- Email: [email protected]
Eating behavior in obese BED, obese non-BED, and non-obese control participants: A naturalistic study
Behaviour Research and Therapy 47 (2009) 897–900
Contents lists available at ScienceDirect
Behaviour Research and Therapy journal homepage: www.elsevier.com/locate/brat
Shorter communication
Eating behavior in obese BED, obese non-BED, and non-obese control participants: A naturalistic study Scott G. Engel a, b, *, Kirsten A. Kahler c, a, Chad M. Lystad a, Ross D. Crosby d, b, Heather K. Simonich a, Stephen A. Wonderlich b, a, Carol B. Peterson e, James E. Mitchell a, b a
Department of Clinical Research, Neuropsychiatric Research Institute, USA University of North Dakota School of Medicine and Health Sciences, USA Department of Psychology, Concordia College, USA d Department of Biostatistics, Neuropsychiatric Research Institute, USA e Department of Psychiatry, University of Minnesota, USA b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 13 February 2009 Received in revised form 25 June 2009 Accepted 30 June 2009
Laboratory studies have shown considerable differences between the eating behavior, particularly binge eating behavior, of participants with and without binge eating disorder (BED). However, these findings were not replicated in two field experiments employing ecological momentary assessment (EMA) in which obese BED and obese non-BED participants reported comparable binge eating behavior. In the current study, we examined differences in binge eating with an innovative assessment scheme employing both EMA and a standardized computer-based dietary recall program to avoid some of the limitations of past laboratory and field research. Obese BED, obese non-BED, and non-obese control participants reported significant differences in eating patterns, loss of control, overeating, and binge eating behavior. Of particular importance was the finding that BED participants engaged in more overeating and more binge eating episodes than non-BED participants. These findings suggest that the use of EMA in combination with dietary recall may be a relatively objective and useful approach to assessing binge eating behavior. The findings further suggest that individuals with BED are observably different from those without the disorder, which may have implications for eating disorder diagnoses in DSM-V. Ó 2009 Elsevier Ltd. All rights reserved.
Keywords: Binge eating disorder Binge eating Ecological momentary assessment
Introduction Binge eating disorder (BED) is characterized in the DSM-IV by the consumption of an objectively large amount of food coupled with a sense of loss of control and not accompanied by compensatory behaviors. BED patients are commonly obese and also experience more impaired emotional functioning, lower quality of life, and poorer physical health compared to non-BED obese individuals (de Zwaan et al., 2002). This population also may have higher rates of impairment and distress due to the increased presence of comorbid psychopathology (Wilfley et al., 2000). Further, the behavior of binge eating has been shown to be associated with weight gain and subsequent obesity (Fairburn, Cooper, Doll, & Norman, 2001; de Zwaan, 2001). Objective evidence that the eating behavior of BED patients differs from other obese individuals without BED would add
* Corresponding author. NRI, 120 8th St. S., Fargo, ND 58103, USA. E-mail address: [email protected] (S.G. Engel). 0005-7967/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.brat.2009.06.018
support to the validity of the diagnosis (Walsh & Boudreau, 2003). However, data from two different assessment methodologies are contradictory. Several laboratory studies have shown that obese BED individuals eat significantly more than comparable obese nonBED (NBED) participants during a simulation of a binge eating episode. When asked to ‘‘binge eat’’ or ‘‘let themselves go when eating’’, BED participants eat markedly more than NBED participants. This finding appears in the laboratory consistently, regardless of whether participants are given an array of food (e.g., Yanovski et al., 1992) or a single-item test meal (e.g., Sysko, Devlin, Walsh, Zimmerli, & Kissileff, 2007). Field studies have attempted to examine binge eating patients in their natural settings (e.g., Wegner et al., 2002) employing a methodology called Ecological Momentary Assessment (EMA) and two EMA-based field studies are particularly relevant to the current investigation. Contrary to the laboratory findings, Greeno, Wing, and Shiffman (2000) that both BED and NBED groups reported binge eating frequencies that well exceeded the minimum of two binge eating episodes per week required for diagnosis and that the caloric content of the participant-identified
898
S.G. Engel et al. / Behaviour Research and Therapy 47 (2009) 897–900
binge eating episodes did not differ between the BED and NBED groups (800 vs. 792 calorie, respectively). A second EMA study (le Grange, Gorin, Catley, & Stone, 2001) found that binge eating frequency did not significantly differ between the BED and NBED groups. The two EMA studies described above used a combination of self-identified behaviors (i.e., after a participant engaged in an eating episode, he/she labeled the eating episode as a ‘‘binge’’ or as something else) and food logs. Both of these techniques have proven useful in clinical settings as well as in several recent empirical studies, however they have several limitations, not the least of which are the subjectivity of the participant-labeled behaviors and the validity of the food logs due to poor compliance and retrospective recall biases (Engel, Wonderlich, & Crosby, 2005). Further, the use of traditional food logs likely limits the validity of comparisons of caloric intake across groups. Regardless of the specific reasons for the contradictory findings in the literature, reconciling this discrepancy will provide extremely useful information about the BED diagnosis. In the current study, we attempt to circumvent the above mentioned limitations by employing a novel adaptation of EMA assessment. Described in more detail below, we couple EMA techniques with a standardized computer dietary recall system that minimizes reliance on retrospective recall, avoids relying on participant definitions of behaviors, and enables us to place EMA and eating information in correct temporal space. Due to this more comprehensive assessment method, we hypothesized that both groups of obese samples (with BED and without BED) would have greater rates of binge eating and overeating than non-obese controls (NOC) while in their natural environment. Also, we hypothesized that the BED group would have greater rates of binge eating compared to the NBED obese group. Methods Participants Participants included 40 individuals recruited through community and university flyers and by referral from an eating disorders treatment facility. Flyers posted in the community and local universities sought individuals who were ‘‘normal weight or overweight and over the age of 18‘‘. Because we were primarily looking to recruit BED patients from the eating disorders treatment facility, clinicians were told that we were seeking individuals who were obese, binge ate, and did not engage in compensatory behaviors. Two participants were excluded from the analyses because they provided EMA data that appeared to be invalid. Of the remaining 38 participants, 16 were NOC (BMI 20-25 and no eating disorder as determined by the Structured Clinical Interview for DSM-IV Axis I disorders [SCID-IV; First, Spitzer, Gibbon, & Williams, 1995]), 13 were obese NBED (BMI>30 and no eating disorder diagnosis as determined by the SCID-IV), and 9 were obese BED (BMI>30 and a diagnosis of BED as determined by the SCID-IV). Participants with a BMI of 25–30 were excluded from the study. Participants could be male or female, and all participants were over the age of 18. Exclusion criteria included being pregnant or currently breastfeeding, having a current diagnosis of a psychotic disorder, previous gastrointestinal surgery, any medical illness requiring dietary modification, the use of any medication associated with weight or eating change, suicidal ideation, purging, and inability to read English. Participants were compensated $100 for participation in the study with the potential to earn another $50 for compliance by attending all scheduled appointments. This protocol was approved by the University of North Dakota Institutional Review Board.
Description of the sample Demographic information can be seen in Table 1. Measures Phone screen The eating disorder module of the SCID-IV (First et al., 1995) was administered by phone to determine eligibility and diagnosis and was used to determine group membership. The phone screen was also supplemented with probes from the Eating Disorder Examination (EDE; Fairburn & Cooper, 1993). This phone screen was the primary instrument used to diagnose the presence of BED in each participant, and was conducted by an SCID-IV/EDE trained Master’s level assessor. Ecological momentary assessment For the EMA protocol, each participant carried a handheld computer for 7 days and was asked to rate mood, stress, hunger, and level of control overeating just before they began any eating episode. Ratings for loss of control were made on a 1-5 Likert-type scale. A rating of 1 on the scale signified ‘‘No Control’’ and a 5 signified ‘‘Complete Control’’. Eating behavior The Nutritional Data System for Research (NDS-R) was used as to gather nutritional intake data on each of the eating episodes recorded by participants (Schakel, Sievert, & Buzzard, 1988). The NDS-R is a Windows-based, interviewer administered assessment that allows for the nutrient intake calculation of foods eaten over a 24 h time interval. It is considered by many to be the gold standard method of assessment of food intake (Feskanich, Sielaff, Chong, & Buzzard, 1999), and has been used successfully with research on overweight and obese samples (e.g., Ebbeling et al., 2004). NDS-R reports of caloric intake also correlate significantly with doubly labeled water data, suggesting the instrument is a valid assessment of eating behavior (Raymond, Neumeyer, Warren, Lee, & Peterson, 2003). Caloric data were collected for each eating episode recorded on the handheld computer. Each eating episode was classified by caloric amount into the categories of less than 1000 calorie, overeating, and binge eating (described below). Past research findings show that 1000 calorie or more exceeds what most people would consume in a typical eating episode (Keel, Cogley, Ghosh, & Lester, 2002, April; Mitchell, Crow, Peterson, Wonderlich, & Crosby, 1998). Therefore, overeating was defined as consuming 1000 calorie or more in one eating episode. Consistent with the DSM-IV definition, binge eating was defined as consuming 1000 calorie or more and a sense of loss of control (LOC). Loss of control was defined as a rating of 1–3 on the 1-5 Likert-type scale for LOC. Procedures After phone screening, qualified participants attended an informational meeting at the research facility. At this meeting, participants first gave their informed consent. Participants next provided descriptive and demographic information. Each participant was given thorough instruction on how to use the handheld computer and completed a one day practice period to ensure an understanding of EMA procedures. Practice data were not included in the analyses. Participants came to the research facility every day during the data collection period. At that time, information from the handheld computer was uploaded and monitored for compliance, and feedback was provided by the research coordinator to the participant about the quality of the data. The eating recordings were then used
S.G. Engel et al. / Behaviour Research and Therapy 47 (2009) 897–900
899
Table 1 Eating behavior by group.a
BMI Age % Caucasian % Never Married Eating Episodes/Day Kcals/Day Kcals/Eating Episode Loss of Control Overeatinge Binge Eatinge a b c d e
BED
NBED
Non-Obese Control
Mean (SE)
Mean (SE)
Mean (SE)
42.3 (3.4) 37.3 (4.9) 100 44.4 3.5 (.2) 2536.0 (136.1) 837.6 (53.3) 3.51 (.07) .24 (.03) .16 (.02)
36.5 (1.9) 34.6 (2.8) 84.6 38.5 3.0 (.2) 2005.0 (107.3) 739.4 (42.0) 3.55 (.06) .16 (.02) .09 (.02)
23.1 (.3) 22.7 (1.0) 92.9 68.8 4.0 (.2) 1606.0 (103.3) 414.6 (40.4) 4.18 (.06) .04 (.01) .01 (.01)
p
Effect Size
Post-hocb
.001 .001 .51 .02 <.001 <.001 <.001 <.001 <.001 <.001
.637c .337c .382d .748d .059c .083c .135c .040c .077c .170c
BED, NBED > NOC BED, NBED > NOC BED, NBED < NOC NBED < NOC BED > NBED > NOC BED, NBED > NOC BED, NBED < NOC BED > NBED > NOC BED > BNED > NOC
All reported means are age-corrected. NOC ¼ Non-obese Control, BED ¼ Binge Eating Disorder, NBED ¼ Non-Binge Eating Disorder. Proportion of variance explained. Phi coefficient. Percentage of episodes meeting criteria for overeating and binge eating is determined by multiplying the value provided by 100 (e.g., .24 100 ¼ 24%).
to interview the participants about their dietary intake. The time points of each eating episode were entered into the NDS-R system, and participants were asked to recall the details of each episode as well as if they had forgotten to report any eating episode. Following each NDS-R interview, the nutritional data of interest were merged with the EMA data in order to create a temporal picture of the eating and loss of control ratings for the previous day. At the final visit, participants returned their handheld computer, completed a payment form, and were debriefed. Statistical analysis Because groups were found to differ by age (p ¼ .001), with NOC participants being younger than BED and NBED participants, all inferential statistics were performed using age as a covariate and means reported are corrected for age. Binge eating, overeating, and loss of control variables were examined using one way analysis of covariance (ANCOVA), with age serving as the covariate. The three groups were compared on overeating (1000 calorie in one eating episode) and binge eating episodes (1000 calorie and a loss of control in an eating episode) using a two-level hierarchical generalized linear model (HGLM) based on a binomial sampling distribution (Raudenbush & Bryk, 2002). Level 1 observations were represented by momentary reports of overeating and binge eating behaviors. Level 2 observations were represented by the group variable: NOC, NBED, and BED. Analyses were performed using SPSS version 16.0.1 (SPSS Inc, 2008) and HLM Version 5.04 (Raudenbush, Byrk, Cheong, & Congdon, 2001). Pairwise ageadjusted, Bonferroni-corrected contrasts were used for post-hoc comparisons. Results Loss of control When examining loss of control immediately before any eating episode, a significant effect for group was identified, F(2,928) ¼ 28.03, p < .001. See Table 1 for more details. Eating behavior Table 1 shows the number of eating episodes per day, the number of kilocalories per day, and the number of kilocalories per eating episode (all age-corrected) for each of the three groups. Analyses showed a significant group difference in the number of
eating episodes per day, F(2,280) ¼ 9.25, p ¼ .001. The total number of kilocalories reported eaten each day also differed by group, F(2,280) ¼ 12.68, p ¼ .001. Analyses also showed a significant group difference on the number of kilocalories eaten at each eating episode, F(2,280) ¼ 19.99, p ¼ .001. See Table 1 for more details. Participants reported a total of 131 overeating episodes (>1000 kcal in an eating episode). After controlling for age, there was a significant effect for group on frequency of overeating, Wald Chi-Square(2) ¼ 51.69, p ¼ .001. A total of 57 binge eating episodes (>1000 kcal in an eating episode and a loss of control) were reported by all participants in the study. After controlling for age, there was a significant effect for group on frequency of binge eating, Wald Chi-Square(2) ¼ 52.47, p ¼ .001. Again, see Table 1 for more details.
Discussion In the current study, we found a number of important differences between the obese participants (BED and NBED) and the NOC group. Further, we found several differences between BED and NBED participants and, importantly, showed that the BED group reported more overeating and binge eating than the NBED group. We believe that the findings of the current study help clarify the contradictory conclusions of laboratory studies (e.g., Yanovski et al., 1992) and field studies (le Grange et al., 2001; Greeno et al., 2000). In spite of our modest sample size, we were able to find significant group differences between BED and NBED groups in overeating and binge eating. We believe this suggests the assessment strategy used in the current study is more sensitive than the assessment strategies previously used in field studies (e.g., food logs and binge eating episodes identified by participants). While these assessments have proven useful clinically, we believe the method used in the current study may be more objective. An obvious limitation of the current study is the modest sample size. However, in spite of the small number of participants per group, we had adequate power to detect the hypothesized group differences. Related to this concern, one must consider the representativeness of the small sample sizes in the current study. Another limitation is that despite the fact that we consider the dietary recall data collected to be more objective and less reliant on recall biases and interpretation than past field studies, these data are still based on self-report. Additionally, BED participants were recruited primarily through a treatment facility while obese nonBED participants primarily were recruited from the community via
900
S.G. Engel et al. / Behaviour Research and Therapy 47 (2009) 897–900
flyers. There is no way to know if these different recruitment techniques impacted the findings. Lastly, when defining binge eating, LOC ratings made immediately before the eating episode were utilized. It is possible that individuals may have developed a sense of LOC while eating and the pre-meal rating may have missed this. However, assessing LOC during the course of an eating episode (rather than immediately before and after) would very likely have reactive effects and this limitation precluded a LOC assessment during the middle of the meal. According to Walsh and Boudreau (2003), demonstrating that individuals with and without BED differ on objective measures of binge eating would provide important support for the validity of the diagnostic category. Also, we would add to this statement that demonstrating that these differences exist in the participants’ natural environment further support the construct of BED. However, the current findings should not be mistaken as support for the predictive validity of BED (i.e., course, treatment outcome, etc.), which need to be assessed with studies that follow BED patients over longer periods of time with longitudinal designs. Acknowledgements We thank the National Eating Disorders Association for providing funding for the current project. References Ebbeling, C. B., Sinclair, K. B., Pereira, M. A., Garcia-Lago, E., Feldman, H. A., & Ludwig, D. S. (2004). Compensation for energy intake from fast food among overweight and lean adolescents. Journal of the American Medical Association, 291, 2828–2833. Engel, S. G., Wonderlich, S. A., & Crosby, R. D. (2005). Ecological momentary assessment. In J. E. Mitchell, & C. Peterson (Eds.), The assessment of patients with eating disorders. New York: Guilford Press. Fairburn, D., & Cooper, Z. (1993). The eating disorder examination. In C. G. Fairburn, & G. T. Wilson (Eds.), Binge eating: Nature, assessment and treatment (pp. 317– 360). New York: Guilford. Fairburn, C., Cooper, Z., Doll, H., & Norman, P. M. (2001). The natural course of bulimia nervosa and binge eating disorder in young women. Archives of General Psychiatry, 57, 659–665. Feskanich, D., Sielaff, B. H., Chong, K., & Buzzard, I. M. (1999). Computerized collection and analysis of dietary intake information. Computer Methods Programs in Biomedicine, 30(1), 47–57.
First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1995). Structured clinical interview for DSM-IV Axis I disorders (SCID-IP). Washington, DC: American Psychiatric Press. le Grange, D., Gorin, A., Catley, D., & Stone, A. A. (2001). Does momentary assessment detect binge eating in overweight women that is denied at interview? European Eating Disorders Review, 9, 309–324. Greeno, C. G., Wing, R. R., & Shiffman, S. (2000). Binge antecedents in obese women with and without binge eating disorder. Journal of Consulting and Clinical Psychology, 68, 95–102. 2002 Keel, P. K., Cogley, C. B., Ghosh, S., & Lester, N. A. (2002, April). What constitutes an unusually large amount of food for defining binge episodes? Paper presented at the Academy for Eating Disorders 10th Interntional Conference on Eating Disorders, Boston, MA. Mitchell, J. E., Crow, S., Peterson, C. B., Wonderlich, S. A., & Crosby, R. D. (1998). Feeding laboratory studies in patients with eating disorders: a review. International Journal of Eating Disorders, 24, 115–124. Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data analysis methods (2nd ed.). Thousand Oaks, CA: Sage. Raudenbush, S. R., Byrk, A. S., Cheong, Y. F., & Congdon, R. (2001). HLM-5: Hierarchical linear and nonlinear modeling. Lincolnwood, IL: Scientific Software International. Raymond, N. C., Neumeyer, B., Warren, C. S., Lee, S. S., & Peterson, C. B. (2003). Energy intake patterns in obese women with binge eating disorder. Paper presented at the annual Eating Disorder Research Society meeting, Port Douglas, Australia. Schakel, S. F., Sievert, Y. A., & Buzzard, I. M. (1988). Sources of data for developing and maintaining a nutrient database. Journal of American Dietitian’s Association, 88, 1268–1271. SPSS for Windows. (2008). Rel. 16.0.1. Chicago, IL: SPSS Inc. Sysko, R., Devlin, M. J., Walsh, B. T., Zimmerli, E., & Kissileff, H. R. (2007). Satiety and test meal intake among women with binge eating disorder. International Journal of Eating Disorders, 40, 554–561. Walsh, B. T., & Boudreau, G. (2003). Laboratory studies of binge eating disorder. International Journal of Eating Disorders, 24, S30–S38. Wegner, K. E., Smyth, J. M., Crosby, R. D., Wittrock, D., Wonderlich, S. A., & Mitchell, J. E. (2002). An evaluation of the relationship between mood and binge eating in the natural environment using ecological momentary assessment. International Journal of Eating Disorders, 32, 352–361. Wilfley, D. E., Friedman, M. A., Dounchis, J. Z., Stein, R. I., Welch, R. R., & Ball, S. A. (2000). Comorbid psychopathology in binge eating disorder: relation to eating disorder severity at baseline and following treatment. Journal of Consulting and Clinical Psychology, 68, 641–649. Yanovski, S. Z., Leet, M., Flood, M., Yanovski, J. A., Gold, P. W., Kissileff, H. R., et al. (1992). Food intake and selection of obese women with and without binge eating disorder. American Journal of Clinical Nutrition, 56, 975–980. de Zwaan, M. (2001). Binge eating disorder and obesity. International Journal of Obesity and Related Metabolic Disorders, 25, S51–S55. de Zwaan, M., Mitchell, J. E., Howell, L. M., Monson, N., Swan-Kremeier, L., Roerig, J. L., et al. (2002). Two measures of health-related quality of life in morbid obesity. Obesity Research, 10, 1143–1151.
Contents lists available at ScienceDirect
Behaviour Research and Therapy journal homepage: www.elsevier.com/locate/brat
Shorter communication
Eating behavior in obese BED, obese non-BED, and non-obese control participants: A naturalistic study Scott G. Engel a, b, *, Kirsten A. Kahler c, a, Chad M. Lystad a, Ross D. Crosby d, b, Heather K. Simonich a, Stephen A. Wonderlich b, a, Carol B. Peterson e, James E. Mitchell a, b a
Department of Clinical Research, Neuropsychiatric Research Institute, USA University of North Dakota School of Medicine and Health Sciences, USA Department of Psychology, Concordia College, USA d Department of Biostatistics, Neuropsychiatric Research Institute, USA e Department of Psychiatry, University of Minnesota, USA b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 13 February 2009 Received in revised form 25 June 2009 Accepted 30 June 2009
Laboratory studies have shown considerable differences between the eating behavior, particularly binge eating behavior, of participants with and without binge eating disorder (BED). However, these findings were not replicated in two field experiments employing ecological momentary assessment (EMA) in which obese BED and obese non-BED participants reported comparable binge eating behavior. In the current study, we examined differences in binge eating with an innovative assessment scheme employing both EMA and a standardized computer-based dietary recall program to avoid some of the limitations of past laboratory and field research. Obese BED, obese non-BED, and non-obese control participants reported significant differences in eating patterns, loss of control, overeating, and binge eating behavior. Of particular importance was the finding that BED participants engaged in more overeating and more binge eating episodes than non-BED participants. These findings suggest that the use of EMA in combination with dietary recall may be a relatively objective and useful approach to assessing binge eating behavior. The findings further suggest that individuals with BED are observably different from those without the disorder, which may have implications for eating disorder diagnoses in DSM-V. Ó 2009 Elsevier Ltd. All rights reserved.
Keywords: Binge eating disorder Binge eating Ecological momentary assessment
Introduction Binge eating disorder (BED) is characterized in the DSM-IV by the consumption of an objectively large amount of food coupled with a sense of loss of control and not accompanied by compensatory behaviors. BED patients are commonly obese and also experience more impaired emotional functioning, lower quality of life, and poorer physical health compared to non-BED obese individuals (de Zwaan et al., 2002). This population also may have higher rates of impairment and distress due to the increased presence of comorbid psychopathology (Wilfley et al., 2000). Further, the behavior of binge eating has been shown to be associated with weight gain and subsequent obesity (Fairburn, Cooper, Doll, & Norman, 2001; de Zwaan, 2001). Objective evidence that the eating behavior of BED patients differs from other obese individuals without BED would add
* Corresponding author. NRI, 120 8th St. S., Fargo, ND 58103, USA. E-mail address: [email protected] (S.G. Engel). 0005-7967/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.brat.2009.06.018
support to the validity of the diagnosis (Walsh & Boudreau, 2003). However, data from two different assessment methodologies are contradictory. Several laboratory studies have shown that obese BED individuals eat significantly more than comparable obese nonBED (NBED) participants during a simulation of a binge eating episode. When asked to ‘‘binge eat’’ or ‘‘let themselves go when eating’’, BED participants eat markedly more than NBED participants. This finding appears in the laboratory consistently, regardless of whether participants are given an array of food (e.g., Yanovski et al., 1992) or a single-item test meal (e.g., Sysko, Devlin, Walsh, Zimmerli, & Kissileff, 2007). Field studies have attempted to examine binge eating patients in their natural settings (e.g., Wegner et al., 2002) employing a methodology called Ecological Momentary Assessment (EMA) and two EMA-based field studies are particularly relevant to the current investigation. Contrary to the laboratory findings, Greeno, Wing, and Shiffman (2000) that both BED and NBED groups reported binge eating frequencies that well exceeded the minimum of two binge eating episodes per week required for diagnosis and that the caloric content of the participant-identified
898
S.G. Engel et al. / Behaviour Research and Therapy 47 (2009) 897–900
binge eating episodes did not differ between the BED and NBED groups (800 vs. 792 calorie, respectively). A second EMA study (le Grange, Gorin, Catley, & Stone, 2001) found that binge eating frequency did not significantly differ between the BED and NBED groups. The two EMA studies described above used a combination of self-identified behaviors (i.e., after a participant engaged in an eating episode, he/she labeled the eating episode as a ‘‘binge’’ or as something else) and food logs. Both of these techniques have proven useful in clinical settings as well as in several recent empirical studies, however they have several limitations, not the least of which are the subjectivity of the participant-labeled behaviors and the validity of the food logs due to poor compliance and retrospective recall biases (Engel, Wonderlich, & Crosby, 2005). Further, the use of traditional food logs likely limits the validity of comparisons of caloric intake across groups. Regardless of the specific reasons for the contradictory findings in the literature, reconciling this discrepancy will provide extremely useful information about the BED diagnosis. In the current study, we attempt to circumvent the above mentioned limitations by employing a novel adaptation of EMA assessment. Described in more detail below, we couple EMA techniques with a standardized computer dietary recall system that minimizes reliance on retrospective recall, avoids relying on participant definitions of behaviors, and enables us to place EMA and eating information in correct temporal space. Due to this more comprehensive assessment method, we hypothesized that both groups of obese samples (with BED and without BED) would have greater rates of binge eating and overeating than non-obese controls (NOC) while in their natural environment. Also, we hypothesized that the BED group would have greater rates of binge eating compared to the NBED obese group. Methods Participants Participants included 40 individuals recruited through community and university flyers and by referral from an eating disorders treatment facility. Flyers posted in the community and local universities sought individuals who were ‘‘normal weight or overweight and over the age of 18‘‘. Because we were primarily looking to recruit BED patients from the eating disorders treatment facility, clinicians were told that we were seeking individuals who were obese, binge ate, and did not engage in compensatory behaviors. Two participants were excluded from the analyses because they provided EMA data that appeared to be invalid. Of the remaining 38 participants, 16 were NOC (BMI 20-25 and no eating disorder as determined by the Structured Clinical Interview for DSM-IV Axis I disorders [SCID-IV; First, Spitzer, Gibbon, & Williams, 1995]), 13 were obese NBED (BMI>30 and no eating disorder diagnosis as determined by the SCID-IV), and 9 were obese BED (BMI>30 and a diagnosis of BED as determined by the SCID-IV). Participants with a BMI of 25–30 were excluded from the study. Participants could be male or female, and all participants were over the age of 18. Exclusion criteria included being pregnant or currently breastfeeding, having a current diagnosis of a psychotic disorder, previous gastrointestinal surgery, any medical illness requiring dietary modification, the use of any medication associated with weight or eating change, suicidal ideation, purging, and inability to read English. Participants were compensated $100 for participation in the study with the potential to earn another $50 for compliance by attending all scheduled appointments. This protocol was approved by the University of North Dakota Institutional Review Board.
Description of the sample Demographic information can be seen in Table 1. Measures Phone screen The eating disorder module of the SCID-IV (First et al., 1995) was administered by phone to determine eligibility and diagnosis and was used to determine group membership. The phone screen was also supplemented with probes from the Eating Disorder Examination (EDE; Fairburn & Cooper, 1993). This phone screen was the primary instrument used to diagnose the presence of BED in each participant, and was conducted by an SCID-IV/EDE trained Master’s level assessor. Ecological momentary assessment For the EMA protocol, each participant carried a handheld computer for 7 days and was asked to rate mood, stress, hunger, and level of control overeating just before they began any eating episode. Ratings for loss of control were made on a 1-5 Likert-type scale. A rating of 1 on the scale signified ‘‘No Control’’ and a 5 signified ‘‘Complete Control’’. Eating behavior The Nutritional Data System for Research (NDS-R) was used as to gather nutritional intake data on each of the eating episodes recorded by participants (Schakel, Sievert, & Buzzard, 1988). The NDS-R is a Windows-based, interviewer administered assessment that allows for the nutrient intake calculation of foods eaten over a 24 h time interval. It is considered by many to be the gold standard method of assessment of food intake (Feskanich, Sielaff, Chong, & Buzzard, 1999), and has been used successfully with research on overweight and obese samples (e.g., Ebbeling et al., 2004). NDS-R reports of caloric intake also correlate significantly with doubly labeled water data, suggesting the instrument is a valid assessment of eating behavior (Raymond, Neumeyer, Warren, Lee, & Peterson, 2003). Caloric data were collected for each eating episode recorded on the handheld computer. Each eating episode was classified by caloric amount into the categories of less than 1000 calorie, overeating, and binge eating (described below). Past research findings show that 1000 calorie or more exceeds what most people would consume in a typical eating episode (Keel, Cogley, Ghosh, & Lester, 2002, April; Mitchell, Crow, Peterson, Wonderlich, & Crosby, 1998). Therefore, overeating was defined as consuming 1000 calorie or more in one eating episode. Consistent with the DSM-IV definition, binge eating was defined as consuming 1000 calorie or more and a sense of loss of control (LOC). Loss of control was defined as a rating of 1–3 on the 1-5 Likert-type scale for LOC. Procedures After phone screening, qualified participants attended an informational meeting at the research facility. At this meeting, participants first gave their informed consent. Participants next provided descriptive and demographic information. Each participant was given thorough instruction on how to use the handheld computer and completed a one day practice period to ensure an understanding of EMA procedures. Practice data were not included in the analyses. Participants came to the research facility every day during the data collection period. At that time, information from the handheld computer was uploaded and monitored for compliance, and feedback was provided by the research coordinator to the participant about the quality of the data. The eating recordings were then used
S.G. Engel et al. / Behaviour Research and Therapy 47 (2009) 897–900
899
Table 1 Eating behavior by group.a
BMI Age % Caucasian % Never Married Eating Episodes/Day Kcals/Day Kcals/Eating Episode Loss of Control Overeatinge Binge Eatinge a b c d e
BED
NBED
Non-Obese Control
Mean (SE)
Mean (SE)
Mean (SE)
42.3 (3.4) 37.3 (4.9) 100 44.4 3.5 (.2) 2536.0 (136.1) 837.6 (53.3) 3.51 (.07) .24 (.03) .16 (.02)
36.5 (1.9) 34.6 (2.8) 84.6 38.5 3.0 (.2) 2005.0 (107.3) 739.4 (42.0) 3.55 (.06) .16 (.02) .09 (.02)
23.1 (.3) 22.7 (1.0) 92.9 68.8 4.0 (.2) 1606.0 (103.3) 414.6 (40.4) 4.18 (.06) .04 (.01) .01 (.01)
p
Effect Size
Post-hocb
.001 .001 .51 .02 <.001 <.001 <.001 <.001 <.001 <.001
.637c .337c .382d .748d .059c .083c .135c .040c .077c .170c
BED, NBED > NOC BED, NBED > NOC BED, NBED < NOC NBED < NOC BED > NBED > NOC BED, NBED > NOC BED, NBED < NOC BED > NBED > NOC BED > BNED > NOC
All reported means are age-corrected. NOC ¼ Non-obese Control, BED ¼ Binge Eating Disorder, NBED ¼ Non-Binge Eating Disorder. Proportion of variance explained. Phi coefficient. Percentage of episodes meeting criteria for overeating and binge eating is determined by multiplying the value provided by 100 (e.g., .24 100 ¼ 24%).
to interview the participants about their dietary intake. The time points of each eating episode were entered into the NDS-R system, and participants were asked to recall the details of each episode as well as if they had forgotten to report any eating episode. Following each NDS-R interview, the nutritional data of interest were merged with the EMA data in order to create a temporal picture of the eating and loss of control ratings for the previous day. At the final visit, participants returned their handheld computer, completed a payment form, and were debriefed. Statistical analysis Because groups were found to differ by age (p ¼ .001), with NOC participants being younger than BED and NBED participants, all inferential statistics were performed using age as a covariate and means reported are corrected for age. Binge eating, overeating, and loss of control variables were examined using one way analysis of covariance (ANCOVA), with age serving as the covariate. The three groups were compared on overeating (1000 calorie in one eating episode) and binge eating episodes (1000 calorie and a loss of control in an eating episode) using a two-level hierarchical generalized linear model (HGLM) based on a binomial sampling distribution (Raudenbush & Bryk, 2002). Level 1 observations were represented by momentary reports of overeating and binge eating behaviors. Level 2 observations were represented by the group variable: NOC, NBED, and BED. Analyses were performed using SPSS version 16.0.1 (SPSS Inc, 2008) and HLM Version 5.04 (Raudenbush, Byrk, Cheong, & Congdon, 2001). Pairwise ageadjusted, Bonferroni-corrected contrasts were used for post-hoc comparisons. Results Loss of control When examining loss of control immediately before any eating episode, a significant effect for group was identified, F(2,928) ¼ 28.03, p < .001. See Table 1 for more details. Eating behavior Table 1 shows the number of eating episodes per day, the number of kilocalories per day, and the number of kilocalories per eating episode (all age-corrected) for each of the three groups. Analyses showed a significant group difference in the number of
eating episodes per day, F(2,280) ¼ 9.25, p ¼ .001. The total number of kilocalories reported eaten each day also differed by group, F(2,280) ¼ 12.68, p ¼ .001. Analyses also showed a significant group difference on the number of kilocalories eaten at each eating episode, F(2,280) ¼ 19.99, p ¼ .001. See Table 1 for more details. Participants reported a total of 131 overeating episodes (>1000 kcal in an eating episode). After controlling for age, there was a significant effect for group on frequency of overeating, Wald Chi-Square(2) ¼ 51.69, p ¼ .001. A total of 57 binge eating episodes (>1000 kcal in an eating episode and a loss of control) were reported by all participants in the study. After controlling for age, there was a significant effect for group on frequency of binge eating, Wald Chi-Square(2) ¼ 52.47, p ¼ .001. Again, see Table 1 for more details.
Discussion In the current study, we found a number of important differences between the obese participants (BED and NBED) and the NOC group. Further, we found several differences between BED and NBED participants and, importantly, showed that the BED group reported more overeating and binge eating than the NBED group. We believe that the findings of the current study help clarify the contradictory conclusions of laboratory studies (e.g., Yanovski et al., 1992) and field studies (le Grange et al., 2001; Greeno et al., 2000). In spite of our modest sample size, we were able to find significant group differences between BED and NBED groups in overeating and binge eating. We believe this suggests the assessment strategy used in the current study is more sensitive than the assessment strategies previously used in field studies (e.g., food logs and binge eating episodes identified by participants). While these assessments have proven useful clinically, we believe the method used in the current study may be more objective. An obvious limitation of the current study is the modest sample size. However, in spite of the small number of participants per group, we had adequate power to detect the hypothesized group differences. Related to this concern, one must consider the representativeness of the small sample sizes in the current study. Another limitation is that despite the fact that we consider the dietary recall data collected to be more objective and less reliant on recall biases and interpretation than past field studies, these data are still based on self-report. Additionally, BED participants were recruited primarily through a treatment facility while obese nonBED participants primarily were recruited from the community via
900
S.G. Engel et al. / Behaviour Research and Therapy 47 (2009) 897–900
flyers. There is no way to know if these different recruitment techniques impacted the findings. Lastly, when defining binge eating, LOC ratings made immediately before the eating episode were utilized. It is possible that individuals may have developed a sense of LOC while eating and the pre-meal rating may have missed this. However, assessing LOC during the course of an eating episode (rather than immediately before and after) would very likely have reactive effects and this limitation precluded a LOC assessment during the middle of the meal. According to Walsh and Boudreau (2003), demonstrating that individuals with and without BED differ on objective measures of binge eating would provide important support for the validity of the diagnostic category. Also, we would add to this statement that demonstrating that these differences exist in the participants’ natural environment further support the construct of BED. However, the current findings should not be mistaken as support for the predictive validity of BED (i.e., course, treatment outcome, etc.), which need to be assessed with studies that follow BED patients over longer periods of time with longitudinal designs. Acknowledgements We thank the National Eating Disorders Association for providing funding for the current project. References Ebbeling, C. B., Sinclair, K. B., Pereira, M. A., Garcia-Lago, E., Feldman, H. A., & Ludwig, D. S. (2004). Compensation for energy intake from fast food among overweight and lean adolescents. Journal of the American Medical Association, 291, 2828–2833. Engel, S. G., Wonderlich, S. A., & Crosby, R. D. (2005). Ecological momentary assessment. In J. E. Mitchell, & C. Peterson (Eds.), The assessment of patients with eating disorders. New York: Guilford Press. Fairburn, D., & Cooper, Z. (1993). The eating disorder examination. In C. G. Fairburn, & G. T. Wilson (Eds.), Binge eating: Nature, assessment and treatment (pp. 317– 360). New York: Guilford. Fairburn, C., Cooper, Z., Doll, H., & Norman, P. M. (2001). The natural course of bulimia nervosa and binge eating disorder in young women. Archives of General Psychiatry, 57, 659–665. Feskanich, D., Sielaff, B. H., Chong, K., & Buzzard, I. M. (1999). Computerized collection and analysis of dietary intake information. Computer Methods Programs in Biomedicine, 30(1), 47–57.
First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1995). Structured clinical interview for DSM-IV Axis I disorders (SCID-IP). Washington, DC: American Psychiatric Press. le Grange, D., Gorin, A., Catley, D., & Stone, A. A. (2001). Does momentary assessment detect binge eating in overweight women that is denied at interview? European Eating Disorders Review, 9, 309–324. Greeno, C. G., Wing, R. R., & Shiffman, S. (2000). Binge antecedents in obese women with and without binge eating disorder. Journal of Consulting and Clinical Psychology, 68, 95–102. 2002 Keel, P. K., Cogley, C. B., Ghosh, S., & Lester, N. A. (2002, April). What constitutes an unusually large amount of food for defining binge episodes? Paper presented at the Academy for Eating Disorders 10th Interntional Conference on Eating Disorders, Boston, MA. Mitchell, J. E., Crow, S., Peterson, C. B., Wonderlich, S. A., & Crosby, R. D. (1998). Feeding laboratory studies in patients with eating disorders: a review. International Journal of Eating Disorders, 24, 115–124. Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data analysis methods (2nd ed.). Thousand Oaks, CA: Sage. Raudenbush, S. R., Byrk, A. S., Cheong, Y. F., & Congdon, R. (2001). HLM-5: Hierarchical linear and nonlinear modeling. Lincolnwood, IL: Scientific Software International. Raymond, N. C., Neumeyer, B., Warren, C. S., Lee, S. S., & Peterson, C. B. (2003). Energy intake patterns in obese women with binge eating disorder. Paper presented at the annual Eating Disorder Research Society meeting, Port Douglas, Australia. Schakel, S. F., Sievert, Y. A., & Buzzard, I. M. (1988). Sources of data for developing and maintaining a nutrient database. Journal of American Dietitian’s Association, 88, 1268–1271. SPSS for Windows. (2008). Rel. 16.0.1. Chicago, IL: SPSS Inc. Sysko, R., Devlin, M. J., Walsh, B. T., Zimmerli, E., & Kissileff, H. R. (2007). Satiety and test meal intake among women with binge eating disorder. International Journal of Eating Disorders, 40, 554–561. Walsh, B. T., & Boudreau, G. (2003). Laboratory studies of binge eating disorder. International Journal of Eating Disorders, 24, S30–S38. Wegner, K. E., Smyth, J. M., Crosby, R. D., Wittrock, D., Wonderlich, S. A., & Mitchell, J. E. (2002). An evaluation of the relationship between mood and binge eating in the natural environment using ecological momentary assessment. International Journal of Eating Disorders, 32, 352–361. Wilfley, D. E., Friedman, M. A., Dounchis, J. Z., Stein, R. I., Welch, R. R., & Ball, S. A. (2000). Comorbid psychopathology in binge eating disorder: relation to eating disorder severity at baseline and following treatment. Journal of Consulting and Clinical Psychology, 68, 641–649. Yanovski, S. Z., Leet, M., Flood, M., Yanovski, J. A., Gold, P. W., Kissileff, H. R., et al. (1992). Food intake and selection of obese women with and without binge eating disorder. American Journal of Clinical Nutrition, 56, 975–980. de Zwaan, M. (2001). Binge eating disorder and obesity. International Journal of Obesity and Related Metabolic Disorders, 25, S51–S55. de Zwaan, M., Mitchell, J. E., Howell, L. M., Monson, N., Swan-Kremeier, L., Roerig, J. L., et al. (2002). Two measures of health-related quality of life in morbid obesity. Obesity Research, 10, 1143–1151.