Introduction

Rapidly increasing use of smart phones, tablets, and computers has made internet an indispensable part in modern society. The negative impact of excessive, maladaptive or addictive internet use has attracted much research attention. In particular, internet addiction (IA) has become a major public health issue worldwide and brought about a dramatic proliferation of research in this area [1–7]. IA is defined as a pathological pattern of internet use, which is also described as internet dependence, compulsive internet use, problematic internet use, internet abuse, and pathological internet use [8]. The user cannot self-control the use of internet, resulting in significant impairments at school, home, work, health or interpersonal relationships [1]. They may find it difficult to stop using the internet due to its anonymity, convenience and accessibility and may use it as a way to escape reality [9]. The types of activity involved in IA include online gaming, social networking, online gambling, online shopping, virtual sex and information overload.

Internet addiction in young people is especially recognized as a social problem. Previous epidemiological studies using community self-report survey reported that the prevalence of IA in adolescents ranged from 38% to 90% [2, 3, 10–19]. Reasons for a wide range of prevalence rates may be related to different study designs, diverse assessment measures, or various diagnostic criteria, cultural backgrounds, and study samples. The conceptualization and definition of IA as a specific independent psychiatric disorder are still in debate by health professionals. One of the central criticisms is that IA was found to be highly co-morbid with other mental health conditions, such as attention-deficit hyperactivity disorder (ADHD) [20–22], depression [23], anxiety disorders [21], low self-esteem [24], impulsivity [25], social anxiety [26], and suicide. [27–28]. Moreover, some researchers have proposed to focus on addiction to particular internet “activity" rather than “internet” per se. For example, the American Psychiatric Association (APA) proposed the “Internet Gaming Disorder” in 2013 as a health condition for further study [29]. There is room for future studies of IA in order to build further evidence to inform the prevention and treatment of IA. However, few studies have considered personality factor concurrently in multiple regression analysis, especially personal trait of anxiety that may correlate with frequent internet use [21]. Most previous studies on IA centered on the young population, particularly the adolescents [7]. But internet use has made a great and wide influence on the society for different age groups. Therefore, the present study aimed to investigate the prevalence of IA and its psychosocial determinants as well as co-morbid psychiatric morbidities among internet users across different age groups.

Materials and Methods

Results

Discussion

Although the conceptions and definitions of IA differ in the literature, the associated psychosocial determinants found in this study were consistent with other studies. A few demographic factors were identified to have close relationship with IA. First, gender was frequently-mentioned risk factors of IA; significantly more males were identified as internet addicts than females [2–4, 12]. One of the reasons reported elsewhere was that males are more likely to play online gaming or gambling, engage in cybersex, and watch cyberporn; all of which were considered as highly associated with addictive internet use. Second, adolescents and young adults were more likely to become internet addicts as compared to other age groups [3]. It was suggested that adolescents are in the process of psychological development, thus less self-regulative, more susceptible to media influence and vulnerable to developing addictive behaviors. Third, students were more likely to be identified as IA in this study. Other than the acceptable internet use for school or job tasks, researchers indicated that college students with an insecure attachment are vulnerable to develop IA [3–12]. Specifically, the internet provides a certain psychological distance from the persons with whom insecurely attached students interact With little effort, people can make new friends and get immediate source of emotional support in the faceless cyberspace community, yet vulnerable youths are prone to addicted behavior. Thus, appropriate time for internet use and prevention of its life impairment should be emphasized among young adults, also certain key predictors of IA to be paid attention, including neurotic personality and psychological distress, as discussed below.

One of the most important psychosocial determinants for IA identified in this study was personality trait of neuroticism as assessed by the MPI- Neuroticism Scale (NS). This result was highly comparable to the findings of previous studies [2, 7, 39, 40]. Neuroticism is characterized by high levels of negative and unstable affect, such as feeling depressed, feeling restless, feeling guilty, mood swings as well as poor concentration or easy distraction. People with a higher NS score often experience negative affects when facing minor stressors. IA has been reported to closely link to stressful life events and psychological symptoms [41, 42]. Based on these characteristics, it was reasonable to infer the close association between neuroticism and IA. Our findings particularly pointed out significant associations between individual items of NS and IA, indicating that problems with concentration, distractibility during conversation, moodiness (e.g., feeling restless, mood swing and feeling guilty) and difficulty with sleep were important predictors of IA. Moreover, we noticed that co-morbidity of various psychiatric conditions with IA has been frequently reported. Our study demonstrated that individuals with IA had a significant higher rate of psychiatric morbidity (65.0%) in comparison with those without IA (32.5%), supporting by the significant correlation between CIAS-R and BSRS-5 scores. Considering the time reference of the psychometrics measured, as shown in Fig A in S1 Fig, path analysis of SEM model on IA revealed that neuroticism was the major contributing predictor. This could directly influence the severity of CIAS-R as well as indirectly contributed to the severity of psychopathology (BSRS-5 score), excessive internet use and severity of life impairment, finally contributing to the prediction of CIAS-R score.

It is evident that there are much in common for the aforementioned significant predictors for IA and the symptoms listed in the DSM5 criteria of ADHD (e.g., distractibility, difficult focusing, and restlessness) [29]. Previous reports stated a close link between IA and ADHD or ADHD symptoms [4, 6, 29]. Among the manifesting ADHD symptoms, attention deficit was reported as the most associated factors with IA [4, 6]. This study found both concentration and attention were specifically important factors related to IA severity. The neuroticism factor is a long-term stable trait; the onset of ADHD often begins before age of seven. Possibly, they share some similar mechanism to contribute to the development of IA. According to a follow-up study on ADHD, it was found that the adolescents with ADHD were more likely to develop IA in two years [43]. Another follow-up study also found that a higher proportion of the students with a higher NS score developed IA during the 2 years of follow-up in comparison to the control group [39]. These similar findings highly suggest that deficits of concentration and attention as well as moodiness or trouble with sleep were strong predictors for development of IA. Most previous reports just focused on neuroticism in general and did not mention about the roles of concentration or attention and moodiness in the development of IA, which makes one of the advantages of this study.

In terms of internet use habits, the findings showed that life impairment and increased amount of time online were important independent predictors for IA next to neuroticism. The items related to these two factors constituted a major part of CIAS-R in the subscales of "compulsive use" and "interpersonal and health problems" (12 items). Overuse of internet might also consequently create problems with time management and work together to induce negative and significant life impairment. Thus, these two factors could be causes or manifestations of IA. In addition, engagement in online gaming was particularly found to be highly correlated with the presence of IA comparing to other types of online activities like social networking. Therefore, our findings supported the concept that IA should focus on specific online activities engaged by internet users, instead of internet per se. For example, the term “internet gaming disorder” proposed by the APA in DSM-5 clearly points out “online gaming” as the media of the disordered behavior [29].

More recently, the issue of suicide among internet addicts attracted much attention in the research fields [44–47]. In this study, subjects with IA were more likely to have suicide ideation and attempts. Given that IA has been found to be highly associated with neurotic traits, psychiatric morbidity, and life impairments, all of these factors could increase suicide risks. High rates of psychiatric morbidity and suicide risks among the subjects with IA indicated that earlier identification of high risk individuals and provision of timely and pertinent management are necessary for suicide prevention and mental health promotion for the internet users. Results in this respect provide the basis for interested researchers to investigate its mechanisms and prevention for suicide behaviors among internet addicts.

In summary, the estimated prevalence of IA (10.6%) in this study is close to the worldwide reported online survey data (4–10%) for all-age internet users [48]. Our findings revealed the prevalence and severity of IA were significantly associated with several psychosocial determinants. We postulated that neurotic personality was a major predisposing factor, reacting to stressful life events as a precipitating factor to cause psychological distress. Under triggers of life events, such personality could eventually cause maladaptive coping behaviors through excessive engagement in specific internet activities, which resulting in internet addiction given significant life impairment. All these factors interacted together and progressed to internet addiction.

In this online survey of internet use and health, the findings should be interpreted under several limitations. Firstly, the female predominance of this sample may affect result interpretations and limit its generalization to female online users. However, this feature also supports recent findings that reveal the gender differences in online health information searching behaviors [49], reflecting the strong female social motive for health-related online activities in the general public. So the female-dominant sample should not be a serious problem to this study. Secondly, there were more middle-to-old-aged general public rather than youth participants in this survey, limiting the comparability between ours and previous ones; however, this can also be the advantage of this study due to its unique age constitution. Our findings filled the gap beyond prior findings focusing on IA in young people. Thirdly, generalization of the findings is confined to frequent internet users or those who are interested in the research topic because we did not have control groups of those non-responders to this online questionnaire or comparable participants for comparison. Last but not the least, causal effects between the study factors and internet addiction may not be confirmed based on our findings due to cross-sectional study design. Future studies are suggested to combine solid and acceptable diagnostic criteria as well as longitudinal design to make comprehensive and integrative formulation to the development of IA treatment and preventive strategies.

Supporting Information

Acknowledgments

The authors were thankful to the statistics consultant of the Taiwan Suicide Prevention Center, Dr Jin-hua Chen, for her opinions toward data analysis and result interpretations. We would also like to acknowledge the assistants at the Center for their administrative and practical help throughout the study procedures. The study was supported financially by the Ministry of Health and Welfare (DOH102B3030).