SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Disclosure
  8. REFERENCES

Objective: To determine the frequency and characteristics of energy intake underreporting in African-American preadolescent girls as part of the Girls health Enrichment Multi-site Studies (GEMS).

Methods and Procedures: Energy intake was summarized using the Nutrition Data System for Research software and computed as a 3-day average of 24-h dietary recalls. Physical activity was assessed by an accelerometer, basal metabolic rate (BMR) was estimated using the World Health Organization's prediction equation, and underreporting of caloric intake was based on the Goldberg equation.

Results: Using a conservative criterion for determining energy underreporting, we classified 54.8% of the girls as underreporters; 45.2% were classified as plausible reporters. Factors related to underreporting included higher BMI (β = −0.506, P ≤ 0.001), older age (β = −0.159, P = 0.001), greater unhealthy eating behaviors (β = −0.118, P = 0.025), and higher self-efficacy for diet (β = −0.098, P = 0.033).

Discussion: Underreporting of dietary intake, specifically energy, is common in African-American preadolescent girls and can be partially explained by weight status and psychosocial variables. The extent of dietary underreporting in specific and high-risk populations is largely unknown and could be evaluated by routinely including a report of such an index in future research studies.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Disclosure
  8. REFERENCES

An unhealthy diet has been associated with three of the leading causes of adult deaths in the United States (i.e., coronary artery disease, cancer, and stroke) and has been linked to a number of preventable chronic diseases and disorders (1). Among these are cardiovascular disease (2,3), obesity (2), and type 2 diabetes (4). Type 2 diabetes is becoming increasingly more common in children. In addition, certain types of cancers and osteoporosis (3,5) have been linked to unhealthy diets. Accurate assessment of dietary intake is important in estimating unbiased associations with health outcomes. Evaluating prevention approaches to obesity and other chronic diseases also relies on accurate assessment of dietary intake.

Three main methods of assessing dietary intake in clinical trials and epidemiologic studies include diet records, food frequency questionnaires, and dietary recalls. Although these three approaches for collecting dietary intake data have their strengths and weaknesses, in terms of accuracy and feasibility, several studies have reported that multiple dietary recalls on nonconsecutive days is the preferred and most reproducible method of dietary assessment in children (6,7). Johnson (8) recommends a minimum of three nonconsecutive days of dietary recall consisting of at least one weekend day for adults.

Assessing dietary underreporting in populations of young adults and children has revealed several general predictors. Children who underreport energy intake tend to be obese (9). Across age, sex, and race, increased body weight is the single-most significant predictor of energy underreporting (9). Results related to influences of socioeconomic status on underreporting are equivocal with some studies finding an association with lower socioeconomic status (9) and others an association with higher socioeconomic status (9,10).

There is a limited literature related to psychosocial predictors of energy underreporting in the child and adolescent population and validation methods vary among studies. Ventura et al. (11) found that underreporters had greater weight concerns and dietary restraint when using calculated predicted energy requirements. In a biracial cohort of female adolescents and young adults using doubly labeled water, Kimm et al. (12) examined psychosocial predictors of underreporting and found only one predictor, the drive for thinness, to be predictive in their sample. Furthermore, African Americans were more likely to underreport their dietary intake than their white counterparts. Specifically, among African-American women, the drive for thinness and BMI were the only factors significantly related to underreporting (12).

African-American preadolescent girls represent an important group in which dietary intake measures have not been adequately studied specifically related to energy intake. Little is known about the frequency of dietary underreporting or the demographic, physical, and psychosocial factors related to dietary underreporting in this population. Therefore, the demographic, physical, and psychosocial correlates of dietary intake underreporting were evaluated in African-American preadolescent girls.

Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Disclosure
  8. REFERENCES

Participants

Data for the current analyses were drawn from the baseline data collected during the phase 2 portion of the Girls health Enrichment Multi-site Studies (GEMS). GEMS is a National Heart, Lung, and Blood Institute-sponsored cooperative agreement testing 2-year obesity prevention interventions among African-American girls. Four clinical centers (University of Minnesota, Baylor College of Medicine, University of Memphis, and Stanford University) were involved in the phase 1 feasibility study, and two clinical centers (University of Memphis and Stanford University) proceeded to phase 2. All the participants in this study were from the Memphis GEMS site. Of the 303, 8–10-year-old girls enrolled in the phase 2 study, 284 had complete data for inclusion in the analytic sample.

Procedures

The recruitment materials, questionnaire assessment procedures, and informed consent documents were approved by the Institutional Review Board at the University of Memphis before the study was initiated. Girls were enrolled over 18 months in a series of recruitment phases. Each girl and one parent/caregiver underwent baseline physical and psychosocial assessments.

Measures

The major variables of interest included the girls' total reported energy intake (repeated dietary recalls), objectively assessed physical activity, BMI, family demographics, and psychosocial assessments completed by the girl and her parent/caregiver. Estimated basal energy expenditure was estimated based on height, weight, and age using the World Health Organization's equation (13,14,15). Detailed descriptions of data collection and assessment methods are available elsewhere (16).

Dietary intake. Dietary intake data were collected and analyzed using Nutrition Data System for Research software (2005) developed by the Nutrition Coordinating Center, University of Minnesota, Minneapolis, MN. Three dietary recalls were scheduled for each participant. Quality-control procedures and methods used in this current study were validated during the phase 1 trial of GEMS (17). Not all participants were available for three interviews, so intake was averaged over the available days. Of the 284 included in the sample, 267 (94%) provided 3 days of data and 17 (6%) provided 2 days of recall information.

Physical activity. The MTI actigraph (Model 7164 WAM; Manufacturing Technologies, Fort Walton Beach, FL) was used to objectively assess physical activity levels. The total number of counts and total number of minutes worn were recorded in 60-s increments and summed across the 3 days. The counts recorded between 7:00 am and 10:59 pm were used for analysis to establish consistency across the variable. Intensity was defined using 2 × the threshold counts, as previously established (18). In this study, we used moderate-to-vigorous physical activity.

Demographic information. Parent education was defined as the highest level of education obtained by the parent/caregiver participating in the study. The categories of total household income ranged from <$20,000 to ≥$80,000. The total number of adults and children living in the household were also assessed.

BMI. Weight was assessed using a calibrated Scaletronix 5602 model scale (Scale-Tronix, White Plains, NY) and height was obtained using the Schorr Height measuring board (Olney, MD). The mean weight and height were computed. BMI was calculated by dividing the mean weight in kilograms by the mean height in meters squared.

Psychosocial variables

Few studies have reported psychosocial variables related to dietary misreporting in children, and none specific to African-American girls. Therefore, the variables included in this analysis have been reported in adult studies namely social desirability, body dissatisfaction/body image, measures of restrained eating, and eating disorders. Other variables were collected to determine potential associations with obesity, namely, self-efficacy for healthy eating.

Social desirability. Social desirability was measured using the 9-item “Lie Scale” from the Revised Children's Manifest Anxiety Scale (19). In the GEMS phase 1 study, the internal consistency of the scale was excellent with a Cronbach's α of 0.78 and test-retest reliability of 0.62 (20). In the GEMS phase 2 study, the internal consistency of the scale was 0.72. Higher scores reflect greater social desirability.

McKnight unhealthy eating behaviors subscale. The elementary school version of the McKnight Risk Factor Survey was used to assess the risk factors for eating disorders. In the GEMS phase 1 study, Sherwood et al. (21) found that this subscale had an α coefficient of 0.76; the internal consistency for the measure in the GEMS phase 2 study was 0.75. Higher scores reflect greater levels of disordered eating.

Body image. The body silhouettes used in this study were adapted from Stunkard et al. (22) and modified to resemble African-American girls. Eight silhouettes ranging from underweight to overweight were presented. These silhouettes were used in the GEMS phase 1 study and showed test-retest reliabilities of 0.75 for “looks like you,” 0.36 for “like to look,” and 0.55 for the discrepancy score between these items (21).

Self-efficacy for healthy diet. A modified version of the original High 5 scale was used in the study. Modifications were made to include questions about drinking water in general and drinking water instead of sweetened beverages (23). The original instrument generated a Cronbach's α of 0.86 (23). The internal consistency for the modified measure in the GEMS phase 2 study was 0.82.

Dietary intake underreporting

We followed a strategy for estimating underreporting that was similar to the approach used in a previous adult study by Klesges et al. (24). To summarize, basal metabolic rate (BMR) was estimated by prediction equations based on height and weight. The minimum energy intake required for survival was then estimated using methods described by Goldberg et al. (25).

In this analyses, calculations of energy intake requirements were based on Goldberg et al. (25) estimates of minimum intake to maintain energy balance. These investigators determined that the average ratio of total energy expenditure to BMR was 1.35 (for studies of whole-body calorimetry) and 1.67 (for studies of doubly labeled water). Values from 1.35 to 1.67 represent the range of energy needed to maintain body weight beyond metabolic costs typically supported by consumption between 35 and 67% more kilocalories. We chose to estimate a conservative minimum cutoff of 0.92, the lower 95% confidence interval of Goldberg's 1.35 estimate; that is, a reported intake of only ∼90% of estimated BMR. Bear in mind that this is quite conservative because this includes some individuals reporting at an intake ratio lower than what is defined as starvation diets by the World Health Organization (13).

The calculated BMR was subtracted from the girl's actual reported intake (kcal) to assess the extent of discrepancy between reported energy intake (EI) and calculated minimum energy requirement for survival. This discrepancy was divided by the estimated minimum energy requirement (ER) (based on Goldberg's method) ((EIBMR)/ER) (25). The resulting estimate of reporting discrepancy was a continuous variable. Negative scores indicated underreporters, and positive scores were considered plausible reporters.

Statistical analysis

All analyses were conducted using the Statistical Package for the Social Sciences (SPSS Version 13.0 for Windows; SPSS, Chicago, IL). Univariate analyses were conducted on all potential demographic, physical, and psychosocial explanatory variables to compare underreporting with plausible reporting of dietary intake. t -Tests were used to analyze continuous variables, and χ2-analyses compared categorical factors. Statistical significance was considered at P ≤ 0.05.

A multiple regression analysis was conducted to further explore independent associations of the psychosocial variables to underreporting of dietary intake. The dependent variable for the analysis was reporting estimates, and the independent variables were age, BMI, McKnight Unhealthy Eating Behaviors Subscale, body image discrepancy score, and self-efficacy for healthy diet.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Disclosure
  8. REFERENCES

Descriptive and psychosocial statistics of underreporting vs. plausible reporting

More than half (54.8%) of the cohort was categorized as underreporters. Table 1 presents the descriptive, demographic, and psychosocial results comparing the underreporters with the plausible reporters. Significant differences were observed between the groups in terms of age, height, weight, and BMI, i.e., all these variables were significantly higher (P ≤ 0.05) in underreporters vs. others. Furthermore, univariate analyses revealed significant differences between groups in three of the four psychosocial variables. No significant differences were found in the demographic variables and so were not included in subsequent models. Underreporters showed significantly higher self-efficacy for diet and unhealthy eating behaviors (P ≤ 0.05) and lower body image (P ≤ 0.001).

Table 1.  Physical and psychosocial characteristics among preadolescent African-American girls and their parent/caregiver on the basis of the girls' dietary intake reporting
inline image

Dietary and physical activity of underreporters vs. plausible reporters

Table 2 presents dietary intake variables (e.g., percentages of dietary intake of carbohydrates, proteins, and fat) and levels of objectively measured physical activity in underreporters and plausible reporters.

Table 2.  Dietary intake and physical activity in preadolescent African-American girls
inline image

Not surprisingly, the underreporters showed lower total energy intake, but the only significant difference between the groups in terms of their diet was that underreporters' diet included a higher percentage of calories from protein (P = 0.017) and lower intakes of calcium, (P ≤ 0.001), sodium (P ≤ 0.001), and vitamin C (P ≤ 0.001) than plausible reporters. The group of plausible reporters reported significantly more daily servings of fruit (P = 0.050) and vegetables (P = 0.017) than did underreporters. Regarding beverages, the underreporters reportedly consumed significantly fewer servings of sweetened beverages (P ≤ 0.001) than did the plausible reporters, but no significant difference was seen in the number of the servings of water (P = 0.791). Therefore, it appears that although certain foods, micronutrients, and total energy is lower in dietary underreporters, the percentage of total energy intake as fat and carbohydrate was not significantly different.

The plausible reporters demonstrated significantly higher levels of moderate-to-vigorous physical activity (P = 0.011) and higher average total activity counts per minute (P = 0.003) (Table 2). Physical activity may be included as a covariate in exploring predictors of underreporting, to account for lower dietary intake (despite very high BMIs) related to a lack of an active lifestyle. Separate models to investigate bivariate associations between reporting discrepancies and physical activity levels are useful to demonstrate unadjusted correlations.

Predictors of underreporting

To determine the independent correlates of dietary underreporting, we conducted a multiple linear regression with the level of underreporting as the dependent variable (lower scores represented greater underreporting and higher scores represented more accurate reporting) and family demographics and psychosocial variables as the independent variables (Table 3). The five independent variables that entered into the equation were BMI, age, McKnight Unhealthy Eating Behaviors Subscale, body image discrepancy score, and self-efficacy for healthy diet. Preliminary examination of the results indicated that all assumptions were met and multicollinearity was not present in the model (highest variance inflation factor is 1.74). Furthermore, two-way interactions between BMI and the other variables were not significant and so not retained in the final models.

Table 3.  Standardized coefficients of dietary intake underreporting in preadolescent African-American girls
inline image

Results from our regression analyses indicated that the five independent variables explain 41.0% of the variance in the underreporting of dietary intake. Four of the five variables significantly contributed to the final model (Table 3). Standardized regression coefficients were tested to determine the level of significance of the independent variables in the regression. In order of magnitude of associations with underreporting were higher BMI (β = −0.506, P ≤ 0.001), older age (β = −0.159, P = 0.001), greater unhealthy eating behaviors (β = −0.118, P = 0.025), and higher self-efficacy for healthy diet (β = −0.098, P = 0.033).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Disclosure
  8. REFERENCES

Obtaining an accurate assessment of dietary intake can be a difficult and daunting task, because many research participants underreport dietary intake. This study is the first to explore the psychosocial and demographic correlates of dietary underreporting in African-American preadolescent girls. Our results indicated that the majority of the cohort underreported their dietary intake, and underreporting of energy intake was associated with higher BMI, older age, disordered eating, and a greater self-efficacy for healthy eating.

Obesity status is a common predictor of dietary underreporting in adults (26) and children (27,28). Our results support previous reports that overweight status is a predictor of diet underreporting in African-American girls; BMI was also predictive of underreporting.

Bandini et al. (10) reported that older children were more likely to underreport their dietary intake, and our results supported that finding. In children, this finding does not make immediate intuitive sense, given that one would think that younger children would be more forgetful or less accurate than older children. However, perhaps as children get older, they may be more distracted socially during meal time. Another possibility is that as children get older, they may become more aware of “good” vs. “bad” foods and selectively report their intake. Studies have shown that underreporting may not be consistent across all food items; participants are more likely to underreport their consumption of unhealthy foods (29,30). However, our data suggest that underreporters in our cohort tended to underreport across a range of nutrients and food groups and did not differentially omit certain foods; thus, our findings are inconsistent with this theory.

In the multivariate equation, two psychosocial variables were independently related to dietary underreporting: self-efficacy for healthy diet and the unhealthy eating behaviors. The girls who believed they were able to eat a healthy diet were less likely to report plausible intake. This finding has interesting implications for the role of self-efficacy in dietary reporting. It would be interesting in future studies to determine whether modification of children's self-efficacy for diet would improve their accuracy of reporting. Greater levels of reported unhealthy eating behaviors in this study were also related to greater underreporting. This finding is similar to that in adults and recent findings by Ventura et al. (11), where higher levels of restrained eating and unhealthy eating behaviors were related to a greater likelihood of underreported dietary intake.

When considering parental education level and household income, we found no significant univariate relations with underreporting. In previous studies, both these variables were frequently cited as correlates of underreporting and were used to distinguish accurate reporters from underreporters. However, this relationship has not been explored in previous samples of African-American preadolescent girls. The majority of our cohort was from lower income families (64% at <$40,000/year); therefore, our group may not represent a sufficient range of incomes to detect a difference.

Although this study has several strengths, including the exploration of psychosocial predictors in a unique sample of African-American girls, limitations warrant discussion. First, BMR was estimated and not objectively measured. The World Health Organization calculation of BMR has been consistently reported as the most accurate for calculating children's BMR, but most of the samples used in those comparisons were from nonminority populations (9,13). It is important to note that Goldberg's cutoffs are based on multiple iterations on BMR. Further, the equation used to calculated BMR does not account for the effect of activity level on the estimate and may not account for possible ethnic differences in resting energy expenditure (31). Tershakovec et al. (32) found that African-American children have a lower resting energy expenditure than white children. This difference may place African-American children at higher risk for being misclassified as underreporters, because their BMR may be overestimated. Another limitation of this study is the limited variability of the sample. This study was restricted to only African-American girls aged 8–10 years. Although this approach allowed us to evaluate this group closely, it also limited the generalizability of our findings. Finally, this study is a cross-sectional view of the prevalence and correlates of dietary underreporting. Patterns of dietary underreporting or biases related to demand characteristics in prevention interventions are understudied for potential effects in children. Future studies of longitudinal relations would increase our knowledge in this area.

In summary, dietary underreporting appears to be prevalent in African-American preadolescent girls, and the reporting bias is systematically explained by several physical and psychosocial factors. Interventional strategies designed to improve accurate reporting of dietary intake are needed so that more precise relations between diet and health outcomes can be established. Routine inclusion of an estimate of dietary underreporting in research reports would help clarify the extent of reporting bias in particular studies, and provide a means to accumulate trends across high-risk populations. Until these strategies are in place, it is essential that there is a clear understanding of the potential bias present in the dietary data when evaluating prevention approaches to obesity and other chronic diseases.

REFERENCES

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Disclosure
  8. REFERENCES
  • 1
    Variyam JN, Blaylock J, Smallwood D, Basiotis PP. USDA's Healthy Eating Index and Nutrition Information. USDA: Washington, DC, 1998.
  • 2
    Kelley C, Krummel D, Gonzales EN, Neal WA, Fitch CW. Dietary intake of children at high risk for cardiovascular disease. J Am Diet Assoc 2004; 104: 222225.
  • 3
    Rockett HR, Colditz GA. Assessing diets of children and adolescents. Am J Clin Nutr 1997; 65 (Suppl 4):S1116S1122.
  • 4
    Ahrens W, Bammann K, de Henauw S et al. Understanding and preventing childhood obesity and related disorders—IDEFICS: a European multilevel epidemiological approach. Nutr Metab Cardiovasc Dis 2006; 16: 302308.
  • 5
    Dwyer J. Starting down the right path: nutrition connections with chronic diseases of later life. Am J Clin Nutr 2006; 83: 415S420S.
  • 6
    Field AE, Peterson KE, Gortmaker SL et al. Reproducibility and validity of a food frequency questionnaire among fourth to seventh grade inner-city school children: implications of age and day-to-day variation in dietary intake. Public Health Nutr 1999; 2: 293300.
  • 7
    Nicklas TA. Dietary studies of children and young adults (1973–1988): the Bogalusa Heart Study. Am J Med Sci 1995; 310 (Suppl 1):S101S108.
  • 8
    Johnson RK. Dietary intake—how do we measure what people are really eating? Obesity Res 2002; 10 (Suppl 1):S63S68.
  • 9
    Livingstone MB, Black AE. Markers of the validity of reported energy intake. J Nutr 2003; 133 (Suppl 3):S895S920.
  • 10
    Bandini LG, Cyr H, Must A, Dietz WH. Validity of reported energy intake in preadolescent girls. Am J Clin Nutr 1997; 65 (Suppl 4):S1138S1141.
  • 11
    Ventura AK, Loken E, Mitchell DC, Smiciklas-Wright H, Birch LL. Understanding reporting bias in the dietary recall data of 11-year-old girls. Obesity (Silver Spring) 2006; 14: 10731084.
  • 12
    Kimm SYS, Glynn NW, Obarzanek E, Aston CE, Daniels SR. Racial differences in correlates of misreporting of energy intake in adolescent females. Obesity (Silver Spring) 2006; 14: 156164.
  • 13
    Tverskaya R, Rising R, Brown D, Lifshitz F. Comparison of several equations and derivation of a new equation for calculating basal metabolic rate in obese children. J Am Coll Nutr 1998; 17: 333336.
  • 14
    World Health Organization. Energy and protein requirements. Report of a Joint FAO/WHO/UNU Expert Consultation. Technical Report Series 724; World Health Organization: Geneva, Switzerland, 1985.
  • 15
    Finan K, Larson DE, Goran MI. Cross-Validation of Prediction Equations for Resting Energy Expenditure in Young, Healthy Children. J Am Diet Assoc 1997; 97: 140145.
  • 16
    Klesges RC, Obarzanek E, Klesges LM et al. Memphis Girls health Enrichment Multi-site Studies (GEMS), phase 2: design and baseline characteristics. Contemp Clin Trials 2008; 29: 4255.
  • 17
    Cullen KW, Watson K, Himes JH et al. Evaluation of quality control procedures for 24-h dietary recalls: results from the Girls health Enrichment Multisite Studies. Prev Med 2004; 38 (Suppl):S14S23.
  • 18
    Murray DM, Stevens J, Hannan PJ et al. School-level intraclass correlation for physical activity in sixth grade girls. Med Sci Sports Exerc 2006; 38: 926936.
  • 19
    Reynolds CR, Paget KD. Factor analysis of the revised Children's Manifest Anxiety Scale for blacks, whites, males, and females with a national normative sample. J Consult Clin Psychol 1981; 49: 352359.
  • 20
    Klesges LM, Baranowski T, Beech B et al. Social desirability bias in self-reported dietary, physical activity and weight concerns measures in 8- to 10-year-old African-American girls: results from the Girls health Enrichment Multisite Studies (GEMS). Prev Med 2004; 38 (Suppl):S78S87.
  • 21
    Sherwood NE, Beech BM, Klesges LM et al. Measurement characteristics of weight concern and dieting measures in 8–10-year-old African-American girls from GEMS pilot studies. Prev Med 2004; 38 (Suppl):S50S59.
  • 22
    Stunkard AJ, Sorenson T, Schulsinger F. Use of the Danish adoption register for the study of obesity and thinness. In: Kety I (ed). Genetics of Neurological and Psychiatric Disorders. Raven Press: New York, 1983, pp 115120.
  • 23
    Reynolds KD, Yaroch AL, Franklin FA, Maloy J. Testing mediating variables in a school-based nutrition intervention program. Health Psychol 2002; 21: 5160.
  • 24
    Klesges RC, Eck LH, Ray JW. Who underreports dietary intake in a dietary recall? Evidence from the Second National Health and Nutrition Examination Survey. J Consult Clin Psychol 1995; 63: 438444.
  • 25
    Goldberg G, Black A, Jebb S et al. Critical evaluation of energy intake data using fundamental principals of energy physiology: 1. Derivation of cut-off limits to identify under-recording. Eur J Clin Nutr 1991; 45: 569581.
  • 26
    Johansson G, Wikman A, Ahrén AM, Hallmans G, Johansson I. Underreporting of energy intake in repeated 24-hour recalls related to gender, age, weight status, day of interview, educational level, reported food intake, smoking habits and area of living. Public Health Nutr 2001; 4: 919927.
  • 27
    Livingstone MB, Robson PJ, Wallace JM. Issues in dietary intake assessment of children and adolescents. Br J Nutr 2004; 92 (Suppl 2):S213S222.
  • 28
    Fisher JO, Johnson RK, Lindquist C, Birch LL, Goran MI. Influence of body composition on the accuracy of reported energy intake in children. Obesity Res 2000; 8: 597603.
  • 29
    Lafay L, Mennen L, Basdevant A et al. Does energy intake underreporting involve all kinds of food or only specific food items? Results from the Fleurbaix Laventie Ville Santé (FLVS) study. Int J Obes 2000; 24: 15001506.
  • 30
    Hise ME, Sullivan DK, Jacobsen DJ, Johnson SL, Donnelly JE. Validation of energy intake measurements determined from observer-recorded food records and recall methods compared with the doubly labeled water method in overweight and obese individuals. Am J Clin Nutr 2002; 75: 263267.
  • 31
    McDuffie JR, Adler-Wailes DC, Elberg J et al. Prediction equations for resting energy expenditure in overweight and normal-weight black and white children. Am J Clin Nutr 2004; 80: 365373.
  • 32
    Tershakovec AM, Kuppler KM, Zemel B, Stallings VA. Age, sex, ethnicity, body composition, and resting energy expenditure of obese African American and white children and adolescents. Am J Clin Nutr 2002; 75: 867871.