Substance Use Disorder Artificially Intelligent chatbot for screening, assessment & referral: SUD Bot

点击翻译按钮获取中文摘要

ABSTRACT The opioid epidemic is considered one of the most severe public health crises we are facing in the U.S. and is worsened by use of stimulants such as methamphetamine. The US Preventive Services Task Force (USPSTF) recommends screening for unhealthy drug use accompanied by offers of and referrals to services that include accurate diagnosis, effective treatment and appropriate care of a substance use disorder (SUD) for persons aged 18 and older. Automating screening and assessment for SUD and removing stigmas and other barriers to treatment can improve the scale, efficiency, and equitable access to SUD care. While artificial intelligence (AI) chatbots using natural language processing (NLP) and Machine Learning (ML) are increasingly common, they (a) fall short of systems that mimic social conversations with persuasive responses to motivate action (b) are not interoperable with service delivery to link users immediately with clinic appointments and (c) aren’t FDA approved under mandates to regulate medical devices that complete assessments for medical outcomes. Specific to SUD, there is a compelling need for AI chatbots that minimize stigma associated with SUD care and resources. In this Phase I STTR, Clinic Chat, LLC, will build on prior research showing the success of their chatbots to support improved access to chronic illness medication in partnership with Be Well Texas, a provider of SUD services to develop and beta-test the feasibility, navigability, and acceptability of using a next generation AI chatbot, called SUD Bot, to facilitate access to and utilization of SUD services and resources. We aim to 1: Enhance existing Clinic Chat AI Chatbots with (a) persuasive SUD screening and treatment messaging and (b) infrastructure with capacity to simulate human conversation and be accessible in English and Spanish via multiple platforms, i.e., text messaging, including voice and video; 2: Build fast healthcare interoperative resource (FHIR) linkages that will allow users to self-schedule appointments for treatment or to access peer support through the recovery network within Be Well Texas; and 3: Conduct a beta-test to determine the functionality and navigability of the FHIR-enabled SUD Bot system. We will recruit 200 users to use the FHIR-enabled SUD Bot system in three weeklong waves interspersed with iterative system refinement based on user feedback. Completion of this Phase I study will generate an FHIR-enabled minimum viable product (MVP) with initial functionality and navigability feedback ready for a Phase II STTR randomized clinical trial testing system efficacy. Our goal is to have an effective and scalable tool that can be adapted for use in any organization delivering SUD services and commercialized through fees for service or Medicaid/Medicare reimbursement.