• How deep can we get our heads underground? – ctd.

    Yesterday’s post was about doctors’ not talking to parents about their overweight children. But in the comments, I saw a usual response, namely that this finding isn’t important, because the parents already know their kids are overweight.

    I wish.

    My anecdotal experiences in clinic are conerning. I can’t tell you how many kids I see who are overweight, but whose parents tell me they won’t eat. But anecdotes aren’t data, and so let’s go to the literature. I’m just going to give you a sample. I’m also going to stick to the findings, so I encourage you to follow the links for the methods if you are interested. Emphasis will always be mine.

    Here’s a good place to start: “Do parents accurately perceive their child’s weight status?

    RESULTS: Misclassification occurred 25% of the time (95% confidence interval: 21.4-28.5). All parents of children with a BMI greater than or equal to the 95th percentile classified their child in a category other than “extremely overweight,” and 75% of children with a BMI from the 85th to less than the 95th percentile were misclassified as “about right” or “underweight.” Boys were more likely to be misclassified than were girls (29% vs 21%, P = .03).

    CONCLUSIONS: The majority of parents of obese and overweight children underestimate their child’s weight status. Parents of boys are more likely to perceive their child’s weight incorrectly.

    Not a single parent of an obese child thought their children was “extremely overweight”. Not one. Three-quarters of parents of overweight children thought their kids were just fine, or weighed too little.

    Next up, “Parents underestimate their child`s overweight.

    RESULTS: The prevalence of overweight (including obese) was 17.4% and 21.6% in 5- and 11-year-old children in this study. Only a few parents misclassified their normal weight children as overweight. By contrast, a majority of parents to the 5-year-old children and about half of the parentsto the 11-year-old children misclassified them as normal weight. Using waist circumference for body size classification did not improve parents’ performance. Mothers performed best when estimating own weight class.

    CONCLUSIONS: A majority of parents fail to recognize overweight or obesity in their 5- and 11-year-old children. The underestimation of overweight may impair the motivation of the parents to adopt weight control.

    Most parents think their overweight or obese children are normal weight. They need to be told!

    Here’s “Underestimation of children’s weight status: views of parents in an urban community“.

    RESULTS: 193 children were included (response rate 87%, 18 months-9 years, 70% black); 31% of parents underestimated their child’s weight status (46% of overweight children, 24% of normal weight). Parents of normal-weight children who underestimated were more likely to be concerned about their child’s weight (39% vs 2.9%, P < .001) than those who did not underestimate. Parents of overweight children who underestimated were less likely to be concerned about their child’s weight (7.7% vs 59%, P < .001) than those who recognized their children as overweight.

    CONCLUSIONS: Many parents continue to underestimate their child’s weight status. These perceptions may present a barrier to the prevention of childhood obesity.

    I could do this all day. Here’s “Parents’ perceptions of their child’s weight and health“.

    RESULTS: Of the 223 children, 60% were <6 years old, 42% were male, 17% were black, 35% were Hispanic, 42% were white, and 7% were other; 19% were AROW [at risk for overweight], and 20% were overweight. Few parents (36%) identified their overweight or AROW child as “overweight” or “a little overweight” using words, but more (70%) selected a middle or heavier sketch. Among parents of overweight and AROW children, 18% recalled a doctor’s concern and 26% were worried about their child’s weight. If the overweight or AROW child was age > or =6 years, parents were more likely to identify their child as “overweight” or “a little overweight” using words, select a middle or heavier sketch, and to be worried. Parents of older children were more likely to be worried if they perceived their child as less active/slower than other children or recalled a doctor’s concern.

    CONCLUSIONS: Few parents of overweight and AROW children recognized their child as overweight or were worried. Recognition of physical activity limitations and physicians’ concerns may heighten the parent’s level of concern. Sketches may be a useful tool to identify overweight children when measurements are not available.

    So parents know what their kids look like, but they don’t think that’s “overweight” or a problem.

    Here’s “Childhood obesity: do parents recognize this health risk?

    RESULTS: Of the 83 parents surveyed, 23% (19/83) had overweight children (> or = 95th percentile of age- and gender-specific BMI growth charts). These parents did not differ from other parents in their level of concern about excess weight as a health risk or in their knowledge of healthy eating patterns, but the two groups of parents did differ in the accuracy of their perceptions about their children’s weight. Only 10.5% of parents of overweight children (2/19) perceived their child’s weight accurately compared with 59.4% of other parents (38/64; p < 0.001). Parents of overweight children invariably underestimated their children’s weight. The median difference between their perception and the growth chart percentile was -45 points.

    DISCUSSION: Given that most parents of overweight children fail to recognize that their child has a weight problem, pediatricians should develop strategies to help these parents correct their misperceptions.

    I’m going to stop now, not because I’m out of studies, but because I’m hoping by now you’ve gotten the idea.

    We have a problem here. We all seem to grasp that obesity and overweight are a real health issue, even in children. Unfortunately, we also seem unable to recognize the problem when it’s occurring in our own kids. I know this is hard to believe. But there’s a lot of evidence that it’s happening all the time. It’s compounded by the fact that most pediatricians don’t seem to be adequately pointing out the facts to their patients and families.

    The first step is recognizing that we have a problem.


    • Since so-called overweight – BMI 25-30 – is the longest-lived cohort, parents aren’t the only ones with screwed up perceptions, IMO.

      • You really believe that’s causal?

        • Aaron I’m a long reader of TIE and am seriously considering no longer coming to this site. In fact I already unsubscribed from my rss reader. My post yesterday never stated that the study you posted “isn’t important” I did not mean to be hostile. But your post comes off snarky. I would like an apology. Also your “I can do this all day” comment came off as being completely elitist. You had no right to treat me that way. I wish you luck in your studies.

          Get off your high horse and maybe I’ll return. By this I mean a formal email’ed apology.

          • Um, what? How did you take anything I wrote as directed at you personally? Unless I respond to someone by name, or link to an individual comment, there’s no reason to believe I’m directly addressing someone specific.

            • Eh I was the first one and I guess I assumed it was directed towards me. Considering you did state you would find further evidence to refute my point. Subsequently you listed a myriad of articles to display your idea. Thats good and well but I think I’m moving on from this site. Have a good day. If I appear overly sensitive I apologize.

          • Ah….Aaron was presenting evidence to back up his point. If you are upset with his counter argument and the evidence he presents then you should come back with your own evidence. He doesn’t owe you a thing. He puts his name out there. We have no idea who you are. Nice work with this blog Aaron, keep it up.

        • I am starting to suspect that a BMI of between 25 and 30 might actually be causal in in terms of all-cause mortality. Or at least the association is pretty robust and so I would be skeptical about claims that BMIs in this range are clearly harmful. Especially in an active individual (where muscle mass could also be an important component of BMI), I am hesitant to see this as concerning.

          That being said, it is clearly much worse to have a BMI > 30 than a BMI of under 25. Given that BMI tends to rise with age, it might be worth erring on the side of caution. It is true my expertise here is entirely with adults, and I freely admit that these observations may not generalize to children.

          But I agree it is hard to directly dismiss the BMI 25 to 30 and mortality findings. I am happy to start citing papers if that would be useful.

          • If your studies show which kids over the 95% BMI end up as adults with a 30% or more BMI then please post them. Or if your studies show which parents are aware the child has a weight problem and which are unaware (as a practicing pediatrician, I can agree with what Aaron says on an anecdotal basis) then please post them. In either situation, your studies will help save me time and energy. Please. Post them so I know which kids to focus on. Thank you!

            • -If the data doesn’t materialize, I’d just look at the parents. If they’re obese as grown-ups, the odds are high that their kids will be too.

              -The only kids who were borderline obese and seemed to grow out of it were kids with non-obese parents. Not sure if it was primarily nature or nuture at work (or both) but it seemed to be a pretty consistent pattern and I’d be surprised if the formal studies support a different conclusion.

            • I’m sorry. That does not help at all. Given the evidence of beta-cell proliferation in fetal pancreas development of pregnant mothers who are following the low-fat and high-carb diet recs given over the last 27 years or so. In which case, looking at the parents will give me zero guidance on the outcome of their offspring. I am looking for evidence that is more concrete.

            • PedsDocinVA:

              I was responding directly to the discussion between Scott Kurland and Aaron Carroll about BMI 25 to 30 kg/m^2. I definitely do not know aqnything about the transition of children at the 95% percentile to adults. I just want to make sure that we do not over-treat ADULTS who are ina safe weight range.

              Recent good evidence is in:

              Body-Mass Index and Mortality among 1.46 Million White Adults: N Engl J Med 2010; 363:2211-2219

              Among white men the bottom oft he curve seems to be close to identical between BMIs of 22 to 30. It’s a little less clear among white women where the curve breaks a bit earlier (around 27 kg/m^2).

              The effect modification by smoking is still something that I am trying to really understand.

              However, my main poinbt was the assymmtry of the curve. Among non-smokers, very low BMIs have hazard ratios of ~1.5. Very high BMIs have hazard ratio of ~2.5. So I can see some clinical caution as it is less risky to undershoot than overshoot.

              I am very worried about pediatric obesity, I just don’t have data to understand it well. My main worry is the youngest group (20 to 49 years of age) has HR of almost 4 (3.7) and the curve is a lot steeper. So what I think is fine for a 55 year old male (e.g. BMI 27.5) might be very concerning in a 20 year old.

              So I think pediatric obesity is a major problem.

              But I don;t want to medicalize the adult overweight (BMI 25 to 30) population given the limited evidence of increased risk and the low absolute risks in this category.

              I am find with medicalizing BMI > 30 kg/M^2 in adults due to the evidence available.

              I wish I had an answer for you re: peds. 🙁 Maybe I should poke around and see who has children in their cohorts. The main problem is that you’d need really long follow-up times to really study this, and time trends make older cohorts less useful.

              Medical research is, sadly, not for people who want perfection.

            • Sorry! I thought we were talking about whether or not we should tell parents their kids are overweight and which cutoff to use.

    • Part of the problem is that childhood obesity is a contagious illness in the sense that parents compare their children to other children they know. They won’t perceive their children as overweight if they compare them to other overweight children and the more overweight children there are the more likely parents are to compare their overweight child to another overweight child.

      • I have a simpler test. Its called survival of the fittest, if your child can outrun you or the average bully you cant kick it up the ass to make it run.harder. Its called the ‘can you escape from zombies simulation’

    • A scientific definition of the term “overweight” would be a good place to start.

    • – While taking survey results from clinical settings with a grain of salt (still think offering $10K for the correct classification would *dramatically* improve the accuracy of parental perceptions) due to the fact that there’s a distinction between social realities and aggregated survey data (e.g. the map is not the territory) …I think you’ve established that parents misperceive both their children’s weight and don’t understand the adverse health consequences. I have a bit more special pleading to do regarding the capacity for humans to selectively reason and believe two logically conflicting propositions, but I’ll leave that aside for the moment.

      -What significance, if any, do you believe that the paper below has for this discussion? IMO if $300K+ worth of education and training can’t overcome the misreporting/misperception problem, nothing can. It’s also worth pondering the nursing population when contemplating the effect that specific training and education have on these matters (they don’t appear to be significantly healthier than folks from the same demographic cohort who work in fields that have nothing to do with healthcare)

      “Pediatricians’ Own Weight: Self-perception, Misclassification, and Ease of Counseling**

      Results: The unadjusted response rate was 62%, and the adjusted response rate was 71% (n = 355). Nearly one-half (49%) of overweight pediatricians did not identify themselves as such.


    • Im sorry but BMI i sa terrible measure. When i was a kid i played a LOT of sport and trained. I was WAY over my BMI, but was far stronger and fitter than the majority of people. as a result i would be ‘overweight’ by your ruling yet it was in fighting shape. Now I have a similar BMI and am slightly overweight – I now have a nice gut 😉 but am strong still with some flabbyness so i am unhealthy. I do not deny that … the point though is that my BMI is almost the same as when i was 18.

      The figure is worse than useless because you have people trying to make the kind of statements you are. Which is why people dont use it at all when judging health. Are there too many fat kids … yeah … but the reason is that we give everyone sedentary or highly competative lifestyles with no HEALTHY middle ground.

    • @PedsDocinVA

      My apologies if I derailed the thread. You are correct, that was the point of the post. I saw the 25-30 kg/M^2 figure and jumped to adults. I once spent several months trying to show a higher risk in the 25-30 kg/m^2 group (in GPRD data so it wasn’t as high quality as the prospective cohort study). Aaron asked about causality and I figured that, since randomization isn’t possible, prospective cohorts were the best evidence that we’d have (and the 25-30 kg/m^2 was only observed in adults, so far as I know).

      I completely agree with you that identifying children in the 95% percentile who would grow up to be under BMI of 30 kg/m^2 would be of great value and allow better targeting of interventions. In the absence of such evidence, the shocking mortality rates for obese and morbidly obese young adults are . . . sobering; and I agree that ignoring this issue cannot possibly be helpful.