Skip to main content
Download PDF
Download ePub

Reliability and validity of the modified Yale Food Addiction Scale 2.0 (mYFAS 2.0) in a sample of individuals with depressive disorders

Journal of Eating Disorders volume 12, Article number: 144 (2024) Cite this article

Abstract

Background

Food addiction (FA) is strongly associated with depressive symptoms. The reliability and validity of the Modified Yale Food Addiction Scale 2.0 (mYFAS 2.0) were not previously determined in clinical samples in Brazil. This study aimed to assess the psychometric properties of the Brazilian version of the mYFAS 2.0 in adult individuals with depressive disorders.

Methods

The data stems from a survey investigating FA in a convenience sample of subjects diagnosed with a depressive disorder. Participants answered mYFAS 2.0 and scales for binge eating, depressive and anxiety symptoms, and alcohol and nicotine use. Height and weight were measured to calculate the Body Mass Index (BMI). We evaluated the factor structure, reliability, convergent, discriminant, criterion, and incremental validity.

Results

The sample encompassed 303 participants with a mean age of 37.03 ± 11.72 years, 84.16% of whom were women. The Cronbach’s alpha for the mYFAS 2.0 was satisfactory (alpha = 0.915). The best goodness-of-fit model was a single factor, and mYFAS 2.0 showed convergent validity with binge eating and discriminant validity with the alcohol and nicotine use measures. Food addiction presented a weak positive correlation with depressive and anxiety symptoms and BMI. Three food addiction symptoms provided the best balance between sensitivity (80.95%) and specificity (74.81%). Incremental validity over binge eating symptoms was confirmed (t = 4.040, β = 0.681, p < 0.001).

Conclusions

The Brazilian mYFAS 2.0 performed satisfactorily in this clinical sample of participants with a depressive disorder. These findings suggest it may be a brief, useful, and valid food addiction screening tool for this group.

Plain English Summary

Food addiction is a dysfunctional consumption of energetically dense, hyper-palatable, and ultra-processed foods that may lead to addictive behaviors. It is associated with mental disorders such as eating, mood, and anxiety disorders, which negatively impact the quality of life for individuals affected. Therefore, healthcare providers need to assess food addiction. The Modified Yale Food Addiction Scale 2.0 (mYFAS 2.0) is a brief instrument consisting of 13 questions developed to assess FA. Although it was previously adapted for Brazilian Portuguese in a non-clinical sample, this is the first study in Brazil to investigate this tool in a psychiatric sample. The main aim of our study was to evaluate the psychometric properties of the Brazilian version of the mYFAS 2.0 in individuals with a Depressive Disorder. The results suggested that mYFAS 2.0 had satisfactory psychometric properties in this sample, and it may be a brief, useful, and valid scale to screen food addiction in individuals with depressive states.

Background

Food addiction (FA) is characterized by a dysfunctional eating pattern of energetically dense, hyper-palatable, and ultra-processed foods for susceptible individuals who present symptoms analogous to those found in Substance-related and Addictive Disorders from the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) [1, 2]. Evidence from animal, imaging, and epidemiologic studies has suggested an addictive role for ultra-processed foods, which are usually rich in sugar, salt, and/or fat [1, 3]. Once FA is highly prevalent among subjects with overweight/obesity (weighted prevalence 24,9%: 95%CI 14.2–40.1) [4], it could be a contributing factor in the development of obesity. Therefore, FA could have implications beyond mental health, and its prevention and management have a nexus with other public health policies [1, 5].

The Yale Food Addiction Scale (YFAS) emerged in 2009 as an attempt to operationalize a standard characterization of FA, initially based on the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders-Text Revision (DSM-IV-TR) criteria for substance dependence [6]. A new version of the YFAS, referred to as YFAS 2.0, sought to align with changes in the criteria for Substance Use Disorder (SUD) in the DSM-5 (e.g., substance abuse and dependence symptoms combined in a single list, the inclusion of craving and definition of severity levels) [2]. Similarly to the YFAS, YFAS 2.0 has two scoring options: (a) a continuous symptom count, representing the number of endorsed symptoms at a clinical threshold, and (b) a diagnostic scoring reflecting the DSM-5 criteria (≥ 2 symptoms and clinically significant impairment or distress) [2]. The YFAS 2.0 has demonstrated satisfactory internal consistency and convergent validity concerning eating-related measures. In addition, FA has been shown to correlate with measures of impulsivity and emotional dysregulation [2, 7], as well as it is significantly associated with a range of mental disorders, notably eating, mood, anxiety, and post-traumatic stress disorders [8, 9].

A modified version of YFAS 2.0 (mYFAS 2.0) was developed and performed similarly to YFAS 2.0 [10]. The mYFAS 2.0 is a brief scale with 13 items that reflects DSM-5 criteria for SUDs and has the same scoring method as the YFAS 2.0. It has been validated in some countries outside the USA [11]. In Brazil, the mYFAS 2.0 was translated and evaluated in a large nonclinical sample of internet users (n = 7.639) [12]. The Brazilian mYFAS 2.0 has adequate internal consistency reliability, a consistent one-factor structure, and a positive correlation with impulsivity scores. Furthermore, FA showed a strong association with depressive, bipolar spectrum, and skin-picking disorders and a history of early-life psychological and sexual abuse in the Brazilian context, but not for nicotine and alcohol use disorders [13].

Several studies have pointed to relevant associations between depressive symptoms and FA, but fewer surveys have investigated FA in clinical samples of individuals with mood disorders specifically [4, 8]. In addition, major depressive disorder (MDD) is the second most common mental disorder reported alongside FA [8], and some studies showed FA prevalence ranges from 25 to 29% in depressed participants [14, 15]. In a nationwide Danish study, it was found a weighted prevalence of 29.4% for FA between affective disorders, 25.3% in unipolar depression, and 43.4% in bipolar disorder [9]. It is essential to evaluate how the mYFAS 2.0 performs in individuals with depressive symptoms, which would provide additional evidence about its psychometric properties.

When investigating FA prevalence and other psychopathological correlates in a cross-sectional study enrolling participants with depressive disorders, an opportunity to provide additional evidence on the psychometric properties of the mYFAS 2.0 was glimpsed based on the collected data. The first aim of this paper is to confirm the reliability and assess the factor structure. Second, we intend to investigate the convergent and discriminant validity of FA with other psychopathologies (binge eating, depression, anxiety, and alcohol and nicotine use disorder) and BMI. Third, this study explores the accuracy properties of mYFAS 2.0 in relation to screening for binge eating (BE) as an external criterion. Finally, the incremental validity of the mYFAS 2.0 will be examined using a hierarchal regression model to predict the variance of the BES scores.

Methods

Participants and study design

This cross-sectional study collected data from a convenience sample in Fortaleza city (Ceara State, Northeast of Brazil) from November 2021 to July 2023. Volunteers were recruited from the routine appointments of two psychiatric outpatient public services (Walter Cantídio University Hospital and Professor Frota Pinto Mental Health Hospital) and via social media announcements.

All volunteers underwent an in-person eligibility assessment. The inclusion criteria required participants to be 18 years or older, be at least literate, and be currently in a depressive episode from major depressive disorder or bipolar disorder or remission. The exclusion criteria were the presence of psychotic symptoms; the current or most recent mood episode was (hypo)manic; a medical history of dementia, intellectual disability, or another disease with cognitive impairment; endocrine and/or metabolic pathologies not currently treated; current cancer treatment; attending under the influence of a psychoactive substance or with withdrawal syndrome; and inability to read or understand instrument questions. A structured interview based on DSM-5 was conducted [16] to assess mood psychopathology, and a checklist was used to verify compliance with the inclusion and exclusion criteria. Then, participants who met the eligibility criteria responded to the other research instruments.

Trained administrators (undergraduate medical students) applied the survey instruments under the supervision of an experienced psychiatrist and collected the anthropometric body mass and stature measurements. All instruments were available on REDCap (Research Electronic Data Capture), an online platform for electronic data registration [16].

Measures

Sociodemographic variables, including sex, age, marital status, self-identification of race, and educational level, were collected for the entire sample. Participants were also asked about the use of psychopharmacotherapy.

The Mental International Neuropsychiatry Interview 7.02 (MINI 7.02) is a gold-standard structured interview for diagnostic confirmation in research [17]. Each module has dichotomous screening questions (yes or no type) that, if answered positively, lead to further questions. Depression and (hypo)mania modules from MINI were used to define the presence or absence of a mood episode (current or past) based on DSM-5 criteria. The MINI has a validated Brazilian version with satisfactory psychometric properties [18].

The mYFAS 2.0 is a self-report questionnaire with 13 questions for the assessment of FA reflecting the DSM-5 criteria for SUDs [10]. The instrument has 11 items addressing each diagnostic criterion, and two refer to clinical distress and/or impairment in psychosocial functioning [10]. Each question is scored from 0 to 7, and there is a threshold to endorse symptoms as described elsewhere [10, 11]. In the diagnostic scoring option, the presence of FA is established by completing a minimum of two endorsed symptoms, and the presence of suffering or clinically significant psychosocial impairment. The symptom count scoring option is made by summing the endorsed symptoms. In addition, based on severity levels for SUDs from the DSM-5, FA severity can be defined according to the number of symptoms: mild (2–3), moderate (4–5), and severe (≥ 6). The Brazilian version of the mYFAS 2.0 demonstrated satisfactory psychometric properties in a non-clinical sample [12].

The BES is a self-report questionnaire developed primarily to screen BE behaviors and their severity [19]. The scale has 16 items, each rated from 0 to 3, and answers are summed to obtain the final score. A positive screening for BE occurs if BES > 17 (BES+). The BES was adapted to Brazilian Portuguese and has been widely used in Brazil [20,21,22,23]. Its study validation showed adequate internal consistency and satisfactory sensitivity and specificity to identify Binge Eating Disorder (BED) when compared to SCID-I/P in a sample of women with obesity [24].

The Montgomery-Asberg Depression Rating Scale (MADRS), one of the most commonly used clinician-rated measures of depression severity, assessed depression severity [25]. It consists of a ten-item questionnaire, which can be scored from 0 to 6 after an interview and summed to provide the total score. Its scores have shown a strong correlation with those of the Brazilian version of the Hamilton-Depression Rating Scale (HDRS) [26].

The 7-item Generalized Anxiety Disorder Scale (GAD-7) screened for Generalized Anxiety Disorder [27]. The GAD-7 is a 7-item self-report questionnaire answered on a 4-point Likert scale on how often they have been bothered by anxiety symptoms in the past two weeks. The sum of the answers is used to obtain the final score. The Brazilian version of the GAD 7 has shown good reliability [28].

Alcohol, Smoking, and Substance Involvement Screening Test 3.0 (ASSIST 3.0) is a structured interview that aims to screen the consumption of legal and illicit substances in primary care services, general medical care, and other settings [29]. It encompasses eight questions addressing the use pattern of nine groups of substances in the last three months, attempts to control use, health and psychosocial harms. The sum of the response scores for each substance category falls into a health risk range. Version 3.0 of the ASSIST is available in Brazilian Portuguese and has good psychometric properties for use in the general population [30]. For this paper, we counted scores from the nicotine and alcohol subscales.

BMI is an indirect index of adiposity based on average body composition, which is widely used to screen obesity worldwide for its convenience, safety, and minimal cost [31, 32]. It correlates highly with direct measures of adiposity (e.g., dual-energy x-ray absorptiometry) despite its limitations in assessing specific phenotypes of body composition (e.g., bodybuilders) and not to be able in distinguishing lean body mass from fat mass [33,34,35]. BMI score is obtained by dividing body mass in kilograms (kg) by stature in meters squared (m2). Stature and body mass were measured using an anthropometric scale, and subjects were classified as underweight (< 18.5), normal weight (18.5–24.9), overweight (25-29.9), and obese (≥ 30) [32].

Statistical analysis

Data analysis was performed on Jamovi® [36], an open statistical platform based on the R statistical language. Sociodemographic categorical variables are described as percentages. Continuous variables are displayed with standard deviation (mean ± SD). The Kolmogoroff-Smirnoff Test assessed the normal distribution of the variables.

Cronbach’s alpha coefficients were used to measure reliability with a 95% confidence interval (95% CI). mYFAS 2.0’s alpha was calculated when each item was dropped out and for the full scale. Additionally, Cronbach’s alpha coefficients for the following scales were calculated: BES, MADRS, GAD-7, and nicotine and alcohol ASSIST 3.0 subscales. Coefficients equal to or greater than 0.7 are considered satisfactory [37].

An Exploratory Factor Analysis (EFA) extracted the better factor solution. A Minimum residual extraction method in combination with an Oblimin rotation was carried out to extract the factors. Kaiser-Meyer-Olkin (KMO) statistic and Bartlett’s test for sphericity were used to assess the factorability of the correlation matrix. The impairment and distress items were excluded from the factor exploration analysis because they express the clinical significance of FA as a syndrome and do not reflect specific symptomatic criteria. A criterion for the better factor solution was based on the presence of eigenvalues greater than 1.0, as proposed by the Kaiser-Guttman rule [38].

We proceed with a Factor Confirmatory Analysis (CFA) to test the suggested model’s fitness from the EFA. The Chi-square test (χ2), the Root Mean Square Error of Approximation (RMSEA), the Standardized Root Mean Square Residual (SRMR), and the Comparative Fit Index (CFI) were used to assess the goodness-of-fit of the model. The Chi-square Test for exact adjustment suggests good fitness if the null hypothesis is accepted. This study considered an RMSEA ≤ 0.06, an SRMR ≤ 0.09, and a CFI ≥ 0.90 acceptable, although there is no consensus in the literature.

Spearman’s correlation coefficient (ρ) was calculated between the continuous symptom count and the diagnostic scoring of the mYFAS 2.0 and measures putatively convergent (BES, BMI, MADRS, GAD7) and discriminant (nicotine and alcohol ASSIST 3.0 subscales). The strength of the correlations was interpreted as proposed elsewhere [39]: negligible: 0.00-0.10; weak: 0.10–0.39; moderate: 0.40–0.69; strong: 0.70–0.89; very strong: 0.90-1.00. Convergent validity was found when ρ ≥ ׀ 0.4 ׀. A negligible (ρ ≤ ׀ 0.1 ׀) or nonsignificant correlation (p < 0.05) was considered evidence for discriminant validity.

The Receiver Operating Characteristics (ROC) methodology allowed us to compare continuous mYFAS 2.0 scoring to the positive screening of the BE as a reference. The BES is not a FA gold-standard instrument but is a widely used validated scale to screen for BE [19, 40]. The Youden’s Index and AUC (Area Under the Curve) were estimated across all cut-offs, as well as sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). The best cut-off was chosen based on Youden´s Index. The following thresholds were used to classify the AUC: 0.9-1.0: excellent; 0.8–0.9: very good; 0.7–0.8: good; 0.6–0.7: sufficient; 0.5–0.6: bad; < 0.5: test not useful.

We performed a hierarchical regression model to assess mYFAS 2.0’s incremental validity in predicting BES score variance, controlling for sex, age, and the use of psychopharmacotherapy at each step. Measures were entered step by step in the following sequence: BMI, MADRS, GAD-7, and symptom count score of the mYFAS 2.0. All analyses considered a significance level of p < 0.05.

Results

Descriptive statistics of the sample are shown in Supplementary Table 1. The final sample encompassed 303 participants, predominantly young (37.03 ± 11.72) and female (84.16%; n = 255). Most had at least a secondary educational level (49.17%; n = 149) and mixed ancestry (51.16%; n = 155). A total of 122 participants (40.26%) screened positive for FA. The sample had a mean of 3.67 ± 3.14 endorsed symptoms of FA. A current depressive episode was detected in 78.55% (n = 238) of the sample, and 21.45% (n = 65) of the participants were in remission. Regarding the mood spectrum, 206 (67.99%) were classified as unipolar, and 97 (32.01%) were bipolar.

Table 1 presents the mYFAS 2.0 Cronbach’s alpha for each deleted item and the full-scale coefficient. The Cronbach’s alpha value was 0.915 for the mYFAS 2.0, and all individual items scored greater than 0.7. The other scales obtained the following Cronbach’s alpha values: BES: 0.914; MADRS: 0.862; GAD-7: 0.85; Nicotine ASSIST subscale: 0.918; and Alcohol ASSIST 3.0 subscale: 0.829. Therefore, mYFAS 2.0 and other scales’ reliability was considered satisfactory.

Table 1 Cronbach´s Alpha for full mYFAS 2.0 and if each item dropped
Full size table

Assumption checks for sample factorability were filled. The sampling adequacy was satisfactory (KMO = 0.913). Bartlett’s sphericity test confirmed the instrument’s factorability (p < 0.001).

Table 2 contains factor loadings and eigenvalues. All the items had factor loadings > 0.3, ranging from 0.408 to 0.756. Eigenvalue inspection revealed a single-factor solution because only one was > 1.0. CFI = 0.915 and SRMR = 0.053 suggested a good fit for this model, but the RMSEA = 0.093 (CI 90%: 0.078–0.110) and a significant Chi-square Test did not in the CFA (See supplementary information).

Table 2 Factor loadings and eigenvalues of the mYFAS 2.0
Full size table

All putatively convergent scales showed positive correlations and statistical significance, however only BES showed a strong correlation (ρ = 0.716; p < 0.001) with FA count symptoms and a moderate correlation with the diagnostic scoring of the mYFAS (ρ = 0.579; p < 0.001) (Table 3). BMI evidenced a positive weak correlation both to continuous (ρ = 0.302, p < 0.001) and diagnostic mYFAS scoring (ρ = 0.346, p < 0.001). Depression and anxiety scores had significant weak positive correlations with FA count symptoms (MADRS: ρ = 0.294, p < 0.001; GAD-7: 0.344, p < 0.001) and mYFAS 2.0 diagnostic scoring (MADRS: ρ = 0.226, p < 0.001; GAD-7: 0.251, p < 0.001).

Table 3 Spearman´s rho (ρ) of the mYFAS 2.0 scoring regarding BMI, MADRS, GAD-7, and the ASSIST 3.0 subscales
Full size table

Regarding discriminant validity, the mYFAS 2.0 and ASSIST 3.0 subscales for nicotine and alcohol use did not exhibit consistent correlations (Table 3). Nicotine use obtained was not correlated with either continuous (ρ = 0.026; p = 0.656) or diagnostic scoring (ρ = 0.008, p = 0.887) of the mYFAS 2.0. Alcohol use correlated positively with FA count symptoms (ρ = 0.136; p = 0.018) and diagnostic scoring option for FA (ρ = 0.113; p = 0.050) but next to the threshold to a negligible correlation.

Supplementary Fig. 1 shows the combined ROC curve where mYFAS continuous scoring is plotted concerning BES+. The AUC was 85.4%, which is considered a “very good” accuracy. The most significant value of Youden’s Index suggested the best cut-off with three symptoms count of FA (Sensitivity: 80.95; Specificity: 74.81%; PPV: 80%; NPV: 75.94%) (Table 4).

Table 4 Indicators of diagnostic test performance of the mYFAS 2.0
Full size table

Incremental validity was assessed by a model using four steps (Table 5). In step 1, BMI alone was significantly associated with BES scores (t = 7.883, β = 0.432, p < 0.001) and accounted for 18.8% of its variance (adjusted R2 = 0.188). In step 2, the inclusion of MADRS in the model accounted for 25.9% of the BES variance (adjusted R2 = 0.259; ΔR2 = 0.072) and predicted significant BES scores (t = 5.438, β = 0.273, p < 0.001). In step 3, the inclusion of the GAD-7 also contributed to explaining the BES variance (adjusted R2 = 0.188; ΔR2 = 0.035) in a significant way (t = 3.905, β = 0.236, p < 0.001). FA count symptoms were an essential predictor of BES scores in step 4 (t = 13.923, β = 0.601, p < 0.001) and explained 57.1% of the BES scores (adjusted R2 = 0.571; ΔR2 = 0.275). BMI was associated with BES variance in all steps (p < 0.001).

Table 5 Hierarchical regression for incremental validity of mYFAS 2.0 continuous scoring
Full size table

Discussion

This study assessed the psychometric properties of the Brazilian mYFAS 2.0 in a sample of participants with a depression diagnosis. Although it was previously adapted for Brazilian Portuguese in a web-based survey, this is the first study in Brazil to investigate the performance of this brief version in a psychiatric sample. Thus, this study adds evidence about how the Brazilian mYFAS 2.0 could work with depressed patients in clinical practice.

Overall, Brazilian mYFAS 2.0 retained previously reported psychometric properties [12]. The Cronbach’s alpha coefficients for both the items and the full scale were > 0.9, demonstrating satisfactory internal consistency reliability [37, 41]. A one-factor solution was confirmed, similar to what was observed in the Brazilian validation of the measure among a nonclinical sample and the mYFAS 2.0 development study [10, 12]. A single-factor structure has been found in different samples and cultures, so food addiction symptoms seem to be consistently linked to one factor [11]. As the mYFAS 2.0 is based on the DSM-5 criteria for SUDs, it is plausible that, similarly to these criteria, the factorial structure is unidimensional [42].

This study tested the convergent validity of constructs linked to eating, depression, and anxiety. The correlations between the BES and mYFAS 2.0 scoring options signaled clear convergent validity between these eating-related measures [43, 44]. This finding suggests convergent validity between the BES and YFAS scores, which aligns with other studies that estimated their linear correlation [43, 44]. Data of a meta-analysis reporting a positive weighted mean correlation of 0.602 for FA and binge eating (BE) (95% CI: 0.557–0.643; I2:76.37, P < 0.05), 0.459 for FA and depression (0.358–0.550; I2:93.74; P < 0.05) and 0.483 for FA and anxiety (95% CI: 0.228– 0.676; I2:96.27, P < 0.001) aligns with our findings [8].

BMI had a significant, positive, and weak linear correlation with FA. Some positive correlation between BMI and FA is expected because of its relation to eating behavior, but samples of participants with obesity have shown a stronger correlation between them [11]. Mixed findings on the correlation between BMI and FA suggest that future research should investigate the role of compensatory behaviors (e.g., increased physical activity) in preventing weight gain in people with FA [11].

A weak positive correlation between mYFAS 2.0 scoring options and depression and anxiety scores was observed in this sample, similar to the findings of other studies [43,44,45]. Despite a positive correlation, the mYFAS 2.0 had no convergent validity with the MADRS and GAD-7 according to the threshold adopted. Some studies used a threshold correlation ≥ 0.3 to define convergent validity, while this study adopted a correlation level ≥ 0.4, as described earlier [39, 43, 44]. Notwithstanding, some degree of covariation between the scores of these scales is understandable because of the affective dysregulation and the emotional eating related to depression and anxiety disorders, with potential mutual influence between these syndromes [46].

ASSIST 3.0 scores for nicotine and alcohol did not reveal relevant covariance with FA in this sample, suggesting a discriminant validity for the mYFAS 2.0. These findings are similar to those observed in previous studies [6, 45, 47]. Although significant comorbidity of FA with SUDs has been reported [48], alcohol and nicotine use have been less associated with FA than other mental health conditions [4, 8]. An increase in SUDs has been observed after bariatric surgeries, arousing a hypothesis of “transfer addiction” [49]. Conversely, FA may play a concurrent role with other addictions given food and drugs share reward circuits [3, 49, 50]. Therefore, it sounds expected to find a weak or no correlation between FA and alcohol and nicotine use disorder.

We employed BE screening as an external criterion regarding the mYFAS 2.0 scoring option to perform the ROC curve analysis. This approach has been adopted in psychometric studies when no gold-standard external criterion is available for comparison [51, 52]. The continuous scoring option of the mYFAS 2.0 seemed to discriminate groups with a positive and negative screening of BE once the AUC was considered “very good,” and Youden’s Index provided the best sensitivity (80.95%) and specificity (74.81%) with a cut-off of three FA symptoms. A similar methodology was applied by Granero et al., who obtained an AUC of 94% (sensitivity: 96.7%; specificity: 77.8%) to discriminate between healthy controls and subjects with positive screening for any eating disorder [53]. It is widely recognized that some eating disorders have high endorsed symptom rates of FA (e.g., bulimia and binge disorder). However, prior studies have demonstrated that FA has unique clinical characteristics (e.g., withdrawal) [1, 4]. In turn, psychopathological interfaces between FA and BE (e.g., loss of control over consumption) may have been reflected in a strong correlation between BES and mYFAS 2.0 scores, which may have influenced the ROC curve’s result in this study [54].

The incremental validity in predicting BES total scores was confirmed over BMI and depressive and anxiety symptoms. This parallels other studies showing the ability of the mYFAS continuous scoring to predict the variability of binge eating scores [10, 43]. Considering the relevant FA input in explaining the variability data of compulsive eating in this sample (adjusted R2 = 0.571; ΔR2 = 0.275) and the strong association between FA and mood disorders [8, 12], early FA detection may aid in weight gain control in this group.

Some limitations of this study must be discussed. First, this was a cross-sectional study, which does not allow causal inferences. Second, our sample, predominantly composed of young adults and females, is not nationally representative, which limits the generalizability of findings to other groups (e.g., men, elderly individuals, and substance users). Third, a positive screening for BE from the BES as an external criterion is not ideal for assessing the criterion validity of the mYFAS 2.0; therefore, sensitivity and specificity data in this study should be interpreted with caution, given that FA and BE do not fully overlap. Our data undoubtedly showed that the mYFAS 2.0 captured its psychopathological interface with BE; however, developing a gold-standard structured interview for FA is an essential step in this field of study. Fourth, this study design was unable to verify the stability of FA diagnosis over time by a test-retest approach. Thus, conducting longitudinal studies with systematic probabilistic sampling and a nationally representative sample could make it possible to estimate the psychometric properties of the mYFAS 2.0 accurately and address these properties in other groups.

The strengths of this study deserve to be highlighted. First, we used a large sample size appropriate for the data analyses. Second, a gold-standard structured interview was used to confirm the depression diagnosis. Third, a stringent exclusion criterion avoided the potential effects of some psychopathological and medical conditions on variables related to eating behavior and BMI. Fourth, all self-report tools were previously validated, and satisfactory reliability was obtained for this sample.

Conclusions

We confirmed that the Brazilian mYFAS 2.0 has good internal consistency, a single-factor dimensionality, and convergent validity with the BES. Discriminant validity between mYFAS 2.0 and the Nicotine and Alcohol ASSIST 3.0 subscales was found. BMI, depression, and anxiety scales showed a weakly positive correlation with FA count symptoms. In addition, mYFAS has demonstrated incremental validity regarding the BES scores.

This work has advanced the study of the psychometric properties of the Brazilian version of the mYFAS 2.0, establishing it as a brief, reliable, and valid scale for use in patients with depressive disorders. Early detection of FA through mYFAS 2.0 may provide opportunities for better management of eating-related psychopathology and mitigate weight gain in patients with depressive disorders.

Data availability

Data analyzed and supporting this manuscript are deposited in the Harvard Dataverse (https://doi.org/10.7910/DVN/LKEFCI) and are available from the corresponding author upon reasonable request.

Abbreviations

ASSIST 3.0:

Alcohol Smoking and Substance Involvement Screening Test Version 3.0

AUC:

Area Under the Curve

BE:

binge eating

BES:

Binge Eating Scale

BMI:

Body Mass Index

CFA:

Factor Confirmatory Analysis

CFI:

Comparative Fit Index

DSM:

Diagnostic and Statistical Manual of Mental Disorders

EFA:

Exploratory Factor Analysis

FA:

food addiction

GAD-7:

7-item Generalized Anxiety Disorder Scale

HDRS:

Hamilton-Depression Rating Scale

KMO:

Kaiser-Meyer-Olkin Test

MADRS:

Montgomery-Asberg Depression Rating Scale

MINI 7.02:

Mental International Neuropsychiatry Interview Version 7.02

mYFAS 2.0:

Modified Yale Food Addiction Scale Version 2.0

NPV:

negative predictive value

PPV:

positive predictive value

REDCap:

Research Electronic Data Capture

RMSEA:

Root Mean Square Error of Approximation

ROC:

Receiver Operating Characteristics

SRMR:

Standardized Root Mean Square Residual

SUD:

substance use disorders

YFAS:

Yale Food Addiction Scale

References

  1. Gearhardt AN, Schulte EM. Is Food Addictive? A review of the Science. Annu Rev Nutr. 2021;41:387–410.

    Article  PubMed  Google Scholar 

  2. Gearhardt AN, Corbin WR, Brownell KD. Development of the Yale Food Addiction Scale Version 2.0. Psychol Addict Behav. 2016;30(1):113–21.

    Article  PubMed  Google Scholar 

  3. Passeri A, Municchi D, Cavalieri G, Babicola L, Ventura R, Di Segni M. Linking drug and food addiction: an overview of the shared neural circuits and behavioral phenotype. Front Behav Neurosci. 2023;17:1240748.

    Article  PubMed  PubMed Central  Google Scholar 

  4. Pursey KM, Stanwell P, Gearhardt AN, Collins CE, Burrows TL. The prevalence of food addiction as assessed by the Yale Food Addiction Scale: a systematic review. Nutrients. 2014;6(10):4552–90.

    Article  PubMed  PubMed Central  Google Scholar 

  5. Cassin SE, Buchman DZ, Leung SE, Kantarovich K, Hawa A, Carter A, Sockalingam S. Ethical, Stigma, and Policy Implications of Food Addiction: a scoping review. Nutrients 2019, 11(4).

  6. Gearhardt AN, Corbin WR, Brownell KD. Preliminary validation of the Yale Food Addiction Scale. Appetite. 2009;52(2):430–6.

    Article  PubMed  Google Scholar 

  7. Maxwell AL, Gardiner E, Loxton NJ. Investigating the relationship between reward sensitivity, impulsivity, and food addiction: a systematic review. Eur Eat Disorders Rev. 2020;28(4):368–84.

    Article  Google Scholar 

  8. Burrows T, Kay-Lambkin F, Pursey K, Skinner J, Dayas C. Food addiction and associations with mental health symptoms: a systematic review with meta-analysis. J Hum Nutr Diet. 2018;31(4):544–72.

    Article  PubMed  Google Scholar 

  9. Horsager C, Faerk E, Lauritsen MB, Ostergaard SD. Food addiction comorbid to mental disorders: a nationwide survey and register-based study. Int J Eat Disord. 2021;54(4):545–60.

    Article  PubMed  Google Scholar 

  10. Schulte EM, Gearhardt AN. Development of the modified Yale Food Addiction Scale Version 2.0. Eur Eat Disord Rev. 2017;25(4):302–8.

    Article  PubMed  Google Scholar 

  11. Meule A, Gearhardt AN. Ten years of the Yale Food Addiction Scale: a review of Version 2.0. Curr Addict Rep. 2019;6(3):218–28.

    Article  Google Scholar 

  12. Nunes-Neto PR, Kohler CA, Schuch FB, Quevedo J, Solmi M, Murru A, Vieta E, Maes M, Stubbs B, Carvalho AF. Psychometric properties of the modified Yale Food Addiction Scale 2.0 in a large Brazilian sample. Braz J Psychiatry. 2018;40(4):444–8.

    Article  PubMed  PubMed Central  Google Scholar 

  13. Nunes-Neto PR, Kohler CA, Schuch FB, Solmi M, Quevedo J, Maes M, Murru A, Vieta E, McIntyre RS, McElroy SL, et al. Food addiction: prevalence, psychopathological correlates and associations with quality of life in a large sample. J Psychiatr Res. 2018;96:145–52.

    Article  PubMed  Google Scholar 

  14. Mills JG, Larkin TA, Deng C, Thomas SJ. Weight gain in major depressive disorder: linking appetite and disordered eating to leptin and ghrelin. Psychiatry Res. 2019;279:244–51.

    Article  PubMed  Google Scholar 

  15. Mills JG, Thomas SJ, Larkin TA, Deng C. Overeating and food addiction in major depressive disorder: links to peripheral dopamine. Appetite. 2020;148:104586.

    Article  PubMed  Google Scholar 

  16. Harris PA, Taylor R, Minor BL, Elliott V, Fernandez M, O’Neal L, McLeod L, Delacqua G, Delacqua F, Kirby J, et al. The REDCap consortium: building an international community of software platform partners. J Biomed Inf. 2019;95:103208.

    Article  Google Scholar 

  17. Sheehan DV, Lecrubier Y, Sheehan KH, Amorim P, Janavs J, Weiller E, Hergueta T, Baker R, Dunbar GC. The mini-international neuropsychiatric interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry. 1998;59(Suppl 20):22–quiz3334.

    PubMed  Google Scholar 

  18. Amorim P. Mini International Neuropsychiatric interview (MINI): validação de entrevista breve para diagnóstico de transtornos mentais. Brazilian J Psychiatry. 2000;22:106–15.

    Article  Google Scholar 

  19. Gormally J, Black S, Daston S, Rardin D. The assessment of binge eating severity among obese persons. Addict Behav. 1982;7(1):47–55.

    Article  PubMed  Google Scholar 

  20. Freitas S, Lopes CS, Coutinho W, Appolinario JC. Tradução E adaptação para o português da Escala De Compulsão Alimentar Periódica. Brazilian J Psychiatry 2001, 23.

  21. Novelle JM, Alvarenga MS. Cirurgia bariátrica E Transtornos alimentares: uma revisão integrativa. Jornal Brasileiro De Psiquiatria 2016, 65.

  22. Caldas NR, Braulio VB, Brasil MAA, Furtado VCS, Carvalho DPd, Cotrik EM, Dantas JR, Zajdenverg L. Binge eating disorder, frequency of depression, and systemic inflammatory state in individuals with obesity – a cross sectional study. Archives Endocrinol Metabolism 2022, 66.

  23. Moraes CEF, Heriseanu A, Mourilhe C, Faller ALK, Hay P, Appolinario JC. Validation of the Brazilian version of the short inventory of Grazing (SIG). Trends Psychiatry Psychother 2024, 46.

  24. Freitas SR, Lopes CS, Appolinario JC, Coutinho W. The assessment of binge eating disorder in obese women: a comparison of the binge eating scale with the structured clinical interview for the DSM-IV. Eat Behav. 2006;7(3):282–9.

    Article  PubMed  Google Scholar 

  25. Montgomery SA, Asberg M. A new depression scale designed to be sensitive to change. Br J Psychiatry. 1979;134:382–9.

    Article  PubMed  Google Scholar 

  26. Dratcu L, da Costa Ribeiro L, Calil HM. Depression assessment in Brazil. The first application of the Montgomery-Asberg Depression Rating Scale. Br J Psychiatry. 1987;150:797–800.

    Article  PubMed  Google Scholar 

  27. Spitzer RL, Kroenke K, Williams JBW, Löwe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med. 2006;166(10):1092–7.

    Article  PubMed  Google Scholar 

  28. Moreno AL, DeSousa DA, Souza AMFLPd, Manfro GG, Salum GA, Koller SH. Osório FdL, Crippa JAdS: factor structure, reliability, and item parameters of the brazilian-portuguese version of the GAD-7 questionnaire. Temas em Psicologia. 2016;24:367–76.

    Article  Google Scholar 

  29. Who Assist Working Group. The Alcohol, smoking and substance involvement screening test (ASSIST): development, reliability and feasibility. Addiction. 2002;97(9):1183–94.

    Article  Google Scholar 

  30. Henrique IF, De Micheli D, Lacerda RB, Lacerda LA, Formigoni ML. [Validation of the Brazilian version of Alcohol, Smoking and Substance Involvement Screening Test (ASSIST)]. Rev Assoc Med Bras (1992) 2004, 50(2):199–206.

  31. World Health Organization (WHO). Physical status: the use and interpretation of anthropometry. Report of a WHO Expert Committee. World Health Organ Tech Rep ser. Volume 854. Geneva, Switzerland;; 1995. pp. 1–452.

  32. World Health Organization (WHO). Obesity: preventing and managing the global epidemic. Report of a WHO consultation. In: World Health Organ Tech Rep Ser. vol. 894. Geneva, Switzerland. 2000: i-xii, 1-253.

  33. Wu Y, Li D, Vermund SH. Advantages and limitations of the body Mass Index (BMI) to assess adult obesity. Int J Environ Res Public Health. 2024;21(6):757.

    Article  PubMed  PubMed Central  Google Scholar 

  34. Yumuk V, Tsigos C, Fried M, Schindler K, Busetto L, Micic D, Toplak H. European Guidelines for Obesity Management in adults. Obes Facts. 2015;8(6):402–24.

    Article  PubMed  PubMed Central  Google Scholar 

  35. Lichtash CT, Cui J, Guo X, Chen Y-DI, Hsueh WA, Rotter JI, Goodarzi MO. Body adiposity index versus body Mass Index and other Anthropometric traits as correlates of Cardiometabolic Risk factors. PLoS ONE. 2013;8(6):e65954.

    Article  PubMed  PubMed Central  Google Scholar 

  36. The jamovi project. jamovi (Version 2.3) [Computer Software]. https://www.jamovi.org.

  37. Nunnally JC, Bernstein IH. Psychometric theory. 3rd ed. edn. New York: McGraw-Hill; 1994.

    Google Scholar 

  38. Kaiser HF. The application of electronic computers to factor analysis. Educ Psychol Meas. 1960;20(1):141–51.

    Article  Google Scholar 

  39. Schober P, Boer C, Schwarte LA. Correlation coefficients: appropriate use and interpretation. Anesth Analg. 2018;126(5):1763–8.

    Article  PubMed  Google Scholar 

  40. Freitas SL, Appolinario CS, Sichieri JC. R: Validação da versão brasileira da binge eating scale - Escala De Compulsão Alimentar Periódica. Brazilian J Psychiatry 2002, 24.

  41. George DM, Paul. SPSS for windows step by step: a simple study guide and reference, 17.0 update, 10/e. Pearson Education India; 2011.

  42. Maccoun RJ. The Puzzling Unidimensionality of DSM-5 Substance Use Disorder diagnoses. Front Psychiatry. 2013;4:153.

    Article  PubMed  PubMed Central  Google Scholar 

  43. Imperatori C, Fabbricatore M, Lester D, Manzoni GM, Castelnuovo G, Raimondi G, Innamorati M. Psychometric properties of the modified Yale Food Addiction Scale Version 2.0 in an Italian non-clinical sample. Eat Weight Disord. 2019;24(1):37–45.

    Article  PubMed  Google Scholar 

  44. Escriva-Martinez T, Galiana L, Herrero R, Rodriguez-Arias M, Fernandez-Aranda F, Gearhardt AN, Banos RM. Food addiction and its relationship with other eating behaviours among Spanish university students. J Eat Disord. 2023;11(1):60.

    Article  PubMed  PubMed Central  Google Scholar 

  45. Koball AM, Borgert AJ, Kallies KJ, Grothe K, Ames G, Gearhardt AN. Validation of the Yale Food Addiction Scale 2.0 in patients seeking bariatric surgery. Obes Surg. 2021;31(4):1533–40.

    Article  PubMed  Google Scholar 

  46. Dakanalis A, Mentzelou M, Papadopoulou SK, Papandreou D, Spanoudaki M, Vasios GK, Pavlidou E, Mantzorou M, Giaginis C. The Association of Emotional Eating with Overweight/Obesity, Depression, Anxiety/Stress, and dietary patterns: a review of the current clinical evidence. Nutrients. 2023;15(5):1173.

    Article  PubMed  PubMed Central  Google Scholar 

  47. Clark SM, Martens K, Smith-Mason CE, Hamann A, Miller-Matero LR. Validation of the Yale Food Addiction Scale 2.0 among a bariatric surgery Population. Obes Surg. 2019;29(9):2923–8.

    Article  PubMed  Google Scholar 

  48. Hoover LV, Yu HP, Cummings JR, Ferguson SG, Gearhardt AN. Co-occurrence of food addiction, obesity, problematic substance use, and parental history of problematic alcohol use. Psychol Addict Behav. 2023;37(7):928–35.

    Article  PubMed  Google Scholar 

  49. Koball AM, Ames G, Goetze RE. Addiction transfer and other behavioral changes following bariatric surgery. Surg Clin North Am. 2021;101(2):323–33.

    Article  PubMed  Google Scholar 

  50. Volkow ND, Wang GJ, Fowler JS, Tomasi D, Baler R. Food and drug reward: overlapping circuits in human obesity and addiction. Curr Top Behav Neurosci. 2012;11:1–24.

    PubMed  Google Scholar 

  51. Kimberlin CL, Winterstein AG. Validity and reliability of measurement instruments used in research. Am J Health Syst Pharm. 2008;65(23):2276–84.

    Article  PubMed  Google Scholar 

  52. DeVon HA, Block ME, Moyle-Wright P, Ernst DM, Hayden SJ, Lazzara DJ, Savoy SM, Kostas-Polston E. A psychometric toolbox for testing validity and reliability. J Nurs Scholarsh. 2007;39(2):155–64.

    Article  PubMed  Google Scholar 

  53. Granero R, Jimenez-Murcia S, Gearhardt AN, Aguera Z, Aymami N, Gomez-Pena M, Lozano-Madrid M, Mallorqui-Bague N, Mestre-Bach G, Neto-Antao MI, et al. Validation of the Spanish Version of the Yale Food Addiction Scale 2.0 (YFAS 2.0) and clinical correlates in a sample of eating disorder, Gambling Disorder, and healthy control participants. Front Psychiatry. 2018;9:208.

    Article  PubMed  PubMed Central  Google Scholar 

  54. Burrows T, Skinner J, McKenna R, Rollo M. Food Addiction, binge eating disorder, and obesity: is there a relationship? Behav Sci (Basel). 2017;7(3):54.

    Article  PubMed  Google Scholar 

Download references

Acknowledgements

The authors thank Antonio Brazil Viana Júnior, a Walter Cantídio University Hospital Researcher Support Center statistician, who supported our analyses.

Funding

PRNN received a grant from the Programa de Pesquisa para o Sistema Único de Saúde /Fundação Cearense de apoio ao Desenvolvimento Científico e Tecnológico (PPSUS, Research Program for the Unified Health System/FUNCAP, Ceará Research Support Foundation) and the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, Brazilian National Council for Scientific and Technological Development).

Author information

Authors and Affiliations

  1. Psychiatry Residency Program, Professor Frota Pinto Mental Health Hospital, Fortaleza, CE, Brazil

    Valéria Portela Lima

  2. Faculty of Medicine, Federal University of Ceará, Fortaleza, CE, Brazil

    Erivan de Olivindo Cavalcante & Mariana Alves Leão

  3. Drexel University Center for Weight, Eating, and Lifestyle Science, Philadelphia, PA, USA

    Erica Marla LaFata

  4. Faculty of Medicine, University of Fortaleza, Fortaleza, CE, Brazil

    Alexandre Menezes Sampaio

  5. Department of Clinical Medicine, Faculty of Medicine, Federal University of Ceará, Prof. Costa Mendes Street, 1608, 4th floor, Fortaleza, Ceará, 60430-160, Brazil

    Paulo Rodrigues Nunes Neto

Authors
  1. Valéria Portela Lima
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Erivan de Olivindo Cavalcante
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Mariana Alves Leão
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Erica Marla LaFata
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Alexandre Menezes Sampaio
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Paulo Rodrigues Nunes Neto
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

PRNN coordinated the study and supervised the planning, data analysis and writing of the article. VPL performed the data analysis and wrote the manuscript. EOC and MAL collected data. EML and AMS participated in the review and discussion. All authors contributed to the final version of this manuscript.

Corresponding author

Correspondence to Paulo Rodrigues Nunes Neto.

Ethics declarations

Ethics approval and consent to participate

This study was carried out in accordance with the Declaration of Helsinki and the standards of the Brazilian National Research Ethics Commission. The Research Ethics Committee from Walter Cantídio University Hospital of the Federal University of Ceará (Committee number 5045) approved the study protocol (protocol number 4.477.006, December 19th 2020). All participants signed a written free and informed consent form.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Supplementary Material 2

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Lima, V.P., de Olivindo Cavalcante, E., Leão, M.A. et al. Reliability and validity of the modified Yale Food Addiction Scale 2.0 (mYFAS 2.0) in a sample of individuals with depressive disorders. J Eat Disord 12, 144 (2024). https://doi.org/10.1186/s40337-024-01108-7

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s40337-024-01108-7

Keywords

Journal of Eating Disorders

ISSN: 2050-2974