Assessment of food addiction and its contribution to eating disorders and body mass index in the general population

Abstract

Background

The modified Yale Food Addiction Scale (mYFAS) was developed to quantify food addiction (FA) symptoms and their level of severity. This study aims to study FA in Israel by validating the Hebrew version of the mYFAS, assess FA prevalence, and test its contribution to eating disorder symptoms and obesity in an Israeli adult sample.

Methods

The Hebrew mYFAS (mYFAS-HEB) was translated and back-checked for accuracy. For validation, we used eating disorder, eating behavior, and depressive symptom questionnaires. We collected data regarding participants’ demographics, body mass index (BMI), and dietary consumption. Reliability was tested via a test–retest method. Confirmatory factor analysis (CFA), internal reliability assessments, and correlational analyses were also conducted, and hierarchical regression models were used to test the unique contribution of FA symptoms to eating disorders and BMI.

Results

Among the 364 participants, the prevalence of FA was 12%. FA symptoms were correlated significantly with all the other measures, particularly bingeing and uncontrolled eating. Reliability testing yielded a Cronbach's α of 0.88 and a Kuder-Richardson 20 coefficient of 0.81. CFA supported a two-factor structure, and standardized factor loadings confirmed the validity of the mYFAS-HEB. FA symptoms demonstrated a distinct and significant association with both eating disorder symptoms and BMI that was not explained by other measured variables.

Conclusions

Our findings suggest that food addiction (FA) is relatively prevalent in Israel, linked to eating disorder symptoms and higher BMI, and can be reliably assessed using the mYFAS-HEB.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40337-025-01346-3.

Plain language summary

Food addiction encompasses eating behaviors that resemble classical forms of addiction. Individuals with food addiction tend to have obesity and experience disordered eating patterns, but how does food addiction differ from obesity and eating disorders? Our study aimed to address this question using a survey that included a validated food addiction questionnaire (the modified Yale Food Addiction Scale), along with eating disorder and eating behavior assessments, and self-reported BMI. Among 364 adults surveyed, 12% met the criteria for food addiction. Notably, our analysis revealed that food addiction was distinctly associated with both eating disorder symptoms and higher body weight, even after accounting for other factors such as emotional eating and uncontrolled eating. These associations may reflect unique characteristics of food addiction that are not fully captured by existing eating disorder diagnoses or obesity criteria. These findings add to the growing evidence that food addiction might be a separate condition requiring its own diagnostic classification and targeted treatment approaches. Our study also validated a Hebrew version of the food addiction scale, enabling further research in this population.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40337-025-01346-3.

Background

Over the past decade, the concept of food addiction (FA) has emerged as a potential contributor to the alarming rise in global obesity rates. In Western countries, where ultra-processed, hyper-palatable foods are readily accessible [1], excessive consumption is prevalent and may lead to dependence over time. Similar to drug and behavioral addictions, which are recognized in the Diagnostic and Statistical Manual of Mental Disorders 5th Edition (DSM-5), FA is described as a pattern of addictive-like eating behaviors, which may result in significant impairment and disruption of daily activities [2]. Core symptoms of FA include frequent and intense cravings for and excessive consumption of foods high in ultra-processed fat, sugar, and salt, often continuing past the point of satiety despite awareness of the negative long-term consequences [3, 4].

However, the validity of FA as a stand-alone clinical construct remains controversial. Critics argue that FA closely overlaps with other eating disorders, such as binge-eating disorder or bulimia nervosa, where compulsive eating is often observed as a form of dysregulated eating [5, 6], complicating its classification as a separate diagnosis. Among the reasons for criticism is that many symptoms of FA, such losing control of eating and recurring episodes of excessive food consumption, are captured by existing DSM-5 criteria for eating disorders [5]. Nevertheless, several studies have shown that FA exhibits unique behavioral and neurobiological patterns that align with those observed in addictive disorders, such as activation of the brain reward circuits and withdrawal-like symptoms, suggesting that it may represent a distinct phenomenon from either eating disorders or addictive disorders [7, 8] and may thus justify its classification as a separate clinical construct. Despite significant progress in defining FA and assessing its prevalence in the general population, it remains unclear how this construct can be effectively differentiated from obesity or other forms of dysregulated eating.

To quantify FA symptoms and associated clinical impairment, the Yale Food Addiction Scale (YFAS) was developed about a decade ago [9]. Its shorter version, the modified YFAS (mYFAS), has demonstrated excellent psychometric properties [10] and has been translated and validated in multiple languages, making it a widely used research tool in the field (see, for example, [11, 12]).

The prevalence of FA in Israel remains unknown, as few studies have used the YFAS to estimate FA among Israeli participants. One study investigated biobehavioral parameters differentiating individuals with obesity and FA, obesity without FA, and healthy control subjects [13]. That study used the original YFAS, which is based on the DSM-4 definition of addiction [9]. A more recent study employed the mYFAS to characterize FA in bariatric patients before surgery and at six and twelve months post-surgery [14]. However, the Hebrew version of the mYFAS has not yet been empirically validated, and the prevalence of FA in the general Israeli population is still undetermined.

Therefore, this study aims to (1) validate the Hebrew mYFAS, (2) estimate the prevalence of FA in Israel, (3) examine the independent association of FA on BMI and eating disorders, providing valuable data that may contribute to future discussions regarding the classification of FA as an independent construct.

Methods

Study design

This cross-sectional study utilized a non-probability sampling method to achieve a comprehensive representation of a variety of backgrounds. The research was advertised in targeted Facebook groups to recruit those from diverse backgrounds in terms of religious affiliation, socioeconomic level, marital status, and age. To be eligible for participation in the study, individuals had to be between the ages of 20 and 65 years old with a good command of both spoken and written Hebrew. Residency within Israel was also required. Individuals with medical conditions affecting food consumption, including liver, kidney, or pancreas disorders, diabetes, thyroid conditions, cancer, and swallowing or chewing disorders, were excluded. Additionally, individuals undergoing chemotherapy or who were taking medications that affect appetite were excluded. Pregnant participants or individuals with dietary restrictions, such as being on a weight loss diet, were not included.

Participants

Participants were recruited using social media advertisements and the snowball sampling approach, wherein existing participants referred potential candidates to the study. The recruitment process involved initial registration and screening with a short questionnaire to assess eligibility for the study. Eligible applicants were contacted and provided with an overview of the study’s objectives, procedures, and terms of participation, and their eligibility to participate was verified. Participants were assured that their data would be kept confidential, and they signed written informed consent forms. Participants were offered a gift voucher as a token of appreciation for their participation in the study. The study protocol was approved by the ethics committee of Tel-Hai Academic College in Israel.

The sample size, roughly 375, was determined using EPICALC for prevalence and G-Power for correlational analysis. An extra 10% was added to compensate for invalid or missing data.

Measures

FA was evaluated using the modified Yale Food Addiction Scale 2.0 (mYFAS 2.0) [10]. To validate the mYFAS 2.0, we used questionnaires that assess eating disorders, eating behaviors, and depressive symptoms. Several additional variables, including demographic characteristics, BMI, and dietary patterns, were also measured. Multiple validation parameters for FA were used to ensure a comprehensive evaluation of the performance of the mYFAS 2.0 and its ability to capture these conditions.

Sociodemographic data

Participants completed a detailed demographic questionnaire that included information on age, sex, weight, height, marital status, religious affiliation, income level, education, physical activity, and lifestyle characteristics.

Yale food addiction scale

As a shorter version of the original YFAS, the mYFAS 2.0 includes 13 items that capture FA symptoms (11 items) and the associated clinical distress (two items) over the last year. Participants provide responses on an 8-point Likert scale, ranging from 0 (“never”) to 7 (“every day”). Each statement has a different threshold that depends on the frequency of the symptoms. To assess FA symptoms, the scores for the 11 criteria are summed, yielding a range of 0–11 symptoms. To assess the clinical severity of FA, two items are scored, yielding four levels of severity. FA is classified as mild, moderate, or severe when clinical severity is present along with two to three, four to five, or six or more symptoms, respectively. FA is not considered present when clinical severity is absent, fewer than two symptoms are reported, or both conditions are not met [10].

To assess convergent validity, we used the Patient Health Questionnaire, Three-Factor Eating Questionnaire, and the Eating Disorders Examination Questionnaire.

Patient health questionnaire

To evaluate depressive symptoms, we used the Patient Health Questionnaire (PHQ-9) [15], which includes nine questions to assess the frequency of emotional states and symptoms experienced in the last two weeks. The score is calculated for each of the nine statements based on 0–3 levels (from 0 = “never” to 3 = “almost every day”). The scores are based on a scale ranging from 0 to 27, with five levels: the absence of depression (a total score of 0–5), sub-symptomatic or low-level depression (5–9), moderate depression (10–14), moderate to severe depression (15–19), and severe depression (20–27). The Cronbach’s α coefficient found in determining the reliability and validity of the online version of the Hebrew questionnaire ranged from 0.79 to 0.83, which indicates a high level of internal consistency [16].

Three-factor eating questionnaire

The Three-Factor Eating Questionnaire (TFEQ-R-18) includes 18 questions that evaluate three subscales of eating behaviors: cognitive restraint, uncontrolled eating, and emotional eating. The raw scores are standardized on a scale of 0–100, and the mean score for each subscale is subsequently calculated. The Cronbach’s α of the TFEQ among the general adult population is 0.84, 0.83, and 0.87 for the three subscales, respectively [17].

Eating disorders examination questionnaire

The Eating Disorders Examination Questionnaire-13 (EDE-Q13) has 13 items that assess eating disorder symptoms but is not intended for clinical diagnosis. Each item has a 7-point scale (ranging from 0 = “not at all” to 6 = “every day”) based on the frequency of symptoms experienced in the last month. The scores are summed into five subscales: (1) restraint in eating, (2) body shape and weight overestimation, (3) body dissatisfaction, (4) compulsive eating, and (5) body purging. The Cronbach’s α coefficient found in the subscales of this questionnaire is 0.92 for eating restraint, 0.99 for shape and weight overestimation, 0.89 for body dissatisfaction, 0.89 for bingeing, and 0.63 for purging, which indicates a high level of internal consistency [18]. Food addiction symptoms closely resemble those seen in binge-eating disorder, particularly the compulsive eating behaviors that characterize this formally recognized eating disorder. This similarity has been extensively discussed in the field [19, 20]. Therefore, the compulsive eating construct of the EDE-Q13 was utilized to assess the convergence validity of the Hebrew mYFAS (mYFAS-HEB).

Food frequency questionnaire

The electronic Food Frequency Questionnaire (FFQ), designed for the Israeli population and validated in Israel [21], includes questions about 115 food products, each with nine frequency choices, spanning from “seldom or fewer than once a month” to “six or more times per day.” Participants are asked to indicate their average consumption frequency over the last year. Serving sizes, which are based on standards determined by the Israeli Ministry of Health [22], accompany each food item on the questionnaire. Completing the questionnaire electronically through self-administration guarantees data integrity, as participants can only finish the questionnaire by completing all the items.

Study procedures

The mYFAS questionnaire was translated into Hebrew and subsequently retranslated into English. Procedural adaptations were conducted according to the CCA (Culture-Centered Approach) protocol [23] to tailor the questionnaire to the Israeli population. This process aimed to ensure conceptual and semantic relevance across questionnaire items and instructions. The questionnaire was submitted to two translators: a native English speaker and a native Hebrew speaker. They translated the questionnaire in consideration of Israeli culture. For instance, the term “bacon” was replaced with “beef” in the Hebrew version, since pork is not consumed in Israel. Following retranslation, a panel of experts assessed the consistency between the two translations and selected the final phrasing that best suited the context. The resulting mYFAS-HEB was used for further analysis and validation.

Participants were recruited between March and June 2022 and asked to complete the questionnaires online through the Qualtrics platform (Qualtrics, Provo, UT). Questionnaires that exhibited inaccuracies, such as those with missing information or exceptionally extreme responses (for example, daily caloric consumption exceeding 10,000), were omitted from the final analysis. Out of 387 respondents, the final sample consisted of 364 participants.

Retest reliability

To assess the reliability of the mYFAS-HEB, a retest was conducted in which 40 participants were asked to complete the questionnaires approximately one month after their initial response. The reliability and consistency of the mYFAS-HEB were assessed using Cronbach's α and Kuder-Richardson 20 (KR-20) indices, as well as a test–retest Spearman correlation. A correlation of 0.7 or greater was considered reliable [24].

Statistical analysis

Confirmatory factor analysis (CFA), internal reliability assessments, data distribution tests, and correlational analyses were also conducted to ensure data integrity. Descriptive statistics were generated by computing the means, standard deviations, and prevalence percentages of the variables analyzed. BMI was calculated from self-reported height and weight. Income level was divided into five classes ranging from “much below average” to “much above average” based on the average income in Israel in 2022 [25]. This categorization was then condensed into three groups (low, middle, high) based on the distribution. Cigarette smoking status was classified dichotomously (smokers/non-smokers), based on an initial division into five groups: non-smoker, 1–3, 4–20, 21–40, and more than 40 cigarettes per day. Caloric intake was determined based on the FFQ. Normally distributed variables were assessed with a T test; otherwise, the Mann–Whitney test was used. The Chi-square test was applied to dichotomous categorical variables. Spearman’s correlations were computed for continuous variables.

CFA was carried out using Jasp 0.17.3 (Amsterdam, The Netherlands). We created a two-factor model, with “Symptoms” as Factor 1 and “Clinical significance” as Factor 2 [10]. We report several goodness-of-fit parameters: Chi-square (χ²), Comparative Fit Index (CFI), Tucker-Lewis Index (TLI), and root mean square error of approximation (RMSEA). CFI and TLI values greater than or equal to 0.90 and RMSEA values less than or equal to 0.08 indicate a good fit of the model [26]. For the reliability test, we used Spearman’s correlation to calculate the correlation between the first and second tests.

Two sets of hierarchical regression models were constructed to examine the association of FA symptoms to the variance in eating disorder symptoms and BMI, separately. Each set included three models with the following sequence of predictor:

1. Model 1: FA symptoms

2. Model 2: FA symptoms and demographic variables (age, gender, and income level)

3. Model 3: FA symptoms, demographic variables, and clinical variables (depressive symptoms, emotional eating, uncontrolled eating, and eating disorder symptoms for set 1; BMI for set 2). The data were processed and analyzed using SPSS version 27.0 (IBM Corp., Armonk, NY).

In the hierarchical regression models, we did not apply corrections for multiple comparisons. This decision follows the rationale presented by Rothman [27], following the recommendations of Rothman [27], and aligns with best practices in theory-driven analyses, where the inclusion of predictors is pre-specified and conceptually grounded rather than determined post hoc [28, 29]. Applying correction procedures in such contexts may increase the risk of false negatives and obscure meaningful associations.

Results

Demographics and participants’ characteristics

The study included 364 participants, 55% of whom were women, and most of whom were young adults (Table 1). The participating women had a higher education level, which corresponds with the distribution of academic education in Israel [30]. There was a greater percentage of men who reported above-average income than women (p < 0.05), consistent with data from the State of Israel indicating that women earn approximately 33% less than men [31]. Additionally, men had a significantly greater BMI compared to women (p < 0.05), with almost 40% of men falling in the overweight category (p < 0.05).

Table 1.

Participants’ demographic and clinical characteristics

	Total	Men	Women	p value1	Effect size
	(n = 364)	(n = 162)	(n = 202)
Age in years, (M ± SD)	34.44 ± 11.92	32.7 ± 10.60	35.80 ± 12.80	0.068	0.192*
Marital status (n, % married)2	197 (54.1)	87 (53.7)	110 (54.5)	0.886	0.007#
Academic education (n, %)3	194 (53.3)	76 (46.9)	118 (58.4)	0.029	0.115#
Household income (n, %)4				0.009	0.162#
Low	228 (62.6)	89 (54.9)a	139 (68.8)b
Middle	84 (23.1)	41 (25.3)a	43 (21.3)a
High	52 (14.3)	32 (19.8)a	20 (9.9)b
BMI (kg/m2)					0.167#
Overall (M ± SD)	25.97 ± 5.43	26.29 ± 4.75	25.72 ± 5.91	0.042
BMI categories (n, %)				0.018
Underweight (< 18.5 kg/m2)	14 (3.8)	2 (1.2)a	12 (5.9)b
Normal weight (18.5–24.9 kg/m2)	167 (45.9)	69 (42.6)a	98 (48.5)a
Overweight (25–29.9 kg/m2)	117 (32.1)	63 (38.9)a	54 (26.7)b
Obese (≥ 30 kg/m2)	66 (18.1)	28 (17.3)a	38 (18.8)a
Smoking (= Yes, n, %)	62 (17.0)	32 (19.8)	30 (14.9)	0.216	0.065#
Energy (Kcal, M ± SD)	2,512 ± 1,583	2,652 ± 1,650	2,399 ± 1,523	0.129	0.160*

¹Mann–Whitney, Chi-squared test

²Marriage was defined as being in a permanent relationship according to the participant’s definition

³The total number of degree recipients from higher education institutions: 61.3% held a bachelor’s degree, 62.5% held a master’s degree, and 50.5% held a third degree [27]

⁴Income categories are defined relative to the average income in Israel (2,054$) [25]

Effect sizes: * indicate Cohen’s d; # indicate Phi or Cramer’s V

^a/_bDifferent letters in the same row represent distinct differences

Food addiction

The prevalence of FA was 12% (n = 43), with severe FA being the most common, occurring 1.5 times more frequently than mild or moderate FA (Fig. 1a). The average symptom count (1.35) reflects a relatively low mean FA score (Fig. 1b). However, over 30% of participants exhibited at least two symptoms of FA. Among those identified with FA (n = 43), 72% were women (n = 31). For both mild and moderate/severe FA, a higher percentage of women were affected compared to men (5% versus 2% and 10.4% versus 6.2%, respectively).

Fig. 1.

Prevalence of food addiction by severity and symptom range (n = 364)

Construct validity of the mYFAS-HEB

CFA was conducted on the two-factor structure of the mYFAS-HEB for the entire sample (n = 364). The Chi-square value was 153.233 (p < 0.001, degrees of freedom [df] = 64), which is expected with a sample size over 150 [32]. The addition of residual covariance between items 1 and 2 of the mYFAS-HEB significantly improved other goodness-of-fit parameters (Table 2), resulting in χ² = 100.327 (p = 0.002), CFI = 0.996, TLI = 0.995, and RMSEA = 0.04 [0.025–0.055] (p = 0.86). Standardized factor loadings are presented in Table 3.

Table 2.

Goodness-of-fit measures of the mYFAS-HEB

Index	Value prior to modification	Value following modification
Comparative Fit Index (CFI)	0.989	0.996
Tucker-Lewis Index (TLI)	0.987	0.995
Root mean square error of approximation (RMSEA)	0.062	0.04

Table 3.

The mYFAS-HEB questions and factor loadings

Factor	Question no. in the mYFAS	Question in the mYFAS	Factor loading in the mYFAS-HEB*
Symptoms	mYFAS_1	I ate to the point where I felt physically ill	0.537
	mYFAS_2	I spent a lot of time feeling sluggish or tired from overeating	0.512
	mYFAS_3	I avoided work, school, or social activities because I was afraid I would overeat there	0.536
	mYFAS_4	If I had emotional problems because I hadn’t eaten certain foods, I would eat those foods to feel better	0.610
	mYFAS_7	My eating behavior caused me a lot of distress	0.793
	mYFAS_8	I had significant problems in my life because of food and eating. These may have been problems with my daily routine, work, school, friends, family, or health	0.865
	mYFAS_9	My overeating got in the way of me taking care of my family or doing household chores	0.668
	mYFAS_10	I kept eating in the same way even though my eating caused emotional problems	0.684
	mYFAS_11	Eating the same amount of food did not give me as much enjoyment as it used to	0.636
	mYFAS_12	I had such strong urges to eat certain foods that I couldn’t think of anything else	0.461
	mYFAS_13	I tried and failed to cut down on or stop eating certain foods	0.633
Clinical significance	mYFAS_5	I was so distracted by eating that I could have been hurt (e.g., when driving a car, crossing the street, operating machinery)	0.885
	mYFAS_6	My friends or family were worried about how much I overate	0.830

^*p < 0.001

Convergent validity with psychobehavioral measures

We used various psychobehavioral measures to evaluate the convergent validity of the mYFAS-HEB (Supplementary Table S1). The average scores for this nonclinical sample were within the normal range across all measures, with the exception of moderate to severe depression, which had a prevalence of 26.6%. This rate is markedly higher than the estimated prevalence of approximately 5–8% in the general population [33, 34].

The mYFAS-HEB symptoms showed significant correlations with all psychobehavioral measures, except for cognitive restraint (Table 4). The strongest associations were observed with binge-eating and uncontrolled eating, and weak to moderate correlations were found with depressive symptoms, emotional eating, BMI, and total energy intake.

Table 4.

Convergent validity of mYFAS-HEB symptoms

Variable	mYFAS-HEB symptoms	Cognitive restraint	Uncontrolled eating	Emotional eating	Bingeing	Depression	BMI
Cognitive restraint	0.085	–
Uncontrolled eating	0.465a	− 0.060	–
Emotional eating	0.355a	− 0.039	0.479a	–
Bingeing	0.485a	0.092	0.569a	0.432a	–
Depression	0.387a	0.065	0.291a	0.370a	0.447a	–
BMI	0.319a	0.121b	0.312a	0.275a	0.329a	0.141a	–
Total energy (Kcals)	0.207a	− 0.091	0.245a	0.056	0.219a	0.031	0.053

^ap < 0.05, ^bp < 0.01 by Spearman test

Test–retest

Cronbach’s α and KR-20 values of 0.88 and 0.81, respectively, indicated high reliability and internal consistency [35, 36]. Additionally, there was a strong correlation between the level of mYFAS-HEB symptoms in the first and second tests (r = 0.74, p < 0.001), as well as between the mYFAS-HEB FA classifications (r = 0.67, p < 0.001) [24].

Food addiction’s impact on eating disorder symptoms and BMI

Table 5 details the results of the hierarchical regression model explaining the variance in eating disorder symptoms within the sample. The findings indicate that FA was significantly associated with eating disorders symptoms (β = 0.520, p < 0.001), even after accounting for demographic (age, gender, income level) and clinical variables (BMI, depressive symptoms, emotional eating, and uncontrolled eating; β = 0.316, p < 0.001). FA symptoms alone were statistically associated with 27% of the variance in eating disorder symptoms, while the full model, including demographic and clinical variables, accounted for 36%.

Table 5.

The independent contribution of food addiction to eating disorder variance

Variables	Model 1		Model 2		Model 3
	β	p value	β	p value	β	p value
Food addiction	0.520	< 0.001	0.511	< 0.001	0.316	< 0.001
Age			0.060	0.214	0.016	0.742
Gender			− 0.083	0.073	− 0.076	0.098
Income level			− 0.006	0.909	− 0.038	0.418
BMI					0.185	< 0.001
Depression					0.203	< 0.001
Emotional eating					0.025	0.634
Uncontrolled eating					0.065	0.219
R2	0.270		0.282		0.355
F(df)	133.910 (1,362)		35.221(4,359)		24.448(8,355)

Model 1 includes food addiction

Model 2 includes food addiction and demographic variables (Age, Gender, Income level)

Model 3 includes model 1, model 2, and clinical covariables (BMI, Depressive symptoms, Emotional eating, and Uncontrolled eating)

BMI (kg/m2); Age (years); Gender (men vs. women); Income level (average, above average, or below average); Food addiction (symptoms); Emotional eating (subscale score); Uncontrolled eating (subscale score)

Similarly, Table 6 presents the results of the hierarchical regression model explaining the variance in BMI. In this model, FA symptoms alone were significantly associated with in BMI (β = 0.343, p < 0.001), and this association remained significant after including demographic and clinical variables (β = 0.170, p = 0.004). FA symptoms alone accounted for 12% of the variance in BMI, while the full model, including demographic and clinical variables, accounted for 28%.

Table 6.

The independent contribution of food addiction to BMI variance

Variables	Model 1		Model 2		Model 3
	β	p value	β	p value	β	p value
Food addiction	0.343	< 0.001	0.348	< 0.001	0.170	0.004
Age			0.261	< 0.001	0.267	< 0.001
Gender			0.114	0.021	0.140	0.004
Income level			− 0.021	0.679	− 0.030	0.545
Eating disorders					0.207	< 0.001
Depression					− 0.064	0.252
Emotional eating					0.177	0.002
Uncontrolled eating					0.088	0.114
R2	0.118		0.196		0.276
F(df)	48.331(1,362)		21.898(4,359)		16.880(8,355)

Model 1 includes food addiction

Model 2 includes food addiction and demographic variables (Age, Gender, Income level)

Model 3 includes model 1, model 2, and clinical covariables (Eating disorders, Depressive symptoms, Emotional eating, and Uncontrolled eating)

BMI (kg/m2); Age (years); Gender (men vs. women); Income level (average, above average, or below average); Food addiction (symptoms); Emotional eating (subscale score); Uncontrolled eating (subscale score)

Physical activity was examined as an additional covariate in the regression models predicting BMI and eating disorder symptoms. Including Physical activity did not change the associations observed between food addiction and either outcome (Supplementary Tables S2 and S3).

Conclusions

In this study, we present the adaptation and validation of the mYFAS for use in the Israeli population. We assessed the prevalence of FA in Israel, and we examined the contribution of FA to related clinical constructs.

Validity and reliability of the mYFAS-HEB

The CFA of the adapted questionnaire demonstrated adequate goodness-of-fit parameters for the two-factor structure of the mYFAS-HEB. Convergence validity against several other psychobehavioral constructs demonstrated positive relationships between FA symptoms and binge-eating, uncontrolled eating, emotional eating, and depression, which is consistent with previous findings [13, 37]. As expected, FA was also significantly related to BMI, which aligns with the positive relationships also found between FA symptoms and caloric intake from food.

The mYFAS-HEB demonstrated discriminant validity against cognitive restraint, which aligns with previous research in this field [10, 38]. Lastly, test–retest results are adequate and suggest that the mYFAS-HEB is a reliable and consistent measure of FA. Taken together, the findings align with previous research [39] and confirm that the mYFAS-HEB is a valid tool for measuring FA.

The prevalence of food addiction

The prevalence of FA observed in the present study was relatively higher than that reported in other countries. For instance, in the United States the prevalence of FA is approximately 15% [40], whereas in European countries it ranges from 5.7 to 9%, as seen in Denmark [41], Germany [42], Italy [43], and France [44]. In Asian countries such as Japan [45] and China [12], the reported prevalence is even lower, at 3.3% and 6.7%, respectively. In Middle Eastern countries like Egypt [46] and Turkey [47], the prevalence rates are similar to those found in Israel (11.0% and 11.4%, respectively). This prevalence (12%) may reflect a combination of cultural, environmental and methodological influences. Israel's unique dietary landscape, which blends traditional Mediterranean and Middle Eastern food patterns with increasing exposure to ultra-processed, Western-style foods, may contribute to a dual-risk environment. While the Mediterranean diet is typically associated with lower food craving prevalence [48], the growing availability of hyper-palatable convenience foods may elevate vulnerability to addictive-like eating behaviors [49]. Additionally, Middle Eastern cultural norms, such as communal eating, generous portion sizes and food-centric hospitality may promote overeating in socially normative contexts [50]. Methodological variation may also explain cross-national differences in reported FA rates. For instance, previous studies have used either the original YFAS or various versions of the mYFAS, which differ in diagnostic thresholds, inclusion of craving criteria, and alignment with DSM-5 standards [9]. Moreover, sample characteristics (clinical vs. general population), recruitment methods, and translation procedures can influence outcomes [39, 48–52].

Notably, in the current study 72% of individuals meeting the criteria for FA were women. The higher prevalence of FA among women may be associated with gender-related differences in emotional regulation, including greater psychological preoccupation with food and eating, a higher likelihood of using food as a coping mechanism for stress or negative affect, and increased tendency to report health-related concerns compared to men [53, 54]. In addition, fluctuations in estrogen and progesterone have been shown to influence craving, appetite regulation, and emotional eating, factors that may contribute to addictive-like eating patterns in women [53], while neurobiological studies suggest sex-based differences in the activation of brain regions involved in reward processing and impulse control [55]. Moreover, societal norms and cultural expectations surrounding thinness and dieting, more pervasive among women, may intensify psychological preoccupation with food and promote emotional regulation through eating, both of which are closely linked to FA symptoms [56].

Discrepancies in reported prevalence rates between research studies may also be attributed to differential FA assessment methods. For instance, the original YFAS was developed based on DSM-4 criteria for substance and behavioral addiction, while newer versions of the YFAS are aligned with DSM-5 criteria. Comparing prevalence rates obtained using different versions of the questionnaire may lead to biased conclusions, as the criteria assessed may differ. For example, the updated YFAS version includes the measurement of cravings and introduces a graded severity scale, but these are not evaluated in the original version. Therefore, research on global FA prevalence should utilize standardized assessment methods to ensure comparability across studies.

The contribution of FA to eating disorders and BMI

The concept of FA as a distinct clinical construct is highly controversial. While neurobiological research has shown similar brain activation patterns between obesity, FA and “classical” addictions [57], clinical evidence suggests that FA shares considerable overlap with binge-eating disorder, particularly in terms of compulsive food consumption and difficulties in regulating intake [58]. To address these issues, we aimed to investigate the unique contribution of FA to eating disorder symptoms, both as an independent construct and in combination with other factors known to be associated with the severity of eating disorders (e.g., age and gender [59], emotional eating [60], and uncontrolled eating [61]).

Our findings indicated that FA was distinctly associated with eating disorder symptoms, explaining 35% of the variance in a comprehensive model that accounted for BMI, depressive symptoms, emotional eating, and uncontrolled eating—factors that are strongly linked to eating disorders [62]. In this model, the effects of emotional and uncontrolled eating were no longer significant [63]. Similarly, we found that FA symptoms were associated with 28% of the variance in BMI in a model with other relevant variables such as age, gender, income level, eating disorder symptoms, depression, emotional eating, and uncontrolled eating.

These results suggest that FA may be one of several factors associated with the presence and persistence of disordered eating behaviors and increased BMI, beyond the influence of other psychological and demographic factors. This finding highlights FA’s potential relevance as a distinct research construct, as it demonstrates unique associations with both eating disorder symptoms and BMI that are not fully explained by overlapping constructs such as emotional and uncontrolled eating. Therefore, gaining a deeper understanding of FA symptoms in clinical settings could provide valuable insights into eating behavior patterns and their potential implications for intervention strategies. However, since FA is not a formal diagnosis but rather a framework for understanding addictive-like eating behaviors, further research is needed to explore its clinical significance, particularly in relation to long-term eating behaviors and treatment outcomes.

Although FA may be present in individuals with Binge Eating Disorder (BED), Bulimia Nervosa (BN), and even Anorexia Nervosa- particularly the binge-purge type- our study suggests that FA encompasses clinical features that are distinct from these eating disorders. For example, while BED and BN are characterized by binge eating episodes [5], FA may also be identified in individuals who do not binge-eat but rather engage in persistent food grazing [64]. All three conditions commonly involve a sense of loss of control over eating—typically involving hyperpalatable ultra-processed highly rewarding food—whether during binge episodes or grazing. Thus, while FA shares certain clinical features with more established eating disorders, our findings support the presence of unique elements specific to FA.

The present study has multiple strengths, as well as limitations. The study results are based on a large sample size, which captures a wide variability of demographic factors, strengthening the external validity of the findings. However, the present study is limited by its cross-sectional design and the use of non-probability sampling, which may introduce selection bias. Our sample overrepresent younger, internet-active individuals and underrepresent groups with limited digital access or different sociodemographic characteristics. As a result, the generalizability of the findings to the broader population may be limited. Particularly, studies on the prevalence and clinical characteristics of food addiction in older adults where food addictions patterns may differ significantly and stem from distinct underlying factors. Studies among older populations may shed light on this important gap in our study [65]. Although participants with known medical conditions affecting food intake were excluded, we did not assess undiagnosed or subclinical conditions, which may have influenced eating behaviors. Additionally, while income level was included as an indicator of socioeconomic status, more detailed SES measures (e.g., education, occupation, or food insecurity) were not collected and should be considered in future studies.

In summary, our findings demonstrate that the mYFAS-HEB is a valid and reliable tool for assessing FA among Israeli adults. The prevalence of FA in Israel is 12%, higher than in Asia and Europe. FA was found to be associated with binge-eating, uncontrolled eating, emotional eating, depressive symptoms, and higher BMI. The distinct associations between FA symptoms and both eating disorder symptoms and BMI observed in this study suggest that FA represent a unique behavioral pattern that shares similarities with binge-eating disorder but is not entirely captured by existing diagnoses. Our results lay the groundwork for further research on FA in Israel and contribute to global efforts aimed at understanding and ultimately treating this condition.

Although research on FA is still in the early stages and does not yet support definitive treatment guidelines, several potential strategies may be explored. Cognitive-behavioral therapy can address psychological and emotional triggers and has shown benefits in reducing compulsive eating and distress associated with FA symptoms. Pharmacological agents, such as GLP-1 receptor agonists, may help reduce cravings and enhance satiety by modulating central and peripheral appetite pathways [66]. Given the multifaceted nature of FA [67], combining behavioral and pharmacological approaches may offer a targeted approach.

Abbreviations

FA

Food addiction

mYFAS-HEB

The modified Yale Food Addiction Scale in Hebrew

Author contributions

Conceptualization: OA, RAF, DRS, ST; Data collection: OA; Analysis: OA, RAF; Supervision: ST; Writing—original draft: OA, RAF, ST; Writing—review & editing: ST, RAF, DRS. All authors have read and approved the final manuscript.

Funding

This study received no external funding. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Availability of data and materials

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

The study was approved by the IRB ethics committee of Tel-Hai College and was conducted in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. All participants gave their informed consent prior to inclusion in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.