Treatment Approaches for Problematic Usage of the Internet (PUI): A Dual-Level Meta-Analysis of Meta-Analyses and Randomized Controlled Trials

Abstract Background Problematic usage of the internet (PUI) is a growing public health concern, with a rapidly expanding literature of primary trials and meta-analyses on treatment approaches. However, this evidence base remains fragmented due to differences in definitions, assessment tools, and methodological approaches, which limit its applicability for clinicians, policymakers, and researchers. To address this gap, we conducted a dual-level meta-analysis: first, a systematic review and meta-analysis of existing meta-analyses; second, a secondary meta-analysis of randomized controlled trials extracted from these studies. Methods Following PRISMA guidelines, we systematically searched PubMed between 2013 and 2024 for meta-analyses assessing treatments for a range of PUI-related conditions, including internet addiction and problems related to online gaming, online pornography, online shopping, social media use, and smartphone use. Eligible studies evaluated behavioral, pharmacological, neuromodulatory, or physical exercise interventions. Studies were excluded if they were not meta-analyses, did not focus on PUI, or lacked sufficient data. Methodological quality was assessed using AMSTAR 2.0 and Cochrane Risk of Bias 2.0 tool for the meta-analyses and RCTs, respectively. A random-effects model was applied to analyze pooled effect sizes, accounting for heterogeneity and publication bias. To enhance rigor, we conducted a secondary meta-analysis of randomized controlled trials (RCTs) extracted from these meta-analyses. This study was registered with the OSF (Open Science Framework) at https://osf.io/8uc2w . Findings Among 402 identified records, 329 were screened, and 20 meta-analyses met the inclusion criteria, comprising 35 units of analysis. Results of the meta-annalysis of meta-analyses confirmed the overall impact of PUI treatments (Standardized Mean Difference [SMD] = −1.41, 95% CI: [−1.54; −1.18], p < 0.0001), with variations across PUI dimensions, treatment categories, and age groups. In the second phase, from the identified 386 unique studies included in the eligible meta-analyses, 44 RCTs were identified and included for the secondary meta-analysis. Behavioral interventions, including CBT and its variants, were the most commonly used approaches (91 units of analysis in 35 RCTs) and showed a significant impact (SMD = –1.91; 95% CI: [–2.22, –1.60]). Neuromodulatory interventions showed the largest effect size (SMD = –3.14; 95% CI: [–3.14, –3.14]) but were underexplored (six units in two RCTs). Pharmacological interventions were also effective (SMD = –1.34; 95% CI: [–1.51, –1.18]) across 14 units in five RCTs. Physical exercise was effective with a lower effect size (SMD = –1.15; 95% CI: [– 1.77, –0.53]) based on eight units in three RCTs. Furthermore, serious concerns were detected with regard to the quality of studies, risks of bias, and methodological rigor at both levels. Interpretation This study provides an overview of the available evidence for PUI treatments and their observed outcomes. Behavioral interventions, particularly CBT, are supported by a substantial body of evidence showing positive results, while pharmacological and neuromodulatory treatments appear promising based on a smaller number of studies. Advancing the field will require RCTs with greater methodological rigor, standardized assessment tools, longitudinal designs, and culturally adapted interventions to improve the accessibility and impact of these treatments.