Emotional Eating Is Associated with T2DM in an Urban Turkish Population: A Pilot Study Utilizing Social Media

Abstract

Lifestyle behaviors and their potential effects on diabetes are being investigated for optimal diabetes management. In patients with type 2 diabetes mellitus (T2DM), the necessary dietary modifications extend to psychological components for consideration. This study aimed to determine the eating behavior of T2DM patients with different sociodemographic characteristics in an urban Turkish population. The Dutch Eating Behavior Questionnaire (DEBQ) was distributed via social media and a smartphone application to 108 T2DM patients, 58 males and 50 females, age 26–40 years (20 individuals, 18.6%) and over 40 years (88 individuals, 81.4 %). Basic component factor analysis varimax rotation was used for the item-total correlation coefficient. The 26–40 years age group exhibited high correlation for both restrained and emotional eating behavior (r > 0.8), while participants over 40 years displayed medium correlation for restrained eating and high correlation for emotional eating (r = 0.6–0.8). Compared to married and single participants, participants with “other” marital status showed significant correlation with all eating behavior categories (r > 0.8). Married participants were less correlated with all categories compared to single participants. Participants with lower education levels exhibited high correlation (r > 0.8) for all forms of eating, more so compared to those with higher levels of education attained. Overweight patients demonstrated moderately high (r = 0.4–0.6) restrictive eating correlation, while normal weight and obese patients exhibited higher correlation (r = 0.6–0.8) for emotional and restrained eating compared to overweight patients. Regardless of demographic factors, when all participants were combined, the strongest correlation was found to be with emotional eating compared to other types of eating.

1. Introduction

According to the International Diabetes Federation (IDF), in 2021 there were 537 million adults worldwide living with diabetes. The global economic burden of diabetes was estimated at a staggering USD 966 billion, representing a 316% increase over the past 15 years [1]. It is projected that by 2045, 668 million people worldwide will have diabetes [1]. Recent estimates suggest that 18.7 million of those with diabetes currently live in low-income countries [1]. Notably, in 2021, three out of four people with diabetes were residing in low- to middle-income countries [1]. In Turkey, a country undergoing epidemiological transition, diabetes diagnosis is at 9 million people and projected to continue to grow [2]. Turkey was higher in terms of adults with diabetes compared to all EU-27 countries in 2019 [2]. Furthermore, according to the nationwide TURDEP-II study held in Turkey, the rate of diabetes in the population over 20 years old is reported at 13.7% [2].

Type 2 diabetes mellitus (T2DM) is a chronic disease that requires everyday monitoring and self-care to prevent multiple acute or chronic complications due to uncontrolled and consistently high blood glucose levels. While T2DM currently remains incurable, patients are primarily relying on disease management through drug therapy and/or lifestyle modifications focused on diet and exercise. However, approaches undertaken such as healthy eating, physical activity, and use of medications, as well as considerations for complications and risks, typically generate psychological distress in patients [3,4,5,6,7,8]. In addition, hypoglycemic and hyperglycemic episodes have been shown to contribute to depression as glucose fluctuations can modify neurological function [9,10,11]. Therefore, signs of depression and anxiety are not uncommon among people with T2DM as patients tend to worry about their condition and its potential implications and outcomes such as heart disease, kidney disease, or nerve damage [12].

Research shows that depression exacerbates diabetes complications such as diabetic neuropathy and cardiovascular disease secondary to diabetes [11,12]. Moreover, higher depressive symptoms are associated with external and emotional eating [13,14], which can then induce a vicious cycle rendering diabetic management more challenging and less effective. Depression and mental health decline in relation to premature cardiovascular disease have been reported among young adults in the US [15], and this is plausibly the case with diabetes (T2DM) as well.

Given the relationship between behavior, mental health-related patterns, and chronic metabolic diseases, it is interesting to investigate the interplay between type 2 diabetes and patient eating behaviors. By characterizing the type of eating behavior, healthcare providers could obtain a more advantageous and robust starting point for effective dietetic counseling and medical nutrition therapy for better T2DM management. This becomes even more critical in a population undergoing epidemiological transition, such as the one in Turkey. Given the less-developed screening and monitoring systems, reduced awareness and education on the topic, financial challenges, and limited access to healthcare in Turkey, T2DM patients may face increased challenges.

Studies on T2DM and eating behavior generally show that there tends to be a stronger association with restrained eating and less of an association with external and emotional eating [16]. It is of note, however, that these three types of eating behaviors have been associated with binge eating and, subsequently, can negatively affect weight and T2DM management [16,17]. Definitions of external, emotional, and restrained eating can be found in Section 2.

In the study presented here, the 33-item Dutch Eating Behavior Questionnaire (DEBQ) [17] was used to investigate the effects of diabetes on emotional, external, and restrained eating in patients with T2DM at different sociodemographic levels. Further, we investigated if and how sociodemographic parameters influence the predominant type of eating behavior revealed, based on the DEBQ assessment. Another point of novelty in our work was that the participants were recruited using social media, whereby participating individuals were members of groups/online communities that specialize in diabetes management consultation and information exchange. Results from our work can inform clinical practices in medical nutrition therapy, nutrition education, and counseling. Such findings can help interventions be more targeted, meaningful, and effective towards positive dietary behavioral change, so that patients can optimize glycemic control and T2DM management specifically.

2. Materials and Methods

2.1. Sample

The present study was conducted on 108 patients diagnosed with type 2 diabetes, age 26–40 years (20 participants, 18.6%), and over 40 years (88 participants, 81.4%). Participants filled out the Dutch Eating Behavior Questionnaire survey through a Google form link provided to them. This link was shared via social media specialized groups for T2DM patients, namely: “Tip 2 Diyabet Grubu”, “İZMİR Tip.2 Diyabetliler Grubu”, and “Tip 2 Diyabetliler Toplanıyor” on Facebook.

Facebook was the primary means of recruitment. The authors extensively reviewed preexisting pages on Facebook and chose the ones with the highest number of members. We applied the inclusion and exclusion criteria and did not manipulate the outcome for eligible candidates for participation. This yielded an age distribution of approximately 18% to 82% for 26–40 years and over 40 years old, respectively. The population distribution reasonably closely describes the public health prevalence of T2DM in terms of age distribution. As reasonably expected, the majority of T2DM patients are over the age of 40 years. We aimed, however, to include younger individuals to capture any potential variation and/or deviation from what is expected due to any potential unclassified reason.

The link was also sent to T2DM patients through the WhatsApp mobile application. Demographic information was collected via the same Google form. Of the total 108 participants, 58 (53.7%) were males and 50 (46.3%) females, producing a reasonably well-balanced sample. Inclusion criteria were adults (>18 years old), diagnosed with T2DM, non-smokers, no other medical condition, and no other medication unrelated to T2DM. Exclusion criteria were pregnancy and lactation for women, medical conditions other than T2DM, and medications not directly related to T2DM/blood glucose management. Diagnostic criteria used were in accordance with the WHO for T2DM, Hb1Ac > 6.5% or at least two fasting plasma glucose measurements of >126 mg/dL [18]. BMI was used as a criterion for obesity status with a BMI ≥ 30 classified as obese. Participants’ BMI values were within the following range: 23–32.

The link was also sent to T2DM patients through the WhatsApp mobile application. Demographic information was collected via the same Google form. Of the total 108 participants, 58 (53.7%) were males and 50 (46.3%) females, producing a reasonably well-balanced sample. Inclusion criteria were adults (>18 years old), diagnosed with T2DM, non-smokers, no other medical condition, and no other medication unrelated to T2DM. Exclusion criteria were pregnancy and lactation for women, medical conditions other than T2DM, and medications not directly related to T2DM/blood glucose management. Diagnostic criteria used were in accordance with the WHO for T2DM, Hb1Ac > 6.5% or at least two fasting plasma glucose measurements of >126 mg/dL [18]. BMI was used as a criterion for obesity status with a BMI ≥ 30 classified as obese. Participants’ BMI values were within the following range: 23–32.

2.2. Data Collection and DEBQ Scales

Demographic information was acquired through a standard survey shared with the Facebook groups and through the Dutch Eating Behavior Questionnaire (DEBQ). The DEBQ developed by van Strien et al., is a validated questionnaire that documents 33 self-reported questions associated with distinct eating behaviors, assessing emotional, external, and restrained eating behaviors [17,19]. Emotional eating is defined as the eating behavior an individual exhibits in response to negative emotions [20]. Many studies on emotional eating behaviors have suggested that this eating behavior is related to health and weight problems [21]. External eating is described as the sensitivity to the food-related external cues/stimuli (such as the smell and appearance of food), with recent studies showing that external eaters may be more sensitive/responsive to food cues [22]. The third eating behavior pattern in the DEBQ, restrained eating, is described as how much and how long a person can stay away from food to control their weight [17]. According to the restrictive eating theory, individuals on a restrictive diet tend to eat more compared to those who are not. This leads to weight gain eventually through inducing food consumption [23].

In addition to the scales the DEBQ describes (i.e., restrained, emotional, and external eating behaviors), some of the 33 items belong to an unfitting response category (e.g., “Do you have a desire to eat when you feel bored or restless?”). These questions can be answered in addition to the standard five-point Likert scale with a “non-relevant” option as well, and are indicated in the relevant table below. The use of those types of questions is based on the premise that some individuals rarely experience a certain highly specific negative emotion, rarely eat much, or rarely gain weight (due to that specific emotion). The emotional eating subscale, for example, was associated with individual differences in reward response during negative moods according to functional neuroimaging studies [24]. Thus, there is variability in terms of the response to the same negative emotion. Therefore, by considering another dimension of responses, the DEBQ is more inclusive and comprehensive in its consideration of psychological drivers for dietary behavior. The items in the DEBQ (33-item list) are evaluated via a five-point Likert scale (1: never, 2: rarely, 3: sometimes, 4: often, 5: very often), while the items that belong to the “unfitting category of responses” also include the “non-relevant” option [17].

The Cronbach’s alpha internal consistency coefficients obtained in the original study of the DEBQ were used for the emotional eating behavior subscale: 0.95, the subscale of external eating behavior: 0.81, and the restricted eating behavior subscale: 0.95 [17].

2.3. Qualifiers of Considered Factors: Education Level, Marital Status, and Employment Status

At the time of enrollment in the study, the maximum education level attained was used as a qualifier for participants in the following manner: low: six years of schooling (basic); medium: high school, trade/vocational school or equivalent; high: college degree or higher. Marital status was categorized as follows: legally married, single and not in a relationship (i.e., not married/not in a relationship), other (divorced/widowed). None of the participants reported not being married but in a relationship. Employment status was considered as employed when a person was working either full-time (40 h per week) or over 20 h per week. All participants who identified themselves as employed stated that they were working full-time; those who reported that they were not working did not have any employment, or their employment was incidental, not regular, and well below what would constitute 50% of full-time employment, i.e., 20 h per week. Income ranges for Turkey were evaluated as previously described based on the official classification system used by the Turkish Ministry of Finances [25].

2.4. Data Analysis—Statistics

All data were analyzed through SPSS 20.0 statistical software. Explanatory factor analysis was used to determine the divergence of eating behaviors among the participants at different education and income levels in the survey. Explanatory factor analysis is used to reveal if there is a possible relationship between variables. In this study, basic components factor analysis varimax rotation is used for item-total correlation coefficients. With this technique, it has been determined which item belongs to which subscale. In the original DEBQ study, the varimax rotated factor matrix was used to correlate four factors (women, men, obese, and non-obese) with the questionnaire [17]. Normality tests examined normal distributions of the data. The results of skewness and kurtosis analysis were found in the assumed normal range. In the test of the reliability of the scale, the internal consistency coefficient, Cronbach’s alpha coefficient, was calculated. The Cronbach’s alpha internal consistency coefficients obtained in the original study of the DEBQ were found to be 0.95 for the emotional eating behavior subscale, 0.81 for the subscale of external eating behavior, and 0.95 for the restricted eating behavior subscale [17] and were used as benchmarks for our study. We calculated Cronbach’s alpha coefficient in our study and compared those values to the respective values reported in the original DEBQ study. We assessed the level of correlation between a characteristic and a type of eating behavior through r values. Specifically, r values in the 0.4–0.6 range were deemed low, while r values in the 0.6–0.8 range were considered strong, and r > 0.8 highly strong.

3. Results

Demographic Information Related to Study’s T2DM Participants

This study included a total of 108 T2DM patients, of whom 58 were men (53.7%) and 50 women (46.3%), all formally diagnosed with T2DM and classified in two age groups, one between 26 and 40 years (18.6%) and one >40 years old (81.4%). Characteristics of the participants regarding their age, body mass index (BMI), marital status, education level, income level, and employment status are provided in

Table 1. Distribution of T2DM patients participating in the study by demographic information.

	N	%
Sex
Women	50	46.3
Men	58	53.7
BMI (kg/m2) *
Normal weight (18.5–24.9)	15	13.9
Overweight (25–29.9)	49	45.4
Obese (≥30)	44	40.7
Age (years)
26–40	20	18.6
>40	88	81.4
Marital Status
Married	88	81.5
Single	13	12.0
Other	7	6.5
Education Level **
Low	27	24.9
Medium	35	32.4
High	46	42.6
Employment Status
Employed	60	55.6
Non-employed	48	44.4

* BMI according to WHO classification and units. ** Educational level classification: low: six years of schooling (basic); medium: high school, trade/vocational school or equivalent; high: college degree or higher.

below. Our study participants were well balanced between the two sexes (female and male). The BMI distribution was overwhelmingly in the overweight and obese categories (>86%), while the age of the participants was over 40 years old for >81% of our study population. Over 80% of our participants were married, and over 42% of our participants had attained higher education degree(s). Moreover, over 55% of our participants were employed (Table 1).

To evaluate the type of eating characterizing our participants’ eating behavior, the DEBQ was distributed. When all participants were assessed collectively, Cronbach’s alpha values were 0.88 for restrained eating, 0.96 for emotional eating, and 0.87 for external eating. Cronbach’s alpha value for all subscales considered combined was 0.909, which indicates that the DEBQ that is applied to type 2 diabetes patients is valid in terms of its significance (

Table 2. Cronbach’s alpha coefficients for DEBQ.

DEBQ Subscales	Cronbach’s Alpha Coefficient
Restrained eating	0.88
Emotional eating	0.96
External eating	0.87
All-subscales	0.91

). In the original study introducing the DEBQ tool [17], Cronbach’s alpha values for restrained, emotional, and external eating were 0.95, 0.94, and 0.80, respectively for the population initially tested [17].

One of the basic concepts related to factor analysis is the “Factor Matrix” or “Factor Loading”. The factor matrix shows the correlation between the subscales and the underlying factors. In our study herein, a rotated factor matrix was used to determine the factors that formed each subscale and the items under it. The factor loadings of the subscales obtained from the rotated factor matrix applied to the questionnaires from type 2 diabetes patients are given in

Table 3. DEBQ factor analysis and subscales items classification (all participants, N = 108).

	Factor/Eating Type
ITEMS	Factor 1 Restrained	Factor 2 Emotional	Factor 3 External
Restrained Eating
(1) If you have put on weight, do you eat less than you usually do? *	0.521
(2) Do you try to eat less at mealtimes than you would like to eat?	0.592
(3) How often do you refuse food or drink offered because you are concerned about your weight?	0.602
(4) Do you watch exactly what you eat?	0.533
(5) Do you deliberately eat foods that are slimming?	0.583
(6) When you have eaten too much, do you eat less than usual the following days? *	0.593
(7) Do you deliberately eat less in order not to become heavier?	0.693
(8) How often do you try not to eat between meals because you are watching your weight?	0.516
(9) How often in the evening do you try not to eat because you are watching your weight?	0.724
(10) Do you take your weight into account when you decide what to eat?	0.649
Emotional Eating
(11) Do you have the desire to eat when you are irritated? *		0.573
(12) Do you have a desire to eat when you have nothing to do? *		0.627
(13) Do you have a desire to eat when you are depressed or discouraged? *		0.718
(14) Do you have a desire to eat when you are feeling lonely? *		0.663
(15) Do you have a desire to eat when somebody lets you down? *		0.816
(16) Do you have a desire to eat when you are cross? *		0.836
(17) Do you have a desire to eat when you are approaching something unpleasant that will happen?		0.686
(18) Do you get the desire to eat when you are anxious, worried, or tense?		0.468
(19) Do you have a desire to eat when things are going against you or when things have gone wrong?		0.789
(20) Do you have a desire to eat when you are frightened? *		0.475
(21) Do you have a desire to eat when you are disappointed? *		0.742
(22) Do you have a desire to eat when you are emotionally upset? *		0.782
(23) Do you have a desire to eat when you are bored or restless? *		0.713
External Eating
(24) If food tastes good to you, do you eat more than usual?			0.717
(25) If food smells and looks good, do you eat more than usual?			0.635
(26) If you see or smell something delicious, do you have a desire to eat it?			0.575
(27) If you have something delicious to eat, do you eat it straight away?			0.520
(28) If you walk past the baker, do you have the desire to buy something delicious?			0.683
(29) If you walk past a snack bar or a cafe, do you have the desire to buy something delicious?			0.641
(30) If you see others eating, do you also have the desire to eat?			0.572
(31) Can you resist eating delicious foods? **			0.595
(32) Do you eat more than usual when you see others eating?			0.488
(33) When preparing a meal are you inclined to eat something?			0.598

* Items with a “non-relevant” response category in addition to the categories 1: never, 2: rarely, 3: sometimes, 4: often, and 5: very often [17]. ** For this item scoring has to be reversed [17]. The DEBQ was translated and modified accordingly for culturally appropriate use in Turkish language.

and provide a measure (intensity) of significance regarding correlation with a given eating type/behavior for a given question of the DEBQ.

The correlation coefficient analysis for BMI and sex is shown in

Table 4. Correlation coefficient (r values) analysis for BMI and sex (all participants, N = 108).

Subscales	Normal Weight	Overweight	Obese	Men	Women
	(N = 15)	(N = 49)	(N = 44)	(N = 58)	(N = 50)
Restrained eating	0.783	0.442	0.696	0.635	0.576
Emotional eating	0.833	0.777	0.685	0.644	0.768
External eating	0.852	0.711	0.701	0.598	0.632

. The correlation between the restrained eating subscale with normal weight patients showed a higher correlation compared to overweight and obese patients. Overweight patients demonstrated a moderately high (r = 0.4–0.6) restrictive eating correlation, while normal weight and obese patients exhibited a higher correlation (r = 0.6–0.8) for restrained eating compared to overweight counterparts. Regarding sex-related correlation, the rate was higher in men than women in restrictive eating behaviors. In women, a higher correlation with emotional eating behavior and external eating behavior compared to men was demonstrated (Table 4).

In terms of age as a factor determining the type of eating behavior,

Table 5. Correlation coefficient (r values) analysis for age (all participants, N = 108).

Subscales	Age 26–40 yrs (N = 20)	Age >40 yrs (N = 88)
Restrained eating	0.869	0.598
Emotional eating	0.852	0.704
External eating	0.754	0.600

shows the correlation coefficient analysis results for two different age groups. The 26–40-year-old age group exhibited a high correlation for both restrained eating and emotional eating behavior (r > 0.8) (Table 5). On the other hand, older participants in the age range of 40+ years old displayed a medium correlation for restrained eating and ahigh correlation for emotional eating (r = 0.6–0.8). Interestingly, both age groups showed a high correlation for external eating behavior (r = 0.6–0.8).

The effect of marital status on the determination of eating type was assessed in our participants. More specifically, it was observed that there is a correlation between marital status and eating behavior type (

Table 6. Correlation coefficient (r values) analysis for marital status (all participants, N = 108).

Marital Status
Subscales	Married (N = 88)	Single (N = 13)	Other (N = 7)
Restrained eating	0.592	0.763	0.888
Emotional eating	0.679	0.888	0.978
External eating	0.614	0.739	0.859

).

Participants identifying themselves as in the “other” marital status category demonstrated a significant difference compared to both married and single patients for all eating behavior categories (r > 0.8). More specifically, participants in the “other” marital status category show a very high correlation for each of the eating behaviors. Single participants exhibited a high correlation for all eating behaviors but were slightly lower when compared to the “other” marital status subscale (r = 0.6–0.8). Married participants were in the medium correlation level for all eating behaviors (r = 0.4–0.6) (Table 6).

Education level was shown to be correlated with the type of eating behavior in our participants. The correlation coefficient analysis for participants of different education levels revealed that participants with lower education levels exhibit high correlation (r > 0.8) for all forms of eating, more so compared to those with higher levels of education attained (

Table 7. Correlation coefficient (r values) analysis for education level (all participants, N = 108).

Education Level
Subscales	Low Education (N = 27)	Medium Education (N = 35)	High Education (N = 46)
Restrained eating	0.850	0.634	0.650
Emotional eating	0.867	0.811	0.715
External eating	0.945	0.691	0.543

). Patients with lower and medium education levels specifically demonstrated high correlation with emotional eating when compared to the higher education groups (r > 0.8). There was also a significant difference between patients with a lower education level and patients with a higher education level for external eating behaviors. External eating appears to be the least correlated with high education, while interestingly the most highly correlated eating behavior regardless of education level is consistently that of emotional eating. Moreover, lower education levels appear mostly correlated with external eating, which is the least strongly correlated with high education level (Table 7).

Employment status did not appear to correlate with a particular eating behavior. More specifically, both employed and non-employed participants demonstrated a medium correlation for restrained eating behavior (r = 0.4–0.6). Also, no difference was found in external eating behaviors when comparing employed and non-employed participants, while the correlation with such behavior was medium in both categories (r = 0.4–0.6). However, the level of correlation for emotional eating behavior was high for both employed and non-employed participants (r = 0.6–0.8).

Similarly to employment status, the income level did not seem to affect the eating behavior of our participants. Correlation coefficient analysis for our participants from different income levels revealed no statistically significant difference among groups when income was categorized as low, medium, and high. All participants demonstrated a high correlation for all subscales of the DEBQ (r = 0.6–0.8).

4. Discussion

Diabetes requires multidisciplinary treatment and mindfulness regarding daily lifestyle choices including diet. The American Diabetes Association (ADA), in addition to diet, includes approaches such as physical activity, medication, stress management, and hydration level monitoring to optimize treatment and management of the disease [26]. Diet, being a critical lifestyle component with a strong element of choice informing behavior, in addition to direct effects on diabetes management, can further modulate the effects of other lifestyle choices either positively or negatively for the individual diagnosed with T2DM. Compared to individuals without diabetes, T2DM patients have higher-level activation in brain regions as reported by Chechlacz and colleagues [27]. More specifically, a higher activation in the insula, orbitofrontal cortex, and basal ganglia was demonstrated in T2DM patients compared to controls when exposed to the same pictures of food [27]. Moreover, such increased activity observed in the insula and orbitofrontal cortex was correlated with external eating. Interestingly, higher responses observed within the subcortical brain regions (amygdala and basal ganglia) were correlated with emotional eating positively and negatively correlated with dietary self-care [27]. These results taken together demonstrate a strong impact of emotional state on dietary behavior in people with diabetes.

In this study, we aimed to determine how patients with diabetes in the Turkish setting vary their eating behaviors as per the DEBQ norms, considering the BMI, sex, age, marital status, employment status, and education level of our participants. The DEBQ has been specifically developed to identify restrained, emotional, and external eating behaviors by van Strien et al., while the translation of the scale/questionnaire to the Turkish language and the validation study in a Turkish population living in Turkey was previously performed on a sample of 508 university students, 346 females and 162 males [28].

BMI: In the original version of the DEBQ, questions were administered to normal weight and overweight participants [17]. After the data were collected, they were factor analyzed, with the resulting factor pattern used for revision of the DEBQ. These steps were repeated and the final version of the DEBQ was obtained. As a result of the Cronbach factor analysis, three factors were found in the original study. These factors were divided into three subscales as restrained eating, emotional eating, and external eating. In the original study, Cronbach’s alpha coefficients were 0.95 for the emotional eating behavior subscale, 0.81 for the subscale of external eating behavior, 0.95 for the restricted eating behavior subscale, and 0.90 for the entire scale [17]. In our study, Cronbach alpha coefficients were found to be 0.88 for restrained eating, 0.96 for emotional eating, 0.87 for restrained eating, and 0.91 for the whole scale. The latter set of values shows a similarly high reliability as the original study [17]. In this sense and given that the DEBQ is validated and deemed culturally appropriate for the Turkish population [28], it seems that the DEBQ is a reliable tool to assess the various types/categories of eating behaviors and assess how these are influenced by certain parameters in patients with diabetes in Turkey.

While there are numerous studies on eating behavior in T2DM patients indicating that overall (all factors combined) there is less association with external and emotional eating and more association with restrained eating behavior [29,30,31,32,33,34], in our study described herein, we found that overall, there is a higher association with emotional eating and less relation with external and restrained eating behavior in Turkish T2DM patients as per the DEBQ measurements.

Kargar et al. found that when the BMI increases, emotional eating behavior decreases in a population in Iran [35]. Similarly, in our study normal weight patients showed higher emotional eating correlation compared to obese patients. According to previous studies a lower BMI is associated with higher external eating behavior [35,36,37,38,39,40]. The results obtained from our study produce similar findings supporting this notion. Normal weight individuals may be more sensitive to external cues as their weight status plausibly does not significantly impact their eating pattern neither from a body image nor from a hormonal perspective.

Sex and Education: Several researchers have reported that emotional eating behavior is higher in women [41,42,43] and in individuals who have attained a higher education level [31]. Interestingly, in our study we observed that, in individuals with T2DM and a higher education level, emotional eating behavior was lower when compared to the other subscales. Possibly this is reflective of the different cultural and societal setting, heavily influenced by tradition and family structures as well as often religion, in the general Turkish population. These attributes may still be more influential overall and combined in Turkish society, given the fact that the country is undergoing epidemiological transition also affected by economic challenges. We have previously shown that traditional practices may influence dietary behavior independent of other factors [44]. Thus, in more traditional settings, whereby people tend to gather more and closely socialize in groups, external cues may be more prevalent for eating as well as restrictive eating. A higher levels of peer pressure or judgement can be applied as part of the communal approach that is arguably more prevalent and impactful in such traditional cultural settings.

Sex and Age: For restrained eating, the strongest correlation has been typically reported in women and younger individuals, and those with a higher education level [31]. In the original DEBQ study, obese individuals showed higher correlation with restrained eating behavior than normal weight individuals [17]. Other studies, however, demonstrate no difference between obese and normal weight individuals regarding restrained eating [45]. In our study here, for restrained eating, women exhibited lower correlation compared to men, while more educated participants showed lower correlation than those participants with a lower education level.

Marital Status: Regarding marital status, Wedin et al. reported, in the case of post-weight loss surgery, that married individuals exhibited a higher correlation for emotional eating behavior, while no information was reported on marital status and external eating or restrained eating [46]. In our study, we found that married participants displayed a weaker correlation for each subscale of the DEBQ. Patients from the “other” marital status category had a significantly higher correlation than their married counterparts, indicative of a potentially positive effect of marriage on the overall eating behaviors of participants. We have reported previously that traditional settings have shown to exert positive effects on healthy dietary choices, to some extent counteracting otherwise negative factors, which tend to induce unhealthy choices [44]. Further, a study on Korean immigrants in the US reported that family support had a positive impact on glucose regulation among patients with T2DM [47], illustrating the importance support systems play in the development of positive dietary habits.

Employment: To the best of our knowledge, there is no study to date investigating eating behaviors assessed through the DEBQ tool with employed and non-employed type 2 diabetes adult patients, conducted in Turkey. Emotional eating behavior appeared to have a high level of correlation regardless of employment status, thus emphasizing an overall need for targeted educational efforts and support systems addressing this eating behavior subscale.

Diabetes management constitutes a complex challenge to tackle; however, providing approaches and strategies that utilize positive modifications in diet, physical activity, and self-care can improve management effectiveness and reduce risk for complications [48,49]. While practically challenging, investigating the relationship between the eating behavior and the food consumption of patients with diabetes across varying demographics in future studies can further our understanding of T2DM and improve our capacity for better diabetes management via focused personalized recommendations.

The present study was limited to the DEBQ only and the food consumption records of the participants were not taken. However, the study was focused on eating behaviors of a particular subset of patients with diabetes, hence, the targeted recruitment of participants with a T2DM diagnosis via specialized Facebook groups. The engagement of patients with diabetes through specialized social media platforms constitutes a strength of the study based on a higher predisposition and arguably motivation of these participants in disease management, evident by their involvement with online groups focusing on their disease [50].

To the best of our knowledge, this is the first study utilizing the DEBQ in a Turkish urban population of adults with T2DM, considering various demographic characteristics including sex, BMI, age, marital status, education level, and employment status. Turkey is a country undergoing epidemiological and nutritional transition. In this region, secular and religious-derived customs and habits, including diet behavior, coexist, contributing to a significant prevalence and incidence of T2DM, thus rendering T2DM a significant public health concern. Additionally, the DEBQ translated in Turkish was administered as a Google form and distributed via a social media platform (Facebook) and a smartphone application (WhatsApp). These characteristics of the study constitute elements of novelty. The online version of the DEBQ offers a tool for identifying effectively emotional drivers of eating behaviors. For the idiosyncratic, in terms of customs and norms, and highly diverse Turkish population, this may be considered a suitable tool for emotional eating assessment, thus helping generate more relevant information for more effective prevention and/or management of T2DM.

5. Conclusions

Type 2 diabetes is a chronic disease where patients rely on management; hence, lifestyle and particularly dietary habits play a significant role. Our work aimed to assess the type of dietary eating among patients with T2DM in Turkey, an understudied topic. By understanding the main dietary behavior characteristics, the healthcare-providing team can identify the optimal approach to maximize effectiveness and efficiency in diabetes care and management. Our results showed that emotional eating, more so than external or restrained eating, is associated with T2DM patients in a Turkish population. Age and income did not appear to associate with a particular type of eating behavior. However, participants with lower education levels exhibit a high correlation for all forms of eating, more so compared to those with higher levels of education attained. Also, we documented that normal weight participants demonstrated higher emotional eating. Married participants exhibited a lower correlation with all categories of the DEBQ, while single participants showed higher correlation with all categories. This finding illustrates a potential protective effect of marriage in terms of eating behavior in patients with diabetes. Overall, these types of studies can be useful in countries such as Turkey, where the literature is limited, and the socioeconomic structures tend to be secular yet often highly traditional.