Investigating Associations Between the Diagnostic Specifiers for Binge-Eating Disorder, Other Clinical Features, and the Presence of a High Body Mass Index: A Population-Based Study

Abstract

Binge eating is the cardinal feature of Binge-Eating Disorder (BED) and is known to be associated with obesity with bidirectional causality. This study aimed to investigate the association of diagnostic specifiers of binge eating, as mandated in the DSM-5 definition of BED, i.e., Criteria B and C (presence of marked distress) and weight/shape overvaluation with body mass index (BMI); and to examine the associations of high BMI with distress, pain, anxiety, and physical and mental health-related quality of life (PHRQoL and MHRQoL). Data for a sub-sample of 255 adults with recurrent binge eating (Criterion A) and without anorexia or bulimia nervosa diagnoses were extracted from the 2017 South Australian Health Omnibus Survey. Bivariate analyses were used to explore the association of DSM-5 BED specifiers with BMI and other measures. This study found that specific BED diagnostic specifiers related to binge eating were associated with higher BMI and distress levels. Additionally, individuals with BED who experience weight/shape overvaluation and higher BMI levels were associated with heigh levels of pain and anxiety, and had poorer PHRQoL and MHRQoL. These findings in relation to the association of higher BMI with different BED specifiers support the clinical significance of the DSM-5 Criterion B and C for BED.

1. Introduction

For over six decades [1], binge eating and Binge-Eating Disorder (BED) have been recognised as common health issues among individuals with higher body weights, as well as with associated medical and psychological problems, and this relationship is understood to be bi-directional [2,3,4,5,6,7,8,9]. BED is defined only in the context of the presence of binge-eating episodes with several specifiers in Criteria B and C of the fifth edition of the American Psychiatric Association Diagnostic and Statistical Manual of Mental Diseases (DSM-5; [10]). Whilst these specifiers—namely, rapid eating, uncomfortably full eating, non-hungry eating, eating in isolation due to embarrassment over the amount of food being consumed, feelings of disgust, guilt, or depression (Criterion B) and marked distress (Criterion C)—can be traced back to the earliest descriptors of binge-eating [7], and are mandatory [10], there has been very limited research into their validity [11]. In one study, the inclusion of Criterion C to the diagnosis of BED was found to be associated with higher eating disorders and depressive symptoms [12]. Klein, Forney, and Keel [13] found that only the Criterion B specifier ‘feeling disgusted with oneself, depressed or very guilty afterwards’ had clinical significance.

Although obesity (BMI ≥ 30) is associated with BED, it is not included as a diagnostic criterion for BED, and the DSM-5 indicates that BED is distinct from obesity [10]. BED is reported across BMI categories, but is most common in obese individuals (36.2–42.4%) [7,14]. Individuals with BED—often comorbid with higher body weight (HBW), e.g., with a high BMI [4,5,6]—show significant attentional biases toward food-related cues, which can impair cognitive control and eating behaviour [15]. Studies also indicate that higher frequencies of binge eating episodes in those with high BMI are associated with greater cognitive deficits, especially in executive functions such as inhibitory control and attention [16,17]. A study has shown that individuals with very high BMI experience cognitive impairments, with a 20% decrease in cognitive performance compared to those with normal BMI [18]. Other studies reported that mental health comorbidities, such as depression, anxiety, and disordered eating, are prevalent among individuals with higher BMI [5,19,20,21].

Despite the association of BED with high BMI, to our knowledge, none of the published studies have examined the association of BMI with the DSM-5 BED mandatory diagnostic specifiers (Criterion B and Criterion C), and other clinical features and relevant constructs in a representative general population setting. Hence, this study aims to investigate the association of DSM-5 BED mandatory diagnostic specifiers with BMI in a general population sample. Additionally, this study also explored the associations of high BMI with gender, age, pain, anxiety/depression, physical health-related quality of life (PHRQoL) and mental health-related quality of life (MHRQoL).

2. Methods

2.1. Study Design

Data for this study were drawn from the 2017 South Australian Health Omnibus Survey (HOS) to ensure adequate power for the study. The HOS was an annual face-to-face cross-sectional survey of the general population that had been conducted since 1991 [22]. The HOS was administered by the Population Research and Outcomes Studies Unit of the University of Adelaide, in collaboration with Harrisson Health Research, under the authority of the South Australian government, representing both governmental and external health organisations [23,24], which provided funding to include their own health-related questions [22]. The HOS was utilised to collect valid and reliable health data to support the formulation of health policies and programmes [24], and was used in this study because it includes questions on BED.

2.2. Sample and Procedure

Participants were selected from South Australia’s urban and rural areas located within Statistical Area Level 1 (SA1s) in 2017, using a strict, stratified random sampling technique based on the 2016 Australian Bureau of Statistics census data [24]. Within each SA1, a starting point was randomly selected and 10 dwellings were chosen using a pre-specified selection procedures based on a ‘skip pattern’ of every fourth household [24].

HOS participants had to be at least 15 years old as of their last birthday. Individuals with a terminal illness, mental disabilities, or poor English proficiency were excluded from the survey. Nursing homes, hospitals, and prisons were also excluded. Towns with a population of <1000 were not included in the randomisation algorithm. Between September and December 2017, trained interviewers conducted face-to-face interviews in each participant’s home, and only one person from each household was randomly selected to be interviewed. If the selected participant was not at home, a maximum of six separate visits were undertaken to attempt to contact them. Non-respondents were not replaced. Participation in the survey was voluntary, and participants could withdraw at any time.

A total of 5300 households were selected for the HOS, resulting in 2977 completed interviews, which reflects a participation rate of 65.3% [24]. The primary reasons for non-participation included refusal, inability to contact, non-English speaking individuals, vacant house, illness/incapacity, no access to property, aggressive dogs, and terminated interviews [22]. Among the 2997 interviewers, data from a sub-sample of 230 participants (weighted sample n = 255) aged 18 years or older (BMI available for 18 years and older) were included in this analysis. These participants were identified as having recurrent binge-eating disorder (Criterion A of BED in the DSM-5) and did not have a diagnosis of bulimia nervosa (BN) or anorexia nervosa (AN), and they provided data regarding their height and weight. Considering standard deviation 6.6 for BMI (mean 30.3, found in these 255 participants), the retrospective power analysis indicate that this study would require a sample size of 171 to estimate a mean with 95% confidence and a precision of 1.

2.3. Ethics Approval

Verbal informed consent was recorded from all adult participants, while written consent was obtained from the parents or guardians of adolescent participants aged 15–17 years old. The Human Research Ethics Committee of the University of Adelaide granted ethics approval for the HOS 2017 (approval number H097-2010) [24].

2.4. Outcome Measures and Assessment Tools

The HOS data were weighted by applying the gender, age, geographical area profile of the target population, as well as the probability of selection within households, to ensure that the survey findings were representative of the entire population [24]. The weighting process utilised the Estimated Residential Population data from the 2016 Australian Bureau of Statistics [22,24].

This study concentrated on the eating disorder questions (EDQ) within the HOS, which were derived from the Eating Disorder Examination (EDE) interview, a standardised tool used for evaluating the specific psychopathology associated with eating disorders [25]. This study encompassed the sociodemographic characteristics, current symptoms of eating disorder, the DSM-5 specifier questions for BED, including the presence of any distress, along with additional questions regarding pain and depression (see Appendix A). Each respondent was asked if and how often they had experienced episodes of binge eating in the last 3 months and whether these episodes coincided with any of the five behavioural indicators for binge eating behaviour (Criterion B) and distress (Criterion C). In DSM-5 diagnostic ‘Criterion C’— i.e., ‘distress’—distress refers to significant psychological discomfort or unhappiness experienced by an individual as a result of specific symptoms or behaviours that might interfere with daily functioning [10]. The responses to the distress question (Is this smaller overeating you experience usually associated with distress?) were reported on a 3-point Likert scale (not at all; yes—a little; and yes—a lot). Additional questions were also asked to assess features of anorexia nervosa and bulimia nervosa such as body weight and occurrence of compensatory vomiting. Questions regarding pain, anxiety and depression were also assessed using a 5-point Likert scale (absent, slight, moderate, severe, and extreme). The overvaluation of one’s body weight and shape is a diagnostic criterion in eating disorders and it considered a central element in the development of BED [26]. In this study, the overvaluation of weight and shape was assessed through the question: How important an issue has your weight and/or your shape been to how you think about (judge or view) yourself as a person in the past three months? This subscale is rated on a 7-point forced-choice format (0 to 6), with higher scores indicating greater severity; for analytic purposes, weight/shape overvaluation subscale scores ‘0–3’ are categorised as ‘no’, and ‘4–6’ are categorised as ‘yes’ (concern of weight/shape related severity). BMI for adults aged 18 years and older is computed by dividing a person’s weight in kilograms by the square of their height in metres.

The 12-item Short Form Health Survey (SF-12) was used in an interview format to evaluate MHRQoL and PHRQoL. The SF-12 is a well-established self-report questionnaire designed to assess how an individual’s mental and physical health status influences their overall quality of life [27]. It comprises 12 questions that evaluate physical and mental health across eight domains within two summary scales, the Physical Health Component Summary Scale (PCS) and the Mental Health Component Summary scale (MCS) [27]. The PCS domains include questions related to general health perception, physical health, limitations in physical roles, and physical pain, while the MCS encompasses questions about vitality, mental health, limitations in emotional role functioning, and social functioning. Respondents receive a score from 0 to 100 based on their answers; scores 50 or above indicate a quality of life that is better than average, and a score below 50 suggest a quality of life that is below average [27,28]. In this study PCS score was employed to evaluate PHRQoL, and the MCS score was used for MHRQoL. The SF-12 instrument has been validated globally across a variety of chronic diseases and conditions [27,29,30], demonstrating strong psychometric properties, and has been validated for use within the Australian population.

2.5. Statistical Analysis

Initially, socio-demographic characteristics (gender, age, marital status, country of birth, employment and schooling status) were compiled for both the unweighted sample (n = 230) and weighted sample (n = 255). Subsequently, using weighted data, descriptive statistics for all outcomes and associated measures, including BED diagnostic specifiers, presence of any distress, weight/shape overvaluation, BMI, pain/discomfort and anxiety/depression, PHRQoL, and MHRQoL are estimated. Categorical variables were expressed as percentages, while the continuous variables were reported as means and standard deviations (SD). Bivariate analyses were performed to investigate the association of participants gender, age, and BED diagnostic specifiers with BMI. Results of bivariate analyses are reported as percentages and means; statistical tests of significance such as chi-square (χ²), t, and F tests were used to examine significant (a priori threshold set at p value < 0.05) differences across sub-groups. Lastly, correlation coefficients were estimated to explore the association of high BMI with pain, anxiety, PHRQoL, and MHRQoL scores. All the analyses were performed using the Statistical Package for Social Sciences (SPSS), version 28 [31].

3. Results

Socio-demographic characteristics, including gender, age, marital status, country of birth, employment, and highest qualification status, for both the unweighted (n = 230) and weighted sample, (n = 255) are presented in

Table 1. Socio-demographic characteristics based on unweighted (n = 230) and weighted sample (n = 255 ^#).

	Unweighted Sample		Weighted Sample		Comments About Weighted Sample
Socio-Demographic Characteristics	Number	Col%	Number	Col%
All	230	100.0	255	100.0	11% Over estimated
Gender
Male	105	45.7	142	55.6	Over estimated
Female	125	54.3	113	44.4
Age group (years)
18–29	31	13.5	54	21.4	Over estimated
30–44	78	33.9	87	34.1
45–64	75	32.6	85	33.4
65 and above	46	20.0	28	11.2	Underestimated
Mean age (St. deviation)	47.7 (16.1)		43.6 (15.6)
Marital status
Married	106	46.1	132	51.8	Over estimated
De Facto	27	11.7	34	13.3
Separated/divorced	36	15.7	21	8.3	Underestimated
Widowed	19	8.3	10	3.9	Underestimated
Never Married	42	18.3	58	22.8
Country of birth
Australia/New Zealand	168	73.0	186	73.1
Asia	23	10.0	31	12.2
Europe/North America	35	15.1	33	13.0
Other	4	1.8	4	1.7
Employment status
Working full-time	91	39.6	114	45.0	Over estimated
Working part-time	43	18.7	49	19.1
Home duties	14	6.1	17	6.6
Student	8	3.5	13	5.2
Unemployed/retired/other	60	26.1	47	18.6	Underestimated
Not working due to injury/disability	14	6.1	14	5.5
Schooling status
Left school after before or after age 15	68	29.6	72	28.4
Left school after 15 but still studying	9	3.9	18	7.0	Over estimated
Trade qualification/apprenticeship	35	15.2	45	17.7
Certificate/diploma	65	28.2	62	24.6
Bachelor’s degree or higher	53	23.0	57	22.3

^# The 2016 Australian Bureau of Statistics Estimated Residential Population data were used for the weighting process by incorporating various population characteristics (e.g., gender, age, area of residence). Notes: According to research on BED, more proportion of female than male diagnosed with BED (roughly 60% of individuals with BED being female and 40% male). Since all the associations are related to BED specifiers measured by weighted sample, caution must be maintained in the interpretation of the findings specific to gender or age.

. The percentage of females in the weighted sample (55.6%) was almost 10% higher than in unweighted sample (45.7%). Additionally, the average age of participants varied between the unweighted sample (47.7 years) and weighted sample (43.6 years). The proportion of individuals aged 65 years or older was notably lower in the weighted sample (20.0% vs. 11.2%). Furthermore, the percentage of the population 65-years-old or older in the weighted sample was significantly lower than South Australian general population (11.2% vs. 22%). More than half of the participants were reported to be married or in a de facto relationship; nearly three quarters (73%) were born in Australia or New Zealand; 45% were employed full-time at the time of the survey; and one in five had completed a bachelor’s degree or higher.

3.1. Findings in Relation to BED Diagnostic Specifiers and Other Clinical Features

Table 2. Distribution of participants by BED diagnostic specifiers and mental health measures (weighted sample n = 255).

BED Diagnostic Specifiers and Other Mental Health Measures	Total Participants	Reported (No or Yes)
		No		Yes
		Number	Row %	Number	Row %
Categorical measures (reported as n, %)
BED diagnostic specifiers (reported multiple responses)
B1. Eating much more rapidly than normal (no, yes)	255	214	84.2	40	15.8
B2. Eating until feeling uncomfortably full (no, yes)	255	129	50.5	126	49.5
B3. Eating large amounts of food when not feeling hungry (no, yes)	255	162	63.6	93	36.4
B4. Eating alone embarrassed about how much are eating (no, yes)	255	248	97.5	6	2.5
B5. Feeling disgusted, guilty or very depressed after eating (no, yes)	255	222	87.1	33	12.9
Any types of distress associated BE (no = not at all, yes = a little or a lot)	253	152	59.6	101	39.7
Weight/shape overvaluation scale (no = 0–3 score; yes = 4–6 score)	255	117	46.1	137	53.9
Number of BED specifiers reported by individuals				Number	Col%
None of the BET of B1 to B5 reported				59	23.0
One BED specifier reported				121	47.4
Two BED specifiers reported				58	22.8
Three specifiers reported				9	3.6
4 to 5 specifiers reported				8	3.2
Total cases				255	100
Continuous measures [presented as mean, standard deviation]	Number	Mean (SD)	Range (min, max)	Skewness	Kurtosis
Body mass index (BMI)	255	30.3 (6.6)	19.0, 57.9	0.9	0.9
Pain/Discomfort (1 to 5)	255	1.8 (0.9)	1, 5	1.2	1.1
Anxiety/Depression (1 to 5)	255	1.4 (0.7)	1, 5	1.8	3.4
MHRQoL [based on MCS in SF-12 scale score 0 to 100]	255	49.7 (9.7)	15.8, 64.8	−1.2	0.6
PHRQoL [based on PCS in SF-12 scale score 0 to 100]	255	47.9 (10.4)	14.6, 65.1	−1.1	0.6

Notes: SD: standard deviation. Pain/discomfort measured as 5-point Likert scale: (1) I have no pain or discomfort, (2) I have slight pain or discomfort, (3) I have moderate pain or discomfort, (4) I have severe pain or discomfort, (5) I have extreme pain or discomfort. Anxiety/depression measured as 5-point Likert scale: (1) I am not anxious or depressed, (2) I am slightly anxious or depressed, (3) I am moderately anxious or depressed, (4) I am severely anxious or depressed, (5) I am extremely anxious or depressed. MHRQoL: Mental health-related quality of life; MCS: Mental component summary of SF-12 using Australian code. PHRQoL: Physical health-related quality of life; PCS: Physical component summary of SF-12 using Australian code. Normality: Data are considered to be normal if skewness is in between −2 to +2 and kurtosis between −7 to +7 [30]. The values of skewness and kurtosis for continuous measures for this study indicates that MHRQoL, PHRQoL, Pain, and Depression scores satisfy the normality criteria.

(weighted sample, n = 255) illustrates the distribution of participants across the five primary diagnostic specifiers of Criterion B (specifiers B1 to B5) and Criterion C (indicating any distress), as well as weight/shape overvaluation and various health metrics, including the average scores for BMI, pain/discomfort, anxiety, depression, PHRQoL, and MHRQoL. Among the five BED diagnostic specifiers, the highest percentage, 49.5%, responded ‘yes’ for B2 (eating until feeling uncomfortably full), followed by 36.3% for B3 (consuming large quantities of food when not hungry), 15.8% for B1 (eating significantly faster than usual), 12.9% for B5 (experiencing feelings of disgust, guilt, or severe depression after eating), and the lowest at 2.5% for B4 (eating alone due to embarrassment about the amount consumed). Approximately 23% of the participants (59 out of 255) did not respond positively to any of the five diagnostic specifiers, while 47.4% indicated a positive response to one specifier, and the remaining 29.6% responded positively to two or more specifiers. Among the participants, 39.7% reported experiencing any form of distress, and 53.9% were classified under the weight/shape overvaluation category. The average scores for other health outcome measures were as follows: BMI 30.3 (SD, 6.6), pain/discomfort 1.8 (SD, 0.9), anxiety/depression 1.4 (SD, 0.7), MHRQoL 49.7 (SD, 9.7), and PHRQoL 47.9 (SD, 10.4; Table 2).

3.2. Association of Higher BMI with Age, Pain, Anxiety/Depression, MHRQoL and PHRQoL

Results presented in

Table 3. Association of gender, age, and BED specifiers, presence of any distress (no = not at all, yes = a little or a lot) and weight/shape overvaluation (score 0–3 = no, score 4–6 = yes) with body mass index (BMI) (weighted sample n = 255).

Gender, Age and BED Diagnostic Specifiers, Distress, Weight/Shape Overvaluation	Body Mass Index (BMI)
	Mean	SD
All participants	30.3	6.6
Gender
Male	29.5	5.6
Female	31.3	7.5
p-values from t-test	0.036
Age groups (years)
15–29	27.8	5.9
30–44	30.2	6.3
45 and above	31.5	6.8
p-values from F-test	0.003
B1. Eating much more rapidly than normal
No	30.7	6.7
Yes	28.0	5.7
p-values from t-test	0.009
B2. Eating until feeling uncomfortably full
No	30.1	6.4
Yes	30.5	6.8
p-values from t-test	0.673
B3. Eating large amounts of food when not feeling hungry
No	29.0	5.9
Yes	32.5	7.1
p-values from t-test	0.001
B4. Eating alone because embarrassed how much you are eating
No	30.1	6.6
Yes	35.5	5.3
p-values from t-test	0.047
B5. Feeling disgusted, guilty or very depressed after eating
No	29.9	6.6
Yes	32.8	6.3
p-values from t-test	0.019
Number of BED specifiers (B1–B5) reported by individuals
None of the BED specifiers of B1 to B5 reported	29.7	5.8
One BED specifier reported	29.4	6.4
Two BED specifiers reported	31.7	7.4
Three to Five specifiers reported	33.4	6.3
p-values rom F test	0.030
Presence of any distress with binge-eating
No	29.3	6.2
Yes	31.6	6.9
p-values from t-test	0.008
Weight/shape overvaluation
No	29.5	6.5
Yes	31.0	6.6
p-values from t-test	0.067

Notes: p-values for comparison of two groups ‘no’ vs. ‘yes’ are calculated based on t-test; and F-test was used for comparisons of more than two groups.

shows that female and older aged participants are more likely to have significantly higher BMI as compared to others. Individuals who responded ‘yes’ for diagnostic specifiers B3 (eating large amount of food when not feeling hungry), B4 (eating alone embarrassed about how much eating), B5 (feeling disgusted, guilty or very depressed after eating) and those reported having presence of any distress have significantly (p = 0.03) higher BMI than others. Conversely, those who responded ‘yes’ to diagnostic specifier B1 (eating rapidly) have significantly (p = 0.009) lower BMI as compared to others.

The analysis of the correlation coefficient (r) between BMI and various outcome measures revealed that higher BMI is significantly associated with increased pain (r = 0.24, p = 0.001) and anxiety score (r = 0.14, p = 0.040). Furthermore, the results also demonstrate that a higher BMI is significantly correlated with poorer PHRQoL (r = −0.45, p = 0.001). However, a higher BMI did not show a significant association with a poorer MHRQoL (r = −0.114, p = 0.105).

4. Discussion

This study aimed investigate any associations between DSM-5 BED mandatory diagnostic specifiers (B1–B5) and other eating disorder diagnostic features, weight/shape overvaluation and BMI. The mean BMI was found to be significantly higher for those who responded ‘yes’ to the diagnostic specifiers B3 (eating large amount of food when not feeling hungry), B4 (eating alone embarrassed about how much eating), B5 (feeling disgusted, guilty or very depressed after eating) and Criterion C (distress); and mean BMI was found to be lower among those who responded ‘yes’ for diagnostic specifier B1 (eating rapidly) as compared to others. This study also found that a higher BMI was significantly associated with female gender, older age, increased pain and anxiety scores, and with a poorer PHRQoL.

The findings support the value of the majority of DSM-5 binge eating specifiers and other clinical features in informing the relationship between BED and high weight. The exceptions were ‘eating much more rapidly than usual’, which was associated with a lower body weight, and MHRQoL and weight/shape overvaluation associations with BMI, which did not reach significance. The first exception is difficult to explain as eating more rapidly is generally considered a feature since the earliest descriptions of BED [11] and eating slowly a feature found in people with restrictive eating disorders [32]. These latter findings may be a false positive due to low power and require replication. Further investigation of the data revealed that the prevalence of ‘eating much more rapidly than usual’ among young adults aged 18–29 is markedly greater than that of individuals aged 30 and above (27.3% compared to 13.0%; p = 0.011). Moreover, the average BMI for young adults in the 18–29 age group is also considerably lower than that of those aged 30 and older (mean BMI: 27.8 versus 31.0; p = 0.002). This study indicates that young adults exhibit a higher frequency of ‘eating much more rapidly than usual’ and possess a lower BMI compared to older age groups, which may contribute to the lower BMI observed in those who reported. There may also be differences in how speed of eating is perceived for younger people of higher weight, and objective observation of eating speed would be a more accurate indicator in investigating this Criterion B feature. Further research and research with larger sample sizes is necessary to investigate this matter.

Furthermore, the more BED specifiers that were present, the higher participant’s weight was, and, as expected, the higher their weight was, the higher their pain and anxiety/depression would be, and the poorer their PHRQoL would be. This supports other studies which have reported high levels of mental and physical health comorbidity in people with obesity complicated by BED [4,5,33].

4.1. Strengths and Limitations

Study strengths are the use of a representative population minimising the bias of self-selected or clinical samples, and employing a diagnostic interview with trained and supervised interviewers. The assessment of anxiety, depression and pain were limited to subjective single item self-reports which, although pragmatic in a largescale survey, is a major limitation and cannot be equated with a diagnostic interview determining the presence of a mood or anxiety disorder. Low statistical power due to small sample sizes limited some analyses (in particular that of the correlation between MHRQoL and BMI) and only a small proportion (7%, n = 17) of subjects fulfilled the three or more DSM-5 BED diagnostic specifiers. Finally, as this study was based on a single wave of cross-sectional survey data, causal inferences could not be made.

4.2. Clinical Implications

Clinicians should investigate the diagnostic specifiers for BED rather than solely rely only on self-reported binge-eating episodes. This is not only to verify the diagnosis, but also to have an understanding of associated features for clinical attention, e.g., address facilitate exposure to eating with others, in order to minimise the adverse effects from co-morbid high body weight. The findings underscore the importance of assessing both behavioural and affective dimensions of binge eating in implementing therapy. For example, in Fairburn’s [34] ‘transdiagnostic treatment model plan’, the therapist and patient collaborate to identify the patient’s goals, pinpoint cognitive and emotional processes that contribute to the patient’s distress, and then develop strategies for managing these processes.

5. Conclusions and Recommendations for Future Research

The present study supports an association with most diagnostic specifiers of BED and high body weight and some of the latter’s associated mental and physical health impacts. Future research should aim to replicate these findings in other larger community representative samples and in clinical populations. Clinicians should assess these dimensional features of binge-eating to inform their treatment.