Sage Journals: Discover world-class research

Abstract

Objective:

Problematic Internet and smartphone use is associated with serious mental health problems. We thus investigated the effectiveness of psychological interventions on problematic Internet and smartphone use.

Methods:

We conducted a systematic review and meta-analysis and searched in PubMed, PsyArticles, PsycInfo, Medline and Google Scholar for peer-reviewed experimental and quasi-experimental research published between 2008 and 2021.

Results:

A total of 10 studies accompanied by 14 effect sizes were included in the meta-analyses for problematic Internet use and 3 studies accompanied by 4 effect sizes for problematic smartphone use. The total sample size was n = 1439 with a mean age of 20.34 (SD = 6.63) years. Psychological interventions were effective for both problematic Internet use (g = −1.41, 95% confidence interval = [−1.90, −0.91], p < 0.001) and problematic smartphone use (g = −0.40, 95% confidence interval = [−0.79, −0.01], p < 0.001).

Conclusion:

Current evidence indicates that psychological interventions can be effective in reducing problematic Internet use/problematic smartphone use. However, the results must be considered preliminary due to the limited number of studies available in the respective fields.

Keywords

Psychotherapy mental health smartphone Internet behavioral addictions

Introduction

The negative mental health effects of excessive technology use have been increasingly discussed in the public domain and in research since the beginning of widespread Internet use more than 20 years ago (Young, 1998).

The main areas of interest are problematic Internet use (PIU) and—more specifically—problematic smartphone use (PSU). PIU has been associated with a wide range of negative mental health consequences such as symptoms of distress, social withdrawal, familial conflicts and clinical conditions like social anxiety disorders or depression (Cerniglia et al., 2017). There is a convincing relationship between PSU and serious mental problems and symptoms (Grant et al., 2019; Winkler et al., 2020).

However, the main function of smartphones is the availability of the Internet. Thus, recent research has shown that PIU and PSU overlap in some key features, although not in all. Tateno et al. (2019) revealed a highly significant correlation of r = 0.67 (p < 0.0001) between Internet addiction measured by the Internet Addiction Test (IAT) and smartphone addiction as measured by the Smartphone Addiction Scale–Short Version (SAS-SV) in a Japanese sample. In their study, Lee et al. (2020) found a significant correlation of r = 0.49 (p < 0.001) between Young’s Internet Addiction Test (Y-IAT) and the Smartphone Addiction Scale (SAS).

Research has used the terms Internet addiction and smartphone addiction extensively to express similarities in key features of other substance-related and behavioral addictions such as excessive use, problems with impulse control, continuation of certain behavior despite conflicts at work and in the family, and withdrawal symptoms (Cash et al., 2012).

Especially, excessive online gaming has been addressed in recent reviews dealing with diagnosis and treatment (Costa and Kuss, 2019; Zajac et al., 2020). Gaming disorder has been included in the International Classification of Diseases, 11th Revision (ICD-11) in the group ‘disorders due to addictive behavior’, and Internet gaming disorder is part of the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (American Psychiatric Association (APA), 2013). Recent research reports insist that ‘general Internet addiction’ and ‘online gaming’ should be addressed separately from each other (Petry et al., 2018; Zajac et al., 2017). Thus, a ‘general Internet addiction’ is still not part of a disease classification system, but can be assigned to ‘other specific disorders due to addictive behaviors’ in the ICD-11 (Brand et al., 2020).

Since recent studies have disputed the assertion that mental health problems associated with excessive technology use can be regarded as addictions without playing down other addictions (Panova and Carbonell, 2018), current discussion focuses on development of valid criteria for considering if a problem behavior is an addictive behavior in line with ICD-11 (Brand et al., 2020). In this study, we use the designations PIU and PSU. In analogy to problem vs disordered gambling, PIU and PSU also include behaviors that may not fulfill all criteria to be considered as ‘addiction’ (Pfund et al., 2020). Thus, we understand PIU/PSU in this meta-analysis as the inability to control screen time, which is accompanied by serious consequences in daily life. Recently, we have clarified that these consequences can imply symptoms of dependence or dysfunctional social interactions (Augner and Hacker, 2012).

Despite these diagnostic controversies, it is unquestionable that PIU and PSU are a key public health concern given the ubiquitous dissemination of the Internet in general and, more specifically, of smartphone use. However, evidence of the effects of psychological interventions or psychotherapy is surprisingly fragmented and thus limited.

Winkler et al. investigated 16 pharmacological and psychological intervention studies on Internet addiction. The meta-analysis revealed a high level of effectiveness of the interventions in reducing levels of Internet addiction, cutting time spent online, as well as reducing depression and anxiety scores from pre- to post-treatment (Winkler et al., 2013). However, online activities may have changed and spread since this review was published.

Another review assessed psychological e-health interventions for PIU. However, since the authors identified only three articles meeting the study selection criteria, no meta-analysis was calculated and no convincing conclusions about effectiveness could be drawn (Lam and Lam, 2016).

Another narrative review identified studies from a wide range of interventions including prevention programs, psychosocial treatments (e.g. cognitive behavioral therapy [CBT]), pharmacotherapy and treatment camps. Unfortunately, the authors did not include information about their literature search strategy and did not calculate a meta-analysis. The studies concerned were partly rather old, and with very heterogeneous backgrounds (e.g. video gaming) (Nakayama et al., 2017).

Zajac et al. analyzed 13 intervention studies on Internet gaming disorder and 13 more on Internet addiction. However, the authors did not calculate a meta-analysis. They reported on methodical flaws in the available studies, such as small sample sizes, a lack of control groups and a lack of consistent definitions of Internet addiction (Zajac et al., 2017).

A recent review by Kim and Noh included five randomized controlled trials (RCTs) and six quasi-experimental studies. A meta-analysis indicated a significant reduction of symptom severity as a result of psychological interventions. Unfortunately, its validity is very limited due to the mixture of Internet addiction and Internet gaming disorder (Kim and Noh, 2019).

A recent meta-analysis supports the assumption that psychological interventions can be effective for Internet and smartphone addiction (Malinauskas and Malinauskiene, 2019). However, only six studies were included in the calculation, which combined Internet and smartphone addiction. In relation to PSU or smartphone addiction, there was only one review which dealt with exercise-based interventions in exclusively Chinese studies, and none with psychological interventions (Liu et al., 2019).

Due to a lack of robust evidence, the objective of this meta-analysis is to identify the available studies focusing on psychological interventions for PIU and PSU. Since PIU and PSU are related, but not identical, we aimed to identify studies for both, but to analyze them separately.

Methods

In accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) method (Moher et al., 2009; Walther et al., 2011), we prepared a review protocol that included a working title, review objective, search strategies, study selection criteria, interventions and end points.

Eligibility criteria and study selection

The objective of the study was to identify evidence of the effectiveness of psychological interventions on PIU and PSU. The inclusion criteria were that the studies:

Evaluated the effect of a psychological intervention and/or treatment and/or psychotherapy on PSU (including ‘smartphone addiction’) and PIU (including ‘Internet addiction’);

Used standardized questionnaires to assess PIU/PSU as the outcome;

Used a (quasi-) experimental design that included an intervention/control group;

Published or online first in peer review journals.

We excluded studies that:

Used psychopharmacological interventions.

Focused on Internet gaming or Internet pornography use.

Focused on clinical samples with significant comorbidities such as depression or anxiety; first, psychological interventions for other mental health problems have a different focus; second, even for significant comorbidities of substance abuse, therapeutic effectiveness and implications remain unclear and cannot be compared to treatment of each health issue alone (Hides et al., 2019).

Did not report relevant data in English.

Literature search

We conducted a systematic literature search in PubMed, PsyArticles, PsycInfo, Medline and Google scholar, whereby we used the following combinations of terms:

((addiction) OR (smartphone addiction) OR (cellular phone addiction) OR (problematic Internet use) OR (problematic smartphone use) OR (problematic cellular phone use)) AND ((treatment) OR (intervention) OR (psychotherapy)). We defined a time span for publication date in order to have at least somehow comparable Internet/smartphone disseminations. We chose 2008 as our earliest reference point, because proportion of worldwide Internet users crossed the 20% threshold shortly before 2008 and cellular phone subscriptions reached about 60% (ITU, 2019). We applied a time filter including studies published from 1 January 2008 to 1 January 2021 only.

Furthermore, we used backward snowballing from the reference lists of existing reviews on PIU or PSU (Kim and Noh, 2019; Lam and Lam, 2016; Malinauskas and Malinauskiene, 2019; Nakayama et al., 2017; Winkler et al., 2013; Zajac et al., 2017).

Screening abstracts

Two authors of the study screened the research literature. We identified relevant articles via three steps: (1) title, (2) abstract and (3) full text and imported them into Endnote X9. Two authors screened the records independently of each other. A consensus was reached via discussion or—if not resolvable—by a senior author’s decision. Duplicates were removed from the reference list (see Figure 1).

Figure 1.

Flow chart study selection process.

Data extraction and quality assessment

Via a coding scheme which we developed, the data extraction was carried out by three authors. An inter-rater reliability of 0.97 based on mean rating and a two-way mixed-effects model (k = 3, absolute agreement) was computed and differences were solved by consensus. The variables of data extraction consisted of the authors of the publication, year of publication, effect sizes as well as the characteristics of the samples. If studies reported results for subgroups (e.g. gender) rather than for the total sample, groups were entered into the meta-analysis as independent samples. The quality of the methodology of the studies included was evaluated by all three authors using the Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies (QATOCCSS) of the National Heart, Lung, and Blood Institute of the National Institutes of Health (NIH) (n.d.). This assessment tool consists of 14 questions regarding various methodological aspects (e.g. research process, sample size justification) and provides an overall score of bad, fair or good quality. An inter-rater reliability of 0.93 was estimated for study quality.

Data synthesis and analysis

Since the studies reported the mean difference between exposure and control groups for both outcome variables, Hedge’s g was selected as the effect size for the meta-analysis. In comparison to Cohen’s d, Hedge’s g leads to a more precise estimation of true effect sizes due to less bias, particularly in small sample sizes (Hedges, 1981). Cujipers et al. (2017) suggest avoiding pre-post effect sizes in meta-analysis focusing on the effects of therapeutic interventions. Since correlations between pre-post scores are hardly given, and the therapeutic setting and process as well as the patients influence the effect of the intervention, results are more biased with considerable errors in estimation. Due to this, Cuijpers et al. (2017) recommend using between-group mean differences since they control for these biased outcomes. Therefore, the meta-analytic results are based on the post-intervention between-group mean differences between the exposure and control groups. Regarding the direction of effect, positive effect sizes indicate higher scores in the exposure group, while negative effect sizes reflect higher values in the control group for Internet and smartphone addiction. Outliers in effect size were identified before calculating the average study effects by applying Cook’s distance (Viechtbauer and Cheung, 2010). A random-effects model with restricted likelihood estimator using Hedge’s invariance weighting was carried out for pooling the effect sizes with the appropriate 95% confidence intervals separately for Internet and smartphone addiction. For the quantification of the homogeneity of pooled effect sizes, τ² was used, while Q statistic indicated significance levels (Cochran, 1954). Sensitivity analysis was carried out when heterogeneity of the effect sizes was given, while possible publication bias was examined by funnel plots and Egger’s tests (Egger et al., 1997), and p-hacking was addressed by p-curve diagrams and analyses (Simonsohn et al., 2014). Weighted mixed-effects meta-regressions were carried out regarding potential moderating effects, and subgroup analyses following a random-effects model further examined the heterogeneity of effect sizes. In general, statistical analyses were carried out using the software R with the metafor package (Viechtbauer, 2010).

Measurement of PSU/PIU

All studies included in this meta-analysis used questionnaires to assess PSU and PIU (see Table 1).

Table 1.

Raw data for the meta-analysis.

Author	Outcome	Measured by	g	Se	EG/CG	Intervention	Mean age	Stakeholder	Quality	Sessions in minutes
Hou et al. (2019)	PIU	Social Media Addiction Scale (BSMAS)	–1.29	0.35	21/17	1	19.71	0	7	30
Hou et al. (2019)	PIU	Social Media Addiction Scale (BSMAS)	–1.26	0.35	21/17	1	19.71	0	7	30
Li and Dai (2009)	PIU	Chinese Internet Addiction Scale (CIAS)	–0.95	0.24	38/38	1	–	0	6	–
Lan et al. (2018)	PSU	Mobile Phone Internet Addiction Scale (MPIAS)	–0.16	0.24	41/29	1	21.3	0	8	480
Lan et al. (2018)	PSU	Mobile Phone Internet Addiction Scale (MPIAS)	–0.92	0.25	41/29	1	21.3	0	8	480
Yu and Son (2016)	PSU	Smartphone Addiction Proneness Scale for Adults (SAPS-A)	–0.60	0.48	9/9	1	–	0	7	120
Cardak et al. (2009)	PIU	Criteria for Problematic Internet Use (Young)	–1.75	0.48	12/12	1	–	0	7	500
Du et al. (2010)	PIU	Internet Overuse Self-Rating Scale	0.13	0.20	32/24	2	16.01	1	6	960
Liu et al. (2015)	PIU	Adolescent Pathological Internet Use Scale	–2.35	0.38	21/25	2	14.5	1	8	720
Liu et al. (2015)	PIU	Adolescent Pathological Internet Use Scale	–1.70	0.34	21/25	2	14.5	1	8	720
Yang and Kim (2018)	PIU	Internet Addiction Proneness Scale (K-Scale)	–1.44	0.25	38/41	2	14	0	7	450
Choi et al. (2020)	PSU	Korean Smartphone Addiction Proneness Scale	0.00	0.28	24/25	3	16	0	6	480
Quinones and Griffiths (2019)	PIU	Compulsive Internet use scale (CIU)	–5.75	0.17	343/350	3	40.5	0	8	140
Quinones and Griffiths (2019)	PIU	Compulsive Internet use scale (CIU)	–3.49	0.12	343/350	3	40.5	0	8	140
Kim (2008)	PIU	K-Internet Addiction Scale (K-IAS)	–1.23	0.43	13/12	4	24.2	0	7	500
Su et al. (2011)	PIU	Young’s Diagnostic Questionnaire score	–1.68	0.40	17/16	4	21	0	6	28.5
Su et al. (2011)	PIU	Young’s Diagnostic Questionnaire score	–1.43	0.42	12/16	4	21	0	6	28.5
Su et al. (2011)	PIU	Young’s Diagnostic Questionnaire score	–1.09	0.39	14/16	4	21	0	6	28.5
Khazaei et al. (2017)	PIU	Young Internet Addiction Test	–4.88	0.58	24/14	2	27.5	0	7	500
Pan (2020)	PIU	Adolescent Internet Addiction Diagnostic Scale	–0.21	0.10	10/10	2	–	0	8	–

EG: exposure group; CG: control group; PIU: problematic Internet use; PSU: problematic smartphone use; CBT: cognitive behavioral therapy.

Intervention type: (1) CBT or mainly CBT-based interventions, (2) systemic or multimodal intervention approaches (incl. Family therapy, etc.), (3) mindfulness-based intervention approaches, (4) psychoeducation and self-help tools; Stakeholder: (0) participants only (1) participants and teacher or parents; Du et al. (2010) included teachers in intervention, Liu et al. (2015) included parents in treatment.

Instruments varied between the studies. However, for PIU most common symptoms assessed were compulsive Internet use, negative outcomes, mood alterations (withdrawal) and tolerance (Cardak et al., 2009; Khazaei et al., 2017; Kim, 2008; Liu et al., 2015; Pan, 2020; Quinones and Griffiths, 2019; Su et al., 2011; Yang and Kim, 2018). Where reported, reliability of measurements was high and Cronbach’s alpha values of the instruments above 0.80 (Hou et al., 2019; Kim, 2008; Liu et al., 2015; Pan, 2020; Quinones and Griffiths, 2019; Yang and Kim, 2018).

For PSU only one study reported psychometric details about the instrument used. The Korean Smartphone Addiction Proneness Scale includes subscales ‘daily life disturbance’, ‘directivity to virtual world’, ‘tolerance’ and ‘withdrawal’. Reliability in the primary study was high (Cronbach’s alpha = 0.88) (Choi et al., 2020).

Treatment classification

Two authors defined the intervention types used in the studies and assigned one intervention type to each study (see Table 1):

(a) CBT or mainly CBT-based interventions, i.e. interventions that involve elements of cognitive reconstruction and/or behavioral changes. Hou et al. (2019) for example used reflection questions to cognitive reconstruct social media use (e.g. ‘Why did they use social media and were there alternative way to achieve the purposes?’) and behavioral change strategies to reduce online time (Hou et al., 2019).

(b) Systemic or multimodal intervention approaches (including family therapy, etc.), i.e. interventions that include family environment of participants and/or used different modalities to address PIU or PSU. Interventions in Liu et al. (2015) focused on parent–adolescent interaction, while Pan (2020) for example used a set of psychological and sports interventions (Liu et al., 2015; Pan, 2020).

(c) Mindfulness-based intervention approaches, i.e. interventions that use meditation as essential part of the therapeutic process. Choi et al. (2020) for example used mind subtraction meditation to reduce PSU (Choi et al., 2020).

(d) Psychoeducation and self-help tools, i.e. intervention that focused mainly on providing information about PSU or PIU, e.g. Su et al. (2011) used an online expert system named Healthy Online Self-helping Center (HOSC) (Su et al., 2011).

Differences in classification among the authors were resolved by discussion.

Stakeholder inclusion

Furthermore, we identified whether the intervention only addressed the target group (stakeholder = 0) or included teachers and/or parents (stakeholder = 1) in any way.

Results

Study characteristics

Overall, we included 13 studies accompanied by 18 effect sizes, achieving a total sample of n = 1439 participants with a mean age of 20.34(SD = 6.63) years, while the publication date ranged from 2008 to 2020. In general, seven effect sizes (38.89%) are based on cognitive behavioral interventions, five effect sizes (27.78%) followed systemic intervention approaches while two effect sizes (11.11%) come from mindfulness-based interventions and four effect sizes (22.22%) are based on psychoeducation. Overall, the study quality was considered fair to good, while in most studies information about the characteristics of the sample, intervention process and statistical analyses was lacking. The total mean duration of therapeutic intervention across all studies was 356.00 (SD = 301.43) minutes.

Meta-analytic results

PIU

For PIU, the test for homogeneity indicated heterogeneous effect sizes of the included studies, Q(13) = 504.84, p < 0.001, with τ² = 0.78. The random-effect model meta-analysis across 10 studies including 14 effect sizes showed significant differences between exposure and control groups with g = −1.41 (95% confidence interval from −1.90 to −0.91, p < 0.001). Therefore, individuals participating in psychotherapeutic treatment show significantly lower scores in PIU than their controls (see Figure 2).

Figure 2.

Forest plot PIU.

PSU

The homogeneity test showed homogeneous effect sizes regarding the included studies for PSU, Q(3) = 7.28, p = 0.06, with τ² = 0.06. A random-effect model meta-analysis including three studies with four effect sizes showed significant differences between exposure and control groups with g = −0.40 (95% confidence interval from −0.79 to −0.01, p < 0.001). According to this, individuals exposed to psychotherapeutic treatment show significantly lower scores in PSU than individuals in the control group (see Figure 3).

Figure 3.

Forest plot PSU.

Sensitivity analysis

Outliers

Before computing the results, we excluded two effect sizes from two different studies (Khazaei et al., 2017; Quinones and Griffiths, 2019) from entering the meta-analysis for PIU, since their magnitude of effect differs substantially from the remaining studies (g = −5.75 for the mindfulness intervention in Quinones and Griffiths, 2019, and g = −4.88 for a range of positive psychology interventions in Khazaei et al., 2017, see further details in Table 1).

In general, three effect sizes from three studies (Du et al., 2010; Pan, 2020; Quinones and Griffiths, 2019) were detected as outliers by the analysis. Removing these outliers resulted in a homogeneity of effect sizes of, Q(10) = 12.24, p = 0.27, regarding PIU across 7 studies with 11 effect sizes while reducing τ² from 0.78 to 0.04. Exclusion of outliers increased the magnitude of the effect size and the meta-analytic model remained significant with g = −1.43 (95% confidence interval from −1.67 to −1.204, p < 0.001). Further examination revealed that the reason for detection as outliers was either a disproportionate sample size or magnitude of effect in comparison to the remaining studies. Therefore, heterogeneity seems to be driven by these outliers.

No outliers could be detected in the included studies for PSU.

Subgroup and moderator analysis

For the subgroup analysis, PIU interventions based on mindfulness were excluded since only two studies, Choi et al. (2020) and Quinones and Griffiths (2019), were comparable. No significant subgroup differences could be found in relation to the type of intervention, Q(2) = 0.92, p = 0.63. Regarding moderator analysis, while mean age of the participants, χ²(1) = 6.70, p = 0.01; r² = 0.42, b = −0.08, showed a significant moderation of effect sizes, neither publication year, χ²(1) = 0.64, p = 0.42; r² = 0.00, b = −0.05, quality of the study, χ²(1) = 2.65, p = 0.10; r² = 0.12, b = −0.48, stakeholder, χ²(1) = 0.09, p = 0.76; r² = 0.00, b = 0.19, nor the duration of the intervention, χ²(1) = 0.86, p = 0.35; r² = 0.02, b = 0.00, did. However, when entering all variables as multiple moderators in a meta-regression, χ²(5) = 31.77, p < 0.001; r² = 0.78, b_{publicationyear} = 0.03, b_meanage = −0.04, b_stakeholder = −0.17, b_studyquality = −0.92, b_{durationintervention} = 0.00, the moderating effects of mean age of participants did not remain significant (p = 0.62).

Due to the small sample size, no meaningful subgroup and moderator analysis could be carried out for PSU.

Publication bias and p-hacking

The funnel plot does not visually suggest publication bias regarding the included studies and Egger’s test does not show significant results (p = 0.94) (see Figure 4). When testing for p-hacking by conducting p-curve analysis, effect sizes showed significant right-skewness (p < 0.001) and non-significant flatness (p = 0.99), indicating robustness against p-hacking (see Figure 5). Therefore, the results regarding the overall effect size of Internet addiction do not seem to be influenced by either publication bias or p-hacking. Again due to the small sample size, no meaningful analysis of publication bias or p-hacking could be carried out.

Figure 4.

Funnel plot PIU.

Figure 5.

P-curve diagram PIU.

Discussion

PIU and PSU manifest a clear public health issue. Psychological interventions could possibly help to avoid or reduce negative behavioral and health effects of Internet and smartphone use. Indeed, the results of this meta-analysis indicate that psychological interventions can significantly reduce PIU/PSU scores measured by standardized questionnaires. Overall, the effect size can be regarded as high for PIU (Hedges g = −1.41, p < 0.001) and moderate for PSU (Hedges g = −0.40, p < 0.001).

The effects of psychotherapeutic interventions on different mental health problems are relatively well established. A recent meta-analysis reported significant effects of psychotherapy on depression (standard mean difference (SMD) = 0.71) (Munder et al., 2019). Furthermore, a review showed that CBT can reduce Internet gaming disorder symptoms (g = 0.92; [0.50, 1.34]) (Stevens et al., 2019).

We did not find significant differences between different intervention methods for PIU. Recent evidence rejects the assumption that special types of psychotherapy (e.g. CBT) may be more effective than other types. On the contrary, so-called non-specific aspects of psychotherapy and psychological interventions (e.g. format of the intervention, frequency, empathy or the motivation of the therapist) are decisive. Palpacuer et al. (2017) reported in a meta-analysis that the effects of different psychotherapy methods on depression failed to be significant in relation to a waiting-list control group when adjusted for non-specific aspects. If specific aspects of psychotherapy methods are of less importance, it is plausible that a wide range of different psychological interventions may be effective in reducing PIU/PSU scores.

Although PIU/PSU cannot be classified as disorders, addictions or illnesses, we can make analogies from psychotherapeutic research on the mode of functioning of psychological interventions that more or less reduce the problematic issues of Internet and smartphone use.

First, a convincing body of evidence suggests that the core symptoms of addictive behaviors in general include impaired self-control and increased stress reactivity (Tang et al., 2015). A wide range of psychological interventions focus directly or indirectly on these self-control and self-regulation functions.

Thus, several studies included in this meta-analysis have used elements that address self-control and self-regulation in the psychological intervention. Yang et al. integrated self-regulation strategies into their intervention. The experimental group improved significantly—not only in Internet addiction scores but also in self-control (Yang and Kim, 2018). Hou et al. (2019) based their interventions on considerations about self-regulation. Further evidence indicates that impaired self-control is the best predictor of problematic social media use (Blachnio and Przepiorka, 2016).

The physiological correlates of self-control problems are located in different areas of the brain. Decreased activity of the anterior cingulate cortex (ACC) and prefrontal cortex (PFC) and reduced connectivity between the ACC and striatum seem to be decisive for lower self-control and higher proneness to addictive behaviors (Tang et al., 2015). Moreover, in relation to Internet addiction, recent research indicates that dysfunctions in ACC and PFC as part of the default mode network (DMN) might lead to deficits in cognitive control and to reduced cognitive efficiency. This might also be associated with attention problems. Deficits of the inhibitory control network as identified in persons with a high PIU via task-related functional magnetic resonance imaging (fMRI) are responsible for the inability to stop or control online behavior (Darnai et al., 2019). There is neuropsychological evidence that different forms of therapy, e.g. mindfulness-based therapy and CBT, can significantly modulate the DMN (Fahmy et al., 2019; Li et al., 2018).

Besides individual and neuropsychological aspects of self-control and PIU/PSU, there are some social and cultural factors to be taken into account. A recent commentary underscores the relevance of social determinants of addictive behaviors (e.g. weak social ties, social exclusion, hyper-individualism) and hypothesizes that our society is somehow organized to consume and therefore to undermine our self-control (Van der Linden, 2015). This seems to be especially true for Internet and smartphone use and closely related behaviors like social media use.

However, this meta-analysis has some limitations. We evaluated exclusively short-term effects, since most studies did not include follow-ups. Actually, one main problem is that we have surprisingly little information about the long-term natural course of PSU/PIU which would determine the necessity of psychological intervention. One study reported that in a group of adolescent Internet addicts, a total of 71 were regarded as remissions and 67 as persistent addicts after 1 year (Ko et al., 2014). Another paper reports that Internet gaming disorder is relatively unstable over time; however, persistent rates vary between 26.5% after 2 years and 50% or 86% after 1 year (Mihara and Higuchi, 2017). Thus, there is a clear need for more research in this area, since a significant reduction in PSU/PIU scores might take place without any intervention.

Furthermore, there are some methodological issues to be taken into account. First, treatment types varied between studies. In some studies, superficial descriptions of procedures made it difficult to classify the intervention (classifications see Table 1). Second, we did not include studies in our meta-analysis that focused on clinical samples with significant comorbidities such as depression or anxiety. Although comorbidity rate between PSU/PIU is high, we aimed to isolate effects of psychological treatments on PIU/PSU, since psychological interventions for other mental health problems have a different focus. Furthermore, recent evidence for substance abuse disorder and depression as comorbidity indicate that therapeutic effectiveness and implications cannot be compared to treatment of each health issue alone (Hides et al., 2019). Nevertheless, this selection process may have influenced our results. Most probably, our sample is mentally healthier than clinical samples are. Thus, meaningfulness of our results on clinical samples may be limited. Third, the heterogeneity of effect sizes was given for the outcome ‘problematic Internet use’. Sensitivity analysis revealed that the inhomogeneous results are mainly driven by outliers. Although removal of these outliers reduced the magnitude of this effect to some degree, the pooled effect size remained significant while showing robustness against publication bias and p-hacking. In contrast, we found homogeneous effect sizes for PSU. However, due to the small sample size of studies, no meaningful sensitivity analysis could be carried out regarding the outcome. The interpretation of our results concerning PSU can thus only be regarded as preliminary. Consequently, we would recommend that future studies should increasingly focus on evaluating the effects of psychological interventions on PSU.

The present meta-analysis indicates that psychological interventions can be effective in reducing PIU/PSU scores in the short term. Against the background of the available evidence, interventions that directly or indirectly affect self-control functions may be especially effective. However, future research should focus on long-term effects of interventions. Moreover, the general lack of studies with high-quality designs in this research area should be addressed.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

ORCID iD

Christoph Augner

Data Availability

This study is a meta-analysis, and all data are available in the public domain.

References

American Psychiatric Association (APA) (2013) Diagnostic and Statistical Manual of Mental Disorders, 5th Edition. Arlington, VA: American Psychiatric Publishing.

Augner

Hacker

(2012) Associations between problematic mobile phone use and psychological parameters in young adults. International Journal of Public Health 57: 437–441.

Blachnio

Przepiorka

(2016) Dysfunction of self-regulation and self-control in Facebook addiction. Psychiatric Quarterly 87: 493–500.

Brand

Rumpf

Demetrovics

, et al. (2020) Which conditions should be considered as disorders in the International Classification of Diseases (ICD-11) designation of ‘other specified disorders due to addictive behaviors’? Journal of Behavioral Addictions. Epub ahead of print 30 June 2020. DOI: 10.1556/2006.2020.00035.

Cardak

Koc

Colak

(2009) The effect of a rational emotional behavior therapy (REBT) group counseling program on the internet addiction among university students. Available at: omeka.ibu.edu.ba/items/show/3395 (accessed 2 February 2021).

Cash

Rae

Steel

, et al. (2012) Internet addiction: A brief summary of research and practice. Current Psychiatry Reviews 8: 292–298.

Cerniglia

Zoratto

Cimino

, et al. (2017) Internet addiction in adolescence: Neurobiological, psychosocial and clinical issues. Neuroscience & Biobehavioral Reviews 76: 174–184.

Choi

Chun

Lee

, et al. (2020) The effect of mind subtraction meditation intervention on Smartphone addiction and the psychological wellbeing among adolescents. International Journal of Environmental Research and Public Health 17: 3263.

Cochran

(1954) The combination of estimates from different experiments. Biometrics 10: 101–129.

10.

Costa

Kuss

(2019) Current diagnostic procedures and interventions for gaming disorders: A systematic review. Frontiers in Psychology 10: 578.

11.

Cuijpers

Weitz

Cristea

, et al. (2017) Pre-post effect sizes should be avoided in meta-analyses. Epidemiology and Psychiatric Sciences 26: 364–368.

12.

Darnai

Perlaki

Zsido

, et al. (2019) Internet addiction and functional brain networks: Task-related fMRI study. Scientific Reports 9: 15777.

13.

Jiang

Vance

(2010) Longer term effect of randomized, controlled group cognitive behavioural therapy for internet addiction in adolescent students in Shanghai. Australian and New Zealand Journal of Psychiatry 44: 129–134.

14.

Egger

Smith

Schneider

, et al. (1997) Bias in meta-analysis detected by a simple, graphical test. British Medical Journal 315: 629–634.

15.

Fahmy

Wasfi

Mamdouh

, et al. (2019) Mindfulness-based therapy modulates default-mode network connectivity in patients with opioid dependence. European Neuropsychopharmacology 29: 662–671.

16.

Grant

Lust

Chamberlain

(2019) Problematic Smartphone use associated with greater alcohol consumption, mental health issues, poorer academic performance, and impulsivity. Journal of Behavioral Addictions 8: 335–342.

17.

Hedges

(1981) Distribution theory for Glass’s estimator of effect size and related estimators. Journal of Educational Statistics 6: 107–128.

18.

Hides

Quinn

Stoyanov

, et al. (2019) Psychological interventions for co-occurring depression and substance use disorders. Cochrane Database of Systematic Reviews 2019: CD009501.

19.

Hou

Xiong

Jiang

, et al. (2019) Social media addiction: Its impact, mediation, and intervention. Cyberpsychology: Journal of Psychosocial Research on Cyberspace 13: 4.

20.

International Telecommunication Union (ITU) (2019) Measuring Digital Development—Facts and Figures—2019. Geneva: ITU.

21.

Khazaei

Ghanbari

(2017) Positive psychology interventions for internet addiction treatment. Computers in Human Behavior 72: 304–311.

22.

Kim

(2008) The effect of a R/T group counseling program on the internet addiction level and self-esteem of internet addiction university students. International Journal of Reality Therapy 27: 4–12.

23.

Kim

Noh

(2019) The current status of psychological intervention research for internet addiction and internet gaming disorder. Issues in Mental Health Nursing 40: 335–341.

24.

Liu

Wang

, et al. (2014) The exacerbation of depression, hostility, and social anxiety in the course of internet addiction among adolescents: A prospective study. Comprehensive Psychiatry 55: 1377–1384.

25.

Lam

(2016) eHealth Intervention for Problematic Internet Use (PIU). Current Psychiatry Reports 18: 107.

26.

Lan

Ding

, et al. (2018) A pilot study of a group mindfulness-based cognitive-behavioral intervention for Smartphone addiction among university students. Journal of Behavioral Addictions 7: 1171–1176.

27.

Lee

Chung

Lee

, et al. (2020) Investigation of correlated internet and Smartphone addiction in adolescents: Copula regression analysis. International Journal of Environmental Research and Public Health 17: 5806.

28.

Dai

X-Y

(2009) Control study of cognitive-behavior therapy in adolescents with internet addiction disorder. Chinese Mental Health Journal 23: 457–470.

29.

Yang

Greenshaw

, et al. (2018) The effects of cognitive behavioral therapy on resting-state functional brain network in drug-naive patients with obsessive–compulsive disorder. Brain and Behavior 8: e00963.

30.

Liu

Fang

Yan

, et al. (2015) Multi-family group therapy for adolescent internet addiction: Exploring the underlying mechanisms. Addictive Behaviors 42: 1–8.

31.

Liu

Xiao

Yang

, et al. (2019) Exercise as an alternative approach for treating Smartphone addiction: A systematic review and meta-analysis of random controlled trials. International Journal of Environmental Research and Public Health 16: 3912.

32.

Malinauskas

Malinauskiene

(2019) A meta-analysis of psychological interventions for internet/Smartphone addiction among adolescents. Journal of Behavioral Addictions 8: 613–624.

33.

Mihara

Higuchi

(2017) Cross-sectional and longitudinal epidemiological studies of internet gaming disorder: A systematic review of the literature. Psychiatry and Clinical Neurosciences 71: 425–444.

34.

Moher

Liberati

Tetzlaff

, et al. (2009) Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. British Medical Journal 339: b2535.

35.

Munder

Fluckiger

Leichsenring

, et al. (2019) Is psychotherapy effective? A re-analysis of treatments for depression. Epidemiology and Psychiatric Sciences 28: 268–274.

36.

Nakayama

Mihara

Higuchi

(2017) Treatment and risk factors of internet use disorders. Psychiatry and Clinical Neurosciences 71: 492–505.

37.

National Institutes of Health (NIH) (n.d.) Quality assessment tool for observational cohort and cross-sectional studies (QATOCCSS). Available at: www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools (accessed 26 February 2021).

38.

Palpacuer

Gallet

Drapier

, et al. (2017) Specific and non-specific effects of psychotherapeutic interventions for depression: Results from a meta-analysis of 84 studies. Journal of Psychiatric Research 87: 95–104.

39.

Pan

(2020) Psychological and exercise interventions for teenagers with internet addiction disorder. Revista Argentina de Clínica Psicológica 29: 226–231.

40.

Panova

Carbonell

(2018) Is Smartphone addiction really an addiction? Journal of Behavioral Addictions 7: 252–259.

41.

Petry

Zajac

Ginley

(2018) Behavioral addictions as mental disorders: To be or not to be? Annual Review of Clinical Psychology 14: 399–423.

42.

Pfund

Peter

Whelan

, et al. (2020) Is more better? A meta-analysis of dose and efficacy in face-to-face psychological treatments for problem and disordered gambling. Psychology of Addictive Behaviors 34: 557–568.

43.

Quinones

Griffiths

(2019) Reducing compulsive internet use and anxiety symptoms via two brief interventions: A comparison between mindfulness and gradual muscle relaxation. Journal of Behavioral Addictions 8: 530–536.

44.

Simonsohn

Nelson

Simmons

(2014) P-curve and effect size: Correcting for publication bias using only significant results. Perspectives on Psychological Science 9: 666–681.

45.

Stevens

MWR

King

Dorstyn

, et al. (2019) Cognitive-behavioral therapy for internet gaming disorder: A systematic review and meta-analysis. Clinical Psychology & Psychotherapy 26: 191–203.

46.

Fang

Miller

, et al. (2011) Internet-based intervention for the treatment of online addiction for college students in China: A pilot study of the healthy online self-helping center. Cyberpsychology, Behavior, and Social Networking 14: 497–503.

47.

Tang

Posner

Rothbart

, et al. (2015) Circuitry of self-control and its role in reducing addiction. Trends in Cognitive Sciences 19: 439–444.

48.

Tateno

Kim

D-J

Teo

, et al. (2019) Smartphone addiction in Japanese college students: Usefulness of the Japanese version of the Smartphone addiction scale as a screening tool for a new form of internet addiction. Psychiatry Investigation 16: 115–120.

49.

Van der Linden

(2015) Commentary on: Are we overpathologizing everyday life? A tenable blueprint for behavioral addiction research Addictions as a psychosocial and cultural construction. Journal of Behavioral Addictions 4: 145–147.

50.

Viechtbauer

(2010) Conducting meta-analyses in R with the metafor package. Journal of Statistical Software 36: 1–48.

51.

Viechtbauer

Cheung

(2010) Outlier and influence diagnostics for meta-analysis. Research Synthesis Methods 1: 112–125.

52.

Walther

Schuetz

Hamm

, et al. (2011) Quality of reporting of systematic reviews and meta-analyses: PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses). RöFo—Fortschritte auf dem Gebiet der Röntgenstrahlen und der bildgebenden Verfahren 183: 1106–1110.

53.

Winkler

Jeromin

Doering

, et al. (2020) Problematic Smartphone use has detrimental effects on mental health and somatic symptoms in a heterogeneous sample of German adults. Computers in Human Behavior 113: 106500.

54.

Winkler

Dörsing

Rief

, et al. (2013) Treatment of internet addiction: A meta-analysis. Clinical Psychology Review 33: 317–329.

55.

Yang

Kim

(2018) Effects of a prevention program for internet addiction among middle school students in South Korea. Public Health Nursing 35: 246–255.

56.

Young

(1998) Internet addiction: The emergence of a new clinical disorder. Cyberpsychology & Behavior 1(1): 237–244.

57.

H-G

Son

(2016) Effects of ACT on Smartphone addiction level, self-control, and anxiety of college students with Smartphone addiction. Journal of Digital Convergence 14: 415–426.

58.

Zajac

Ginley

Chang

(2020) Treatments of internet gaming disorder: A systematic review of the evidence. Expert Review of Neurotherapeutics 20: 85–93.

59.

Zajac

Ginley

Chang

, et al. (2017) Treatments for internet gaming disorder and internet addiction: A systematic review. Psychology of Addictive Behaviors 31: 979–994.

Tackling the ‘digital pandemic’: The effectiveness of psychological intervention strategies in problematic Internet and smartphone use—A meta-analysis

Abstract

Objective:

Methods:

Results:

Conclusion:

Keywords

Introduction

Methods

Eligibility criteria and study selection

Literature search

Screening abstracts

Data extraction and quality assessment

Data synthesis and analysis

Measurement of PSU/PIU

Treatment classification

Stakeholder inclusion

Results

Study characteristics

Meta-analytic results

PIU

PSU

Sensitivity analysis

Outliers

Subgroup and moderator analysis

Publication bias and p-hacking

Discussion

Footnotes

Declaration of Conflicting Interests

Funding

ORCID iD

Data Availability

References