SBIR Phase I: An Automated Tool to Measure and Predict Substance Use Recovery

The broader impact of this Small Business Innovation Research (SBIR) Phase I project is to improve access to high-quality, responsive, and evidence-based treatments for risky users of alcohol and other drugs. Current standards of care are cost-prohibitive, time-intensive and overly reliant on community support groups lacking a scientific basis or a mechanism for accountability. The research proposed here will add to an interactive online platform providing tools and treatment for individuals seeking to change their relationship to substances. The technological innovation supported by this research will result in a graphical meter that provides users of the platform with a simple-to-interpret interface. This tool is designed to strengthen motivation, enhance capacity for self-assessment, and allow for the better allocation of care provision resources. This development will give the proposing organization a major competitive advantage as it contributes to the growing field of digitally-provided behavioral healthcare products. Employers, insurers, and consumers have communicated a clear demand for new methods of treating individuals with substance use disorders, and this technology will equip the proposing organization to meet that demand. The proposed project addresses the "black box" of recovery from risky use of alcohol and other drugs. Ample research demonstrates that behavioral interventions can be effective ways to mitigate risk for individuals with unhealthy use of alcohol and other drugs. Monitoring, assessing, and quantifying that recovery, however, remains out of reach for behavioral interventions delivered in the current standard-of-care. Instead of present-moment substance use recovery, the typical measure remains previous substance consumption. By statistically analyzing a range of available data from users of an online substance use recovery platform, this research will identify the data most correlated with successful recovery from substance use. Regression modeling and bivariate correlations will explore the relationships between different data within the platform and several validated outcome measures already integrated into the intervention. This research will use discourse analysis and natural language processing to identify and categorize user-submitted content to identify linguistic elements related to substance use recovery or risk of relapse. The principal outcome will be a holistic measure of recovery from substance use aggregating multiple instruments and domains of substance health. It will be developed into a graphical interface that communicates to both users and other behavioral healthcare stakeholders in-the-moment recovery from risky substance use.