MCHB/EPI Miami Conference — December 7 - 9, 2005
Childhood Obesity: Weighing in for Improved Health — Transcript
DAVID LANGMEYER: Hello. Hello. There's no slide up there so you're going to have to listen to me. The live presentation, my stand-up act, is different than the handout, but none of the content that I'm going to present is not in your handout. If that makes sense. But not everything that's in your handout is going to be in my presentation, blessedly. But the handout is complete.
In looking at the other presentations throughout the day--I haven't been to a conference in quite a while--I wonder, especially with PowerPoint being so popular, whether one of the reasons is that they save money on interpreters for the deaf, since almost everybody just reads what's on the screen. Actually, my first idea was to do this in a different mode, not PowerPoint, that has audio and times the slides based on the audio, and I would just sit in the chair. No, I'm really serious. Sit in the chair and give the speech. But I didn't have time to do that. I think I'll try that the next time.
Since I'm not an epidemiologist nor is my primary work with maternal and child health, coming here is a little presumptuous and daunting. But once you start, you just have to commit and follow through. I'm on the bottom, by the way. It's not a two-way, but I'm on the bottom and that was my 60th birthday present from my ex-wife. What I say now, which is not true, is that she bought me the jump but didn't buy me the parachute. That is not true. She is very generous, even after we separated.
My topic is breastfeeding as a predictor of later body mass index in children two to 12 . Dr. Daniels said that obesity or high BMIs are a problem and she also asked what can we do about it? So this is a feeble attempt at doing something about it. And I'll talk a little bit more about how it might be used.
I want to give a lot of recognition to my co-author Ziya Gizlice, who at the time we started working was working for the state. I had worked for the state, in my past, for 12 and a half years. I have almost no respect for state employees. I have great respect for Ziya. He's one of a handful of state who employees who deserve it. I won't go on and on, he gets embarrassed. But the database that a lot of us are using--there are nine papers that are all based on a database that Ziya helped, more than helped, but start--the CHAMP. And he's going to present the details about that tomorrow morning. And it's really worth listening to. And it was done totally in-house. You know, North Carolina CDC didn't pay for it. It's wonderful that it's been done.
I don't think you need too much of this but I was told I really should put this in--facts about children's overweight. It's clearly and epidemic and I'll give more data on that. But one of the quotes is that, "Because of obesity and overweight our current teen generation may be the first in the nation's history to have a shorter life expectancy than its parents." We'll see if that occurs but it's certainly problematic. The rest of--all right. And then tying in--there's a little confusion here because I'm going to talk about obesity and breastfeeding, as Ziya said, that this can confuse you. I give you more credit than that. I think you can take both. But the literature on breastfeeding and obesity is already established but it's a little confusing about what the exact mechanisms are. And this is a quote, but I also paraphrased it a little bit. For those who actually know what the quote is you may recognize there's some sentences left out. Such is journalism these days. But you can read it for yourself. But there is already established a relationship between breastfeeding and later obesity in children. This is unpublished so I want to spend a little time on it. The North Carolina --can I go ahead? Thank you.
This is actually, I think, done in every state. I'm not sure. But this is from clinics, health clinics, children seen in health clinics, so it's a special population. And the BMIs by age group represents 30 percent, 39 and 44 as you go up, which is a little bit different than what we found in the CHAMP where the younger age group had higher BMIs. Apparently this is more common, though, that rates of obesity go up with age. But as I said, this is not a random study, this is a study of children who come into health clinics.
Wayne County itself, which is a fairly large county, about 700,000, with about six universities. We're smart people. So we do our own stuff. And we have a couple of studies that are not published that are specific to Wayne County . One of them, the first one, is not a scientifically valid study. It was not random it was opportunistic. There were six public schools in the eastern region which is generally a poor area. That doesn't mean all the kids come from that region. We have a lot of mixing geographically, busing. But this was third grade and it was measured by the physical education instructor, much to their kicking and screaming. But it was measured, height and weight. And we had a large enough sample so we did by ethnicity. And it goes up from white, you know, Caucasian, we have to call them in Wayne County . Caucasian, African-American, Latino, go up. This is 85th percentile and above. Is there anybody not familiar with the BMI? Okay, so you know what all that means.
The next slide was a follow-up to the first. The first got our interest and people started saying, "Well, let's do a more valid study, you know, one that will stand up to public criticism." We weren't thinking of this as research but as something that might be asked by County Commissioners , you know, how valid is this, et cetera.
So I won't go into great detail but we did a stratified random sample of kindergarten class--of schools I'm sorry--of kindergarten schools in Wayne County and there are quite a few, well over 60. And so this could be presented as, you know, a scientific or refereed study. It was kindergarten we used, as our source of data, the required medical examination, so it may or may not be actually weighed. We hope it is but there may be some flippage in that. Again, it goes up by race--ethnicity. And this is an older population, I mean, I'm sorry, a younger population. And if you go back and look at the rates, for third grade they're higher which is more, again, typical, that the rates go up, as I think you'll hear in our third presentation.
I want to make some points. This has come up a number of times in other presentations and--not the first one so much. There was no statistical difference in kindergarten between gender. Period. There was just none. There starts to be as you get older. But the second point is very, very important to me because I've heard it now in three or four presentations. The differences in ethnicity or anything was all due to the 95 percent and above. Let it sink in. The statistical significance between white, African-American, Latino, all of the difference was in the 95th percent and above. It was flat across for at-risk for children. And I've heard that--you know, I didn't expect that, this isn't my field. I mean, I just looked and said, oh, that's interesting. Except I've seen it four times now in these presentations. So I'm dwelling on it because there's something here and it's very different. You'd intervene very differently, I think, if, you know, at the highest level versus sort of the lower level. Okay. Maybe it's obvious to all of you and I shouldn't be making such a fuss.
So why am I interested in this? I claim I'm not--I claim, I no longer do evaluation if I can help it. I don't want to be a researcher anymore. So why get interested? Really? My, my, my. It's because we're thinking of using breastfeeding as an intervention within Wayne County . We have a pretty good participation rate for WIC for North Carolina . It could be better. And our six months duration rate is 30 percent. Again, that's very good for North Carolina , but it could be better. And breastfeeding, unlike other kinds of behavioral interventions for obesity, is passive on the part of the person who's going to be obese. For any of you who've tried to lose weight or have lost weight, stop smoking, do anything--not easy. Breastfeeding is done to you. If it really works, we've got something. I think I'll go through this. Yeah. Okay.
I'm going to skip over--we use the CHAMP, but it's already been discussed what it is so I'm going to skip those.
The dependent variables that were in my analysis are listed. I make a point very strongly that all data is manipulated. All data is manipulated. And by that I mean that you don't lie and cheat and steal--well, some of you might; some of us might--but what I mean is the model. What you get as a result is dependent on the model you use. I could've used other variables. We picked them on purpose. I mean, there was a reason to pick them. But that's a manipulation. If you only used race and you didn't use economic level it turns out race blows you away. When you start to throw in economic information, race washes out. Or may wash out, as Ziya demonstrated in an earlier paper. So, but these were the ones that we used.
We had a total of 1,844 in this particular study, two through 12. So let me get into the numbers. This is actually, again, I'm going to focus on this for a second. My interest, my hope, was that initiating breastfeeding was going to be it. You know, if you just initiated breastfeeding that was your effect because that's the easy one. Keeping them going is hard. Harder. Initiating is easy. So you do a bivariate and the Chi square is down in the lower right. It's really hard to see but it's 0.001. It's highly significant. Initiating, non-initiating. But if you look--I really liked what Dr. Daniels had done and the way she had the logistic regression association on the same chart. I wasn't able to--didn't have time to, make it look like that. But I wrote up at the top no association. Later on I'll skim it, but there was a logistic regression. No association. No association. But if you just did the bivariate, if you stayed with that, you win.
Again, all data is manipulated. You know, if I wanted to win I could've just shown you the bivariant. But it wouldn't have been true. All right. Just, again, you have this in front of you so you can see for yourself, but duration of breastfeeding did have an association, number one. And you can see the percents of overweight and at-risk based on not initiated. This was a different data point, by the way, that's why you may not have an agreement between the first slide and this one. This was a different question. Go on to the next.
By race, there was some association, wasn't very strong. We also didn't have a very good--poverty. Age group was there, and associated. Gender had some association. I don't know if I'm--you're keeping up with this. Parent education had some association. So it matches what was found before. For those interested, these are the--okay. You can look at those later on.
Let me focus on this: initiation versus non-initiation did not show an independent significant effect, much to my disappointment. Duration of breastfeeding did matter. That's been in previous literature as well. And duration of six months or longer was suggested as necessary for an independent effect on overweight, at-risk and overweight.