Proactive, real-time digital augmentation of behavior therapy shows promise for anticipating and deflecting children’s severe temper tantrums, a randomized clinical pilot study found.

In the Pistachio trial of 50 children with disruptive behavior disorder, psychiatrist Magdalena Romanowicz, MD, Department of Psychiatry and Psychology, Mayo Clinic, Rochester, Minnesota, and colleagues showed that intervention via alerts from smartwatches worn by the child is feasible.

In families that completed parent-child interaction therapy (PCIT), children given a smartwatch showed good adherence to wearing the devices, and parents responded very quickly to behavior prompts sent to their smartphones, providing proactive strategies to ward off behavior outbursts.

Conducted from March 2022 to December 2023, the study, presented in JAMA Network Open, enrolled children aged 3-7 years with externalizing behavior problems rated above the clinically significant range (I score ≥ 120; T score ≥ 60) on the Eyberg Child Behavior Inventory (ECBI).

The impetus for the study stemmed from a clinical challenge. “PCIT skills are typically applied reactively, after a tantrum begins, and parents often struggle to recall and implement strategies under stress,” Romanowicz told Medscape Medical News. “While parents didn’t specifically request an alert system, many expressed difficulty in anticipating outbursts, and this inspired us to explore whether real-time digital cues could enable proactive parenting and improve outcomes.”

Romanowicz and co-lead author Arjun P. Athreya, PhD, MS, Mayo Clinic electrical and computer engineer, had hypothesized that digital augmentation would be feasible and that adherence to the wearable would exceed 70%, enabling proactive PCIT implementation. “This was borne out: Smartwatch wear time averaged 75.7%, and parents responded to alerts in a median of 3.65 seconds,” Athreya said. While behavioral rating scale improvements did not reach statistical significance in this pilot study, tantrum durations were significantly shorter in the intervention arm (10.4 vs 22.1 minutes; P < .001).

The Study

The median age of children was 5.0 years (interquartile range, 4.0-6.0 years), and 34 (68%) participants were boys. Randomization assigned 28 to the PCIT-artificial intelligence (AI) arm and 22 to the PCIT-treatment as usual (TAU) arm. Of these, 37 (74%; 21 in the PCIT-AI arm and 16 in the PCIT-TAU arm) completed PCIT, with children wearing the watch a median of 75.7% (67.8%-82.9%) of the time.

The protocol involved the Garmin vívosmart 4 commercial fitness tracker to continuously monitor a child’s heart rate. When the mean heart rate over a 10-minute window rose into a predefined range of 105-129 beats/min, the system sent an alert to the parent’s smartphone via a secure app. Parents were then prompted to engage PCIT strategies immediately, with reminders sent at 20-minute intervals, if needed.

Outburst duration was based on parental logs of start and end times of observed tantrums, again using a smartphone application.

The study involved 12 weekly sessions of PCIT with either TAU or AI-enhanced therapy (PCIT-AI) augmented with real-time behavioral outburst alerts.

The primary outcome was achieving the a priori feasibility benchmark of greater than 70% compliance for both completion of PCIT sessions and daily smartwatch wear. Parental reaction time to behavior prompts was measured in seconds. Secondary outcomes were percentage changes in ECBI and scores on the Pediatric Sleep Questionnaire, which gauges such indices such as sleep disturbance, daytime sleepiness, and hyperactivity.

“Sleep disturbances are common in children with disruptive behavior disorders and can exacerbate emotional dysregulation,” Romanowicz said. “Assessing sleep provides insight into overall well-being and helps determine whether interventions have broader benefits beyond tantrum management.”

Despite the positive results, cost and access remain considerations, especially for families with limited resources. Future studies may explore lower-cost or insurance-supported options, Romanowicz said. Her group is now planning a fully powered, multicenter randomized trial. Key changes will include using pediatric-specific wearables for better fit and data accuracy and stratifying randomization by comorbidities and social determinants of health. The new trial will also feature integration with telehealth platforms and collaboration with schools and other stakeholders.

Commenting as a nonparticipant in the study, Cheryl B. McNeil, PhD, professor of psychiatry at the University of Florida in Gainesville, Florida, called it exciting to see research that foreshadows the future of evidence-based interventions for children’s mental health, particularly PCIT. “The study results demonstrate that parents can shorten child tantrums by being cued in advance to biological indicators of child dysregulation and prompted to use proven parenting responses to address children’s emotional needs,” she told Medscape Medical News. “This study is on the cutting edge of emerging mental health trends because it demonstrates the ability to use machine learning in naturalistic settings to expand the scope, impact, and reach of traditional mental health services.”

In an accompanying editorial, Jennifer N. Bress, PhD, and Clare C. Beatty, MA, Department of Psychiatry, Weill Cornell Medicine, New York City, agreed that digital health technologies offer welcome possibilities for extending treatment beyond traditional settings. “Bridging the gap between therapy room skills and their practical implementation is a long-standing challenge in mental health care,” they wrote. “Augmentative interventions such as PCIT-AI represent a promising avenue for addressing this gap and enhancing the effectiveness of existing evidence-based approaches.”

In their view, the Mayo study is “an elegant illustration of the use of digital health tools to augment existing interventions.” Such expansion is needed since even the most effective mental health interventions available are far from perfect. Cognitive-behavioral therapy, for example, considered the first-line treatment for depression and anxiety, has yielded remission rates of only about 50%.

But while the feasibility study suggested digital tools can indeed deliver therapeutic principles to daily life and provide support at real-life critical moments, they cautioned, “Realizing the potential of this approach will depend on understanding which patients benefit most, sustaining long-term engagement, and ensuring successful integration into varied clinical and cultural contexts.”

In the meantime, said Romanowicz, digital augmentation of PCIT is feasible, well tolerated, and associated with high adherence and faster parental response. “While preliminary, this approach shows promise for reducing tantrum severity and duration, offering a scalable tool to support families, especially in settings with limited access to behavioral health resources.” Larger trials are needed to confirm efficacy, optimize technology, and evaluate cost-effectiveness.

Funding for this study was provided by Mayo Clinic research programs. Romanowicz and Athreya reported having a pending patent. Romanowicz disclosed having nonfinancial, in-kind support from Kiddo outside of the submitted work. Co-authors Shekunov and Bobo also reported having pending patents. Bobo reported receiving research support from the National Institute of Mental Health, the Agency for Healthcare Research and Quality, the National Science Foundation, the National Institute of Nursing Research, the Myocarditis Foundation, the Watzinger Foundation, the Blue Gator Foundation, and the Mayo Foundation for Medical Education and Research, as well as making contributions to UpToDate. McNeil and the editorial commentators reported having no competing interests to disclose.