A School-Based Digital Game Modestly Reduced Adolescent Depression: What Clinicians Treating Cannabis-Using Teens Should Know
| Audience | Family physicians, pediatric and adolescent medicine clinicians, and cannabis clinicians managing adolescents who use cannabis, alcohol, or nicotine alongside depressive or anxiety symptoms |
| Primary Topic | A prespecified secondary analysis of a randomized clinical trial testing a codesigned digital behavioral health game against an active control among 532 adolescents recruited from Connecticut school-based health programs, with past-30-day cannabis use as one of several eligibility criteria, published in JAMA Network Open, August 2026 |
| Source | Read the study on PubMed |
A School-Based Digital Game Modestly Reduced Adolescent Depression: What Clinicians Treating Cannabis-Using Teens Should Know
A randomized trial of 532 adolescents recruited in part for past-30-day cannabis, alcohol, or nicotine use found that a codesigned digital behavioral health game modestly reduced depressive symptoms over 12 months compared with an active control, with no cannabis-use outcome measured and no benefit detected for anxiety.
| Study Type | Prespecified secondary analysis of a parallel-group, active-controlled randomized clinical trial (CONSORT-reported) |
| Population | 532 adolescents ages 16 to 19 (mean age 16.6, SD 0.7) recruited from 15 Connecticut school-based health programs, October 2021 to February 2024 |
| Eligibility | No prior opioid misuse, plus past-30-day use of alcohol, cannabis, nicotine, or other nonopioid substances, or a positive screen for depressive or anxiety symptoms |
| Intervention | PlaySmart, a codesigned digital behavioral health game delivered on tablets in 40- to 60-minute sessions over up to 6 weeks (n=269) |
| Control | Active control using commercially available games (e.g., The Sims, Papers Please) matched for time and attention (n=263) |
| Measures | Depressive symptoms (PHQ-8) and anxiety symptoms (GAD-7); secondary psychosocial outcomes included beliefs about psychological services (BAPS), help-seeking intentions (GHSQ), and emotion regulation (S-DERS) |
| Follow-Up | Baseline, 6 weeks, and 3, 6, and 12 months |
| Depression Finding | Adjusted PHQ-8 scores 1.12 points lower in the intervention group at 12 months (95% CI, -1.85 to -0.38; P=.002; Cohen d=0.21) |
| Anxiety Finding | No significant group-by-time interaction for GAD-7 (F3,1308=1.69; P=.17) |
| Adverse Events | 2 mild adverse events (brief gameplay-related discomfort); no serious harms |
| Journal | JAMA Network Open |
| Published | August 3, 2026 |
| DOI | 10.1001/jamanetworkopen.2026.29173 |
| PMID | 42606863 |
Researchers conducted a prespecified secondary analysis of a randomized clinical trial testing PlaySmart, a codesigned digital behavioral health game originally built to prevent opioid misuse, against an active control among 532 adolescents recruited from 15 school-based health programs in Connecticut.
Eligible teens were 16 to 19 years old with no history of opioid misuse, and qualified by reporting past-30-day use of alcohol, cannabis, nicotine, or other nonopioid substances, or by screening positive for depressive or anxiety symptoms, meaning cannabis use was one of several possible entry points rather than a study focus.
Depressive symptoms, measured with the 8-item Patient Health Questionnaire (PHQ-8), showed a significant group-by-time interaction (F3,1308=2.93; P=.03), with intervention participants scoring 1.12 points lower than control participants at 12 months (95% CI, -1.85 to -0.38; P=.002; Cohen d=0.21).
The effect size was small, consistent with universal prevention programs aimed at broad populations rather than diagnosed clinical depression, and the authors describe the pattern, a benefit that widened over a year rather than appearing immediately, as consistent with a possible sleeper effect in which skills learned during gameplay become more usable once initial skepticism about a game-based intervention fades.
The group-by-time interaction for anxiety symptoms (GAD-7) was not significant (F3,1308=1.69; P=.17), and no between-group differences in anxiety emerged at any follow-up assessment.
Help-seeking intentions and emotion regulation also showed no significant intervention effect, and moderation analyses found no differential benefit by sex, grade level, food insecurity, familial substance misuse, or peer substance-use norms.
Intervention participants reported more favorable beliefs about psychological services (BAPS) immediately after the 6-week gameplay period (adjusted difference, 0.03; 95% CI, 0.01-0.05; P=.003), and this difference remained similar at later assessments even though the overall group-by-time interaction for BAPS was not significant.
An exploratory, post hoc mediation analysis found that this early improvement in BAPS partially explained the depressive symptom reduction observed at 6 months (average causal mediation effect, -0.13; 95% CI, -0.30 to -0.01; P=.02), though no mediation effect was detected at 3 or 12 months, and the authors describe this analysis as hypothesis-generating rather than confirmatory.
This trial did not measure cannabis use as an outcome, did not report results stratified by baseline cannabis use, and did not test whether the game changed cannabis use, craving, or related behaviors; past-30-day cannabis use was only one of several qualifying eligibility criteria, alongside alcohol, nicotine, other nonopioid substances, or elevated mood or anxiety symptoms alone.
Readers should not interpret this trial as evidence that the game reduces cannabis use or that it was designed with cannabis specifically in mind. It is a general adolescent mental health intervention studied in a population that, by design, regularly includes substance-using teens.
This trial adds to a growing but still limited literature on serious games and interactive digital media for adolescent mental health, most of which has tested short-term effects or used passive comparison groups that leave open questions about whether apparent benefits reflect a game’s specific content or simply attention and novelty.
By using an active, time-matched control of commercially available games, this trial offers a more rigorous test of content-specific effects than many earlier digital mental health studies, though its authors are explicit that the psychosocial mechanism findings remain exploratory and require replication in preregistered studies.
What I find useful about this trial is the population it was run in. School-based health programs reach exactly the mix of teens I worry about most, adolescents who may be using cannabis, alcohol, or nicotine and who also carry real depressive symptoms, often without ever reaching specialty mental health care. A scalable, low-cost digital tool that modestly reduces depressive symptoms in that setting, tested against an active control rather than nothing, deserves attention.
I want to be direct about what this study is not. It is not a cannabis trial. Cannabis use got a teenager into the study alongside several other possible entry points, but nothing here was measured, analyzed, or reported about cannabis use itself. I would not tell a patient or a parent that this game treats cannabis use or reduces the urge to use it. What I would say is that if a cannabis-using adolescent in my practice is also dealing with depressive symptoms, a scalable, evidence-supported digital tool like this is a reasonable option to know about, alongside, not instead of, direct conversation about their cannabis use.
How to Read a Secondary Analysis With an Exploratory Mediation Finding
This trial’s headline result, a modest depressive symptom reduction against an active control, is genuinely well supported. Its secondary claim, that improved beliefs about seeking help partly explain that benefit, rests on much thinner statistical ground.
Four checks keep the trial’s real contribution from being overstated in either direction, toward cannabis relevance it does not have, or away from the depression benefit it does show.
A Four-Step Reading Frame
Separate the primary trial from this secondary analysis
Depression and anxiety were prespecified secondary outcomes; the original trial was powered for opioid-related outcomes, not these.
Note what was never measured
Cannabis use was an eligibility criterion only. No cannabis outcome, craving, or use-frequency data appears anywhere in this analysis.
Treat the mediation finding as hypothesis-generating
The BAPS-to-depression mediation effect was exploratory, post hoc, and only significant at one of three follow-up points (P=.02 at 6 months).
Weigh the effect size, not just the p-value
A statistically significant 1.12-point PHQ-8 difference (Cohen d=0.21) is real but small, typical of universal prevention programs.
A Depression Benefit, Not a Cannabis Finding: Eight Ways to Read This Trial
This trial modestly improved adolescent depression in a population that regularly includes cannabis-using teens, but it never measured cannabis use itself. These eight perspectives keep that distinction clear while still drawing out what is genuinely useful here.
What This Means If Your Teen Uses Cannabis and Also Seems Down
This trial found that a school-delivered digital game modestly reduced depressive symptoms over a year among teens recruited in part for recent cannabis, alcohol, or nicotine use, but it did not test whether the game changes cannabis use itself, so it is not a cannabis treatment or a substitute for talking directly about substance use.
If your teenager is struggling with low mood and also uses cannabis, this study suggests that scalable, school-based digital tools can help with the mood piece, while the cannabis use still needs its own direct conversation with a clinician who understands both issues.
A Scalable Option Worth Knowing, Not a Cannabis Intervention
This trial supports a codesigned digital game as one low-cost, scalable option for reducing depressive symptoms in adolescents seen in school-based health programs, a setting that overlaps heavily with teens who use cannabis, alcohol, or nicotine.
Clinicians should be precise with patients and families about what this study does and does not show: it measured mood outcomes, not cannabis use, so it belongs in a broader treatment plan rather than as a stand-alone response to adolescent cannabis use.
Why a Non-Cannabis Study Still Belongs on a Cannabis Clinician's Radar
Adolescents who self-medicate anxiety or depression with cannabis are a familiar population in cannabis-focused practice, and this trial was conducted in a setting, school-based health programs, that regularly serves exactly those teens, even though cannabis use itself was never analyzed.
Knowing that a validated, active-controlled digital tool can modestly improve depressive symptoms gives cannabis clinicians another option to discuss with families who want to address mood symptoms without adding medication or without cannabis being the only tool on the table.
A Real Effect, but a Small One, With an Exploratory Mediation Claim
The 12-month depression finding is statistically solid, a significant group-by-time interaction (P=.03) and a between-group difference with a tight confidence interval (95% CI, -1.85 to -0.38; P=.002), though the effect size itself was small (Cohen d=0.21).
The mediation analysis linking early belief change to later depression reduction was exploratory, post hoc, and significant at only one of three follow-up points (P=.02 at 6 months), which means it should be read as a plausible mechanism worth testing, not a confirmed pathway.
School-Based Health Programs Are the Right Place to Test This
Delivering a mental health intervention inside school-based health programs reaches adolescents, including those from historically underserved racial and ethnic groups who reported some of the highest rates of sadness and hopelessness nationally, who might not otherwise access mental health care.
The trial’s diverse sample, 45.3% Black or African American and 38.0% Hispanic or Latinx participants, and its use of an active rather than passive control, make this a stronger test of real-world feasibility than many earlier school-based digital mental health studies.
Secondary Outcomes, Self-Report, and Multiple Comparisons
Depression and anxiety were prespecified secondary outcomes in a trial originally powered around opioid-related endpoints, and every outcome relied on adolescent self-report without biological or collateral verification.
The trial tested several secondary and exploratory outcomes without a described correction for multiple comparisons, which raises real questions about how many of these findings, particularly the borderline mediation result, would hold up in a study purpose-built and powered to test them. A study purpose-built and adequately powered to test that mediation pathway directly would give clinicians far more confidence before treating it as a real mechanism.
A Template for Testing Digital Mental Health Tools Rigorously
This trial’s active, time-and-attention-matched control group is a methodological strength many digital mental health studies lack, and it allows a more credible test of whether the game’s specific content, rather than novelty or attention, drove the depression benefit.
The authors’ own framing points to the clear next step: a study purpose-built and powered to test the BAPS-mediation pathway directly, ideally with objective corroboration of self-reported outcomes and preregistered subgroup analyses.
Scalable Tools Matter Even When Effects Are Small
A 1.12-point PHQ-8 difference will not feel dramatic to any one adolescent, but delivered at scale through existing school-based health infrastructure, a modest, low-cost, low-burden intervention like this can meaningfully shift population-level depressive symptom burden.
The relevant public health caution is framing: this trial should not be described as a cannabis intervention or as evidence that digital games reduce substance use, since neither was tested, and overstating its scope would misrepresent a genuinely useful but narrowly defined finding.
Join the Conversation
Have a question about how this applies to your situation? Ask Dr. Caplan
Want to discuss this topic with other patients and caregivers? Join the forum discussion
Frequently Asked Questions
What did this trial test?
Researchers tested PlaySmart, a codesigned digital behavioral health game, against an active control of commercially available games among 532 adolescents recruited from Connecticut school-based health programs, following them for 12 months to assess depressive and anxiety symptoms and related psychosocial outcomes.
Was this a study about cannabis?
No. Cannabis use was one of several possible eligibility criteria, alongside alcohol, nicotine, other nonopioid substance use, or elevated depression or anxiety symptoms alone. The trial never measured cannabis use, craving, or related outcomes.
Did the digital game reduce depression?
Yes, modestly. At 12 months, adjusted PHQ-8 depression scores were 1.12 points lower in the intervention group than the control group (95% CI, -1.85 to -0.38; P=.002; Cohen d=0.21), with a significant group-by-time interaction (P=.03).
Did the game help with anxiety symptoms?
No significant benefit was found. The group-by-time interaction for anxiety symptoms (GAD-7) was not significant (P=.17), and no between-group differences emerged at any follow-up assessment.
How was the game different from a typical screen-time control?
The control group played commercially available games, such as The Sims and Papers Please, matched for time and attention, which is a more rigorous comparison than a passive or no-treatment control because it accounts for the effects of novelty and screen engagement alone.
What might explain why depressive symptoms improved?
An exploratory, post hoc mediation analysis found that early improvements in beliefs about seeking psychological help, measured 6 weeks after gameplay, partially explained the depression reduction seen at 6 months, though this mediation effect was not detected at 3 or 12 months.
Who were the adolescents in this trial?
532 teens ages 16 to 19 (mean age 16.6) recruited from 15 Connecticut school-based health programs; 45.3% identified as Black or African American, 33.6% as White, and 38.0% as Hispanic or Latinx, with 53.4% identifying as male.
Were there any safety concerns?
Two mild adverse events were documented, both involving brief discomfort during or after gameplay judged as possibly related to the intervention. No serious or unexpected harms occurred.
What are the main limitations of this trial?
Depression and anxiety were prespecified secondary outcomes rather than the trial's original powered endpoint, all measures relied on adolescent self-report, the mediation analysis was exploratory and post hoc, and the trial did not describe correction for multiple comparisons across its several secondary outcomes.
Should this change how clinicians manage cannabis-using adolescents?
This trial offers a scalable, evidence-supported option to consider for adolescents with depressive symptoms, including those recruited from substance-use-inclusive populations, but it does not address cannabis use directly and should not replace direct clinical conversation about a teen's cannabis use.
