Problematic Smartphone Use Is Associated With Current and Previous Major Depressive Disorder

Abstract

Objective

Although many studies have demonstrated the association between depression and problematic smartphone use (PSU), temporal precedence of this relationship remains controversial. This study aimed to investigate the relationship between present PSU and both past and current major depressive disorder (MDD) among adults.

Methods

We utilized data from the Korean Epidemiologic Catchment Area study, a nationwide epidemiologic survey that employed a multistage cluster sampling method. Participants were assessed using the Korean version of the Composite International Diagnostic Interview to diagnose MDD and the Smartphone Overuse Screening Questionnaire to identify PSU.

Results

Of a total of 916 subjects, 139 (15.2%) belonged to the PSU group. The prevalence of PSU in older adults was substantial: 9.3% in the age group of 36–50 years and 5.4% in the age group of over 50 years. We found a significant association between MDD and PSU. Both current and previous MDD episodes were associated with an increased risk of PSU.

Conclusion

This study confirmed the association of PSU and depression among general adult population and the precedence of MDD to PSU. Even previously resolved depression was found to affect current PSU.

INTRODUCTION

Recently, smartphones have become indispensable in modern society worldwide, particularly in developed countries. In 2023, worldwide smartphone shipments totaled 1.17 billion units [1] and 69% of the global population was estimated to have access to smartphones [2]. The widespread use of smartphones with their penetration into daily life has raised concerns about problematic smartphone use (PSU). PSU refers to a dysfunctional and pathological pattern of smartphone usage [3,4] that results in significant problems in daily life [5]. It is often accompanied by symptoms of dependence. While some researchers argue that there is insufficient evidence to consider PSU as a genuine addiction [6,7] and suggest that it should be considered as a separate category with less severe symptoms or different aspects [7], others insist on approaching it within the framework of behavioral addiction [5,8]. Many studies have defined PSU and smartphone addiction similarly as “excessive or uncontrolled use of smartphones.” Thus, many researchers have mixed concepts of PSU and smartphone addiction [8,9].

Many scales for evaluating PSU assess salience, tolerance, withdrawal, and conflict [10]. These are also included in Griffiths’ classic criteria for behavioral addictions (salience, mood modification, tolerance, withdrawal, conflict, and relapse) [11]. Some researchers have proposed that smartphone addiction should be understood as part of a continuum of problematic behaviors [12]. When the severity of dysregulated behavior exceeds a certain threshold, it can be classified as an addiction.

PSU has not yet been officially recognized as a diagnosis in the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition, Text Revision (DSM-5-TR) or the International Classification of Diseases, 11th Revision. The underlying pathology of PSU remains unclear. However, PSU is gaining recognition as a significant public health concern due to its association with various negative outcomes. PSU is known to increase the risk of traffic accidents and pedestrian incidents [13], possibly due to distraction. PSU can also lead to physical health problems (such as musculoskeletal pain [14], headaches, fatigue [15], and sleep disturbances [9]) and mental adversities (such as stress, anxiety, and depression [3-5,8,9,16-18]). In particular, the association between PSU and depression has been repeatedly reported [3-5,9].

However, many researchers have stated that current evidence cannot establish a causal relationship between PSU and depression [19], let alone an antecedent relationship [5]. Most previous studies have utilized cross-sectional designs, making it difficult to determine temporal precedence [4,8,9,20]. Several longitudinal studies have reported conflicting results [21-25]. Previous reports also have several limitations that hinder generalization. For example, many studies have primarily focused on the young population, limiting the applicability of findings to other age groups. Furthermore, there is a lack of studies examining the potential impact of previously resolved depression on subsequent PSU.

Thus, the aim of this study was to investigate the relationship between major depressive disorder (MDD) and PSU. Specifically, this study intended to elucidate the antecedent relationship between PSU and major depressive episodes by analyzing the association of present PSU with both past and current depressive episodes.

METHODS

Participants and procedure

This study utilized data from the Korean Epidemiologic Catchment Area (KECA) study conducted between April and November 2016. The KECA study is a nationwide epidemiologic survey that examines the prevalence, associated sociodemographic factors, and comorbidities of major psychiatric disorders among community-dwelling adults. The survey has been conducted every 5 years since 2001 by the Ministry of Health and Welfare of South Korea. To specify participants for the survey, a multi-stage cluster sampling method was employed in reference to the 2010 national census of population and housing data. Sampling units were selected from 21 catchment areas of 16 regions based on metropolitan cities and provinces. Households were randomly selected within these sampling units. The number of participants from each region was determined proportionally to the population of that region [26]. Trained investigators visited selected households and interviewed one person aged 18 years or above per household using the last birthday method, which involved selecting the individual with the most recent birthday.

Out of a total of 5,102 participants in the national survey, 2,508 subjects from randomly selected region were asked whether they had used a smartphone for more than three hours per week in the past month. Among them, 959 participants who answered “yes” completed the questionnaire assessing PSU. We excluded 20 subjects who had missing responses for depressive episodes and 23 subjects who had missing responses for the PSU scale from the analysis.

Assessment

The Korean version of the Composite International Diagnostic Interview (CIDI) version 2.1 was utilized for diagnosing MDD. The CIDI is a comprehensive and fully standardized interview tool designed for non-experts to assess and diagnose psychiatric disorders. The Korean version of CIDI was developed according to the guidelines provided by the WHO. It has been proven to be valid and reliable [27]. Participants were required to answer if they had experienced symptoms that met the diagnostic criteria outlined in the DSM-IV. They were also asked about when they experienced those symptoms and when they initially and most recently met the diagnostic criteria. Based on their responses, we determined the diagnosis of MDD for both lifetime and the last month. Based on the timeline of depression diagnosis, participants were categorized into three groups: no depression (ND) for those who had never experienced MDD, past depression (PD) for those who had a history of MDD but did not experience it in the last month, and current depression (CD) for those who had depressive symptoms in the recent month.

The Smartphone Overuse Screening Questionnaire (SOS-Q) was used to identify PSU. The SOS-Q is a self-rated scale consisting of 28 items regarding six factors: preoccupation, loss of control, craving, insight, overuse, and neglect of other areas. Participants indicated the frequency of experiencing each statement regarding smartphone use in the previous month. Responses were scored on a scale of 0 to 3 (0: not at all, 1: sometimes, 2: often, 3: always). A total score of 49 or higher was classified as PSU. It demonstrates high sensitivity (0.814) and specificity (0.861). The SOS-Q exhibited excellent internal consistency (Cronbach’s alpha=0.95) and validity, as indicated by an area under the curve value of 0.877 in Korean adults [28].

Statistical analysis

Sociodemographic variables were compared between the PSU group and the control group using chi-square tests. Binary logistic regression was then performed to examine risk factors for PSU, including depressive history and other sociodemographic factors. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated for all variables. To explore whether the pattern of PSU varied based on the history of depression, we compared average scores for each of the six factors (preoccupation, loss of control, craving, insight, overuse, and neglect of other areas) comprising the SOS-Q across ND, PD, and CD groups. A p-value of less than 0.05 was considered statistically significant. All statistical analyses were conducted using IBM SPSS Statistics for Windows, version 24 (IBM Corp.).

Ethics

All procedures of this study were conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board (IRB) of Seoul National University College of Medicine (IRB No. C-1104-092-359). All participants were fully informed about this study. Informed consent was obtained from each participant prior to participation.

RESULTS

Of 916 subjects, 139 (15.2%) belonged to the PSU group. Table 1 presents sociodemographic characteristics of PSU and non-PSU groups. Gender, age, marriage status, employment status, and history of depression exhibited significant associations with PSU. The prevalence of PSU was higher in women, single individuals, and the unemployed. PSU was more prevalent in individuals under 35 years old (21.9%) than in the 36–50 years age group (9.3%) and the over 50 years age group (5.4%). The prevalence of PSU was also higher in those with a history of depression. Specifically, 62.5% of the CD group and 35.4% of the PD group had PSU, while 13.2% of the ND group exhibited PSU (Table 1).

Table 1.

Socio-demographics factors of subjects with or without PSU

	No PSU	PSU	p
Gender*			0.001**
Man	346 (89.6)	40 (10.4)
Woman	431 (81.3)	99 (18.7)
Age* (yr)			<0.001***
≤35	363 (78.1)	102 (21.9)
36–50	291 (90.7)	30 (9.3)
>50	123 (94.6)	7 (5.4)
Marriage status			<0.001***
Married	437 (89.0)	54 (11.0)
Single†	340 (80.0)	85 (20.0)
Employment status			0.001**
Employed‡	344 (89.4)	41 (10.6)
Unemployed	433 (81.5)	98 (18.5)
Education			0.057
Graduate	554 (83.4)	110 (16.6)
High school or less	223 (88.5)	29 (11.5)
Income level ($)			0.363
<2,000	185 (82.2)	40 (17.8)
2,000–4,000	257 (86.5)	40 (13.5)
>4,000	281 (85.7)	47 (14.3)
History of depression			<0.001***
Current depression	3 (37.5)	5 (62.5)
Past depression	42 (64.6)	23 (35.4)
No depression	732 (86.8)	111 (13.2)
Total	777 (84.8)	139 (15.2)

Values are presented as number (%). The prevalence of PSU was higher in women, single individuals, and the unemployed. PSU was most prevalent in individuals under 35 years old. The prevalence of PSU was also higher in those with a history of depression.

^*p<0.05;

^**p<0.01;

^***p<0.001;

^†single: not married, divorced, bereft, living independently;

^‡employed: full time or part time, unemployed: student, housewife, unemployed. PSU, problematic smartphone use

Binary logistic regression analysis revealed that PSU was associated with woman gender (OR: 2.350, 95% CI: 1.507–3.662) and age under 35 years (OR: 4.719, 95% CI: 1.875–11.875). After adjusting for sociodemographic factors (gender, age, marriage status, employment status, education, and income level), depression was significantly associated with PSU. The CD group had a higher likelihood of PSU (OR: 6.366, 95% CI: 1.253–32.353) than the ND group. The PD group also exhibited a greater risk of PSU (OR: 2.826, 95% CI: 1.528–5.227) than the ND group (Table 2).

Table 2.

Risk factors of PSU

	OR (95% CI)	p
Gender
Man	1
Woman	2.350 (1.507–3.662)	<0.001***
Age (yr)
≤35	4.719 (1.875–11.875)	0.001**
36–50	1.741 (0.685–4.424)	0.244
>50	1
Marriage status
Married	1
Single†	1.218 (0.740–2.003)	0.438
Employment status
Employed‡	1
Unemployed	1.220 (0.792–1.879)	0.366
Education
Graduate	1
High school or less	0.957 (0.572–1.600)	0.866
Income level ($)
<2,000	0.952 (0.555–1.633)	0.858
2,000–4,000	0.836 (0.518–1.351)	0.465
>4,000	1
History of depression§
No depression	1
Current depression	6.366 (1.253–32.353)	0.026*
Past depression	2.826 (1.528–5.227)	0.001**

Binary logistic regression analysis revealed that PSU was associated with woman gender, age under 35 years, and history of depression. Both current-depression and past-depression groups had a higher likelihood of PSU than those without depression history.

^*p<0.05;

^**p<0.01;

^***p<0.001;

^†single: not married, divorced, bereft, living independently;

^‡employed: full time or part time, unemployed: student, housewife, unemployed;

^§adjusted for gender, age, marriage status, employment status, education, and income level.

PSU, problematic smartphone use; OR, odds ratio; CI, confidence interval

To determine whether the pattern of PSU might vary depending on the history of depression, we compared average scores for each of the six factors (preoccupation, loss of control, craving, insight, overuse, and neglect of other areas) in the SOS-Q among ND, PD, and CD groups.

With the exception of “neglect of other areas,” all factors exhibited similar patterns. Average scores were the lowest in the ND group and the highest in the CD group. However, the PD group had the highest average score in the “neglect of other areas” domain.

Bonferroni post hoc analysis showed that average scores for preoccupation, loss of control, craving, lack of insight, and overuse were significantly higher in both CD and PD groups than in the ND group. The CD group had higher average scores in lack of insight and overuse than the PD group. Notably, the PD group, but not the CD group, demonstrated significantly higher average scores in the “neglect of other areas” than the ND group (Figure 1).

Figure 1.

Pattern of problematic smartphone use by history of depression. *p<0.05.

DISCUSSION

Choi et al. [29] have reported that depression could act as a protective factor, whereas anxiety is a risk factor for smartphone addiction. Other than that, many studies have consistently demonstrated the association between depression and PSU or smartphone addiction. A previous meta-review has indicated a 3.17-fold increase in the risk of depression in the PSU group [4]. Additionally, a recent meta-analysis has found a positive correlation with between PSU and depression [5] (r=0.28, 95% CI: 0.22 to 0.34, p<0.001). Consistent with findings of previous studies, our results once again confirmed that depression and PSU were significantly associated not only in students and young individuals, but also in general adults living in the community selected through random sampling. Additionally, our study found that both CD and previous depression in remission were associated with PSU. The risk of PSU was 6.366 times higher in the CD group and 2.826 times higher in the PD group compared to the ND group.

Traditionally, young people have been the focus of research and public health concern regarding PSU. The association between depression and PSU has predominantly been studied in adolescents and young adults, particularly college students [18,24,30-46]. Previous studies on general adults mainly recruited participants through an open online survey posted on social networking sites [47], website [48], or by sharing online survey links [49,50]. All studies included in a recent meta-analysis paper examining the connection between depression and PSU involved participants with an average age of 35 years or below [5]. Only a few studies have included senior adults [47,50,51]. Consequently, previous studies might have been susceptible to selection bias with limited generalizability. In our study, we found a significant association between PSU and depression in a sample of general adults selected through randomized cluster sampling, addressing these limitations. Notably, we found a substantial prevalence of PSU among individuals aged 35 years and older, although the prevalence showed a decreasing trend with an increasing age. The prevalence of PSU was 9.3% in the age group of 36–50 years and 5.4% in the age group of 50 years and older. This result suggests that middle-aged and older individuals are not free from PSU or its negative consequences. This might be attributed to the widespread popularity of smartphones in Korea, where even people in their 50s and 60s have high smartphone usage rates. Similar patterns might be found in other countries as smartphone usage among older adults continues to rise. Furthermore, it is crucial to recognize that PSU in middle-aged or elderly individuals might have additional side effects or vulnerabilities specific to their age group. For instance, age-related changes in muscle strength and cognitive function may render the elderly more susceptible to negative consequences such as musculoskeletal pain and accidents. Older individuals experiencing social isolation and loneliness may face an increased risk of PSU as social isolation and loneliness are associated with PSU [31,42,52]. Collectively, PSU represents a public health concern that requires attention across all age groups. In the future, more attention should be paid to characteristics of PSU among older adult population.

Our results proved the association between PSU and MDD diagnosed using a standardized tool. To date, there have been no studies utilizing the diagnosis of MDD through structured tools or assessment by psychiatrists. Psychological outcomes including depression have been mainly defined by self-report questionnaires rather than formal diagnoses [4,5,19]. A subgroup analysis of a meta-review has indicated that studies utilizing the Beck Depression Inventory to measure depression show a more pronounced association between depression and PSU than the studies using other measures [9]. This finding suggests that the strength of the association may vary depending on the method used to assess depression [9].

Moreover, it has been reported that both the risk and severity of PSU are increased as the severity of depression increases [5,8]. Given that a diagnosis of MDD indicates the presence of significant depressive symptoms that can impact daily functioning, it is plausible that there is a stronger association between MDD and PSU or an association of MDD with more severe PSU. This might explain our finding that the risk of PSU in the PD group was significantly higher than that in the ND group, but lower than in the CD group. Mild depressive symptoms, which do not meet the diagnostic criteria, might persist in the PD group even after remission of a major depressive episode.

While evidence strongly supports the association of depression with PSU, their temporal precedence and causality remain controversial with conflicting findings from longitudinal studies. Some studies have suggested that PSU may precede depression. For example, one longitudinal study has shown that smartphone dependency can predict depressive symptoms in late adolescents 3 months later [21]. Another study targeting university students has shown that PSU at baseline can increase subsequent psychological distress after 9 months [22]. High mobile phone use is associated with depressive symptoms one year later among adolescents [53] and young adults [54]. In a four-year study of adult couples, initial compulsive internet use predicted increases in depression, but depression itself did not predict later compulsive use [55]. Furthermore, a randomized trial targeting depressive students demonstrated that reducing social media use led to decreased depressive symptom scores after one month [56]. However, these findings have limitations in determining temporal precedence due to differences in variables studied. Psychological distress may increase the risk of depression but not necessarily lead to it. Similarly, PSU is not equivalent to high mobile phone use, compulsive internet use, or social media use.

Conversely, other studies have indicated that PSU follows depression. Higher levels of depression in children are associated with increased smartphone addiction proneness in the following year [23]. Among college students, initial depression predicted smartphone dependency 1 month later through mindfulness [57] and addictive behaviors after one year [24]. A 6-month longitudinal study also reported depression before smartphone addiction among high school students, but not vice versa [25]. Qiu et al. [58] have suggested that relieving depression could be an effective intervention for PSU in children and adolescents, as the presence of meaning and subsequent PSU are mediated by depression and self-control. According to the mood enhancement hypothesis, individuals experiencing negative emotions tend to engage in entertaining or distracting activities to escape from unfavorable emotional experiences [59] such as stress [16,60,61] and loneliness [42]. In this context, a depressed person may use a smartphone to cope with their mood, making them more vulnerable to PSU [5]. In our study, even participants in the PD group—those who had previously experienced depression but were currently in remission—were found to have a higher risk of PSU. In such cases, maladaptive coping mechanism acquired during the depressive episode may persist even after depressive symptoms improve, thus increasing the risk of PSU.

A more complex relationship rather than a simple causal relationship can exist between depression and PSU. Longitudinal studies targeting Korean adolescents [62] and Chinese college students [31,32,63] have reported bidirectional relationships between PSU and depressive symptoms. Depression can lead to excessive smartphone use as a means of escaping negative emotions [8]. Subsequently, excessive use worsens psychopathology through factors such as decreased sleep [47,54,64], impaired social relationship [47], and functional impairment [65]. Alternatively, common vulnerabilities may increase the risk of both PSU and depression. Zhou et al. [25] have found that social isolation and loneliness are risk factors for both depression and smartphone addiction. Neuroticism, a known risk factor for depression, has been found to be related to PSU [66]. Without these vulnerabilities, PSU would not be sustained. Depression would not occur either. The directionality of the relationship and the impact of preceding issues may vary depending on subject characteristics including gender. A cross-sectional study has reported that depression is a risk factor for PSU in women, but not in men [45]. Bidirectional relationships between depression and PSU are significant only in female but not male college students [32].

It is important to note that PSU would not be a uniform entity. A study of 5,372 smartphone users has identified distinct subtypes of smartphone addiction based on four factors: 1) self-control, 2) anxiety, 3) depression, and 4) dysfunctional impulsivity [67]. Patterns of PSU differed between PD and CD groups in our study. Specific personality traits have been found to be associated with different PSU symptoms [68]. The relationship between PSU and depression can even vary according to PSU subtypes.

Our study has several limitations. Firstly, inherent design limitations exist, preventing us from establishing a causal relationship between depression and smartphone addiction, although the precedence of depression was confirmed. Additionally, our study did not investigate past smartphone addiction, leaving open the possibility of it preceding depression. Moreover, our study was subject to potential memory bias due to its cross-sectional design. Lastly, the relatively small sample size in certain groups should be noted. Despite analyzing a large population, the number of individuals diagnosed with depression in the past month was limited.

Despite these limitations, our study possesses several strengths. Recent meta-reviews have provided suggestions for future research based on an examination of current findings [4,5]. The authors advocate for selecting participants through a sampling method closer to random sampling for improved generalizability, employing more objective measurement tools, and conducting longitudinal studies that encompass both baseline and follow-up data. In our study, we addressed some of these recommendations by utilizing cluster random sampling to select a sample representing the general adult population, spanning all age groups. Furthermore, we employed structured diagnostic tools, offering more objective outcomes than self-report screening measures. Notably, despite being cross-sectional, we successfully confirmed the precedence relationship between PD episodes and present smartphone addiction. Previously resolved depression was found to be associated with current PSU. Consequently, our study significantly contributes to filling the gaps identified in prior studies. However, since we did not assess past PSU, our findings could not exclude the possibility that PSU might precede depression. Our study also highlighted the need to pay attention to PSU across all age groups since PSU was prevalent even among individuals aged 35 years and older. Future research studies are needed to explore characteristics and consequences of PSU in different age groups and investigate the causal relationship between depression and PSU. Understanding the association between depression and PSU is crucial for developing effective interventions for PSU and MDD.