Advancing Scalable Alcohol Brief Interventions in the ED: A Trial of Conversational AI With Digital Reinforcements

NCT07661563 | Not Yet Recruiting

• 1 other identifier: 86953
• Study Type: interventional
• Target: 750
• Locations: 1 country
• Sites: 1

Brief Summary

The goal of this clinical trial is to learn if a novel digital strategy combining a generative AI counseling agent (MICA) and weekly SMS boosters works to reduce risky drinking among Emergency Department patients.

Conditions

Alcohol Misuse

Keywords

alcohol; emergency department; adults; AI; motivational interviewing; text messaging

Outcome Measures

Primary Outcomes (1)

• Mean weekly alcohol consumption

  Participants report (1) the average number of days per week they consumed alcohol and (2) the typical number of standard drinks consumed on a drinking day. Estimated weekly alcohol consumption will be calculated as: Mean weekly drinks = (average drinking days per week) × (typical drinks per drinking day)

  Baseline, 3-, 6-, and 12-month follow-ups

Secondary Outcomes (2)

• Number of heavy drinking days

  Baseline, 3-, 6-, and 12-month follow-ups

• Number of drinks per drinking day

  Baseline, 3-, 6-, and 12-month follow-ups

Study Arms (3)

MICA

EXPERIMENTAL

MICA is a generative AI-powered counseling agent that delivers brief Motivational Interviewing (MI)-consistent sessions, following the validated MI process (engagement, focusing, evoking, planning, and closing). It uses personalized prompts to promote autonomy and evoke change talk, ending with a session summary and satisfaction check. Participants will complete a MICA session (\~5-10 min) in the ED and at 3 months.

Behavioral: MICA

MICA+SMS

EXPERIMENTAL

This condition includes the MICA ED session, a 3-month booster, and 12 weeks of weekly SMS messages supporting self-monitoring, goal-setting, and adaptive strategies. Messages are tailored based on reported alcohol use and goal commitment, offering encouragement, feedback, and harm-reduction messaging.

Behavioral: MICA; Behavioral: TRAC

Standard care

NO INTERVENTION

Standard ED care and as-needed referral to alcohol resources

Interventions

MICA BEHAVIORAL

MICA is an AI chatbot that uses a secure API with ChatGPT and prompted to follow MI principles to support individuals in exploring and resolving ambivalence about their alcohol use. MICA follows a predefined conversational framework and steps: introduction, engagement, focusing, evoking, planning, and closing-adapted from Miller \& Rollnick's Motivational Interviewing: Helping People Change (3rd ed.).

MICA; MICA+SMS

TRAC BEHAVIORAL

Weekly SMS check-ins designed to support goal-setting, self-monitoring, and adaptive self-regulation strategies. Beginning the Sunday after enrollment, participants will engage with an automated SMS program once weekly for 12 weeks. Prompts include reporting on the past week's alcohol use and whether they are willing to commit to a drinking limit goal for the upcoming week. Responses trigger tailored feedback.

MICA+SMS

Eligibility Criteria

• Age: 18 Years - 65 Years
• Sex: all
• Healthy Volunteers: No
• Age Groups: Adult (18-64), Older Adult (65+)

You may qualify if:

• For men: 1+ days with 5 or more drinks OR any week with 14+ drinks/week in the past 3 months
• For women: 1+ days with 4 or more drinks OR any week with 7+ drinks/week in the past 3 months
• Owns a mobile phone
• Reads English language

You may not qualify if:

• For women: current or planned pregnancy in next 12 months
• Prisoner or active military status

Sponsors & Collaborators

• Stanford University: lead
• National Institute on Alcohol Abuse and Alcoholism (NIAAA): collaborator

Study Sites (1)

Stanford University Medical Center

Palo Alto, California, 94304, United States

Location

Related Publications (2)

• Suffoletto B, Prieksaitis C, Kim D, Rose C, Pitre V, Pillai N, Mandal M. AI counseling agent for motivational interviewing: Conversational processes associated with change talk in emergency department patients. J Subst Use Addict Treat. 2026 Jun 12:210053. doi: 10.1016/j.josat.2026.210053. Online ahead of print.

  PMID: 42285418 BACKGROUND

• Suffoletto B, Prieksaitis C, Rose C, Kim D, Pillai N, Pitre V. Feasibility and Acceptability of a Large Language Model-Based Motivational Interviewing Agent in the Emergency Department. Ann Emerg Med. 2026 Jan;87(1):121-123. doi: 10.1016/j.annemergmed.2025.08.021. Epub 2025 Oct 2. No abstract available.

  PMID: 41037040 BACKGROUND

MeSH Terms

Conditions

Emergencies

Interventions

mica

Condition Hierarchy (Ancestors)

Disease Attributes; Pathologic Processes; Pathological Conditions, Signs and Symptoms

Study Officials

• Brian Suffoletto, MS MS

  Stanford University

  PRINCIPAL INVESTIGATOR

Study Design

• Study Type: interventional
• Phase: not applicable
• Allocation: RANDOMIZED
• Masking: DOUBLE
• Who Masked: INVESTIGATOR, OUTCOMES ASSESSOR
• Purpose: PREVENTION
• Intervention Model: PARALLEL
• Sponsor Type: OTHER
• Responsible Party: PRINCIPAL INVESTIGATOR
• PI Title: Professor, Emergency Medicine

Study Record Dates

• First Submitted: June 16, 2026
• First Posted: June 22, 2026
• Study Start (Estimated): August 1, 2026
• Primary Completion (Estimated): August 1, 2030
• Study Completion (Estimated): February 1, 2031
• Last Updated: June 22, 2026

Record last verified: 2026-06

Data Sharing

• IPD Sharing: Will share
• Shared Documents: STUDY PROTOCOL, SAP, CSR, ANALYTIC CODE
• Time Frame: Data will be available within 1 year of study completion and available indefinitely.
• Access Criteria: Individuals with appropriate ethical approval and access to the NIAAADA or a DUA with the PI.

Locations

US (1)