across the country on getting people to grow their own food on common lots, the old idea of allotments.

What are the biggest emotional reactions that you hear from parents?

I wouldn’t say emotional reactions are a big problem. I think the biggest problem we hear is time, and that’s of course related to money, but it’s basically time. People in this society work a huge amount. Most parents, both parents are working. So that leaves relatively little time to implement any of the things that should be implemented. That’s the biggest thing we hear.

What about food fights?

The biggest trigger for food fights is a child who’s a picky eater. We’ve done a bit of research on picky eating, and that provokes great parental anxiety, may provoke differential parental anxiety, and ultimately provokes quarrels between parents about how they should deal with a picky eater.

To what extent do you think parental behavior inadvertently causes over eating?

Just to single out a single cause I think is quite wrong. You’ve got to understand obesity is a multidimensional problem. It starts way up with grocery stores, food producers, government regulations, and so on and so forth. Comes down to schools, universities, comes down to families, and individuals, to the peers of children as they go out into school, enormously complex, all these things effect how we eat.

Then how much can parents do?

They can do their bit! This is the problem in prevention programs. First, they haven’t been terribly effective, nor as we know are the usual treatments for obesity very effective. And they’re not terribly effective because of these multiple influences on eating. Whether or not ultimately we can get a program that deals with most of these influences I think is a question at the present time, but it shouldn’t stop parents from trying.

And the biggest things that the parents can do, in your mind, are what?

The biggest thing is choose the kind of foods they serve to their child, make sure their child gets enough exercise, extra activity levels, less TV watching and computer games. And they do not interfere too much with their child’s eating. That is it’s the child’s responsibility to eat and the parent’s responsibility to serve the food, the kinds of food, at reasonable times.

What, more specifically, can parents do?

Well parents are of course in charge of serving food. So they can decide what they want to serve to the child. The NIH program gives them a certain amount of guidance in that regard. Parents also can interact with their children in ways that don’t impinge on their child’s eating, or at least don’t impinge over much of their child’s eating, and they take their responsibility of serving the right kinds of food at the right kinds of times. It’s very complex because we are dealing with a narrow slice of the problem and there are many things that impinge on this such as family variables which by differ by ethnicity; what happens in schools, school lunch programs and so on, what happens in peer interactions as school, what happens in the supermarket, what happens in the factories, what happens in government regulations. And one might ask what on earth can the parents do? Parents can in fact do quite a lot.

Is there any relationship between parental control and childhood eating or obesity?

There are experiments with children that do show that when parents over-control their children ant then they are released from this control from experimental manipulation of one kind or another that the child will eat more than children that don’t have an over-controlling parent. So that’s good laboratory data. But we have no idea if changing parental control will do anything for the child’s eating. But we are about to look at this with our experiment.

Within our experiment, how are you measuring parental control?

Basically we’re looking at various variables. We are looking at the actual behaviors that parents do at the table: how they serve the food, how much pressure they put on their child to eat, that kind of thing.

How do you do that without cameras on the wall?

We would love to have cameras on the wall, but that’s quite impractical because it would be a huge expense. However, we may do it later on. We do it through questionnaires. Self-report is always a problem. The Satter model encourages parents to not impinge on the child’s feeding practices and for them to take their responsibility of providing the food at reasonable times and preferably, when they can, within a family mean.

So over the course of five meetings [with parents], you are hoping to see significant changes in parent’s--? Tell me what you’re hoping for.

We’re not really hoping to change behaviors of parents, we are examining what behaviors do change with these programs. If there is no behavior change, we know that the programs aren’t very effective. However, we know from previous studies that some behaviors do change. But what we’re much more interested in is whether the behavior changes are the same in both programs, or different. If they are the same in both programs, either program is fine. If they are different, then we might want to combine both programs together. So it has implications for the kind of programs we’d use in future prevention studies.

And even if the behavior of the parents changes in both groups, how big a leap is it from there to the eating behavior of the children changing?

That’s the next study. If we do find that the parents’ behavior changes, then we would want to study both the child’s behavior and the parent’s behavior. This would mean videotaping meals in the home and laboratory--a much more detailed and expensive project.

What is your hunch about the extent to which parents can change their own behavior?

Can parents change their own behavior? Yes I think they can. I think it’s difficult, due to barriers, but I think just from the clinical point of view when we’ve done these things clinically rather than in the research mode, parents do seem to be able to change some of their behaviors.

What’s your hunch about children’s abilities to respond?

I think if parents change their behaviors then children will also change their behaviors.

In what way are the two approaches compatible?

I think these two behavior change applications are quite compatible and probably could be merged very easily into one. But for this particular effort, we are taking each one separately and looking at what behavior parents do in fact change with each one.

In a hypothetical world, how might one program compliment the other?

The two programs are definitely complimentary. One deals with what the parent might be serving to the child, which is part of a parent’s responsibility towards the child, and the other gets a little bit more into the details of the parent’s responsibilities and the child’s responsibilities.

To what extent do you think the problem can be addressed with a purely behavioral approach?

There is a whole technology for changing behavior. We can’t always use all of our expertise in the things that we’re doing, but we can put a lot of it into practice. So there is a behavioral technology. We are resting on a hundred years of the development in psychology.

Can you talk a little bit about mindfulness when it comes to eating?

Now we are looking at a much more complicated concept. And mindfulness is going to be used not so much in prevention as in attempts to get people to change their weight.


The notion is that many people are unmindful of what they eat. Eating in front of the television…rather than concentrating on the food. If we could get people to concentrate on what they’re eating, or at least see in their mind whether they are eating too much, too little, too much of the wrong sort of thing, too little of the right sort of thing, and so on. However, that’s going to be much more useful in ameliorating weight gain than in prevention.

And is it something that you think is useful?

Yes I think it’s useful, it’s a useful piece of technology. Not only does one want to change overt behavior, but one does want to change thinking. That’s a bit more difficult…there are techniques to do it.

What are your predictions for the outcome of this study?

If I knew what the outcome was, I wouldn’t do the study!

Is there a way to approach this without deprivation?

What we are trying to do is prevent children from gaining weight. Is there a way to lose weight without feeling some form of deprivation? I’m afraid there probably isn’t, because all diets tend to deprive one of something or other… desserts or chocolates, and people are going to feel deprived. In fact I think that’s one of the reasons that diets don’t work, is that the deprivation ultimately overcomes the motivation to lose weight. In fact, I don’t think one should be losing weight for aesthetic purposes. I think one should be changing one’s diet, changing ones activity levels, becoming fitter, rather than using dieting for aesthetic purposes. Fitness I think is very important.

What is [a more important motivator]: attractiveness or fitness?

That’s a controversial question. I’m probably going to get in hot water for answering this one! In my opinion, the whole aim of any form of diet, changing one’s diet, changing one’s activity level, should really be focused on fitness, not on what it does to one’s body shape and so on. One’s almost bound to be disappointed in terms of weight loss as far as it affects body shape because even the best weight loss programs probably don’t lead to more than five to eight percent change in body weight, which probably has a very marginal effect on attractiveness. On the other hand, that weight loss will change blood pressure levels, cholesterol levels, and so on, so it will improve fitness, and that I think is a far more important thing to aim for. In fact, we know that fit overweight people do better than unfit overweight people.

What are the medical consequences of the current obesity epidemic? Is it an epidemic?

I don’t like the term epidemic particularly. There’s been a good deal of increase in overweight and obesity over the last thirty years, which has probably doubled in the last fifteen years as far as it goes both in children and in adults. What are the consequences of it? We’re already seeing far more cases of diabetes for example in children and certainly as overweight children grow up we are going to see many more diabetics in the adult population as we see now in the adults who are overweight and obese. So we’ve got diabetes, we’ve got high cholesterol, we’ve got high blood pressure, we’ve got the consequences of those things: heart attracts, stroke. And then we have other consequences of obesity: social consequences, for example stigmatization of the obese. If you’re overweight or obese it’s harder to get that job, it’s harder to get into a prestigious college; it’s been shown. Those sorts of things and then other things, for example, carrying that weight around is not very good for your joints and so one gets arthritis and so on. And carrying that weight around makes one less agile. If a bus comes along you might end up underneath it, or you might trip over something and remember it’s your weight coming down. You’re going to be more likely to break something or to injure yourself worse than if one were normal weight, so it has a multitude of consequences.

I’ve got the sense that what works for one person does not work for another. Is that the case?

Absolutely. People are individuals that have totally individual different lives, different family lives, different circumstances and so on. And so definitely what works for one individual will not work for another. In terms of getting people to change behavior, it’s important to take these differences into account. We call them often barriers to behavior change. We need to know about those for each individual and help them overcome those particular barriers. One study looked at [Mexican immigrants who arrived] with relatively healthful and low-fat diets. It found that, over the course of one generation, they adopted the American diet and that heart disease and diabetes were present [where there they hadn’t been before] - they were getting the same diseases Americans were. Are you seeing [this in]children too? There is no doubt about the immigrant problem for adults. No matter what culture they come from, they tend to gain weight in the American culture. The first generation does. The first generation may have children and those children will be heavier than the offspring from within the culture they come from. Immigration does affect children and therefore there is something within the American diet and the American culture about feeding and eating and diet that does predispose to overweight. But it’s not just America, it’s the whole western world.

In general, how many approaches are there to childhood obesity? Where do the experts agree and disagree?

The traditional way of approaching prevention is to change diet, what we put into our bodies, and activity levels, how we expend calories. So that’s the sort of calorie balance model, and the second one is to attempt to change parental behavior towards their child’s feeding. Where do the experts agree or disagree? Basically we know remarkably little in terms of control studies about the Satter program. That’s actually why we’re [studying] this, we feel it’s a very promising program… to see what in fact does change with the Satter program. The NIH program is very widely used and so we want to see what behaviors are changed with that. Some experts favor the NIH program and some experts, including Doctor Satter, favor the approach of dividing the responsibilities in terms of parental responsibilities over feeding and child responsibilities over feeding and hopefully allowing the sort of natural learning of self control.

Is the increase in obesity occurring disproportionately - are the fat getting fatter? Or are all people moving up in weight?

It looks as though whole distribution has shifted towards the right: towards the obese end. I think with all these things there is probably going to be a balancing act. That is, we have a particular food environment, we have a particular set of genes, and these two things have to interact. At the moment they’re interacting in a way which facilitates growing fatter, becoming over weight, becoming obese. But that’s got to level off at some point, so I would expect as we project forward that there is going to be a leveling in the curve. I predict it’s going to flatten out at some point in the future whether it’s next year or in twenty years time I don’t think anyone at this point can say.