PLOS Mental Health Paper Examines Mindfulness as a Skills Construct in Internet-Delivered DBT for Substance Use Disorders

Emotion dysregulation, not mindfulness practice, was the clearest driver of substance use outcomes in a 12-week digital DBT trial, a finding that reframes where clinicians and self-directed practitioners should focus their attention first.

The study, led by Danielle Downie, Alexander R. Daros, Chelsey R. Wilks, and Lena C. Quilty, enrolled 72 adults with past-year substance use disorders in a self-guided internet-delivered dialectical behavior therapy program. Participants completed assessments at baseline and at weeks 4, 8, and 12, giving the researchers a layered, within-person view of how skill acquisition corresponded to changes in SUD severity and functional disability over the full 12-week arc.

The team tracked three constructs simultaneously: mindfulness, DBT skills use, and emotion dysregulation. Increases in emotion dysregulation correlated with worse SUD severity and disability both within and between participants, making it the most consistent signal in the dataset. Mindfulness fell short of statistical significance after correction for multiple comparisons, but produced a small-to-medium effect size at the between-person level. Put plainly: participants who accumulated more mindfulness gains across the entire program tended to have better outcomes; the connection was less reliable week to week.

That distinction matters, because in DBT, mindfulness is not a mood or a disposition. It is a teachable, measurable set of discrete behaviors. The core "what" skills are observe (attending to present-moment experience without reacting), describe (putting words to that experience), and participate. These pair with "how" skills: nonjudgmental stance, one-mindfully, and effectively. For anyone working with substance cravings, urge surfing is where these skills converge most practically. You treat the craving as a wave, track its physical rise and fall using the observe skill, and let it crest without acting. The Downie team was trying to capture precisely this kind of accumulating, discrete skill use over time, and their results suggest that tracking it week by week is harder than it sounds.

Engagement metrics, including acceptability, usability, and time spent with the program, showed no significant moderation effects after statistical correction. Several effects still approached small-to-medium sizes, a recurring pattern the authors read as promising but unconfirmed. Their recommendation: sharpen the measurement tools for skill acquisition and extend follow-up periods beyond 12 weeks before drawing firm conclusions about digital behavioral therapies.

The between-person mindfulness signal is the result worth sitting with. It suggests that building the skill consistently over time carries a real, if modest, weight. A low-stakes way to start accumulating that kind of practice: the next time a craving or difficult emotion arrives, try one observe-describe cycle. Pause, name what is present in a single sentence ("There is tightness in my chest and a pull toward the refrigerator"), and hold that description without evaluating it for 60 seconds. Do not try to change the experience. That one cycle is the unit of practice the Downie study was trying to measure, and its data suggest those who build it into a consistent pattern carry something forward that the program's weekly check-ins could barely capture.