Impact of Interviewer's Body Mass Index on ... - Wiley Online Library

10 downloads 0 Views 52KB Size Report
a lean or obese interviewer to gather reported energy intake. (rEI). Basal metabolic ..... Our data show. (Table 4) that when the 2 days of recalls (phone and in-.
Impact of Interviewer’s Body Mass Index on Underreporting Energy Intake in Overweight and Obese Women Debra C. McKenzie, Rachel K. Johnson, Jean Harvey-Berino, and Beth Casey Gold

Abstract MCKENZIE, DEBRA C., RACHEL K. JOHNSON, JEAN HARVEY-BERINO, AND BETH CASEY GOLD. Impact of interviewer’s body mass index on underreporting energy intake in overweight and obese women. Obes Res. 2002;10:471– 477. Objective: To determine if overweight and obese women provide more accurate reports of their energy intake by 1) in-person recall with an obese interviewer, 2) in-person recall with a lean interviewer, or 3) telephone recall with an unknown interviewer. Research Methods and Procedures: Eighty-eight overweight and obese women participated in this study. Subjects completed one telephone-administered multiple-pass 24hour recall (MP24R) with an unknown interviewer and were then randomly assigned to an in-person MP24R with either a lean or obese interviewer to gather reported energy intake (rEI). Basal metabolic rate (BMR) was measured using a Deltrac monitor, and physical activity (EEPA) was estimated using a Caltrac accelerometer. Therefore, estimated energy expenditure was determined by: estTEE ⫽ (BMR ⫹ EEPA) ⫻ 1.10. Results: No significant differences were found between the two in-person interview modes for subject age, weight, body mass index, percentage of body fat, total energy expenditure, rEI, and misreporting of energy intake. In-person recall data were combined for comparison with the telephone recalls. No significant difference was found between the in-person and telephone recalls for rEI and misreporting. Mean reported energy intake was significantly lower than estimated total energy expenditure for the telephone recalls

Submitted for publication October 23, 2001. Accepted for publication in final form January 23, 2002. Department of Nutrition and Food Sciences, The University of Vermont, Burlington, Vermont. Address correspondence to Rachel K Johnson, 108 Morrill Hall, CALS Dean’s Office, University of Vermont, Burlington, VT 05405. E-mail: [email protected] Copyright © 2002 NAASO

and combined (lean and obese modes) in-person recalls. Conclusions: This study found that interviewer body mass index had no impact on self-reported energy intake during an in-person MP24R, and that telephone recall data were comparable with in-person recalls. Underreporting was a widespread problem (⬃26%) for all modes in this sample. Key words: dietary assessment, multiple-pass 24-hour recalls, dietary intake

Introduction Since the problem of misreporting energy intake has been uncovered, much research has been done to attempt to answer questions about who underreports (1–19). Researchers have found that underreporting tends to be highest among women (9) and the obese (3– 8). However, the question of “why” certain people are more prone to underreporting remains largely unanswered. It has been suggested that cultural factors may play a role in underreporting (18,19). Among Westernized societies, overweight and obese people are seen as unattractive, sloppy, lazy, and lacking self-control (20 –22). As countries become more Westernized and views of obesity and food change, underreporting of intake increases (18). Obese individuals in Westernized societies live with the stigma attached to being overweight and are well aware of the negative attitudes toward and social unacceptability of obesity (23). Embarrassing incidences related to being overweight could lead to feelings of inadequacy and a need for acceptance, and possibly to the intentional misreporting of dietary data (24). If an obese subject feels the interviewer agrees with society’s critical views of overweight people, it may become difficult to establish trust, which may lead to underreporting. In fact, studies have shown that registered dietitians and nutrition students (25,26), people who often participate in gathering dietary information, do have negative attitudes toward the obese. These negative attitudes, along with the negative opinions obese people may OBESITY RESEARCH Vol. 10 No. 6 June 2002

471

Impact of Interviewer BMI on Underreporting, McKenzie et al.

have regarding their own appearance, may impact an obese person’s desire to portray their diet or other health factors honestly. It has been found that physical characteristics of the interviewer can affect the information received during an in-person interview. Characteristics such as gender (27) and race (28) have been shown to be possible barriers in the forming of a relationship between interviewer and respondent. If the interview contains information that is considered “sensitive” or potentially embarrassing, having a withdrawn or uncomfortable interviewee could interfere with data collection. Studies on interview bias have shown that there is a higher level of trust if the interviewer and respondent are similar with respect to race, social status, or group membership (29). This in turn allows for a greater accuracy of response. Age has not been found to be a significant factor during interviews (30,31). No research was found that specifically looked at the possible impact of interviewer characteristics on food-intake information. It has been theorized that telephone interviews may be more objective because there is no actual physical presence of the interviewer and the psychosocial variables associated with face-to-face contact (32). Because physical appearance is not obvious during phone recalls, it may be assumed that if appearance interferes with data collection, telephone dietary recalls should obtain more accurate results. However, studies researching phone recalls (32,33) have shown that they gather only comparable information with an in-person interview. This may be because the respondents met the interviewer before the actual telephone recall (32), because they perceived social bias, particularly with obese subjects, or because of the technique used. No studies were found that specifically looked at obese women. There is an urgent need to identify strategies to improve the accuracy of dietary intake data in obese women, a population highly susceptible to underreporting. Because dietary recalls may be considered embarrassing or threatening to an obese woman, overweight individuals are considered to be more sympathetic, and subjects who identify with the interviewers provide more accurate information, it is possible that assessment methods may gather more valid data among obese women if the interviewer is overweight as well. The aim of this study was to determine if overweight and obese women provide more accurate reports of their energy intake by 1) in-person recall with an obese interviewer, 2) in-person recall with a lean interviewer, or 3) telephone recall with an unknown interviewer.

mont. The subjects participated in this study before they began a behavioral weight-control program. The research was approved by the Committee on Human Research for the Medical Sciences at the University of Vermont, Burlington, and written informed consent was obtained from each volunteer. To be eligible, subjects had to be between 25 and 45 years of age and have a body mass index (BMI) between 25 and 39.9 kg/m2. Subjects were excluded from participation if they smoked, were pregnant or planning to become pregnant within 18 months after the study began, had a history of major illness (i.e., cancer, heart disease, diabetes) and/or psychosis, regularly used prescription medication, or had any physical limitations that would preclude them from participating in an exercise program. Interviewers Dietary recalls were performed by two interviewers. Interviewer 1 had a BMI of 19.3 kg/m2 and interviewer 2 had a BMI of 33.4 kg/m2. Both interviewers and all subjects were white women. The interviewers were required to undergo intensive training in the use of multiple-pass 24-hour dietary-recall technique and the Food Intake Analysis System (FIAS), described below. All coding of the recalls was completed by interviewer 1 to prevent possible recall differences caused by variability in coding technique. Interinterviewer reliability was assessed before recalls began by having both interviewers obtain recalls from the same test subjects for the same day. Recalls were compared and training continued until similar results were obtained with no significant difference found.

Research Methods and Procedures

Study Protocol Each subject underwent two dietary recalls— one telephone recall and one in-person interview. Telephone interviews were completed before subjects were scheduled for their initial visit at the General Clinical Research Center (GCRC), and the subjects had no prior contact with the interviewer. Physicals were completed and preweights were recorded at least 9 days before the GCRC stay. Also, instructions on the use of the Caltrac accelerometer (Muscle Dynamic Fitness Network, Torrance, CA) were provided. The Caltracs were worn the week before the overnight stay as an objective measure of physical activity. Subjects were randomly distributed between the two interviewers for the in-person recalls. Subjects reported to the GCRC at the University of Vermont on the evening of day 1 for anthropometric measurements, including height and postweight, and the in-person recall. Subjects spent the night at the GCRC. On the morning of day 2, basal metabolic rate (BMR) was measured, and a DXA scan was performed for body-composition measurement.

Subjects The women participating in this study were recruited for an on-going research project studying the role of exercise in the maintenance of weight loss at the University of Ver-

Reported Energy Intake Self-reported energy intake for the previous day was gathered using the multiple-pass 24-hour recall (MP24HR)

472

OBESITY RESEARCH Vol. 10 No. 6 June 2002

Impact of Interviewer BMI on Underreporting, McKenzie et al.

as described by Tippett and Cypel (34). Briefly, the method consists of three passes—the quick list, detailed description, and the review—and is designed to minimize underreporting. Subjects were provided two-dimensional food models to aid in portion size estimation during an orientation session before the study. Interviewers did not attend the orientation. Recall data were coded and analyzed using the FIAS (Human Nutrition Center, University of Texas Health Science Center, School of Public Health, Houston, TX) previously described elsewhere (35). Number of Food Mentions The number of foods reported by each subject was used as another means of evaluating differences between the interview modes. The number of food mentions was determined by counting the total number of foods and beverages mentioned during a 1-day recall. All foods reported were counted as one food without taking portion size into account. Combination foods (i.e., turkey sandwich) were counted as one food mention. Condiments (mayonnaise, ketchup, creamers, etc.) were not counted as separate foods, but as part of the combination food. BMI Body weight was measured to the nearest 0.1 kg using a hospital-quality calibrated digital scale with the subject wearing street clothes and no shoes. Height was recorded to the nearest 0.5 cm using a stadiometer, also without shoes. BMI was calculated using the formula [weight (kilograms) divided by square height (square meters)].

each subject’s gender, height, weight, and age. The programmed values are used by the Caltrac to calculate estimated resting energy expenditure. Energy expended in physical activity is then estimated from the activity counts and recorded continuously in calories for the required record period. Subjects were asked to wear the Caltrac over a 7-day period. Participants recorded any time that the Caltrac was removed (to sleep, shower, etc.) and each 24-hour reading. To identify underreporters, the calories provided by the Caltrac were used to place subjects into low-, medium-, or high-activity levels. Subjects were classified based on the ranges provided by Black (41). If activity levels were determined to be ⬍60% of BMR, the subject was placed into the low category. If activity levels were between 60% and 71.5% of BMR, the subject was considered to have a medium-activity level, and those above 71.5% of BMR were placed in the high-activity category. Calculation of Total Energy Expenditure and Misreporting Total energy expenditure (TEE) is made up of three components: BMR, the thermic effect of food (TEF), and energy expended in physical activity (EEPA). BMR was measured by indirect calorimetry and EEPA was estimated using the Caltrac. TEF accounts for approximately 10% of TEE (42). Hence, estimated TEE was calculated as follows: estTEE ⫽ (BMR ⫹ EEPA) ⫻ 1.10. Self-reported energy intake (rEI) misreporting was calculated by subtracting TEE from the reported energy intake gathered using MP24R.

BMR BMR was measured by indirect calorimetry using a portable Deltrac metabolic monitor (Sensormedics, Inc., Loma Linda, CA) as described by Poehlman et al. (36,37). After an overnight fast, during which the subject slept in the GCRC, BMR was established for 45 minutes, which included a 30-minute measurement period after a 15-minute equilibration period. The monitor was calibrated against standard gases before each test and by ethanol combustion tests on a weekly basis. Expired air was collected using a clear plastic ventilated canopy system, which was placed over the subject’s head. Room air was drawn through the hood, and flow rate was measured. Energy expenditure was calculated using the Weir equation (38).

Identifying Underreporters The Goldberg cut-off, described extensively by Goldberg et al. (43), Black et al. (44), and later by Black (45), was used to identify the most obviously implausible intake values. The lower cut-off ratios (EI:BMR) used to identify underreporters were found to be 0.90, 0.94, and 1.04 for low, medium, and high activity, respectively, for individuals and 1 day of rEI and were 1.00, 1.04, and 1.16, respectively, for individuals and 2 days of rEI (phone and in-person). The upper limits to identify overreporters were found to be 2.72, 2.86, and 3.17 for low, medium, and high activity, respectively, for individuals and 1 day of rEI and 2.45, 2.57, and 2.85, respectively, for individuals and 2 days of rEI.

Physical Activity Physical activity monitors were used as an objective measure of physical activity. Caltrac physical monitors act as accelerometers and measure movement in the vertical plane. Caltracs were found to be strongly correlated (r ⫽ 0.69) to observed physical activity in adults (39). Validation against double-labeled water in adults yield correlations of r ⫽ 0.55 (p ⬍ 0.05) (40). Caltracs were programmed for

Body Composition Body composition, specifically percentage of body fat, was measured using a DPX DXA scanner (Lunar Radiation Corp., Madison, WI) as described in Svendsen et al. (46,47). Subjects laid supine on a padded table for 20 to 45 minutes, during which time a series of transverse scans were made from head to toe with 1-cm intervals. Pregnancy tests were performed before performing the DXA to further ensure safety. OBESITY RESEARCH Vol. 10 No. 6 June 2002

473

Impact of Interviewer BMI on Underreporting, McKenzie et al.

Statistics Data are presented as means ⫾ SD. One-sample Student’s t tests were performed to examine differences between the lean and obese interview modes with respect to physical characteristics of the subjects, energy expenditure, reported energy intake, and misreporting. Paired Student’s t tests were conducted to examine differences between reported energy intake, misreporting, and the number of foods mentioned during the telephone recalls and in-person recalls, as well as to compare the energy intakes reported by the telephone and in-person recalls and total energy expenditure. Data were analyzed using Excel (Microsoft Corp., Redmond, WA) for windows and the Statistical Package for the Social Sciences for Windows (Version X; SPSS Inc., Chicago, IL), with statistical significance set at p ⬍ 0.05.

Results Eighty-eight women between the ages of 25 and 46 years, with a BMI between 24.90 and 43.20 kg/m2 participated in the study (Table 1). As a group, subjects were not weightstable during the 2-week period between the physical and overnight stay, but rather, their mean weight significantly increased by 0.33 kg (p ⫽ 0.02). This suggests that, because the group was in positive energy balance, the mean reported energy intake should be in excess of the group mean total energy expenditure for both the inperson and telephone modes. No significant differences were found between the two modes for age, weight, BMI, percentage of body fat, total energy expenditure, reported energy intake, and misreporting of intake see (Table 2). Therefore, interviewer BMI did not impact the data received during an in-person interview. Hence, the in-person data were grouped for comparison with the telephone interviews. Mean reported energy intake was significantly lower than the estimated total energy expenditure for telephone recalls

Table 1. Characteristics of subjects (n ⫽ 88) Characteristics Age (years) Weight (kilograms) Body mass index Percentage of body fat Estimated TEE*

Mean ⴞ SD 37.8 ⫾ 5.7 87.7 ⫾ 12.3 32.0 ⫾ 4.2 41.6 ⫾ 5.4 2522.6 ⫾ 370.5

* TEE, in kilocalories, was estimated using the equation: estTEE ⫽ (BMR ⫹ EEPA) ⫻ 1.10. TEE, total energy expenditure; BMR, measured basal metabolic rate by Deltrac monitor; EEPA, estimated physical activity by Caltrac; 1.10, accounts for the thermic effect of food.

474

OBESITY RESEARCH Vol. 10 No. 6 June 2002

Table 2. Physical characteristics, energy expenditure, reported energy intake, and misreporting values estimated by the lean and obese interviewer modes (n ⫽ 88) Lean interviewer [mean ⴞ SD (n ⴝ 44)]

Obese interviewer [mean ⴞ SD (n ⴝ 44)]

Age (years) 37.2 ⫾ 6.1 38.3 ⫾ 5.3 Weight (kilograms) 86.1 ⫾ 10.6 89.3 ⫾ 13.8 Body mass index 31.6 ⫾ 3.9 32.4 ⫾ 4.6 Percent body fat 41.6 ⫾ 5.2 41.6 ⫾ 5.7 Estimated TEE* 2528.9.9 ⫾ 395.8 2516.3 ⫾ 348.0 Reported energy intake† 2123.6 ⫾ 887.0 1861.6 ⫾ 690.0 Misreporting of intake‡ ⫺405.4 ⫾ 957.44 ⫺654.7 ⫾ 785.5

p value 0.38 0.24 0.37 0.90 0.87 0.13 0.19

* TEE, in kilocalories, was estimated using the equation: estTEE ⫽ (BMR ⫹ EEPA) ⫻ 1.10. † Reported energy intake, in kilocalories, was gathered with multiple-pass 24-hour recalls. ‡ Misreporting of intake, in kilocalories, was determined by subtracting reported energy intake from estimated TEE. TEE, total energy expenditure; BMR, measured basal metabolic rate; EEPA, estimated physical activity; 1.10, accounts for the thermic effect of food.

and in-person recalls (see Table 3). Thus, there was a bias toward underreporting for both interview modes. No significant difference was found between the telephone recalls and in-person recalls with respect to reported energy intake, misreporting of energy intake, and the number of foods mentioned, indicating that the use of telephone recalls for gathering dietary information provides data that are similar to in-person recalls. There was no association between EI misreporting and BMI in the study sample. This was likely a result of the limited range of the subjects’ BMI. The telephone- and in-person-reported intakes were averaged to get a 2-day mean to determine if multiple days would increase the number of valid reports. No significant difference was found between the three groups for any of the reporting statuses— under-, valid-, or overreporting. Eighteen of the 88 women (20.5%) underreported both recalls. Underreporters (n ⫽ 26) reported significantly fewer foods than valid- and overreporters (n ⫽ 62; 9.4 vs. 11.29, respectively; p ⬍ 0.001) when phone and in-person recalls were averaged.

Impact of Interviewer BMI on Underreporting, McKenzie et al.

Table 3. Reported energy intake gathered using the multiple-pass 24-hour recall method, misreporting value, and number of food mentions estimated by telephone and in-person recalls (n ⫽ 88)

Reported energy intake* Misreporting of energy intake† Number of mentions‡

Phone (mean ⴞ SD)

In-person (mean ⴞ SD)

p value

2045.4 ⫾ 811.3

1992.6 ⫾ 801.0

0.57

⫺477.2 ⫾ 885.4

⫺530.0 ⫾ 879.6

0.57

10.6 ⫾ 3.0

10.8 ⫾ 3.2

0.53

* Reported energy intake, in kilocalories, was gathered with multiple-pass 24-hour recalls. † Misreporting of intake, in kilocalories, was determined by subtracting reported energy intake from estimated TEE. TEE was estimated using the equation: estTEE ⫽ (BMR ⫹ EEPA) ⫻ 1.10, estimated TEE mean ⫾ SD for the sample was determined to be 2522.6 ⫾ 370.5. ‡ The number of mentions was determined by counting the total number of food and beverages mentioned during the 1-day recall. Combination foods (i.e., turkey sandwich) were counted as one mention. Condiments were not counted as a food. TEE, total energy expenditure; BMR, measured basal metabolic rate; EEPA, estimated physical activity; 1.10 accounts for the thermic effect of food.

Discussion The major findings of this study on overweight and obese women were as follows: 1) interviewer BMI had no impact on self-reported energy intake during an in-person multiplepass 24-hour recall; 2) telephone recalls gathered data comparable with in-person recalls; and 3) on a group basis, multiple-pass 24-hour recalls underestimated energy intake. It seems that the rapport built between interviewer and subject are similar between telephone and in-person recalls because both gather comparable results. It is possible that without actual face-to-face contact, subjects may feel less need to report acceptable intakes and be more likely to report accurately. However, we know that telephone recalls are susceptible to the same amount of underreporting as the in-person interview. One reason for the continued problem may be the social stigma of obesity. Obese individuals are well aware of Western societies’ views of obesity. The need for social acceptance has been found to be associated with underreporting (16). This need may be so great that no matter who completes the recall, the need for approval outweighs the need for accuracy. Researchers should assess subjects’ risk of underreporting and should focus on stress-

Table 4. The number and percentage of subjects found to be under-, valid-, and overreporters for the telephone recalls and in-person recalls; and the reported energy intakes for the phone and in-person recalls were averaged (n ⫽ 88)

Underreporters* Valid-reporters† Overreporters‡

Phone recalls

In-person recalls

2-Day mean

21 (23.9%) 65 (73.8%) 2 (2.3%)

24 (27.3%) 62 (70.4%) 2 (2.3%)

26 (29.5%) 60 (68.2%) 2 (2.3%)

* Underreporters was defined as the number of individuals whose reported energy intake to basal metabolic rate ratio was below the determined lower 95% confidence limit. † Valid-reporters was defined as the number of individuals whose reported energy intake to basal metabolic rate ratio was within the determined 95% confidence limits. ‡ Overreporters was defined as the number of individuals whose reported energy intake to basal metabolic rate ratio was above the upper 95% confidence limit.

ing the importance of accurate data and receive training in recall methods to decrease the interviewee’s fear of reporting “bad” foods. The women who took part in our study were volunteers preparing to enter a behavioral weight-control program, and were therefore, aware of their weight problem and ready to take steps to change. It may be assumed that they would be more accurate in their reporting to ensure that they are successful in the program. It was noted that several women did comment on the “therapeutic” effect of telling the interviewer about “bad” foods that were consumed. Although reported intakes were low, it might be assumed that they indeed represent actual intake, and these women were simply eating less to decrease weight or were restrained eaters. However, the group mean weight increased, over an average period of 16 days, suggesting that as a group, the women were in positive energy balance and therefore reported energy intakes should be greater than TEE. Our data show (Table 4) that when the 2 days of recalls (phone and inperson) were averaged, 30% of the subjects had reported energy intakes that were significantly less than estimated TEE. Twenty percent of the sample reported intakes that were well below BMR for both recalls—telephone and in-person. This supports the notion that the subjects were underreporting either by omitting certain foods or underreporting portion sizes. Krebs-Smith et al. (48) reported that 21.7% of adults, both men and women (age ⱖ20 years; BMI ⬎ 25 kg/m2), who took part in the 1994 to 1996 United States Department of Agriculture-Continuing Survey of Food Intakes of Individuals, underreported food intake. The difference in OBESITY RESEARCH Vol. 10 No. 6 June 2002

475

Impact of Interviewer BMI on Underreporting, McKenzie et al.

the percentage of underreporters between the two studies may be because men were included in the CSFII; men have been found to underreport to a lesser degree than women (9,11). A limitation of this study was the use of only one recall per interview mode. However, the one recall was used to look at group differences and not individual differences to determine the best interview method for collecting groupintake data. When the phone and in-person recalls were averaged, providing 2 days of dietary intake data, the mean average of reported energy intake was still significantly lower than estimated TEE. Whereas we expect to see individual reports that are both below and above TEE because of day-to-day variability, group means should equal, or in the case of groups with increasing weights be greater than, the TEE. Means from repeated days of reported energy intake will not get any closer to actual intake if all days reported are below TEE, no matter how many days are gathered. Another limitation of this study was that TEE was not validated with a biomarker, such as double-labeled water. BMR was measured through indirect calorimetry, the Deltrac monitor, which is a well-established method. Because physical activity varies greatly among individuals and is a large component of energy intake, it is important to include a measure of physical activity in studies where self-reported energy intakes are compared with energy expenditure. Because biomarkers, such as double-labeled water, are costly and may not be feasible in all studies, researchers often must estimate activity levels using tools such as accelerometers and physical-activity recalls. For this study, physical activity was objectively measured using the Caltrac accelerometer. Studies that validated the Caltrac against double-labeled water and trained observation in adults found that the technique provides an accurate measurement of mean physical activity for groups (39,40). When underreporting was determined on an individual basis using the Goldberg cutoff, the Caltrac data were simply used to categorize physical activity into low, medium, and high levels and not as a direct measurement. Black and Cole (13) found that underreporting tends to be subject-specific. If a subject is an underreporter on one occasion, it is very likely that they will be an underreporter on other occasions. For example, this study found that 20% of the subjects underreported on both recalls. These subjects may be repeatedly misclassified into improper quartiles of nutrient or food intake (13), and associations between nutrient intake and disease may either be created or ignored (17). It is not always feasible to simply remove underreporters from studies, because they are often at high risk for disease and may make up large portions of groups being studied (e.g., low-income or obese populations). One-third of our sample would have to be removed from data analysis if we were to simply remove the underreporters. If underreporters are removed from a survey, unless the conclusions 476

OBESITY RESEARCH Vol. 10 No. 6 June 2002

remain unchanged, any conclusions based on this altered data may not be applicable to them. Therefore, we must continue to look for ways to determine who is at risk of underreporting and why, and develop methods that will lead to more accurate data. Increased usage of biomarkers in groups with a high risk of underreporting may be warranted until more is known.

Acknowledgments This project was supported by National Research Initiative Competitve Grants Program/United States Department of Agriculture 031CG to Dr. Johnson, National Institute of Diabetes and Digestive and Kidney Diseases Grant 51,517 to Dr. Harvey-Berino, and General Clinical Research Center (GCRC) Grant RR109. The authors would like to thank Betty Mayer, RN, and Rebecca Soultanakis, PhD, RD, for their help with the dietary recalls and interviewer training. References 1. Bandini LG, Schoeller DA, Cyr HN, Dietz WH. Validity of reported energy intake in obese and nonobese adolescents. Am J Clin Nutr. 1990;52:421–5. 2. Livingstone BE, Prentice AM, Coward WA, et al. Validation of estimates of energy intake by weighed dietary record and diet history in children and adolescents. Am J Clin Nutr. 1992;56:29 –35. 3. Prentice AM, Black AE, Coward WA, et al. High levels of energy expenditure in obese women. Br Med J. 1986;292:983–7. 4. Lichtman SW, Pisarska K, Berman ER, et al. Discrepancy between self-reported and actual caloric intake and exercise in obese subjects. N Engl J Med. 1992;327:1893– 8. 5. Livingstone MBE, Robson PJ. Measurement of dietary intake in children. Proc Nutr Soc. 2000;59:279 –93. 6. Kretsch MJ, Fong AKH, Green MW. Behavioral and body size correlates of energy intake underreporting by obese and normal-weight women. J Am Diet Assoc. 1999;99:300 – 6. 7. Samaras K, Kelly PJ, Campbell LV. Dietary underreporting is prevalent in middle-aged British women and is not related to adiposity (percentage body fat). Int J Obes Relat Metab Disord. 1999;23:881– 8. 8. Johnson RK, Soultanakis RP, Matthews DE. Literacy and body fatness are associated with underreporting of energy intake in U.S. low-income women using the multiple-pass 24-hour recall: a doubly labeled water study. J Am Diet Assoc. 1998;98:1136 – 40. 9. Johnson RK, Goran M, Poehlman E. Correlates of overand underreporting of energy intake in healthy older men and women. Am J Clin Nutr. l994;59:1286 –90. 10. Briefel RR, Sempos CT, McDowell MA, Chien SCY, Alaimo K. Dietary methods research in the third national health and nutrition examination survey: underreporting of energy intake. Am J Clin Nutr. 1997;65(suppl):1203S–9S. 11. Price GM, Paul AA, Cole TJ, Wadsworth MEJ. Characteristics of the low-energy reporters in a longitudinal national dietary survey. Br J Nutr. 1997;77:833–51. 12. Pryer JA, Vrijheid M, Nichols R, Kiggins M, Elliot P. Who are the ‘low energy reporters’ in the dietary and nutritional survey of British adults? Int J Epidemiol. 1997;26:146 –54.

Impact of Interviewer BMI on Underreporting, McKenzie et al.

13. Black AE, Cole TJ. Biased over- or under-reporting is characteristic of individuals whether over time or by different assessment methods. J Am Diet Assoc. 2001;101:70 – 80. 14. Blundell JE. What foods do people habitually eat? A dilemma for nutrition, an enigma for psychology. Am J Clin Nutr. 2000;71:3–5. 15. Johnson RK, Soultanakis RP, Matthews DE. Psychological factors and energy intake underreporting in women. FASEB J. 1999;13:A695. 16. Taren D, Tobar M, Hill A, et al. The association of energy intake bias with psychological scores of women. Eur J Clin Nutr. 1999;53:570 – 8. 17. Johnson RK. What are people really eating and why does it matter? Nutr Today. 2000;35:40 – 6. 18. Mennien LI, Jackson M, Cade J, et al. Underreporting of energy intake in four populations of African Origin. Int J Obes Relat Metab Disord. 2000;24:882–7. 19. Harrison GG, Galal OM, Ibrahim N, et al. Underreporting of food intake by dietary recall is not universal: a comparison of data from Egyptian and American women. J Nutr. 2000; 130:2049 –54. 20. Peternelj-Taylor CA. The effects of patient weight and sex on nurses’ perceptions: a proposed model of nurse withdrawal. J Adv Nurs. 1989;14:744 –54. 21. McArthur LH. Nutrition and non-nutrition majors have more favorable attitudes toward overweight people than personal overweight. J Am Diet Assoc. 1995;95:593– 6. 22. Neumark-Sztainer D. The weight dilemma: a range of philosophical perspectives. Int J Obes Relat Metab Disord. 1999; 23:S31–S7. 23. Fuller ML, Groce SB. Obese women’s responses to appearance norms. Free Inq Creat Sociol. 1991;19:167–74. 24. Muhlheim LS, Allison DB, Heshka S, Heymsfield SB. Do unsuccessful dieters intentionally underreport food intake. Int J Eat Disord. 1998;24:259 – 66. 25. Oberrieder H, Walker R, Monroe D, Adeyanju M. Attitudes of dietetics students and registered dietitians toward obesity. J Am Diet Assoc. 1995;95:914 – 6. 26. McArthur LH, Ross JK. Attitudes of registered dietitians toward personal overweight and overweight clients. J Am Diet Assoc. 1997;97:63–5. 27. Benney M, Riesman D, Star S. Age and sex in the interview. Am J Sociol. 1956;62:143–52. 28. Athey KR, Coleman JE, Reitman AP, Tang J. Two experiments showing the effect of the interviewer’s racial background on responses to questionnaires concerning racial issues. J Appl Psych. 1960;44:244 – 6. 29. Alers JO. Interviewer effects on survey response in an Andean estate. Int J Comp Sociol. 1970;11:208 –19. 30. Lin TR. A field study of race and age similarity effects on interview ratings in conventional and situational interviews. J Appl Psychol. 1992;77:363–71. 31. Hagenaars JA, Heinen T. Effects of role-independent interviewer characteristics on responses. In: Dijkstra W, Vanderzowen WD, eds. Response Behavior in the Survey-Interview. New York, NY: Academic Press; 1982, pp. 52– 64. 32. Tran KM, Johnson RK, Soultanakis RP, Matthews DE. In-person versus telephone administered multiple-pass 24-

33.

34.

35.

36.

37.

38. 39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

hour recalls in women: validation with doubly labeled water. J Am Diet Assoc. 2000;100:777– 80. Casey PH, Goolsby SLP, Lensing SY, Perloff BP, Bogle ML. The use of telephone interview methodology to obtain 24-hour dietary recalls. J Am Diet Assoc. 1999;99:1406 –11. Tippett KS, Cypel YS, eds. Design and Operation: The Continuing Survey of Food Intakes by Individuals and the Diet and Health Knowledge Survey, 1994 –1996. Beltsville, MD: United States Department of Agriculture, Agricultural Research Service; 1997. Johnson RK, Driscoll P, Goran MI. Comparison of multiple-pass 24-hour recall estimates of energy intake with total energy expenditure determined by the doubly labeled water method in young children. J Am Diet Assoc. 1996;96:1140 – 4. Poehlman ET, McAuliffe TL, Van Houten DR, Danforth E. Influence of age and endurance training on metabolic rate and hormones in healthy men. Am J Physiol. 1990;259:E66 –E72. Poehlman ET, Gardner AW, Ades PA, et al. Resting energy metabolism and cardiovascular disease risk in resistance-trained and aerobically trained males. Metabolism. 1992;41:1351– 60. Weir JB. New methods for calculating metabolic rate with special reference to protein. J Physiol. 1949;109:1–9. Klesges RC, Klesges LM, Swenson AM, Pheley AM. A validation of two motion sensors in the prediction of child and adult physical activity levels. Am J Epidemiol. 1985;122:400 –10. Gretebeck R, Montoye M, Porter W. Comparison of the doubly labeled water method for measuring energy expenditure with Caltrac accelerometer. Med Sci Sports Exerc. 1991;23:S60. Black AE. The sensitivity and specificity of the Goldberg cut-off for EI:BMR for identifying diet reports of poor validity. Eur J Clin Nutr. 2000;54:395– 404. Johnson RK, McKenzie D. Energy requirement methodology. In: Coulston A, Rock C, Monsen E, eds. Nutrition in the Treatment and Prevention of Disease. San Diego, CA: Academic Press; 2001, pp. 31– 42. Goldberg GR, Black AE, Jebb SA, et al. Critical evaluation of energy intake data using fundamental principles of energy physiology. 1. Derivation of cut-off limits to identify underrecording. Eur J Clin Nutr. 1991;45:569 – 81. Black AE, Goldberg GR, Jebb SA, Livingstone MBE, Cole TJ, Prentice AM. Critical evaluation of energy intake data using fundamental principles of energy physiology. 2. Evaluating the results of published surveys. Eur J Clin Nutr. 1991;45:583–99. Black AE. Critical evaluation of energy intake using the Goldberg cut-off for energy intake: basal metabolic rate. A practical guide to its calculation, use and limitations. Int J Obes Relat Metab Disord. 2000;24:1119 –30. Svendsen OL, Haarbo J, Heitmann BL, Gotfredsen A, Christiansen C. Measurement of body fat in elderly subjects by dual-energy x-ray absorptiometry, bioelectrical impedance, and anthropometry. Am J Clin Nutr. 1991;53:1117–23. Svendsen OL, Haarbo J, Hassager C, Christiansen C. Accuracy of measurements of body composition by dual-energy x-ray absorptiometry in vivo. Am J Clin Nutr. 1993;57:605– 8. Krebs-Smith SM, Graubard BI, Kahle LL, Subar AF, Cleveland LE, Ballard-Barbash R. Low energy reporters vs others: a comparison of reported food intakes. Eu J Clin Nutr. 2000;54:281–7.

OBESITY RESEARCH Vol. 10 No. 6 June 2002

477