 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
SUPPLEMENT ARTICLE |
a Department of Pediatrics, University of Colorado School of Medicine, Denver, Colorado
b Division of Epidemiology and Community Health, University of Minnesota School of Public Health, Minneapolis, Minnesota
c Office of Disease Prevention and Health Promotion, Department of Health and Human Services, Rockville, Maryland
d Department of Pediatrics, Baylor College of Medicine, Houston, Texas
e PS duPont Elementary School, Wilmington, Delaware
f Department of Pediatrics, University of California, Davis, Sacramento, California
 |
ABSTRACT |
|---|
 TOP ABSTRACT MEASUREMENTS OF OVERWEIGHT AND...ASSESSMENT COMPONENTS OF THE...CLINICAL DATA: PHYSICAL...REFERENCES |
|---|
Key Words: obesity • clinical assessment • BMI • diet • physical activity
Abbreviations: IOTF—International Obesity Task Force • CDC—Centers for Disease Control and Prevention • PPV—positive predictive value • DGAC—Dietary Guidelines Advisory Committee • HDL—high-density lipoprotein • LDL—low-density lipoprotein • T2DM—type 2 diabetes mellitus • CVD—cardiovascular disease • WAVE—weight, activity, variety, and excess
Obese and obesity are terms commonly used in the clinic as well as on the street corner, often with a wide range of meanings. For medical purposes, obesity refers to excess body fat; however, the exact meaning of excess has not been defined. Obesity most often is regarded as an excess percentage of body weight that is fat, but no widely accepted diagnostic definitions or cutoff points are available for children. For an understanding of developmental patterns, mean body fat percentages (derived from bioelectrical impedence analyses) are available for US children >12 years of age,1 and percentile curves have been published for British children 5 to 18 years of age.2
|
MEASUREMENTS OF OVERWEIGHT AND OBESITY |
|---|
|
TOPABSTRACTMEASUREMENTS OF OVERWEIGHT AND...ASSESSMENT COMPONENTS OF THE...CLINICAL DATA: PHYSICAL...REFERENCES |
|---|
Several expert and advisory groups have recommended BMI as the preferred measure for evaluating obesity among children and adolescents 2 to 19 years of age.4–7 BMI expresses the weight-for-height relationship as a ratio, that is, weight (in kilograms)/[height (in meters)]2. Experts recommend BMI because it can be obtained easily, it is correlated strongly with body fat percentage (especially at extreme BMI levels), it is associated only weakly with height, and it identifies the fattest individuals correctly, with acceptable accuracy at the upper end of the distribution (eg, 
85th or 95th percentile for age and gender).
In 1994, the Expert Committee on Clinical Guidelines for Overweight in Adolescent Preventive Services recommended that children whose BMI exceeds 30 kg/m2 or is 95th percentile for age and gender (whichever is smaller) should be considered overweight.5 The BMI limit of 30 kg/m2 was recommended because, at the oldest ages of adolescence for which the 95th percentile values exceed 30 kg/m2 (>17 years), elevated BMI is associated with early adulthood patterns of risk for obesity-related disease and death,8,9 as well as to provide continuity with recommendations for adults. The expert committee considered adolescents whose BMI was 85th percentile but <95th percentile to be "at risk of overweight." The committee deliberately avoided the term obese, because of the inference regarding adiposity and body composition and the inability of height and weight data, even as BMI, to measure total body fat specifically. These definitions are considered standard in describing the weight status of children and adolescents 2 to 18 years of age.4,10
In 2005, the Institute of Medicine consciously departed from the previously described terminology and elected to define children with BMI of 95th percentile for age and gender as obese, rather than overweight.6 The Institute of Medicine report conveyed the seriousness, urgency, and medical nature of childhood obesity, as well as the need to take action. The current expert committee endorses the position of the Institute of Medicine report and recommends that individuals 2 to 18 years of age with BMI of >30 kg/m2 or 95th percentile for age and gender (whichever is smaller) should be considered obese (see summary report). We think that the nature of the current epidemic and the need for medical professionals and others to address the problem actively justifies this change.11 Moreover, we recommend that individuals with BMI of 85th percentile but <95th percentile or 30 kg/m2 (whichever is smaller) now be considered overweight and that this term replace the term "at risk of overweight." The Institute of Medicine report published in 2005 was silent regarding this category of BMI.
The expert committee concluded that the scientific data linking elevated BMI to risk factors and morbidity,12,13 as well as the difficulty of changing early trajectories of weight gain, support the change in terminology. The terms overweight and obese also may be easier than "at risk of overweight" for parents to understand. This new terminology will allow US medical practice to parallel the recommendations of the International Obesity Task Force (IOTF)14 and to align with the International Classification of Diseases, Ninth Edition, diagnosis codes. Finally, these changes in descriptions of weight status for children and adolescents will provide continuity with the recommended adult cutoff points of BMI of 25 kg/m2 and 30 kg/m2 for overweight and obesity, respectively.15 We are aware that the BMI categories for adults that are linked to minimal subsequent mortality rates are not without controversy.16 However, the threshold levels of 25 and 30 kg/m2 for adult BMI are still recommended by national and international organizations.7,15
For children 
2 years of age, the weight-for-recumbent length percentiles from the CDC 2000 growth charts are appropriate for evaluating weight relative to linear growth, but the term obese generally should not be applied to children this young. Weight-for-length percentiles of 95th identify these children as overweight.3
Little evidence is available regarding the most effective way to evaluate the severity of obesity for children with BMI of >97th percentile (the highest level on the CDC growth charts). Inge et al17 recommended that bariatric surgery for adolescents should be restricted to those with BMI of 40 and significant comorbidities that may be improved with surgery. In research settings, age-specific z scores or SD scores are used for extreme values of anthropometric measures. These scores describe the number of SD units above or below the median for the individual value. For example, in a normally distributed population, the 99th percentile is equivalent to a z score of 
3.0. Unfortunately, a computer program is needed to calculate BMI z scores, and many clinicians are unfamiliar with their use and interpretation. A BMI z score calculator is available on the Internet (www.kidsnutrition.org/bodycomp/bmiz2.html). Limited available data suggest that BMI-for-age values of 99th percentile are associated strongly with the presence of comorbidities, excess adiposity, and persistence of obesity into adulthood.18 This severity of obesity may well warrant more-aggressive therapeutic interventions. Although they are not available currently on growth charts, more-routine availability of 99th percentile BMI cutoff points would likely be valuable for tailoring optimal treatment approaches.
Anthropometric Methods and Determination of BMI
Weight, height (sometimes referred to as stature), and recumbent length of children are measured routinely in most clinics. Nevertheless, the importance of careful accurate measurements should be emphasized to clinic staff members. Staff members should take particular care when BMI is be calculated, compared to reference data, and made the basis for important decisions regarding the child's health. Detailed protocols are available for measuring recumbent length, height, and weight in a manner comparable to that for reference data.19
BMI may be calculated directly as weight (in kilograms)/[height (in meters)]2 or determined from published tables or nomograms.5,20,21 Many BMI tables, nomograms, and calculator programs are available online (eg, www.cdc.gov/nccdphp/dnpa/bmi/calc-bmi.htm or http://nhlbisupport.com/bmi/bmicalc.htm). The National Heart, Lung, and Blood Institute provides a free program for calculating BMI on hand-held devices (http://hp2010.nhlbihin.net/bmi_palm.htm). If BMI is calculated from height and weight measured in inches and pounds, then the formula is BMI = [weight (in pounds)/[height (in inches)]2] x 703. Some BMI tables and charts are designed for adults and either may not accommodate the smaller heights and weights appropriate for children or may not provide age/gender-specific percentiles.
Development of BMI References and Implications
The developmental pattern for BMI differs somewhat from the more-familiar patterns for height and weight (Fig 1). The normal pattern is for BMI to decrease from 2 years of age until 5 or 6 years of age and to increase thereafter. This early decrease in BMI reflects a corresponding decrease in subcutaneous fat and the percentage of body fat.22 The resulting V-shaped pattern in early childhood has been termed the "adiposity rebound."23 It coincides with the period between the ages of 4 and 7 years when BMI reaches its nadir and then begins to increase through the remainder of childhood and into young adulthood. Early adiposity rebound has been cited as a risk factor for the development and persistence of later obesity.24 More-recent analyses suggest that this primarily is a reflection of rapid weight gain during infancy and early childhood and that it identifies young children with high BMI percentiles and/or children who are crossing percentiles upward.24,25 Rapid weight gain in infancy, including during the first week,26 the first 4 months,27 and the first year,28 has been found to predict later obesity. In one prospective cohort, increased weight gain during the first 3 years of life was associated independently with higher BMI, fat mass, and waist circumference at 17 years of age.29 For clinical purposes, the utility of assessing adiposity rebound is limited, because it is difficult to determine for an individual child and it is, by definition, a retrospective determination. Identification of the age of adiposity rebound as a strategy for clinicians to identify children at risk of overweight or obesity is unlikely to contribute more than plotting of weight and length for age and determination of BMI percentiles for young children.
|
An attractive aspect of BMI is that it correlates closely with total body fat30,31 and other risk factors for obesity-related morbidity in adults.32,33 Such correlations are based on the joint associations of the entire distributions of BMI and related outcomes. Interpretation of assessments of overweight in children using only BMI for age and gender should include the realization that some children may have relatively high weights primarily because of high lean mass, rather than high body fat levels. This is most common among male adolescents, for whom gains in BMI during adolescence may have a large component of lean mass.34,35 At 95th percentile, however, almost all of those who are identified as obese on the basis of BMI have high weights because they have high total body fat levels. In clinical practice, the important question is whether the criteria for overweight and obese that are based only on the upper portion of the BMI distribution (ie, 85th percentile or 95th percentile) identify correctly the fattest children and those at greatest health risk.
This sort of categorical identification often is evaluated by using the same sensitivity/specificity approach that is used to evaluate medical screening procedures.36 Sensitivity (or the true-positive rate) in this case represents the proportion of children who are considered the fattest with a standard method for assessment of total body fat (eg, dual-energy X-ray absorptiometry) and who also are identified correctly with the BMI criteria. The complement of sensitivity is specificity, or the proportion of children who are considered not the fattest with the standard method and who are identified correctly as not overweight or obese with the BMI criteria. Finally, the positive predictive value (PPV) is the proportion of children who are identified as overweight or obese with BMI criteria who are truly the fattest children identified with the standard method. The PPV is important for clinical applications because its complement (1 – PPV) is an estimate of the proportion of children who may be identified incorrectly as overweight or obese when BMI is used. Such children may be labeled, treated, or referred inappropriately.37
The sensitivities of the 85th BMI percentiles on the CDC 2000 growth charts in identifying correctly the fattest children range from 75% to 93% in several studies, and the corresponding specificities range from 67% to 96%.30,38,39 The accompanying PPVs (presented or calculated from sensitivity, specificity, and prevalence values) range from 61% to 98%. The sensitivities of the 95th BMI percentiles on the CDC charts in identifying correctly the fattest children range from 54% to 100%, and the corresponding specificities range from 96% to 99%.38,39 PPVs for the 95th BMI percentile criterion range from 56% to 99%.
Some of the aforementioned estimates of sensitivity, specificity, and PPV are difficult to compare directly across studies, because of the differing samples and standard criteria used. Nevertheless, several general conclusions can be drawn from these and similar studies. Most important is that the BMI criteria, although imperfect, perform reasonably well in identifying correctly children who have the highest percentages of body fat. As the BMI criteria become more restrictive (ie, 95th percentile versus 85th percentile), the sensitivities in identifying the fattest children decrease and the specificities increase. Finally, the specificities and PPVs are almost always higher than the corresponding sensitivities. This means that there should be relatively few children diagnosed incorrectly as overweight or obese by using BMI.
The CDC 2000 BMI-for-age percentiles are recommended for US children from all racial/ethnic backgrounds. Some evidence exists that, in general, black children tend to have relatively less body fat and Mexican American children tend to have relatively more body fat, compared with white children with the same BMI.40 Also, South Asian adolescents living in England have higher percentages of body fat than do their peers of European heritage with the same BMI.41 Because the racial/ethnic differences in body fat/BMI relationships have not been described fully for children, however, the same BMI reference values are currently recommended for assessment of all children. Any racial/ethnic differences in health risks assessed by using the CDC BMI reference values should be small.
The validity of using high BMI to identify children with the highest total body fat levels seems to be approximately the same for healthy children and children receiving growth hormone, children with inflammatory bowel disease, and children treated previously for malignancy.42 Differences between boys and girls in the sensitivities and specificities of BMI for identifying the fattest children are inconsistent and probably not important clinically.30,39,43 Similarly, available data show no consistent age patterns in BMI sensitivities and specificities between the ages of 6 and 18 years.
Pediatric obesity is associated with increased risks of concomitant psychological or psychiatric problems, cardiovascular risk factors, chronic inflammation, type 2 diabetes mellitus (T2DM), and asthma.33,44 In an important study, Katzmarzyk et al45 assessed the validity of BMI and waist circumference criteria for overweight and obesity for identifying correctly youths 5 to 18 years of age who had 3 of 6 risk factors (low high-density lipoprotein [HDL] cholesterol levels, high low-density lipoprotein [LDL] cholesterol levels, high triglyceride levels, high plasma glucose levels, high plasma insulin levels, or high blood pressure). The overall sensitivities and specificities for BMI of 85th percentile were 69% and 76%, respectively, and those for BMI of 95th percentile were 49% and 90%, respectively. These sensitivities and specificities are quite low, and the corresponding PPVs calculated from the authors’ data are 36% and 50%, respectively, for BMI of 85th and 95th percentiles. Therefore, even among children with 3 risk factors, the least-restrictive and therefore most-sensitive BMI cutoff point (BMI of 85th percentile) still identified correctly only approximately two thirds. Moreover, of all children with BMI of 85th percentile who were considered overweight, approximately two thirds did not have 3 risk factors. The authors concluded that waist circumference added substantially to BMI alone for assessment of cardiovascular disease (CVD) risk.45 If these results can be replicated in other samples, they argue strongly that BMI criteria by themselves are insufficient to identify children who are most likely to have clusters of risk factors and that additional screening and assessment criteria should be applied to estimate risks.
Implications for Overweight and Obese Children in Adulthood
Systematic reviews confirm the persistence of obesity from childhood into adulthood.46 Predictably, the higher the BMI is in childhood, the greater the probability is of obesity in adulthood. Guo et al47 analyzed lifelong data from the Fels Longitudinal Study and estimated the probabilities of having a BMI of 30 kg/m2 at 35 years of age. For girls with BMI of 95th percentile during childhood, the probabilities of being obese as an adult were 20% to 39.9% from 3 to 5 years of age, 40% to 59.9% from 6 to 11 years of age, and 60% from 12 to 20 years of age. For boys with BMI of 95th percentile during childhood, the probabilities of being obese as an adult were <20% from 3 to 4 years of age, 20% to 39.9% from 5 to 11.5 years of age, 40% to 59.9% from 11.5 to 16 years of age, and 60% from 17 to 20 years of age.
For children with BMI of 85th percentile during childhood, the probabilities of adult obesity were lower. For girls, the probabilities of being obese as an adult were <20% from 3 to 4 years of age, 20% to 39.9% from 5 to 17 years of age, and 40% to 59.9% from 18 to 20 years of age. For boys with BMI of 85th percentile during childhood, the probabilities of being obese as an adult were <20% from 3 to 16 years of age, 20% to 39.9% at 17 years of age, and 40% to 59.9% from 18 to 20 years of age. On the basis of these data, the odds ratios for being obese (BMI of 30 kg/m2) at 35 years of age were 19.3 for boys and 15.7 for girls if BMI at 18 years of age was >72nd percentile (the most discriminating level). Clearly, if individuals end their adolescence with moderately elevated BMI, then the likelihood of obesity as an adult is high.
Overweight and obesity in childhood and adolescence have been associated with adverse socioeconomic outcomes, increased health risks and morbidities, and increased mortality rates in adulthood.32,33,46 Must et al48 studied children in Boston (13–18 years of age) who were evaluated initially between 1922 and 1935 and were assessed in 1988. Compared with those with BMI of 25th to 50th percentile in adolescence, those with BMI of >75th percentile in adolescence had increased heart disease, atherosclerosis, T2DM, colorectal cancer (men), gout (men), hip fracture (women), arthritis (women), and all-cause mortality (men) rates.
Alternative Reference Data and Measures of Fatness
IOTF Standards
In 2000, reference BMI categories based on 6 pooled international data sets were developed for children 2 to 18 years of age.14 These reference curves have become known as the IOTF standards. They assume that the most-appropriate cutoff points for overweight and obesity in children are those corresponding to the locations of BMI of 25 kg/m2 and 30 kg/m2, respectively, in the BMI distribution for adults, points that are recognized internationally as defining overweight and obesity.49 Particularly outside the United States, the IOTF standards have been widely used to classify overweight and obesity in children. The IOTF charts provide only overweight and obesity categories and not a full array of percentile levels. Therefore, they are not recommended for monitoring the BMI progress of individual children. Sensitivities and specificities of IOTF cutoff points in identifying the fattest children and predicting adult morbidity are similar to those of the CDC 2000 BMI charts.39,50 The IOTF reference values are not recommended for routine clinical use.
In various research settings and in the scientific literature, measures other than BMI that are related to childhood fatness and obesity are used frequently. These were investigated, and consideration was given to their appropriateness for routine clinical use in the assessment of pediatric overweight and obesity, as well as whether each provides important information beyond that available from BMI.
Skinfold Thickness
Skinfolds are double, compressed thicknesses of subcutaneous fat and skin that are measured with standardized calipers at selected sites (eg, triceps, subscapular, and suprailiac sites).51 Skinfold measurements have a long history in studies of nutrition and body composition. They are considered attractive research tools because measurements are noninvasive and specific to subcutaneous fat.52 Previous expert committees considering childhood obesity recommended that skinfold measurements be included in in-depth medical assessments, to distinguish those who are overweight from those who are overfat.5,53
Without question, skinfold thicknesses are predictive of total body fat in children and adolescents.54,55 Moreover, when skinfold measurements are included in regression models, they provide unique information beyond height and weight in accounting for variations in risk indicators, including blood lipid levels, lipoprotein levels, blood pressure, plasma glucose levels, plasma insulin levels, insulin resistance, and inflammation.13,56–59
When categories of skinfold thicknesses or ratios based on percentile cutoff points are used to identify the fattest individuals or those with metabolic syndrome, the skinfold measurements perform as well as BMI or waist circumference values.60–63 Nevertheless, there is little evidence that, once height and weight (or BMI) are known, skinfold thickness categories increase the accuracy of identifying those with the most total body fat or other risk factors.
Therefore, the expert committee does not recommend the routine clinical use of skinfold thickness measurements in the assessment of childhood obesity. The basis for this conclusion includes the lack of readily available reference data on skinfold thicknesses for US children, the considerable potential for measurement errors without rigorous training and regular experience,51,55 and the lack of optimal criteria as a basis for intervention.
Waist Circumference
Waist circumference has attracted much recent attention as an indicator of fatness and health risks in children and adults. The interest in waist circumference stems from research linking accumulated visceral adipose tissue to increased health risks and metabolic disorders in children and adults.45,64,65
Compared with BMI, waist circumference in children provides a better estimate of visceral adipose tissue measured with MRI at the level of the fourth lumbar vertebra (65% vs 56% of variance), whereas BMI is better at estimating subcutaneous adipose tissue (89% vs 84% of variance).66 In multivariate regression models, waist circumference is significantly more efficient than BMI in predicting insulin resistance, blood pressure, serum cholesterol levels, and triglyceride levels.67–69 Consequently, measurements of waist circumference provide unique predictive information regarding health risks, especially for adolescents.
The overall ability of waist circumference percentile cutoff points to identify the fattest boys (as assessed with areas under the receiver operating curves), however, was no greater than that of triceps skinfold or BMI percentiles.63 Also, Moreno et al61 found no overall differences in the ability of BMI, waist circumference, and triceps/subscapular skinfold ratio cutoff points to identify correctly Spanish children with the metabolic syndrome.45
Translation of the available information on waist circumference into meaningful clinical application for the assessment of overweight and obesity in children is difficult. No data are available to identify waist circumference cutoff points that are appropriate for identifying children with the most visceral fat or the greatest risk for cardiovascular or metabolic problems, having been identified as overweight or obese through BMI. Consequently, it is not known exactly which waist circumference percentile clinicians should use and what clinical actions that value would indicate. Nevertheless, clinicians, especially those in subspecialty referral settings, may add waist circumference to the tools they use to assess risk. If they do, clinicians should use a high, age-specific, percentile cutoff point, such as the 90th or 95th percentile, to evaluate risk.
Waist circumference may prove useful in the future, but the expert committee withheld recommending it for routine clinical use at the present time because of incomplete information and the lack of specific guidelines for clinical application. Waist circumference percentiles are now available for US children70 and for other populations.71–73 One possible approach may be to calculate the best waist circumference cutoff points for identifying at-risk children within BMI categories, as has been proposed for adults.74
|
ASSESSMENT COMPONENTS OF THE MEDICAL HISTORY |
|---|
|
TOPABSTRACTMEASUREMENTS OF OVERWEIGHT AND...ASSESSMENT COMPONENTS OF THE...CLINICAL DATA: PHYSICAL...REFERENCES |
|---|
Health Behavior Changes
The history portion of assessment of childhood obesity should be directed, in part, toward identifying modifiable behaviors. Physicians and other health care professionals are more likely to provide successful treatment if they work with patients to target behaviors for change, rather than working from a "top-down" approach. Several approaches are available to negotiate lifestyle behavior changes that can improve health. The principles described below are intended for use with overweight or obese patients, but they apply to any circumstance in which health behavior changes are desired.
Self-efficacy is the personal belief that one can attain or accomplish successfully what one sets out to do. Because patients and families are more likely to do what they perceive to be both pleasant and feasible, providers should assess which activities patients enjoy and think they are capable of performing. Interventions and recommendations should be tailored accordingly.75–79
"Readiness to change" is a behavioral approach that assesses an individual's readiness to adopt a particular behavior (otherwise known as the transtheoretical model).80,81 This approach stresses the interest in and motivation for thinking about, starting, or maintaining a behavior and allows for tailored messages and interventions based on 5 stages of change, as follows: stage 1, precontemplation; the patient is not yet considering the change; stage 2, contemplation; the patient is evaluating reasons for and against the change; stage 3, preparation; the patient is planning for the change; stage 4, action; the patient has made the change (<6 months); stage 5, maintenance; the patient has maintained the change (>6 months).
Individuals may not go through each step sequentially, and they may not spend the same amount of time in each stage. Behavior is seen as a dynamic process and not an "all-or-none" phenomenon. Assessing a person's stage of change acknowledges the patient's attitudes, respects his or her perspective, and is a vital step in ensuring that the behavioral intervention is delivered in a manner that is most beneficial for the patient and/or family.76,78,82,83 For example, recommending that a family change its food choices when the parent is not aware or convinced of the child's weight being problematic (precontemplative) may not be as successful as first identifying the issue and discussing the rationale for concern.
Rollnick et al84 incorporated the principles described above into an approach called motivational interviewing. They defined motivational interviewing as a "client-centered counseling style for eliciting behavior change by helping clients explore and resolve ambivalence." For a brief clinical assessment, they suggest asking 2 questions to gauge a patient's motivation to change an unhealthy behavior, that is, (1) how important (on scale of 1–10) the change in behavior is to the patient and (2) how confident the patient feels in his or her ability to make the change.84 These 2 concepts help direct the focus of the interaction between the clinician and the patient. If a patient does not identify a condition (eg, a child's high BMI) as important, then the discussion may target health-related risks. If the patient or family member recognizes the problem and its importance but is not confident in making a change, then the discussion may usefully target strategies for change, as well as barriers that may interfere with the change. This approach allows health care professionals to collaborate with patients to promote change by using a brief, patient-centered assessment that can be adapted easily to the clinic setting.
A related approach put forth as a general clinical prevention tool is the 5As, that is, ask/assess, advise, agree, assist, and arrange follow-up care.85 The exact wording of the 5As varies slightly among different publications, but the intent and process remain the same. These steps reinforce the concept that health care professionals need to assess behavior patterns and health belief structures to agree on a plan of action or intervention that is most appropriate for each patient.
Preliminary data on successful behavior changes using these approaches in health care settings show mixed results, and these approaches have been applied most often in the adult population86–91 for tobacco, alcohol, and drug use/addiction. Several studies that applied these methods to nutrition and physical activity assessment showed successful short-term results but less-convincing long-term results.92 Ongoing projects are examining the feasibility of these behavior change strategies in primary care settings, including pediatric practices.93
Dietary Assessment
Assessment Methods
Many complex dietary factors are associated with obesity, and age, gender, and genetic predisposition are likely to influence their effects. Although individual nutrients have been linked to obesity,94 few attempts have been made to identify eating patterns that may lead to obesity. Scientists have reached a consensus that obesity results from an imbalance in the energy balance equation; energy intake exceeds energy expenditure. Therefore, assessment should address both sides of the equation (diet and physical activity) in efforts to prevent or to treat obesity. Assessment of energy intake is challenging even under the most-controlled research conditions, and typically assessment includes a combination of assessment methods. Traditional dietary assessment methods include 24-hour recalls, food records, and food frequency questionnaires.95 In a 24-hour recall, the interviewer asks an individual what he or she ate and drank in the past 24 hours. Ideally, this is repeated several times, to obtain a view of the individual's usual dietary intake. To complete food records, patients write down, for several days, the foods, amounts, recipes, and preparation methods for everything they consume. A food frequency questionnaire asks patients how often they consume specific foods and beverages and the sizes of their usual portions.
All of the methods described above have advantages and disadvantages in research settings, but they are impractical for use in most clinical settings. These interventions are time-consuming, expensive, and difficult for health care professionals to administer in the office. Furthermore, the value of estimating energy intake per se is limited because it is virtually impossible to assess energy expenditure accurately and precisely and therefore to determine energy balance. A few rapid assessment methods are available for practitioners to evaluate their patients’ eating behaviors and physical activity, as well as to deliver effective nutrition counseling (Table 1). The weight, activity, variety (in diet), and excess (WAVE) tool allows a quick assessment of the patient's weight status, activity and inactivity patterns, variety of foods, and potential excessive consumption of selected foods. The evidence base for the WAVE tool and other potential assessment tools is presented in Table 2.
|
|
An important consideration in the interpretation of the results of this research is that the percentage variance in the eating pattern/overweight models was extremely small,105,106 which suggests that weight status likely stems from a combination of interrelated eating patterns, rather than a single eating pattern. In addition, the effects of these interrelated patterns on weight status may be cumulative, and they may vary according to gender, ethnicity, and genetic factors. Limitations of the literature include the predominance of cross-sectional studies, rather than prospective longitudinal studies, small sample sizes, and limited study populations (particularly a dearth of studies with children). Therefore, results often are inconsistent, and the findings of many studies have not been replicated. Despite these limitations, studies that identify eating patterns that may contribute to excessive energy intake, and that propose targets for behavior changes are useful for clinicians who are helping their patients prevent excessive weight gain.
Restaurant Food Consumption
Consumption of foods away from home increased considerably in children107,108 and adults109 between 1977 and 1996. The proportion of foods that children consumed from restaurants and fast food outlets increased by nearly 300% during that 19-year period.108 Fast food consumption was reported by 42% of children and 37% of adults,110 although investigators noted that there is no uniform definition of fast food and definitions varied among studies. The percentage of energy obtained from food prepared away from home also increased during that period, from 18% to 32%.111 Portion sizes in restaurants increased from 1970 to 1999,112,113 with the result that soft drinks contained an additional 206 kJ, hamburgers 407 kJ, and French fries 286 kJ.
Portion sizes influence energy intake. Diliberti et al114 found that customers who purchased a larger portion of the entree served at a fast food outlet increased their intake of the entree by 43% and that of the entire meal by 25%, resulting in greater energy intake. Some studies showed that children110,115 and adolescents116–118 who consumed fast food more frequently had higher energy intakes and poorer diet quality, compared with those who did not. Interestingly, overweight adolescents were less likely than their leaner counterparts to compensate for the increased energy in the food by adjusting energy intake throughout the day.119
Studies have reported that the frequency of eating fast food is associated with BMI and body fatness in children120 and adults.121–126 In a longitudinal study of 101 girls 8 to 12 years of age, the frequency of eating quick-service food at baseline was associated positively with changes in BMI z scores at 11- and 19-year follow-up evaluations.120 In young adult women, increases in frequency of fast food restaurant use were associated with increases in body weight over 3 years in a randomized, prospective, intervention trial on weight gain.122 In the Coronary Artery Risk Development in Young Adults study of 3031 young adults, Pereira et al126 reported that changes in fast food frequency were associated with changes in body weight but the changes varied according to racial group. Data from another study suggested that older children who consumed fried foods away from home more frequently over a 1-year period were heavier and had greater total energy intake, compared with children with low frequency of fried food consumption away from home.127 In contrast, French et al116 reported that the frequency of fast food consumption by adolescents was not associated with overweight status. Although not entirely consistent, the data suggest that fast food consumption may be related to BMI. For individuals and families that eat regularly at restaurants or fast food establishments, reducing the frequency of these meals may be a strategy to decrease total energy intake.
Sweetened Beverage Consumption
Experts have raised concerns about high intakes of sweetened beverages and their possible association with the increasing prevalence of overweight and obesity among children.128,129 Over the past 4 decades, national data on individuals 2 years of age showed an increase in sweetened beverage consumption for all age groups.112,130 Soft drink consumption accounts for one third of added sugar intake in the US diet.131 In one study of fourth-grade and fifth-grade children, sweetened beverages constituted 51% of the average daily intake of beverages consumed.132 This large intake of sweetened beverages could contribute to increased energy intake, tilting the energy balance toward excessive weight gain.
Most cross-sectional studies have shown a positive relationship between greater intake of added sugars and total energy intake.128,129,133–137 Energy intake has been reported to be related positively to consumption of sweetened beverages by children and adolescents.133 In another report, children who drank the most sweetened beverages consumed 1390 kJ more per day than did those who did not drink sweetened beverages.132 The Bogalusa Heart Study examined energy intake among 10-year-old children from 1973 to 1994. Findings from the study showed that children who did not consume sweetened beverages did not have increased energy intake. However, energy intake did increase among children who consumed small to moderate to large amounts of sweetened beverages.137 Of interest, mean BMI increased in all categories of sweetened beverage consumption, including children who did not consume sweetened beverages.
Although several studies showed an association between sweetened beverage consumption and risk of obesity,128,129,134,139–142 other studies found no association,105,143–148 and a few indicated a negative relationship.149–151 In a pilot intervention study, Ebbeling et al140 showed that reducing sweetened beverage consumption reduced body weight in adolescents in the upper baseline BMI tertile. In their comprehensive review of studies that examined the relationship between sweetened beverages and adiposity, Bachman et al151 concluded that the association between sweetened beverages and overweight is unclear and that the evidence is inconsistent. In another review, however, the authors came to a completely different conclusion.153
The strongest current evidence supports a positive association between sweetened beverage consumption and energy intake. These conclusions were similar to those made by the 2005 Dietary Guidelines Advisory Committee (DGAC).154 Decreasing sweetened beverage consumption may be one strategy to decrease total energy intake. More intervention studies are needed, particularly in children, for better understanding of the relationship between sweetened beverage consumption and weight gain. However, it may be concluded intuitively that, if an individual consumes excessive sweetened beverages, then the resulting increase in energy intake may lead to weight gain.
Fruit Juice Consumption
Recently, 100% fruit juice has received much attention as a potential culprit in the prevalence of obesity among young children. In 2001, the Committee on Nutrition of the American Academy of Pediatrics concluded that 100% fruit juice had no beneficial effect over whole fruit for infants >6 months of age and children.155 For a number of reasons, the recommendations included limiting 100% fruit juice to 4 to 6 oz/day for children 1 to 6 years of age and 8 to 12 oz for older children. The 2005 DGAC recommended that no more than one third of the total fruit group intake recommended come from fruit juice.154
Limited data are available to assess the relationship between 100% fruit juice consumption and body weight in children. Two separate studies by Dennison et al156,157 showed that consumption of 100% fruit juice (>12 fl oz/day)156 and apple juice only157 was associated positively with BMI in samples of children 2 to 5 years of age. Tanasescu et al158 found that fruit juice and possibly fruit drinks were associated with overweight in 29 obese Puerto Rican children 7 to 10 years of age. In contrast, 6 longitudinal and cross-sectional studies reported either a negative or neutral association between 100% fruit juice intake and weight status in children.105,142,148,159–163 Overall, the current evidence shows only a weak association between 100% fruit juice consumption and excessive weight gain.
Portion Sizes
A number of short-term feeding studies, 1 longitudinal study, and 3 observational studies showed that portion sizes influence energy intake. Adults served large portion sizes consumed more food and more total energy130,164–167 than did individuals who were served smaller portion sizes; there was no evidence of meal-to-meal compensation for higher intakes.
Several well-controlled, laboratory-based studies showed that providing older children and adults with larger food portions could lead to significant increases in food and energy intakes, independent of the energy density of the food.164,165,167 This effect was demonstrated for snacks,168 delicatessen-style sandwiches,169 and entrees.130,164,166,170 The responses to the variations in portion sizes were not influenced by gender or BMI.130,166,169 The energy density of food can have an effect on energy intake when portion sizes are varied.164,167 Therefore, increases in portion sizes and energy density may lead to independent and additive increases in energy intake.165
The responses of young children to portion sizes seemed to be similar to those of adults; presentation of larger portion sizes resulted in increased energy intake.170,171 One study found that larger portion sizes resulted in greater energy intakes for children 5 years of age but not for children 3 years of age.171 Another study by the same group found that, when children 3 to 5 years of age were presented with a large portion size of an entree, they consumed 25% more of that entree and their energy intake increased 15% for the whole meal, compared with children who were presented with an age-appropriate portion size.170 That study also reported that the children consumed 25% less of the entree when they were allowed to serve themselves than when the entree was served to them on individual plates.
Two cross-sectional studies of preschool-aged children, using national data, examined the relationships between portion sizes, energy intake, and body weight. Portion size alone accounted for 17% to 19% of the variance in energy intake, whereas body weight predicted only 4%.172 Body weight was related positively to energy intake and portion size but not the number of different foods or the number of eating occasions.173
More studies with infants and children are needed to understand how larger portion sizes at a single feeding or meal affect total energy intake over a 24-hour period. There are no longitudinal studies with children showing an association between increased portion size and BMI. Data suggest, however, that reducing food portion sizes may be an effective strategy for decreasing energy intake, especially for energy-dense foods. For clinicians, however, determining the appropriateness of portion sizes presented to and consumed by a child is difficult, as is making specific recommendations for age-appropriate portion sizes.
Energy-Dense Foods
Energy density refers to the amount of energy in a given weight of food and depends on the content of fat, carbohydrate, protein, and water. Water has the greatest impact on energy density, because it adds weight without energy. The high-energy content of fat also influences the energy density of food. Fiber can decrease the energy density of foods.
In several feeding studies, ad libitum consumption of foods that were high in energy density resulted in significantly greater total energy intake, compared with foods that were low in energy density.174–177 Delayed satiety may be the reason why some individuals consumed large amounts of energy-dense foods.176
A number of other laboratory studies indicated that energy density was associated with a reduction in energy intake.165,178–180 For example, eating low-density foods such as salad or soup as the first course of a meal reduced total energy intake, compared with eating a meal that consisted entirely of foods high in energy density.165,181 Rolls et al176 showed that adding air to test meals that had similar macronutrient compositions and energy contents reduced energy intake significantly, which suggests that the mass and volume of a meal are important. For foods that are low in energy density, satisfying portions should be encouraged, because they provide little energy and produce satiety.
Low-fat diets have been associated with lower energy intake,15 possibly because of a reduction in energy density. Laboratory studies showed that fat content, independent of energy density, had little influence on energy intake.165,171,174 Because lower-fat diets generally have lower energy density, reducing the intake of total fat may be one strategy for reducing energy intake.
What is the relationship between energy-dense foods and weight? Two clinical trials tested the influence of variations in energy density on body weight. In one study, adults who incorporated 2 servings of soup (which is low in energy density) into a calorie-restricted diet lost significantly more weight than did those who incorporated a similar number of calories as energy-dense snacks.182 In another study, investigators examined how 2 strategies to reduce energy density in the diet affected body weight during a 1-year period.183 One group was counseled to reduce fat intake and to limit portions. The other group was counseled to increase intake of water-rich foods and to choose reduced-fat foods. Both groups succeeded in lowering the energy density of their diets, and they lost significant amounts of weight and kept the weight off over the year. These studies provide promising results, but more long-term intervention studies are needed to understand whether diets with reduced energy density prevent weight gain, particularly in children. One cross-sectional study with children 10 years of age found that consumption of energy-dense foods was a predictor of being overweight.105 However, those results were not confirmed by others.184,185
On the basis of the current studies, insufficient evidence exists to determine the contribution of energy-dense foods to weight gain; no studies of children are available. However, consuming energy-dense foods may contribute to excessive energy intake. The 2005 DGAC came to a similar conclusion in its report.154 Encouraging consumption of foods low in energy density, including those with high fiber and/or water contents and those with modest fat content, may be a useful strategy for individuals who are trying to lose weight or to maintain their current weight.186 Unfortunately, there is no standard calculation method for determination of energy density in foods.187 No published studies, particularly involving children, have examined the impact of consuming high-energy density foods on diet quality and intake of fat-soluble vitamins, essential fatty acids, and amino acids. One adult study showed that low energy density of diets was associated with high diet quality.188 Unfortunately, there were some concerns about the study189 and the definition of the energy density categories. More studies are needed in this area of research, particularly involving children.
Fruit and Vegetable Consumption
The specific relationships between fruit and vegetable consumption, energy balance, and obesity prevention represent an emerging area of research. Fruits and vegetables are high in fiber and water content, and they may play a role in promoting satiety and decreasing total energy intake by displacing energy-dense foods. Despite long-standing recommendations to eat several servings of fruits and vegetables each day, intake among US children remains low.190
Findings from observational studies are equivocal,191 with some studies showing an inverse association125,192–205 and others showing no relationship158,184,206–216 between fruit and vegetable consumption and a measure of body adiposity. The studies showing an inverse association, however, have not been consistent with respect to gender, ethnicity, age group, and type of fruit or vegetable. Two studies reported that fruit consumption was associated inversely with weight status in children,105,198 but a relationship with vegetable intake was not apparent. Several observational studies did not control for potential confounders (physical activity and, in some cases, dietary energy intake). The percentage of the variance in children's BMI explained by fruit and fruit juice consumption was <3%.105
The lack of association between vegetable consumption and weight status may be specific to the type of vegetable consumed. Some vegetables typically are consumed with fat added during preparation, such as fried potatoes. This may explain why the study by Lin and Morrison198 found a positive association between intake of potatoes and weight status among adults. Clearly, more studies are needed to better understand these inconsistencies in the findings across studies.
Numerous interventions have been designed to promote increased consumption of fruits and vegetables, but very few studied weight status or change in BMI as an outcome variable. Some interventions included multiple components, making the identification of an independent effect of fruit and vegetable consumption in the prevention of overweight or weight gain difficult.
A number of adult trials examined the effects of increased fruit and vegetable consumption on weight; those studies were reviewed by Rolls et al.217 The 2005 DGAC concluded that data from those studies showed that, without advice to lose weight, increased fruit and vegetable consumption by itself did not lead to weight loss.154
Intervention studies in children that examined fruit and vegetable consumption targeted mainly changes in intake and not effects on body weight. Encouragement to eat more fruits and vegetables often has been one of several messages aimed at modifying energy balance.218,219 However, efforts to increase knowledge and to improve attitudes toward fruit and vegetable consumption have had modest effects on actual consumption.220,221
In one randomized trial examining weight loss in children, Epstein et al221 reported that a message that targeted specifically increasing fruit and vegetable consumption resulted in greater weight loss than did an intervention message that focused on reducing high-fat and high-sugar food intakes. More studies with children are needed to understand the independent effect of increased fruit and vegetable consumption in randomized, controlled trials on prevention of weight gain. Encouraging greater fruit and vegetable consumption is a sound message in general, and limited evidence suggests that it may be a useful strategy in efforts to achieve and to sustain weight loss.
Breakfast Consumption
Several studies showed that skipping breakfast decreased the nutritional quality of the diets of children107,222–224 and adults.225–227 The average total energy intake was significantly lower for children who did not consume breakfast, and they did not make up the differences in energy intake at other meals.222 The energy content of school breakfasts has increased in the past 15 years.228
A few cross-sectional and longitudinal studies and one randomized, clinical trial have examined the association between breakfast consumption and BMI. A number of cross-sectional studies have shown a positive association between overweight and skipping breakfast among children216,229–232 and adults.233 However, other studies, particularly one with children,107 found no association.
Two longitudinal studies, one each with children and adults, have been reported. The first was conducted with >14000 children 9 to 14 years of age.234 After a 1-year follow-up period, overweight children who never ate breakfast had a greater decline in BMI than did overweight children who ate breakfast. Normal-weight children who never ate breakfast, however, had weight gains comparable to those of normal-weight children who ate breakfast. The adult study found that skipping breakfast was associated with an increased prevalence of obesity.233 In a randomized, clinical trial,235 adults who ate no breakfast at baseline and who were assigned randomly to eat 3 meals per day lost slightly more weight by 12 weeks, compared with those who were assigned randomly to consume no breakfast and to eat 2 meals per day. However, of breakfast eaters at baseline, those who were assigned randomly to eat only 2 meals per day lost more weight than did those who continued to eat breakfast. The authors suggested that the effects might have been influenced by subjects having to make the most-substantial changes to their usual routine.235 Clearly, more studies are needed, because current evidence related to the effect of breakfast consumption and the content of the meal is inconclusive. Children should not be encouraged to skip breakfast. More importantly, skipping breakfast may result in poorer nutritional quality of the diet and may have adverse effects on performance in school.236–238
Meal Frequency and Snacking
Previous studies demonstrated an inverse association between meal frequency and the prevalence of obesity in children239,240 and adults.241,242 Four studies examined this association. Three found no association between the number of eating episodes and overweight in children 10 years of age.105,243 However, a cross-sectional study with 4370 German children 5 to 6 years of age found that the prevalence of obesity decreased according to the reported number of meals consumed each day.240 The prevalence of obesity was 4.2% among children who consumed 3 meals per day, compared with 1.7% among those who consumed 5 meals. Although some studies suggested that a "nibbling" or "grazing" meal pattern may be associated with leanness, those studies were vulnerable to methodologic errors that might have generated spurious relationships because of dietary underreporting and posthoc alterations in eating patterns in response to weight gain. Moreover, the association between increased eating frequency and lower body weight status might have been influenced by increased physical activity and a reduction in the mean energy consumed per eating episode. More longitudinal studies are needed to better understand the association, if any, between meal frequency and overweight in children.
On the basis of national data, the prevalence of snacking has increased for individuals 2 to 18 years of age,244,245 although the average size of snacks and energy per snack have remained relatively constant.244 There has been a shift from meals to snacks in the past 20 years.228 In contrast, one study showed that snacking decreased among children 10 years of age from 1973 to 1994 in Bogalusa, Louisiana,243 although the prevalence of obesity increased. These conflicting findings may reflect differences in age groups studied, regions of the country, methodologic changes over time, or the definition of what constitutes a snack or a snacking occasion. In adults, increased snacking resulted in increased energy intake but was not associated with BMI.246 Other studies showed that obese adults were more frequent snackers247,248 and total energy intake was higher for snackers248 than for reference adults. Two cross-sectional studies of children 10 years of age showed no association between snacking and overweight status.105,243 More longitudinal studies are needed to better understand the associations between snacking, total energy intake, and overweight in both children and adults. The data on meal frequency and snacking are inconclusive and therefore do not represent a priority area of inquiry for all patients.
Summary
A number of studies have been conducted with adults, but far fewer with children, that address the associations between specific eating patterns and weight status. Results are inconsistent, largely because of methodologic limitations and small sample sizes. More well-designed, longitudinal studies and randomized, controlled trials are needed before any definitive statements can be made regarding which eating patterns are associated most strongly with overweight and how age, gender, ethnicity, and geographic location affect these associations. Evidence supports an association between at least some of the eating patterns discussed in this report and increased energy intake for some individuals, and these patterns represent behaviors that can be targeted for change.
Overall Recommendations for Dietary Assessment
The assessment of dietary patterns among children and adolescents should address the following: (1) assessment of self-efficacy and readiness to change, (2) qualitative assessment of dietary patterns, and (3) working in conjunction with patients and families to identify dietary practices that are targets for change. The assessment writing group recommends the following. (1) Qualitative assessment of dietary patterns should be performed for all pediatric patients at each clinic visit, at a minimum, for anticipatory guidance. (2) Assessment should address dietary practices for which evidence supports a positive association with energy intake and behaviors for some individuals and that represent behaviors that can be targeted for change. By decreasing energy intake without increasing energy intake throughout the day or from other foods, changes in these behaviors may prevent excessive weight gain. These behaviors include the frequency of eating outside the home at restaurants or fast food establishments, excessive consumption of sweetened beverages, and consumption of excessive portion sizes for age.
The assessment writing group also suggests assessment of additional dietary practices that have a weaker evidence base for association with energy intake but may be important for some individuals and that represent behaviors that can be targeted for change. The writing group suggests consideration of (1) excessive consumption of 100% fruit juice, (2) breakfast consumption (frequency and quality), (3) excessive consumption of foods that are high in energy density, (4) low consumption of fruits and vegetables, and (5) meal frequency and snacking patterns (including quality). The child version of the WAVE assessment tool (Table 1), which provides a means for quick assessment of both diet and activity, may be useful to clinicians in primary care settings.
Physical Activity Assessment
Levels of Physical Activity
Physical activity is an important component of health and well-being for people of all ages. Children who are physically active may gain immediate and long-term positive effects, such as improved mental health status and self-esteem; increased physical fitness, which enhances performance of daily activities; promotion of bone formation; weight maintenance; and prevention of cardiovascular risk factors.76,249 In addition, physical activity patterns established during childhood may continue into adulthood, establishing healthier choices over the entire lifespan.250,251 Health benefits for physically active adults include lower risks of coronary artery disease, T2DM, hypertension, hyperlipidemia, osteoporosis, certain cancers, and depressive symptoms.76,252,253
Despite these benefits, results from the 2003 Youth Risk Behavior Surveillance Study and the 2002 Youth Media Campaign Longitudinal Survey showed that many children and adolescents do not meet recommended physical activity levels.254,255 Nationwide, 62.6% of students in the ninth through 12th grades met the recommendations for vigorous physical activity (20 minutes on 3 of the past 7 days), and 24.7% of students nationwide met recommendations for moderate physical activity (30 minutes on 5 of the past 7 days). Overall, 33% of this group of students reported some but insufficient levels of physical activity, and 11.5% reported no moderate or vigorous physical activity.255 In addition, 38.2% reported watching >3 hours of television per day, on average. Twenty-three percent of younger children (9–12 years of age) had not engaged in any free-time physical activity outside of school in the past 7 days, and 61.5% had not participated in organized physical activity during nonschool hours.254 Higher levels of physical activity were reported by boys than by girls and by non-Hispanic white youths than by other racial and ethnic groups. Levels of physical activity also decline as children get older. It is estimated that physical activity levels decrease by 1.8% to 2.7% per year for boys 10 to 17 years of age and by 2.6% to 7.4% per year for girls 10 to 17 years of age.77
Diet and physical activity are inextricably linked. Overweight and obesity result when daily energy intake is greater than daily energy expenditure over time. This concept of energy balance is crucial for successful assessment, prevention, and management of overweight and obesity in childhood and adolescence. Energy intake is a relatively easy concept, because it includes all foods and beverages consumed during the day. Energy expenditure is more complex, because it is a combination of resting metabolic rate, the thermic effects of food, and the variety of activities the individual performs during the day.253,256 Therefore, measurement of physical activity is not equivalent to measurement of total energy expenditure; rather, physical activity is one (albeit the most variable and modifiable) element of total energy expenditure. For children and adolescents, a certain amount of positive energy balance is necessary for proper growth and development. The overall energy balance should tip in favor of slightly greater energy intake, relative to expenditure, although the percentage of total energy required for growth is small after infancy.257
Clarification of several terms is necessary to understand what is being measured when physical activity is being discussed. Physical activity is defined as any bodily movement produced by the contraction of skeletal muscles that increases energy expenditure above the basal level.258 Physical activity thus encompasses movement resulting from free play, structured activities such as sports, and general activities of daily living. Exercise is planned, structured, and repetitive bodily movement performed specifically to improve or to maintain physical fitness.258 Children and adolescents often participate in planned activities during physical education classes or in structured sports activities; however, the goal is not necessarily physical fitness. Physical fitness is a set of attributes that people have or achieve, such as cardiorespiratory fitness, muscular strength, flexibility, endurance, and body composition.258,259 This report focuses on the assessment of physical activity for the purpose of preventing or managing overweight and obesity in childhood and adolescence. Total energy expenditure, exercise, and fitness are beyond the scope of this report.
Assessment Methods
Appropriate assessment of physical activity patterns requires valid (accurate) and reliable (reproducible) instruments. Researchers have developed several approaches for measuring physical activity in children and adolescents, and most are reasonably reliable, with low to moderate validity.253 Briefly, these include questionnaires (self-report or interviewer-administered), direct observation, and electronic or mechanical monitoring (with a pedometer, accelerometer, or heart rate monitor). Methods such as double-labeled water testing and calorimetry assess total energy expenditure and resting metabolic rate, respectively, and can be used to estimate physical activity. Each method has strengths and weaknesses, which are described elsewhere.252,253,256,260,261 This report discusses the methods most adaptable to the clinic setting, that is, brief questionnaires and accelerometers or pedometers.
Questionnaires
The most common method for measuring physical activity is a self-report survey or checklist of the frequency, intensity, and duration of specific activities within a defined period (eg, past 24 hours, 1 week, or 1 month). Recurring problems with any self-report survey include recall bias and the documented tendency to overestimate activity levels, compared with observation, movement monitoring, or estimations from total energy expenditure. Depending on the goals of the questionnaire, this limitation may be tolerable. It is also a challenge to determine the lower age limit at which children can recall accurately what they did, as well as the intensity and duration of their physical activity. In general, children <10 years of age are considered too young to give reliable answers to physical activity questions.252,262 Parents should be used for proxy responses; however, they do not always capture accurately the physical activity levels for their children, either at home or in other settings.263
An additional challenge is the sporadic unstructured nature of physical activity among children, especially those <10 years of age. Unlike adolescents and adults, who can sustain 10 to 60 minutes or more of physical activity, young children typically have multiple frequent bursts of activity followed by periods of rest. Questions that aim to assess 30 minutes of moderate-intensity activity or 20 minutes of vigorous-intensity activity are not realistic for children; alternate assessment questions would be more appropriate. For example, proxy measures such as time spent outside or involvement in community sports programs have been shown to be predictive of physical activity in children.261,264
A review of the literature reveals that very few questionnaires have been developed and validated for pediatric age groups. Most focus on adolescents, are quite lengthy, and have not been assessed for use in the clinic setting. Examples include written, verbal, and computer-based questionnaires. More-detailed information about these research questionnaires can be found elsewhere.252,253,265 A few questionnaires with the potential for clinic use have been designed and are discussed in Table 3.
|
Pedometers are easier to use and measure physical activity as steps walked, distance walked, or energy expended. Several studies have shown the reliability (correlation range: 0.51–0.92) and validity (correlation range: 0.49–0.93) of pedometer use for children and adolescents.260,268–272 Two studies found that, on average, children 8 to 10 years of age272 take between 12000 and 16000 steps per day.271,273 Jago et al269 determined that taking 4000 steps in 30 minutes and taking 8000 steps in 60 minutes (fast walking) meet current US physical activity recommendations.
Pedometers could be used at home to assess baseline physical activity levels for children and adolescents, and specific activities could be recorded in conjunction with times the monitor is worn. For example, a clinic visit with BMI screening may prompt a physical activity assessment and counseling on the basis of overweight or obese status. The patient can be instructed to wear a pedometer daily for 1 week, to record specific physical activities in a diary, and to determine a baseline average number of daily steps (with the assistance of a parent, if necessary). Results can be used as a proxy for overall physical activity and compared with documentation in the activity diary. Discussion at a follow-up visit with a designated clinic staff member can determine necessary modifications and can address any barriers to increasing physical activity levels. No studies have assessed the feasibility, reliability, and validity of using pedometers for baseline assessment in this way. For a list of available pedometers, see the report by Bjornson268 (and www.pedometers.com).
Behavior Changes
Approaches such as readiness to change, motivational interviewing,
