Targeted cognitive behavioural therapy shows promise for high-risk adolescents in school-based internet use prevention

A recent study suggests that school-based cognitive behavioural therapy programs are more effective than media literacy lessons for adolescents at high risk of problematic gaming and internet use, signalling a shift towards targeted prevention strategies.

A targeted cognitive behavioural therapy course appears to give schools a more effective option for pupils already showing signs of problematic gaming or internet use, but the same approach does not seem to offer a meaningful advantage when delivered to adolescents as a whole. That is the main conclusion of a randomised clinical trial published in JAMA Network Open on 4 September, with the journal and its media release both framing the result as evidence for indicated prevention – aimed at higher-risk teenagers – rather than a universal programme for every pupil. (jamanetwork.com)

The distinction matters because the headline result rests on a relatively small high-risk subgroup within a much larger school sample. In the trial, 205 adolescents classed as high risk at the outset had lower symptom severity after 12 months if they received PROTECTtraining rather than the media-literacy comparator, PROTECTinfo, with mean modified Video Game Dependency Scale scores of 14.7 against 17.7. In the full cohort of 1,793 students, however, the difference was not statistically significant, at 8.8 versus 9.4. The paper defined the high-risk group as pupils meeting at least two DSM-5 criteria for gaming disorder or unspecified internet use disorder at baseline. (jamanetwork.com)

The study itself was unusually practical in design, which helps explain why the authors think the findings are relevant to real schools rather than ideal laboratory settings. PROTECTconfirm ran from 1 July 2020 to 31 August 2024 across 44 secondary schools in Baden-Württemberg, Germany, with follow-up at one, four and 12 months. Pupils aged 11 to 18 were recruited from 92 classes, and 1,793 from 90 classes entered the analysed sample – 81.6 per cent of the 2,198 students who received one of the interventions. Sessions were delivered during normal school hours by trained local prevention professionals, including social workers, school psychologists, teachers and youth counsellors; four classes moved online during Covid-19 school closures. (jamanetwork.com)

What separated the two programmes was not the amount of contact but what happened inside it. Both arms involved four weekly 90-minute sessions, yet PROTECTtraining focused on motivation to change, maladaptive thinking, dysfunctional coping and problems with emotion regulation. According to the JAMA paper, its core elements included psychoeducation on addictive mechanisms, cognitive restructuring, behavioural activation, problem solving and emotion-regulation skills. PROTECTinfo, by contrast, covered smartphone use, privacy settings, online communication, the risks and benefits of gaming, cyberbullying and the legal aspects of digital behaviour, but did not teach cognitive behavioural skills. Heidelberg University said the two school-based interventions were deliberately matched for duration and implementation effort. (jamanetwork.com)

Katajun Lindenberg, the Heidelberg psychologist who led the work, argued that the study helps answer a live policy question about how best to protect young people’s mental health in an intensely digital environment. In the university’s account of the findings, she said that, in direct comparison, the behavioural therapy programme was “50 percent more effective” than the media education course for the adolescents most at risk. Heidelberg University said the research was funded by the then Federal Ministry of Education and Research and Baden-Württemberg’s former Ministry of Social Affairs and Integration, a detail also carried in News-Medical’s English-language summary of the release. (uni-heidelberg.de)

The results were not a clean sweep for the cognitive behavioural approach, and the paper is fairly restrained about that. Secondary outcomes – including broader internalising, externalising and transdiagnostic symptoms – did not differ significantly between groups. The responder analysis also showed modest absolute rates of reliable improvement in the full sample: 12.7 per cent for PROTECTtraining and 9.9 per cent for PROTECTinfo. In the high-risk subgroup, the improvement rates were 46.4 per cent and 38.6 per cent respectively, while reliable deterioration was least common among high-risk pupils who received the CBT-based programme, at 3.6 per cent. The authors said both interventions were well tolerated, but their own conclusion was that stepped prevention after screening is likely to be the more efficient model. (jamanetwork.com)

The school trial also sits inside a wider Heidelberg research effort that is testing different ways to tackle unhealthy digital habits. A separate entry in the German Clinical Trials Register shows Lindenberg’s group has completed an adult study called PROTECTapp, registered as DRKS00034109 on 8 May 2024 and last updated on 6 July 2026. That project is not the same trial and involves adults rather than school pupils, but it uses the same PROTECT umbrella for a smartphone-based training tool in which participants swipe away harmful internet-related cues and pull healthier activities towards themselves. The registry describes the work as part of a broader programme spanning prevention, early intervention and treatment for internet- and gaming-related problems. (drks.de)

Heidelberg has also linked the school work to a broader European prevention agenda. In a 2024 university briefing, the institution said Lindenberg would lead the German recruitment centre for BootStRaP, a multinational project expected to involve more than 8,300 teenagers in nine countries and coordinated by the University of Hertfordshire across a 14-country collaboration. The plan was to combine app-based records of daily online habits with anonymous passive phone data such as screen time, activity and sleep in order to identify risk profiles and then build more individualised interventions. Against that backdrop, the new JAMA trial offers one immediate lesson for schools: broad digital literacy lessons may have value, but the clearest clinical gain came when support was aimed at the pupils already most likely to need it. (uni-heidelberg.de)