OBJECTIVE: Examine the rate of screening for adolescent overweight and obesity by pediatric health care professionals and the provision of advice on healthy eating and physical activity.
METHODS: Our sample contains adolescents 11 to 17 years old (6911 girls and 6970 boys) from the 2001–2007 Medical Expenditure Panel Survey who reported having at least 1 health provider visit in the previous 12 months. Using logistic regression, we investigated factors associated with whether parents reported that their children were weighed and measured and whether they or their children received counseling on their eating habits and physical activity. All models were estimated separately by gender.
RESULTS: Forty-seven percent of girls and 44% of boys who visited a health provider were advised to eat healthy, and 36% of boys and girls were advised to exercise more. Obese boys and girls were both more likely to be advised to eat healthy (odds ratio [OR] = 2.10, P < .001 and OR = 1.70, P < .001) and exercise more (OR = 2.37, P < .001 and OR = 1.90, P < .001) than adolescents who have normal weight. However, overweight boys and girls were counseled at a much lower rate than those who were obese. Adolescents who were more likely to receive such advice lived in the northeast, were from higher-income households, had parents with at least some college education, and had a usual source of medical care.
CONCLUSIONS: Greater efforts should be made to incorporate guidelines on childhood obesity screening and counseling into clinical practice.
- MEPS —
- Medical Expenditure Panel Survey
- OR —
- odds ratio
What's Known on This Subject:
The rapidly rising prevalence of overweight and obesity among children and adolescents over the past 4 decades is a significant public health concern. Experts urge pediatric health care providers to provide routine obesity screening and counseling.
What This Study Adds:
We provide the first nationally representative estimates of the rate of screening and counseling for adolescent obesity by pediatric health professionals. We also examine how socioeconomic factors and access to health care affect whether adolescents receive these services.
Obesity is a leading cause of premature mortality in the United States.1–3 Overweight and obesity in childhood and adolescence have serious immediate and long-term health consequences,4 and obese adolescents are more likely to remain obese as adults.5 Adolescents with weight difficulties are also more likely to experience psychological and behavioral problems.6–8 As a result, the rapidly rising prevalence of overweight and obesity among children and adolescents over the past 4 decades is of significant concern.9
Recognizing the severity of this problem, pediatric health professionals have been urged by various expert groups to periodically screen for overweight and obesity by using the BMI.10,11 Current clinical guidelines also recommend that health care providers evaluate a pediatric patient’s dietary patterns and activity levels at each well child visit and counsel children in the obese and overweight BMI range to improve their diet and increase their level of physical activity.12
Recent studies show varying degrees of obesity screening and counseling in pediatric practice.10,13–15 Studies that directly survey providers suggest that the rate of obesity screening and counseling10,15 is higher than found when reviewing medical records from selected practices.13,14 The only population-based study uses data from California and indicates that obese (but not overweight) adolescents received more counseling on diet and physical activity than normal weight adolescents in 2003 and 2005, but not in 2007. Specifically, the rate of nutrition counseling fell from 75% to 59% between 2003 and 2007, and the rate of physical activity counseling declined from 74% to 60%.16
By using the 2001–2007 Medical Expenditure Panel Survey (MEPS), we provide the first nationally representative estimates of the rate of screening for adolescent overweight and obesity and the provision of advice about healthy eating and physical activity by pediatric health care professionals. We also examine how body weight as well as socioeconomic factors and access to health care affect whether adolescents receive these services. The American Medical Association expert committee on childhood obesity met in 2005 to revise its earlier 1998 recommendations in light of advancing obesity rates.12 The publication of their conclusions in 2007 coincides with the last year of our sample period. Our results, therefore, reflect clinical practice leading up to the issuance of formal guidelines.
The MEPS is an ongoing nationally representative survey of the US civilian noninstitutionalized population, conducted annually since 1996, with the use of an overlapping panel design. Respondents are interviewed about their family’s medical care use over the course of 2 years through 5 interview rounds. During the interviews, data are collected on demographic characteristics, health conditions, health status, health insurance coverage, and employment for all household members.17 The MEPS also contains parent-reported height and weight information (typically by mothers), which allows the calculation of each child's BMI.
Outcome and Obesity Measures
The corresponding parent(s) were asked: “Has a doctor or other health provider ever measured (PERSON)’s height? When was that?” and “Has a doctor or other health provider ever measured (PERSON)’s weight? When was that?” From the answers to these questions, we constructed a dichotomous measure of whether each child's height and weight were measured within the past year. We did not, however, directly observe whether health providers calculated BMI or conveyed the results to patients.
Parents were also asked “Has a doctor or other health provider ever given you or (PERSON) advice about (PERSON) eating healthy? When was that?” and “Has a doctor or other health provider ever given you or (PERSON) advice about the amount and kind of exercise, sports, or physically active hobbies (PERSON) should have? When was that?” Based on the answers to these questions we derived 2 dichotomous outcome measures for whether each child was given advice about healthy eating and about physical activity within the past year.
We categorized each adolescent’s weight status by comparing his or her BMI to age (in months) and gender-specific percentiles from the 2000 Growth Charts provided by the Centers for Disease Control and Prevention.18 A child is classified as underweight if BMI is <5th percentile; healthy weight if BMI is between 5th and 85th percentile; overweight if BMI is between 85th and 95th percentile; and obese if BMI is ≥ 95th percentile.
Our sample consists of adolescents between ages 11 and 17 years with nonmissing height and weight information who had at least 1 health provider visit within the past year. Health providers may include physicians, physician assistants, nurses, or nurse practitioners. The 2001–2007 MEPS contains 21 834 total adolescents of these ages. After excluding 6560 adolescents because they did not visit a health provider during the previous 12 months, and an additional 1393 with missing height or weight, we arrived at a final sample size of 13 881, composed of 6911 girls and 6970 boys. We focus on this age group because obesity during adolescence persists into adulthood to a larger degree19 and appears to independently predict obesity-related mortality decades later.20 In addition, the nonresponse rate for height and weight in the MEPS is higher for younger children, which may reflect the greater difficulty parents have in accurately reporting weight and height for smaller, faster-growing children.21 We study boys and girls separately because interaction tests based on pooled models suggest differential effects by gender.
We estimated logistic regressions to examine factors associated with a child being measured for height and weight, and receiving dietary and physical activity counseling. The key variables of interest are the clinical weight categories based on BMI–underweight, overweight, obese–with healthy (normal) weight as the reference group. All models control for the child’s age, race, and ethnicity, Census region (Midwest, West, South, reference group Northeast), household poverty status (<100% federal poverty line, 100% to <125%, 125% to <200%, 200% to <400%, reference group ≥400%), insurance status (Medicaid, uninsured, reference group private), and lack of usual source of care (= 1 if there is no health provider that the child usually goes to when sick, = 0 if there is), as well as the responding parent’s marital status, education level (high school, some college, bachelor’s degree, reference group less than high school), and obesity status. We also include indicators for survey year to capture aggregate nonlinear trends in clinical practice. The MEPS sample design uses complex sampling techniques including stratification, clustering, and differential sampling rates for selected population subgroups.17 We use the Stata 11.0 survey logistic command, which accounts for the complex survey design of the MEPS to generate estimates that are nationally representative of the US noninstitutionalized population.
Appendix 1 contains descriptive statistics for our analytical sample. Seventy percent of adolescent girls who visited health providers’ offices were in the healthy weight range, 4% were underweight, 14% were overweight, and the remaining 11% were obese. More boys were overweight or obese (17% each) than girls; 60% of boys were in the healthy weight range and 5% were underweight. Our analytical sample was similar to the subgroups that were excluded because of missing height or weight information (Appendix 2) or because they did not visit a health provider within the previous 12 months (Appendix 3). However, the excluded groups were more likely to be nonwhite (58% and 49%, respectively, vs 33% in the analytical sample), from lower income families (25% and 18% below the poverty line, respectively, vs 13%), and to have parents with less education (14% and 15% with a bachelor’s degree vs 29%). In addition, among girls who did not have a health visit a higher proportion were obese (13% vs 11%) and overweight (16% vs 14%).
The large majority of adolescents had their height and weight measured during health provider visits (Table 1). Forty-seven percent of girls and 44% of boys received advice about healthy eating from their health providers, and fewer were advised to exercise more. The rate at which health providers gave advice did increase after 2005 (Fig 1), particularly for girls. Because the estimates in Fig 1 are based on adolescents who visited a health provider, the overall prevalence rates of advice are lower. Extrapolating to the full population, only 38% of girls and 35% of boys received healthy eating advice, whereas 30% of boys and girls received advice on exercise.
After controlling for child and parent characteristics and aggregate time trends by using logistic regression, we found that adolescents in all ranges of the BMI distribution were equally likely to have their height and weight measured by health providers (Table 2). Both boys and girls without insurance or without a usual source of care were significantly less likely to have their height and weight measured, as were those who lived outside the Northeast.
Table 3 presents the results from logistic regressions of whether adolescents were advised by health providers to eat healthy. Obese and overweight girls had higher odds of receiving dietary advice from their health providers (OR 2.098, P < .001, and OR 1.442, P < .001) compared with their normal weight counterparts. In contrast, only obese boys, but not those who were overweight, received more counseling than normal weight boys (OR 1.703, P < .001). Both overweight boys and girls had statistically significant lower odds of receiving dietary counseling than the obese (P = .003 and P < .001), and the magnitude of this difference (60%–70% lower odds) is substantial.
In Table 4, we report our findings from logistic regressions of whether adolescents were advised to exercise more by health providers. Obese and overweight girls had higher odds of receiving physical activity advice (OR 2.366, P < .001 and OR 1.370, P = .001) than their normal weight counterparts. Both obese boys and overweight boys also received more exercise advice than boys of normal weight (OR 1.903, P < .001 and OR 1.258, P = .045). Overweight boys and girls were significantly less likely to receive physical activity counseling than the obese (P < .001).
Many of the same demographic and socioeconomic factors predict higher odds of receiving obesity counseling for both girls and boys, including higher levels of parental education, living in the Northeast, higher family income, and having a usual source of care. In addition, both black and Hispanic adolescents were more likely to receive advice on proper diet than whites (P = .006 and P < .001), whereas only Hispanic adolescents were more likely to receive advice on physical activity (P = .075 and P < .001).
In alternative model specifications we included a time trend instead of the dichotomous survey year indicators and found that trends in screening and counseling were positive and statistically significant. However, we chose to present specifications with separate year indicators because they reveal that the aggregate upward trend in screening and counseling is nonlinear. This variation in specification does not impact our results, nor does restricting sample to only well-child visits or to visits made exclusively to pediatricians.
To our knowledge, this study is one of the first to examine the state of obesity-related counseling to adolescents over an extended period in the United States. Although they were more likely to receive advice on healthy eating or physical activity than their normal weight peers, we find that a significant number of obese and overweight adolescents did not receive advice from medical professionals during the past year. Furthermore, overweight adolescents were counseled at a much lower rate than obese adolescents. This is troubling because experts believe that obesity is easier to prevent than to treat, which implies that physician counseling for the overweight, but not yet obese may have the greatest potential to prevent obesity later in life.22,23 We do, however, find that overall rates of obesity counseling improved toward the end of our study period. This may reflect providers’ increasing awareness of childhood obesity, possibly from publicity related to the convention of the American Medical Association expert committee on childhood obesity in 2005.12
The only national estimates of obesity counseling are from the National Ambulatory Medical Care Survey15 and suggest that 37% of all general medical visits, and 58% of visits by overweight or obese adolescents, involve some mention of a weight-related topic. These estimates are, however, not directly comparable with ours because the National Ambulatory Medical Care Survey samples doctor visits and the data are reported by health providers, whereas the MEPS samples individuals and contains every visit reported by the individual for the entire year.
Whether adolescents receive obesity-related counseling varies consistently with several factors. One previous study found that low-income children enrolled in public insurance were more likely to receive preventive care.24 Our results suggest that uninsured adolescents were less likely to be screened for obesity, but that, among those with insurance, the source of insurance was not a significant determinant of either screening or counseling. We do find, however, that adolescents were more likely to receive advice on diet and exercise if they lived in a higher-income household and had parents with a college education. One possible explanation is that parents with more education and economic resources have easier access to health providers who are more aware of and compliant with clinical guidelines. This could occur through enrollment in health plans that selectively contract with more compliant health providers, or through community knowledge of which health providers provide higher-quality care. Another interesting finding is that both black and Hispanic adolescents were more likely to receive advice about diet than whites, which could reflect awareness among health providers that minorities are at greater risk for obesity.25
Previous studies have found that individuals without a usual source of care receive less preventive care, and that those without health insurance, minorities, and the poor are more likely to fall into this group.26,27 We find that, even after explicitly controlling for demographic factors, not having a usual source of care independently lowers the probability that adolescents are screened for obesity and receive counseling. This may result from adolescents visiting facilities, such as walk-in clinics or hospital emergency departments, which are more focused on acute care and devote fewer resources to tracking BMI.
Our analysis suggests that health providers may have missed important opportunities during their encounters with adolescent patients to provide counseling on diet and physical activity. Although more research is needed to identify and address barriers to counseling, possible reasons for limited levels of counseling include lack of awareness of the risks of childhood obesity,28 lack of training, and lack of confidence in the efficacy of giving such advice to adolescents.29–31 In addition, third-party payers typically do not reimburse providers for visits that are extended to provide counseling. Nonetheless, the increasing prevalence of obesity counseling coinciding with dissemination of the 2007 guidelines is promising.
One study on adolescents found that physician obesity counseling is associated with attempted weight loss and moderate changes in dietary behavior.32Likewise, advice on exercise given by physicians to adults during primary care visits is associated with more weight-loss attempts,33 improved diet,34 and greater physical activity.35–37 However, physicians recognize the need to receive training on how to best incorporate obesity counseling into routine office visits.38,39 One mechanism for helping physicians develop a specific protocol and operational plan for providing counseling is continuing medical education, and recent evidence suggests that, at least for adult patients, training does improve the quality of obesity-related counseling.40,41
Research also suggests that one of the most important things a physician can do is to make patients aware of their weight status. Significant body size misperception is a barrier to weight management by parents and caregivers of preschoolers and elementary school age children,42,43 adolescents44 and adults,45 and at least 1 study identifies the absence of pediatrician comment on child weight as a strong predictor of misperception.42 Moreover, overweight boys and girls who accurately perceive their weight are more likely to report having attempted weight loss.44 These results highlight the need for health providers to discuss weight management with their patients.
Although multicomponent and high-intensity behavior intervention may be needed for some obese children and adolescents to improve their weight status,46 evidence suggests that innovative interaction between physicians and parents can encourage the behavioral change needed to prevent excessive weight accumulation. For instance, Tarveras et al47 reported that parents were more proactive in preventing their children from becoming overweight when pediatricians evaluated their confidence and readiness to initiate weight-related behavioral changes in their children.
Several limitations of our study must be kept in mind. First, we rely on parent-reported measures of height and weight to determine child BMI, and although these measures are more accurate for adolescents than for younger children, parents have a tendency to underreport adolescent body weight, particularly for girls.21 Furthermore, there may be differential misclassification of BMI if, for example, parents whose adolescents were weighed and measured by providers make more accurate reports than those who were not. In addition, BMI may have changed between the last health provider visit and time of survey. Overall, it is difficult to determine the direction and magnitude of the misclassification of weight status from these different sources, although categorical weight status measures should be more robust than BMI. Second, our measure of physician advice is also parent reported, and some parents may not have heard or accurately remembered the advice. Likewise, if parents of overweight and obese adolescents have better recall than parents of normal weight adolescents with no weight concerns, the differences in counseling rates we report will be overstated. Third, racial minorities and those of lower socioeconomic status were disproportionately represented among the 36% of the sample excluded because they did not report weight or height or did not visit a medical provider in the past year.
Obesity is a serious health concern, and obese youth are more likely to remain obese as adults. Obesity also disproportionally affects Hispanic and black adolescents,25 who are known to have poorer access to health care. Greater efforts should be made to incorporate recommendations from the 2007 guidelines on childhood obesity screening and counseling into clinical practice through targeted training of pediatrics health professionals. In addition, policies designed to provide better continuity of care to adolescents might help improve access to obesity-related counseling, and ultimately, reduce obesity-related health disparities.
- Accepted March 22, 2012.
- Address correspondence to Lan Liang, PhD, Agency for Healthcare Research and Quality, 540 Gaither Rd, Suite 5000, Rockville, MD 20850. E-mail: Liang:
The views expressed in this paper are those of the authors, and no official endorsement by the US Agency for Healthcare Research and Quality or the US Department of Health and Human Services is intended or should be inferred.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
- Strauss RS
- Janssen I,
- Craig WM,
- Boyce WF,
- Pickett W
- Klein JD,
- Sesselberg TS,
- Johnson MS,
- et al
- Whitlock EP,
- Williams SB,
- Gold R,
- Smith P,
- Shipman S
- 12.Barlow SE; the Expert Committee. Expert Committee recommendations regarding the prevention, assessment, and treatment of child and adolescent overweight and obesity: summary report. Pediatrics. 2007;120(suppl 4):S164–S192
- Dilley KJ,
- Martin LA,
- Sullivan C,
- Seshadri R,
- Binns HJ,
- Pediatric Practice Research Group
- Jasik CB,
- Adams SH,
- Irwin CE Jr,
- Ozer E
- 17.↵Cohen J. Methodology Report # 1: Design and Methods of the Medical Expenditure Panel Survey Household Component. AHCPR publication 97-0026. Rockville, MD: Agency for Health Care Policy and Research; 1997
- 18.↵Centers for Disease Control and Prevention, National Center for Health Statistics. CDC Growth Charts: United States. May 30, 2000. Available at: www.cdc.gov/growthcharts/. Accessed July 17, 2007
- Guo SS,
- Chumlea WC
- Dietz WH
- World Health Organization
- Weinick RM,
- Zuvekas SH,
- Cohen JW
- 27.↵Kirby JB, Berdahl T. Trends in Access to Health Care: Disparities by Poverty Status and Health Insurance 1996—2006. Statistical Brief #265. Rockville, MD: Agency for Healthcare Research and Quality; 2009 Available at: www.meps.ahrq.gov/mepsweb/data_files/publications/st265/stat265.pdf. Accessed March 7, 2010
- Barlow SE,
- Dietz WH,
- Klish WJ,
- Trowbridge FL
- Kant AK,
- Miner P
- Story MT,
- Neumark-Stzainer DR,
- Sherwood NE,
- et al
- Hernandez RG,
- Cheng TL,
- Serwint JR
- Zeller MH,
- Ingerski LM,
- Wilson L,
- Modi AC
- Edwards NM,
- Pettingell S,
- Borowsky IW
- Barton M,
- US Preventive Services Task Force
- Taveras EM,
- Mitchell K,
- Gortmaker SL
- Copyright © 2012 by the American Academy of Pediatrics