Clinical decision support for early detection of relapse risk in adolescents with substance use disorders

The National Institute on Drug Abuse awarded Pillsmart Inc. $399,993 on June 1, 2026, under the Drug Use and Addiction Research Programs (CFDA 93.279) to develop and validate the technical feasibility of LEVL, a clinical decision support system for early detection of relapse risk in adolescents with substance use disorders.

LEVL integrates physiological markers (heart rate, heart rate variability, sleep), behavioral data, and environmental context into an FHIR-compliant adaptive platform that proactively alerts clinicians when elevated relapse risk is detected. The system includes an adolescent-facing mobile application to support timely interventions and self-management. The Phase I project will establish prototype usability, clinician adoption, and preliminary validation of the predictive risk model through a structured Delphi process for expert consensus on biomarkers, proprietary analytics for risk detection, iterative usability testing, and a 90-day feasibility trial. Clinical predictive validity and direct clinical effectiveness in reducing relapse will be evaluated in future Phase II trials.

Performance of work occurs in Brooklyn, New York. The period of performance runs from June 1, 2026, through May 31, 2027.

Abstract

Over 60% of adolescents with substance use disorders relapse within 90 days of completing treatment, largely due to fragmented data, delayed intervention, and lack of adolescent-specific clinical guidelines. This phase I sbir project aims to develop and validate the technical feasibility of levl, a novel clinical decision support (CDS) system designed specifically for early detection of relapse risk among adolescents. Levl integrates physiological markers (heart rate, heart rate variability, sleep), behavioral data, and environmental context into an adaptive, fhir-compliant platform that proactively alerts clinicians when elevated relapse risk is detected. The platform also features an adolescent-facing mobile application to support timely interventions and self-management. The primary goal of this phase I study is technical feasibility, establishing the prototype's usability, clinician adoption, and preliminary validation of the predictive risk model. Clinical predictive validity and direct clinical effectiveness in reducing relapse will be rigorously evaluated in future phase II trials. This study will use a structured delphi process for expert consensus on biomarkers, proprietary analytics for predictive relapse risk detection, iterative usability testing, and a 90-day feasibility trial to refine the prototype. Adolescents experiencing untreated or recurrent SUD may incur lifetime healthcare expenditures exceeding $30,000-$40,000, making early intervention a valuable cost-savings strategy. Early identification and proactive intervention have strong potential to significantly reduce healthcare costs associated with relapse. By enabling clinicians to intervene earlier through proactive alerts, levl aligns with federal healthcare priorities emphasizing evidence-based care, targeted cost-savings, and early intervention in substance use disorders.