PlaySmart video game nudged teen depression scores down a point over a year

PlaySmart is a video game that sets teenagers tasks like identifying signs of depression in a friend’s room, then makes them live with whatever they choose.

For six weeks in Connecticut, 269 students played it for about an hour after school once a week while another 263 played ordinary commercial role-playing games with no therapeutic content. Then both groups were left alone for a year.

When the year was up, the PlaySmart players came out lower on depression. At 12 months, adjusted mean scores were 5.30 against 6.42, on a questionnaire that runs from 0 to 24. Barely a point. What makes it worth a second look is that the gap had outlasted eleven months of nobody playing anything.

The teen depression nobody treats

Caroline Barry, the study’s first author, puts the problem at the door of the clinic, before any treatment starts.

“One of the biggest challenges in adolescent mental health is that many young people never receive care because of barriers including stigma, limited access to providers, or just not recognizing they could benefit from help,” she says.

The timing makes it urgent. Depression often begins in the teens or early adulthood, and it is not a mood that passes: depression in teens is a serious medical illness, more than just a feeling of being sad or blue for a few days. It brings an intense feeling of sadness, hopelessness, and anger or frustration that lasts much longer.

Teenagers already using alcohol or cannabis sit at higher risk, since other mental health conditions, including substance use, raise the odds. Those were exactly the teenagers this trial went looking for.

What PlaySmart asks a player to do

The game puts the player inside a story built from things real teenagers described. Behavioral health experts in the play2PREVENT Lab gathered scenarios from teens and other stakeholders while developing it, documenting healthy and unhealthy decisions and the blind spots in what teens knew. Play involves choices with consequences. Tasks include identifying signs of depression in a friend’s room, or deciding to turn to a trusted adult for help with a difficult situation. Better choices lead to more positive outcomes for the player and other characters, while bad choices lead to more negative outcomes.

Players can then rewind. The game lets them work backward to analyze their decision-making and explore what might have happened had they made different choices, which is a mechanic no leaflet and no assembly talk can offer.

Why the control group matters more than the game

Most digital mental health trials compare an app against a waiting list, and a waiting list loses to almost anything. The team, reporting in the American Medical Association journal, JAMA Network Open, did not take that route.

Participants were randomized 1:1 to a digital game of approximately 300 minutes delivered during 6 weeks or to an attention-matched active control. The control group got games too, just commercial ones with no therapeutic content.

Lynn Fiellin, the senior author and a professor of biomedical data science and medicine at Dartmouth’s Geisel School of Medicine, puts the weight on that design.

“Serious games for prevention and mental health promotion are a rapidly growing area, but the evidence base is still emerging,” she says. “Few clinical trials in this space combine a randomized design with commercial games as an active control and a 12-month follow-up. The rigorous design of our study alone sets a high bar for demonstrating impact.”

She is right, and it cuts both ways. Testing against real games strips out the novelty, the screen time and the attention, so whatever survives is closer to the content itself. It also guarantees a smaller number at the end, because the comparison group was not doing nothing.

PlaySmart’s one point, held for a year

Here are the numbers the trial produced. At 12 months, adjusted mean depression scores were 5.30 in the intervention group and 6.42 in the control group, an adjusted difference of about 1.12 points.

Both groups landed in the mild band of that questionnaire. Nobody was cured; a cohort with some symptoms still had some symptoms. What travels is the persistence. Six weeks of gameplay, then eleven months of nothing, and the gap was still measurable. The authors describe the effect as modest and sustained reductions in depressive symptoms during 12 months of follow-up, and both halves of that phrase are doing honest work.

Anxiety, meanwhile, went nowhere: no association detected for the seven-question anxiety scale. That is half the stated target missed, and the authors say so as plainly as this.

The best clue about how the game works

Buried under the headline is a hint about mechanism. At 6 weeks, the PlaySmart players reported higher scores on a measure of beliefs about psychological services: they thought better of the idea of getting help. And exploratory mediation showed that early improvements in those beliefs were associated with depression score reductions at 6 months.

Read that as a chain and it says something unexpected. The game may not be treating anybody’s depression. It may be changing their mind about asking someone else to.

That would fit Barry’s framing of the problem, and it would fit the content, which spends its time on noticing symptoms and telling a trusted adult rather than on therapy technique. It would also explain the flat anxiety result, since nothing in the game goes near anxiety. The paper labels the analysis exploratory, which makes it a lead worth chasing and not yet a mechanism.

What the PlaySmart trial leaves open

This prespecified secondary analysis of a randomized clinical trial named depression and anxiety as outcomes before anyone looked at the data. That is a real distinction from the fishing that “secondary analysis” often implies, and it still is not the question the trial was designed around.

Everything rests on self-report. Teenagers filled in questionnaires about their own symptoms and their own attitudes, and no clinician assessed anybody. The teenagers were also specific: 16 to 19, drawn from 15 school-based health programs in one state, recruited because they were already using substances or already had symptoms. Around 45% were Black and 38% were Hispanic, groups the release notes report among the nation’s highest rates of sadness and hopelessness, so this is evidence about a population most digital health trials leave out, and not about teenagers in general.

Barry argues the consistency is the point: “The benefit held across all subgroups, which matters for an intervention meant to be delivered schoolwide or in communities broadly.”

Where a one-point gain in depression actually counts

The authors never call this a treatment. Their conclusion stops at a sentence worth reading twice: “digital games may support adolescent mental health at scale”. All the weight sits on the last two words.

One point is a poor showing for a therapy. For something that costs almost nothing per extra player, needs no clinician and reaches the teenagers Barry describes as never receiving care, it is a different sum altogether, because the alternative is not a better game but a counselor who does not exist. Connecticut cannot hire its way to 269 more of those, and it can copy a game for nothing.

A parent has less to work with here than the headline suggests. PlaySmart is a lab’s research tool and not something you can buy, and a teenager in trouble needs a person rather than a download. The route to getting teen depression treated still runs through a doctor or a school health program, and nothing in this trial moves it.

What the trial leaves behind is the lead about belief. If part of what keeps a teenager away from help is what they think of help, and six weeks of a game can shift that, then the cheapest thing worth building may be whatever gets them through the door. That analysis was exploratory. One trial, one hypothesis, and no more than that.