Reliability, stability, and factor structure of the Bulimia Test-Revised and Eating Disorder Inventory-2 scales in adolescence

Measurement invariance of the Eating Pathology Symptoms Inventory (EPSI) in adolescents and adults

Adolescence is a common period for eating disorder (ED) onset. The availability of psychometrically sound measures of ED psychopathology enables clinicians to accurately assess symptoms and monitor treatment outcomes continuously from adolescence and adulthood. The purpose of this study was to assess if the Eating Pathology Symptoms Inventory (EPSI) is invariant across adolescents and adults. Participants (N = 29,821) were adolescent (n = 5250) and adult (n = 24,571) users of the Recovery Record (RR) mobile phone application who provided EPSI responses through the application. Measurement invariance testing was conducted to assess invariance of the EPSI Body Dissatisfaction, Restricting, Excessive Exercise, Purging, Cognitive Restraint, and Binge Eating scales across adolescents (age 13 through 17) and adults (age 18 and older). Findings indicated that all EPSI factors administered in the RR app replicated in both adolescent and adult users. The EPSI factor structure was largely equivalent in adolescents and adults, demonstrating evidence for configural and metric invariance, as well as some evidence for scalar invariance. Our results indicated that EPSI scales measured the same constructs across development. Clinicians and researchers may benefit from utilizing the EPSI to measure ED psychopathology in adolescents and for continued progress monitoring into adulthood.

The Bulimia Test-Revised (BULIT-R): Psychometric properties in a non-clinical sample of White and Latina college women

The Bulimia Test-Revised (BULIT-R) is a measure of bulimia nervosa and binge eating disorder symptoms. While this measure can serve as a useful screening tool, the validity of the BULIT-R in White and Latina samples is uncertain. Previous research has proposed several different BULIT-R factor structures, but has yet to find consistent support for a measurement model or to test measurement invariance. To address this gap, the current study attempted to examine the factor structure and measurement invariance of the BULIT-R among 699 non-clinical White (40.9%) and Latina (59.1%) undergraduate women. Analyses indicated that none of the previously published factor structures in adult samples of the BULIT-R were an acceptable fit in either Latina or White undergraduates. These findings further underscore difficulties in replicating the BULIT-R factor structure, even when using similar sample characteristics to those in the literature. Given the lack of an adequate fitting measurement model, tests of measurement invariance were not conducted. In the absence of measurement invariance of the BULIT-R, there is uncertainty regarding group comparisons, such as whether group differences reflect true differences or are artifacts of measurement error. This study highlights the need for additional psychometric investigation of the BULIT-R. The focus should be on diverse groups as well as non-clinical samples, with the latter being less likely to endorse extreme eating behaviors. Special consideration should be given to the number of items included in the measurement model and the number of indicators per latent factor. Caution should be exercised when interpreting scores on this instrument.

Association of etiological factors across the extreme end and continuous variation in disordered eating in female Swedish twins

BACKGROUND

Accumulating evidence suggests that many psychiatric disorders etiologically represent the extreme end of dimensionally distributed features rather than distinct entities. The extent to which this applies to eating disorders (EDs) is unknown.

METHODS

We investigated if there is similar etiology in (a) the continuous distribution of the Eating Disorder Inventory-2 (EDI-2), (b) the extremes of EDI-2 score, and (c) registered ED diagnoses, in 1481 female twin pairs at age 18 years (born 1992-1999). EDI-2 scores were self-reported at age 18. ED diagnoses were identified through the Swedish National Patient Register, parent-reported treatment and/or self-reported purging behavior of a frequency and duration consistent with DSM-IV criteria. We differentiated between anorexia nervosa (AN) and other EDs.

RESULTS

The heritability of the EDI-2 score was 0.65 (95% CI 0.61-0.68). The group heritabilities in DeFries-Fulker extremes analyses were consistent over different percentile-based extreme groups [0.59 (95% CI 0.37-0.81) to 0.65 (95% CI 0.55-0.75)]. Similarly, the heritabilities in liability threshold models were consistent over different levels of severity. In joint categorical-continuous models, the twin-based genetic correlation was 0.52 (95% CI 0.39-0.65) between EDI-2 score and diagnoses of other EDs, and 0.26 (95% CI 0.08-0.42) between EDI-2 score and diagnoses of AN. The non-shared environmental correlations were 0.52 (95% CI 0.32-0.70) and 0.60 (95% CI 0.38-0.79), respectively.

CONCLUSIONS

Our findings suggest that some EDs can partly be conceptualized as the extreme manifestation of continuously distributed ED features. AN, however, might be more distinctly genetically demarcated from ED features in the general population than other EDs.

Body image as a predictor of nonsuicidal self-injury in women: A longitudinal study

OBJECTIVES

To determine whether Body Image Dissatisfaction (BID) predicted NonSuicidal Self-Injury (NSSI) cross-sectionally and longitudinally, independent of comorbidity between NSSI and Disordered Eating (DE). Another aim was to determine whether BID could predict number of NSSI methods present.

METHOD

Adult females completed measures of NSSI and DE (n = 283); and a longitudinal sample (n = 106) completed these measures again one year later.

RESULTS

BID was a small yet significant predictor of NSSI both cross-sectionally and longitudinally. Poorer BID significantly explained a greater number of NSSI methods cross-sectionally and longitudinally.

CONCLUSIONS

BID explains unique variance in NSSI (including increased number of methods), and is not a function of comorbidity with DE. This has the potential to influence theory, as well as inform early intervention initiatives for BID in females. Further research is required to determine other variables implicated in this relationship, as well as whether these findings are applicable to other groups such as adolescents and males.

Test–Retest Reliability of Common Measures of Eating Disorder Symptoms in Men Versus Women

Approximately 10% to 30% of individuals with eating disorders (EDs) are male, yet because measures often have not been tested among male participants, it is unclear whether the psychometric properties of ED measures are equivalent between sexes. The purpose of this study was to compare the test–retest reliability of common ED measures in men versus women. Participants ( N = 227; 58.1% female) completed self-report measures of body dissatisfaction, restrained eating, disinhibited eating, bulimic symptoms, and desire-for-muscularity at baseline and 2-to-4 weeks later. Intraclass correlations were used to compute retest correlations. Spearman’s rho was used to compute retest correlations for skewed and kurtotic variables. We compared 95% confidence intervals for intraclass correlation coefficients to determine whether measures differed in reliability between sexes. Most ED measures had at least acceptable test–retest reliabilities. However, few measures of disinhibited and binge eating demonstrated good reliability in men. Results highlight the utility of several ED measures for assessing symptom change over time, and the need for additional research to identify and correct for sources of gender unreliability among ED self-report measures in men—particularly for assessing constructs that include binge-eating behavior.

Disordered Eating, Compulsive Exercise, and Sport Participation in a UK Adolescent Sample

The sport literature has produced equivocal results as to whether sport participation is a protective or risk factor for disordered eating. One mechanism by which it could be a risk factor is the increased drive or compulsion to exercise. This study compared the levels of disordered eating and compulsive exercise between adolescent sport and non-sport participants. A sample of 417 male and female adolescents, aged 14-16 years old, was recruited from UK secondary schools. Participants completed questionnaire packs that included: the Eating Disorder Inventory; a measure of exercise behaviour; and the Compulsive Exercise Test (CET). Non-sport participants reported significantly greater body dissatisfaction than sport participants, and this was true for boys and girls. Significant group differences were also reported for many of the CET scales, with sport participants generally reporting greater levels of compulsive exercise than non-sport participants. Implications of these results are discussed. Copyright © 2016 John Wiley & Sons, Ltd and Eating Disorders Association.

Bone parameters in relation to attitudes and feelings associated with disordered eating in oligo-amenorrheic athletes, eumenorrheic athletes, and nonathletes

OBJECTIVES

Disordered eating may negatively impact bone in athletes. However, it is not known whether this effect is independent of the associated amenorrhea and relative hypercortisolemia. We aimed to compare attitudes, feelings, and cognitions associated with disordered eating using the Three-Factor Eating Questionnaire (TFEQ) and Eating Disorder Inventory-2 (EDI-2) in normal-weight oligomenorrheic athletes (OA), eumenorrheic athletes (EA), and nonathletes, and determine the associations with bone independent of confounders.

METHOD

109 OA, 39 EA, and 36 nonathletes (14-25 years) completed the TFEQ and EDI-2. Dual-energy X-ray absorptiometry was used to assess spine bone mineral density (BMD), and high-resolution pQCT to assess radius microarchitecture. We measured integrated cortisol (q 20', 11 PM-7 AM), bone formation (procollagen Type 1 N-terminal propeptide, P1NP), and resorption (C-telopeptide, CTX) markers in a subset.

RESULTS

OA had lower spine BMD Z-scores than EA. Cognitive eating restraint (CER), drive for thinness (DT), ineffectiveness, and interoceptive awareness (IA) were higher in OA than EA (p < 0.05); CER was higher in OA versus nonathletes (p = 0.03). Pulsatile cortisol was positively associated with DT, ineffectiveness, and IA (p < 0.03). CER was inversely associated with BMD Z-scores and P1NP, and ineffectiveness with radius cross-sectional area even after controlling for age, BMI, amenorrhea duration, and cortisol (p < 0.03).

DISCUSSION

Higher CER in athletes independently predicts lower BMD.

More than half of high school students report disordered eating: a cross sectional study among Norwegian boys and girls

Disordered eating and eating disorders are of great concern due to their associations with physical and mental health risks. Even if adolescence has been identified as the most vulnerable time for developing disordered eating, few studies have used a broad spectrum of criteria to investigate the prevalence of disordered eating among high school students of both genders, in different programs of study, nor assessed correlates of disordered eating among this important target group. The purposes of this study were therefore to investigate the prevalence and correlates of disordered eating among both male and female high school students in sport-, general and vocational programs. A comprehensive questionnaire was completed by 2,451 students (98.7%), aged 15–17 years. The total prevalence of disordered eating was 54.9%, with 64.3% among girls and 45.0% among boys (p<0.001). The highest prevalence of disordered eating was found among vocational students (60.7%), followed by students in general programs (49.8%) and sport students (38.3%) (p<0.001). Female gender, school program (vocational and general), overweight/obesity and weight regulation were positively associated with disordered eating. The high prevalence indicates the importance of tailored prevention efforts directed at high school students, particularly in vocational programs. Furthermore, a smaller girls–boys ratio than expected indicates that the efforts to identify and manage disordered eating among high school students should include both genders.

Psychometric Properties of the Eating Disorder Inventory in Clinical and Nonclinical Populations in Taiwan

Objective. To examine psychometric properties and investigate factor structures of the Mandarin Chinese version of the Eating Disorder Inventory (C-EDI). Method. The Mandarin C-EDI and other self-administered questionnaires were completed by a group of female eating disorder (ED) patients ( n = 551) and a group of female nursing students ( n = 751). Internal consistency, and convergent and discriminant validities were evaluated. Exploratory and confirmatory factor analyses were conducted to examine the construct validity of the Mandarin C-EDI. Results. The Mandarin C-EDI had good internal consistency and convergent and discriminant validities. With a few exceptions, the original clinically derived eight EDI subscales were clearly identified and the factorial validity of the first-order eight-factor structure and the second-order two-factor structure showed an acceptable degree of fit to our empirical data in clinical patients. Discussion. The findings suggest that the Mandarin C-EDI is a valid tool for clinical use in Taiwan.

A longitudinal investigation of sports-related risk factors for disordered eating in aesthetic sports

Previous studies have indicated a higher risk of disordered eating in certain types of elite sports such as aesthetic sports (e.g., rhythmical gymnastics, figure skating). But even though some studies on risk factors for disordered eating in sports exist, most research on this topic is based on cross-sectional data with limitations on causal inferences. We examined sports-related risk factors for disordered eating in a 1-year longitudinal study with two assessment points. The participants were 65 adolescent athletes from aesthetic sports (mean age 14.0  ±.2.2 years) who completed measures of disordered eating, social pressure from the sports environment, sports-related body dissatisfaction, desire to be leaner to improve sports performance, and emotional distress resulting from missed exercise sessions. All variables were relatively stable in the mean. Individual changes in the desire to be leaner to improve sports performance were associated with individual changes in disordered eating. Furthermore, a cross-lagged partial correlation analysis showed that the desire to be leaner to improve sports performance was predictive of disordered eating and not vice versa. The results of our study indicate that athletes are more at risk for disordered eating if they believe it is possible to enhance their sports performance through weight regulation.

Reduced body size and shape-related symptoms in young adults born preterm with very low birth weight: Helsinki study of very low birth weight adults

OBJECTIVE

To test the hypothesis that being born prematurely with very low birth weight (VLBW) (birth weight<or=1500 g) is associated with subphenotypes of eating disorders (dissatisfaction with body shape and pursuit of thinness) in young adulthood.

STUDY DESIGN

Preterm-born VLBW subjects (n=163) and controls (n=189) born at term completed 3 subscales of Eating Disorder Inventory-2 questionnaire (EDI): Drive for Thinness, Body Dissatisfaction, and Bulimia. Data were analyzed with multiple linear regression adjusted for confounders.

RESULTS

Among both sexes, EDI total scores were lower in VLBW subjects than in controls. The fully adjusted difference was -11.0% (95% CI, -18.4%, -2.2%) for women and -11.2% (95% CI, -20.2%, -1.3%) for men. Among women the lower scores in VLBW adults were observed in each EDI subscale. Results were similar when adjusted also for fat percentage, measured by dual-energy x-ray absorptiometry. Of the covariates, higher body mass index and higher score in Beck Depression Inventory contributed significantly to a higher EDI total score.

CONCLUSIONS

Young adults, particularly women, born prematurely with VLBW have fewer body size and shape-related symptoms and possibly lower risk for eating disorders than their term-born peers.

Psychometrische Gütekriterien des Eating Disorder Inventory (EDI-2) bei Jugendlichen

Ziel der vorliegenden Studie war die Überprüfung der psychometrischen Gütekriterien des deutschsprachigen Eating Disorder Inventory-2 (EDI-2) von Thiel et al. (1997 ) bei Jugendlichen. Es wurden die innere Konsistenz sowie konvergente und diskriminante Validität des EDI-2 an 371 Jugendlichen im Alter zwischen 13 und 18 Jahren überprüft und eine konfirmative Faktorenanalyse durchgeführt. Die Werte der inneren Konsistenzen sind für die Patientengruppe als hoch, für die beiden Kontrollgruppen als noch befriedigend einzustufen. Die Zusammenhänge mit anderen klinischen Messinstrumenten wiesen in die erwartete Richtung und stützen die externe Validität des EDI-2. Mittelwertsvergleiche der einzelnen Skalen des EDI-2 zwischen essgestörten Patientinnen (n = 71), weiblichen (n = 150) und männlichen Kontrollprobanden (n = 150) zeigten, dass der EDI-2 gut zwischen den verschiedenen Gruppen Jugendlicher differenzieren kann. Eine Diskriminanzanalyse ergab, dass insgesamt 86.0 % der Fälle korrekt klassifiziert werden konnten. Die konfirmative Faktorenanalyse bestätigte im Wesentlichen die Sechs-Faktorenstrukur des EDI-2 ( Thiel et al., 1997 ).

A cross-cultural comparison of the Eating Disorder Inventory

OBJECTIVE

A systematic analysis of data collected with the Eating Disorder Inventory (EDI) was made.

METHOD

A literature search identified 94 sources in which the mean values of the EDI or EDI-2 subscales were reported, comprising 310 samples differing by sex, age, diagnosis, language, ethnicity, or some other relevant attribute. The total number of respondents was 43,722, from 25 different countries, having used the EDI in one of the 16 languages (1-94).

RESULTS

The factorial structure of the aggregate means of the EDI subscales, for both clinical versus nonclinical and Western versus non-Western samples, was almost identical suggesting generalizability across languages and cultures. Non-Western participants scored higher than Western participants on virtually all EDI subscales, both in normal and eating-disordered samples. It was shown that age is a risk factor when someone is already diagnosed with an eating disorder but, in the general population, increasing age reduces the likelihood of being afflicted by eating disorders.

DISCUSSION

Symptoms of eating disorders are more pronounced in non-Western than in Western samples.

Readiness for change and short-term outcomes of female adolescents in residential treatment for anorexia nervosa

OBJECTIVE

To determine if readiness for change (RFC) at admission predicted length of stay (LOS) and short-term outcomes among female adolescents in residential treatment for anorexia nervosa (AN).

METHOD

Using a prospective cohort design to collect data from participants (N = 65) at admission and discharge, Kaplan-Meier survival analysis and Cox regression tested whether RFC on admission predicted time in LOS to a favorable short-term outcome--a composite endpoint based on minimum criteria for weight gain, drive for thinness, depression, anxiety, and health-related quality of life (HRQOL).

RESULTS

Participants with low RFC had a mean survival time to a favorable short-term outcome of 59.4 days compared to 34.1 days for those with high RFC (log rank = 8.44, df = 1, p = .003). The probability of a favorable short-term outcome was 5.30 times greater for participants with high RFC.

CONCLUSION

Readiness for change is a useful predictor of a favorable short-term outcome and should be considered in the assessment profile of patients with AN.

Psychological traits and platelet monoamine oxidase activity in eating disorder patients: their relationship and stability

Self-reported behavior and attitudes towards eating [Eating Disorder Inventory-2; Garner DM (1991). Eating Disorder Inventory-2: Professional Manual. Odessa, Fl.: Psychological Assessment Resources; Estonian version Podar I, Hannus A, Allik J (1999). Personality and Affectivity Characteristics Associated With Eating Disorders: a Comparison of Eating Disordered, Weight-Preoccupied, and Normal Samples. J Pers Assess; 73(1), 133-147] and the activity of platelet monoamine oxidase (MAO) was studied in 11 patients with anorexia nervosa (AN), 43 patients with bulimia nervosa (BN) and a healthy control group (n=138). Nineteen patients filled in the EDI-2 questionnaire and donated blood samples three times with three month intervals in order to determine platelet MAO activity. Eating disordered (ED) patients scored higher on all EDI-2 subscales and had lower MAO activity compared to the control group. They also scored higher than the control group on the Neuroticism domain but lower on the Extraversion, Openness, and Conscientiousness domains of the NEO-PI-R questionnaire. The average stability of MAO on different occasions (r=.56) was slightly smaller than the stability of the EDI-2 scores (r=.70). The lack of correlations between personality dispositions and MAO activity indicates that they have independent influence on eating disorders. A possible relationship between neurochemical mechanisms and psychological symptoms of eating disordered behavior is discussed.

Peer influence on bulimic symptoms in college students

This longitudinal study examined the influence of peer selection and socialization on bulimic symptoms in college students. Ninety-eight participants completed measures of bulimic symptoms, self-esteem, perfectionism, and impulsivity in the spring and fall of 2003. Peer influence was assessed by examining similarity among selected peers, unselected peers, and nonpeers over time. Among selected peers, bulimic symptoms demonstrated patterns of socialization, self-esteem and perfectionism demonstrated patterns of selection, and these personality traits longitudinally predicted changes in bulimic symptoms. Unselected peers demonstrated no similarity for bulimia, self-esteem, or perfectionism, but they did evidence socialization for impulsivity. The findings support an etiological model that integrates social and individual risk factors in creating environments that influence disordered eating among college students.

The Minnesota Eating Behavior Survey: a brief measure of disordered eating attitudes and behaviors

This article describes details of the development and psychometric characteristics of a brief self-report inventory for assessing attitudes and behaviors symptomatic of eating disorders that is currently in use in a longitudinal study of over 700 families with 11-year-old or 17-year-old twin girls. The Minnesota Eating Behavior Survey (MEBS), formerly the Minnesota Eating Disorder Inventory, is a 30-item measure developed for use with children as young as 10 years as well as adults. An examination of the MEBS's psychometric properties in a large, community sample of girls, women, and men demonstrated good factor congruence, internal consistency reliability, three-year stability, as well as evidence of concurrent and criterion validity. This questionnaire has promise as a screening and assessment measure for eating disturbance in cross-sectional and longitudinal research involving individuals of a wide range of ages.

Body dissatisfaction, self-esteem, and overweight among inner-city Hispanic children and adolescents

We examined the associations among self-reported body image, self-esteem, and measured body mass index (BMI) in El-Salvadoran American youth. Higher BMI was associated with body size dissatisfaction, lower peer esteem, and attempts to lose weight. Body size dissatisfaction was also significantly related to self-esteem in these El-Salvadoran American youth.

Prevalence of overweight among inner city Hispanic-American children and adolescents

OBJECTIVE

National surveys have pointed to a particularly high risk of pediatric overweight among U.S. Hispanics. However, the data have been primarily from the Mexican-American community. We studied the prevalence of overweight and clinical comorbidities in children and youth of predominantly El Salvadoran ancestry.

RESEARCH METHODS AND PROCEDURES

A sample of 309 Hispanic youth, 6-18 years was surveyed from two inner city Washington, DC, clinics. BMI; triceps skinfold (TSF) and subscapular skinfold thickness (SSSF); bioelectrical impedance analysis (BIA); and blood pressure measures were obtained, along with information regarding physical activity, sedentary behavior, dietary history, family, and personal medical history.

RESULTS

Thirty-eight percent were overweight (BMI > or = 95th percentile) and 22% at risk for overweight (BMI 85-94th percentile). Thirty-four percent had TSF > or = 90th percentile and 29% had SSSF > or = 90th percentile. Fifty-one percent of males and 70% of females had body fat > 30%. Compared to their nonoverweight counterparts, overweight youth had significantly higher systolic blood pressure (111.4 +/- 1.3 vs. 104.5 +/- 0.9 mm Hg, p < 0.0001). Among children younger than 11 years, overweight was associated with onset of adrenarche (23% vs. 10%, p = 0.01). Participation in one or more sports teams was negatively correlated with overweight) p = 0.04).

DISCUSSION

The prevalence of overweight and at risk for overweight in this sample was twice the national average for U.S. children and 1.7 times greater than that of Mexican-American children in national surveys. Overweight was associated with advanced pubertal development, high body fat, elevated blood pressure, and decreased sports participation.