Internet Addiction and its Relationship with Attachment Styles Among Tunisian Medical Students

Internet addiction is a growing addictive behavior and a major public health problem worldwide. Several psychological factors can contribute to the problematic use of the Internet. This study aimed to determine the prevalence of Internet Addiction (IA) in a sample of university students and to examine the relationship between IA, self-esteem, and attachment styles. The present study was a cross-sectional study, involving 135 students from the Faculty of Medicine of Monastir. The participants completed a questionnaire, which contained the socio-demographic data, the reasons for Internet use, the Young Cyberaddiction scale to seek IA, the Relationships-style-questionnaire-RSQ to assess attachment style, and the Rosenberg Self-Esteem Scale (RSES). The mean age of the students was 21.5 ± 1.9 years old. They were 112 (83%) females. The prevalence of Internet addiction was 23.7%. The average connection time was 3.5 ± 1.8 hours/day. The most frequent online activities were chat (online discussion, forums ...) in 94.8% of cases followed by download activities (78.5%), scientiﬁ c research (75.6%), online games (23.7%), and online shopping (13.3%). IA was associated with school failure, alcohol use, online gaming, and low self-esteem. The majority of the participants (84.4%) in this study reported an insecure attachment style. Logistic regression analysis showed a strong association between IA and fearful attachment style. IA was frequent among students. Fearful attachment style was found to be a risk factor for IA. This study highlights the impact of relationships between child and their caregivers on the development of addiction.


Introduction
The Internet has become an integral part of everyday life all over the world.Widespread access to the Internet has given people quick access to online information and has offered new opportunities for entertainment and social communication [1,2].Although the Internet offers many advantages, excessive Internet use may increase the likelihood of problematic Internet use or Internet Addiction (IA).
The mean age of the students was 21.5 ± 1.9 years old.They were 112 (83%) females.The prevalence of Internet addiction was 23.7%.The average connection time was 3.5 ± 1.8 hours/day.The most frequent online activities were chat (online discussion, forums ...) in 94.8% of cases followed by download activities (78.5%), scientifi c research (75.6%), online games (23.7%), and online shopping (13.3%).IA was associated with school failure, alcohol use, online gaming, and low self-esteem.The majority of the participants (84.4%) in this study reported an insecure attachment style.Logistic regression analysis showed a strong association between IA and fearful attachment style.
IA was frequent among students.Fearful attachment style was found to be a risk factor for IA.This study highlights the impact of relationships between child and their caregivers on the development of addiction.
Previous studies [2,6,7] have shown that students are more likely than the general population to use the Internet and to develop IA.A meta-analysis [2] that examined the prevalence of IA among medical students in different countries found that the prevalence of IA among 3651 medical students was 30.1%.Medical students would be more vulnerable to addiction given that they frequently use the Internet to search for medical information [2,7].Additionally, Internet use may help cope with stressors and negative emotions [2].On the other hand, excessive Internet use was associated with more depressive symptoms, a decline in social relationships, and low selfesteem [6,[8][9][10].
Although it has not been formally codi ied in the 5th edition of the Diagnostic and Statistical Manual of Mental Disorders https://doi.org/10.29328/journal.jatr.1001027(DSM5), Internet Addiction (IA) has become a worldwide issue.
Internet use and addiction are in luenced by several factors, among which attachment style is interesting to study [5,11].Attachment theory was originally developed by Bowlby [12], who de ined attachment as the irst and most important relationship that children have with their caregivers.He worked on the attachment of children to caregivers and extended the principles of parent-child attachment to close personal relationships in adulthood [13].Bartholomew and Horowitz [14] proposed a four-item model based on secure, preoccupied, dismissing, and fearful attachment.Secure individuals are comfortable with themselves and their partners in their relationships [14].They have high regard for the self and others in relationships.Preoccupied individuals need constant reassurance and affection from others but they attempt to control the relationship because they are anxious that their partner may reject them.A preoccupied person tends to be high in anxiety but low in avoidance [14].Dismissing individuals have a positive perception of the self but more negative perceptions of others [14].So that, they are more independent and do not prioritize close relationships.Fearful individuals are uncomfortable in close relationships because they are worried about being hurt by others [14].They are anxious and, because they lack assertiveness, tend to avoid relationship issues [15].By studying the relationship between attachment styles and IA, some studies have identi ied a negative relationship between IA and secure attachment.The results are inconclusive concerning the contribution of different attachment styles to this addiction.
From this perspective, we proposed through this work to evaluate the prevalence of IA in a Tunisian student population and to examine the relationship between attachment styles and IA.

Study design
A cross-sectional, questionnaire-based study was conducted among 135 students of the faculty of medicine of Monastir, Tunisia.An anonymous self-questionnaire was distributed after a presentation of the main objectives of the study.The questionnaire consisted of (1) the sociodemographic data, (2) the reasons for Internet use, (3) the Young cyberaddiction scale (4) the Relationships-stylequestionnaire-RSQ, and (5) the Rosenberg Self-Esteem Scale.

Measures
Young's Diagnostic Questionnaire (YDQ): It is an 8-item questionnaire adapted from the diagnostic criteria of pathological gambling in the DSM-IV used to evaluate IA.We used its French version [16].
A person with a score of 5 or more measured by the YDQ during 6 months is classi ied as Internet dependent.
Rosenberg Self-Esteem Scale (RSES): The Rosenberg Self-Esteem Scale [18] is a measure of global self-regard, consisting of 10 items that were measured on a 4-point scale.A total self-esteem score ranged from 10 to 40 with higher scores indicating higher self-esteem.A score lower than 25 suggests low self-esteem [19].

Statistical analyses
Data analysis was performed using SPSS version 20.Descriptive analysis was used to describe the demographic and Internet-use characteristics of the participants, the prevalence of IA, self-esteem, and different types of attachment styles.
The proportions were compared by the "chi2" test or by the exact "Fisher" test.
The Student test was used to compare two means of two independent samples.The signi icance level was set at 0.05.
After the univariate analysis, a logistic regression model was performed to identify the factors independently associated with Internet addiction.The variables included in the model were the signi icant ones at the statistical threshold of 0.25.
Ethics: All subjects were informed about the study and provided informed consent.All subjects were over 18 years of age.

Socio-demographic characteristics
The mean age of the students was 21.5 ± 1.9 years old.They were 112 (83%) females (Table 1).
The majority of the students (98.5%) were single.Nearly half (57%) were enrolled in the second academic year and 11.1% had experienced a school failure.Almost half of the students (48.1%) lived with their families, 41% lived with pairs; 10.4% of the students were followed for chronic diseases; 3.7% reported having psychiatric antecedents.Students reported smoking in 4.4% (n = 6), alcohol use in 4.4% (n = 6), and cannabis use in 1.5% (n = 2) of cases.
All the students have used Facebook.Students have used Facebook to discuss with their friends (68.9%), to meet new friends (9.6%), to look at their friends' photos (48.9%), to follow the news (67.4%), to play on Facebook applications (14.1%), to listen to music (31.9%), and to work (15.6%).
Nearly a quarter of participants (24.4%) have developed romantic relationships through Facebook.
According to the Young Scale, 23.7% of the students were addicted to the Internet.

Attachment styles
Based on attachment theory, 50.4% of the respondents had a dismissive attachment style, 21.5% preoccupied style, 15.6% secure style, and 12.6% fearful style.

Self-esteem
The average RSES score of the study sample was 30.23 ± 5.2; 18.5% of the participants had low self-esteem.

Factors associated with internet addiction (Table 2)
Internet addiction was associated with school failure (p = 0.048) and alcohol use (p = 0.028).It was also associated with low self-esteem (p < 0.001).
The insecure attachment was associated with IA.Participants with IA had higher scores for fearful attachment (p = 0.04) and lower scores for secure attachment (p = 0.03).

Use of internet and attachment styles (Table 3)
By examining the relationship between the different styles of attachment and the activities practiced on the Internet, it appeared that the secure attachment style was associated with establishing romantic relationships (p < 0.001) and chatting (p = 0.02) on Facebook, and making online purchases (p = 0.03).Preoccupied attachment style was associated with checking friends' pro iles and photo albums (p = 0.005) and with knowing new people through Facebook (p = 0.03); fearful attachment style was associated with knowing new people through Facebook (p = 0.038) and listening to music (p = 0.011).We found a negative association between fearful attachment style and chatting on Facebook (p = 0.03).

Logistic regression analysis
To eliminate confounding factors and determine the variables independently associated with IA, we used a multivariate regression model.The variables included in the model were Internet addiction as the dependent variable and sex, smoking, alcohol use, time spent on the Internet, using the internet to know new people through Facebook, establishing a romantic relationship through Facebook, using the Internet to play online, self-esteem, school failure, secure attachment style score, preoccupied attachment style score and fearful attachment style score as covariates.The multivariate regression analysis (Table 4) showed an association between

Discussion
The present study sought to determine the prevalence of IA among a sample of Tunisian medical students and to investigate the association between Internet use and attachment styles.
In this study, the prevalence of IA was 23.7%.Different studies have reported different results for the prevalence of IA ranging from 3.2% to 30% (Table 5).
The differences in reported rates may be explained by the different tools used in the diagnosis of IA and by the fact that studies have been conducted in different populations and at different times [9].According to Byun, et al. [23], this broad range of prevalence rates is due to the lack of a comprehensive standardized de inition covering all ages, genres, and levels of education.The prevalence of IA in this student sample appears to be situated among the highest igures in the literature.Tunisia was the irst Arab and African country to connect to the web, as early as 1991.Facebook experienced a real fashion effect since right after "the revolution", the number of users increased from 2 million in 2010 to 4.6 million in 2014 [24].
In this study, there was no difference between women and men in terms of the level of IA.Many studies have reported that, for social and cultural reasons, IA is more frequent among men than women [9].The lack of a signi icant association in this study between gender and the level of IA may be related to a biased sample of predominantly females.
We found that school failure was associated with IA (odds ratio [OR] = 13.15;con idence interval [CI] = [10.55,15.73]; p = 0.001).International studies have revealed that pathological Internet gamblers were more likely to have poorer academic results [22].
In this study, alcohol use was more often reported with students addicted to the Internet (p = 0.028).Shapira, et al. [25] found that 10% of subjects with IA had alcohol dependence.Several studies have reported an association between IA and substance abuse [6,25].
The time spent on the Internet is an important factor in addiction in this study; the prevalence of IA was higher among students who used the Internet more than 4 hours daily (p = 0.017).It was demonstrated that the level of IA increases with the increasing duration of Internet use [9].
The most practiced online activities for the students in this study were online chat (forums, Facebook…) (94.8%) followed by download activities (78.5%), scienti ic research (75.6%), online gaming (23.7%), and online shopping (13.3%).Online gaming was a signi icant risk factor for IA.Online gaming was similar to substance-related addiction [26].Engaging in online gaming was associated with openness to experience [26]; it led to spending large amounts of time online in search of a rewarding experience.According to Van Rooij, et al. [27], online gaming is the only Internet activity that combines a distinct reward structure, and social interactions in a virtual environment.Many studies [21,22,26,27] suggested that speci ic online activities such as online gaming and online social networking had high addictive potential.Kuss, et al. [21,26] indicated that frequent usage of social online applications especially Facebook and online forums was a strong risk factor for IA.All of the students in this study had a Facebook account.The use of Facebook in Tunisia increased and reached 42.1% of the Tunisian population in 2014 [24].Tunisia ranked 47 th place worldwide in terms of the number of Facebook accounts in February 2013 [28].
Establishing romantic relationships through the Internet seems to be a protective factor against IA.In this study, online romantic relationships were negatively associated with IA (p < 0.001) and insecure attachment style (p < 0.001).We suggested that students, which have established romantic relationships, were more comfortable with social interactions.Extraverts and agreeable people tend to be more satis ied in their close relationships [29].Kuss, et al. [21] reported that agreeableness, extraversion, and emotional stability were negatively associated with IA.
Students addicted to the Internet had lower self-esteem scores (p < 0.001).Results of logistic regression revealed that self-esteem was a predictor of IA ([OR] = 0.83; [CI] = [0.752,0.920]; p < 0.001).This result is corroborated by the literature [30,31]; it was found that self-esteem was a strong risk factor for IA and the amount of time spent online per week.Students with low self-esteem and who have coping dif iculties in their lives away from home might ind the Internet a means of escape [8,32].
Previous studies [33][34][35] showed that attachment style in luences personality traits and different types of behaviors such as romantic love, friendships, social skills, and addictive behaviors.Research on addictive behaviors showed that insecure attachment is associated not only with alcohol addiction and other substance use but also with behavioral addictions such as IA [11].This study revealed that secure attachment style was a protective factor against developing signs of IA ([OR] = 0.21; [CI] = [0.085,0.558]; p = 0.002).This inding was consistent with the results found in the literature [34,36].Ghasempour and Mahmoodi-Aghdam [37] concluded that individuals with a secure attachment style are comfortable with interpersonal interactions, have higher self-esteem, act with more functional coping strategies to face obstacles, and therefore are less likely to turn to addictive behaviors.Insecure attachment is a predictor of IA [36]; addiction can be considered as an adaptive strategy to discharge psychic tension; The Internet, especially social networks, can satisfy the need for security and belonging of insecurely attached subjects [38].Additionally, the Internet may provide them with distant relationships to communicate with others with less emotional stress in the faceless virtual space [22].
Research in the neurobiology of attachment supported the relationship between insecure attachment and addiction and proposed that oxytocin increases resilience against addiction and stress effects by facilitating the processing of social attachment.Low oxytocin levels could be a key predisposing factor for the emergence of IA [39].
The majority of the participants (84.4%) in this study reported an insecure attachment style.Insecure attachment style prevalence was higher than in the general population (49%) [40].
Several studies have tried to explain the role played by attachment styles in IA.However, the speci ic types of attachment associated with IA were different among the studies.Some authors reported that both avoidant and anxious attachment styles explain problematic Internet use [11,35], while others found that attachment anxiety was the only signi icant factor [34,41], and still others found that attachment avoidance was the only signi icant factor [42].
The majority of studies have examined the relationship between IA and attachment styles using a model in 3 styles of attachment: secure, anxious, and avoidant, but they did not discuss a category of people who might exhibit both high avoidance and high anxiety.This category was described by Bartholomew & Horowitz [14] and labeled fearful.The results of this study showed that fearful attachment style was signi icantly associated with IA (p = 0.04).Fearful subjects, who have negative views of both the self and others [14], may present the highest level of interpersonal dif iculties and social avoidance because they see others are unreliable and will turn to virtual life and relations which do not require emotional investing.Furthermore, it was shown that social isolation, loneliness, and depression were associated with addiction to communication technologies such as the Internet [9,31,37,43].It has been found a relationship between depressive symptoms and avoidant attachment style [37].
By exploring styles of attachment according to the different activities on the Internet, we found a negative association between fearful attachment and chatting on Facebook, while this style of attachment was associated with connecting on Facebook to ind new friends.The fearfulavoidant attachment style may be one of the most complicated styles.It is characterized by high avoidance of relationships and fears of abandonment and proximity, with, on the other hand, a strong desire to be in a relationship.We noticed that students with this kind of attachment were the most likely to spend time on the Internet looking for making new friends through social networks but they would have less chance to interact with other individuals [10].It is reasonable to assume that subjects with fearful attachments are uncomfortable with chat functions.
Because people with fearful attachments are both anxious and avoidant, Facebook may provide them the perfect opportunity to monitor their partner and perceived relational threats passively without having to interact with or confront him or her directly [15].Due to their tendency to keep their distance from others, they are attracted to anonymity, less risk, and less responsibility characterizing Internet interactions [35].
In this study, we did not ind an association between IA and preoccupied attachment style.However, this style of attachment was signi icantly associated with meeting online acquaintances.According to attachment theory, the preoccupied style involves low avoidance but high anxiety and is related to an individual's fear of rejection and abandonment in relationships [42].Through social networking sites, they seek approval and reassurance from others [35].People with a preoccupied attachment style might feel more control and closeness by using Facebook [15].It was demonstrated that these subjects have less anxiety about rejection and abundance through Facebook and perceive themselves then more able to initiate virtual relationships [33].

Limitations
The indings should be interpreted under several limitations.First, the sample was small; second, the present study is limited to adolescents attending the Faculty of Medicine of Monastir; third, this sample was characterized by a strong female predominance; and inally, this study employed a cross-sectional rather than a longitudinal design and therefore has inherent limitations for determining causal relations.

Conclusion
The present research aimed to evaluate the prevalence of IA and to examine the relationship between attachment styles, and IA in students.
Our results demonstrated a high prevalence of IA and a high prevalence of insecure attachment styles in Tunisian students.https://doi.org/10.29328/journal.jatr.1001027 We found a negative association between IA and secure attachment.Speci ically, subjects with secure attachment establish better relationships with others, are more con ident, and are less to turn to addictive behaviors.Findings showed that there was an association between fearful style and IA.Those subjects, who tend to avoid social interactions, spend more time on the Internet but they are not the most users of chat functions.
The presented results suggest the potentially high importance of attachment style and more precisely the in luence of caregiver-child interactions in the emergence of addictive behaviors.

Compliance with ethical standards
We have no con lict of interest to declare.All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/ or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.
Informed consent was obtained from all individual participants included in the study.

Table 1 :
Age and sex of total sample and subsamples of not-addicted and addicted students.

Table 2 :
Internet use, substance use, and questionnaire scores among not-addicted and addicted students.

Table 3 :
Internet use among secure and insecure attachment style.

Table 4 :
Logistic regression results for IA.

Table 5 :
Prevalence of IA among students.