AI-DBT For Serious Illness

Official title

Pilot Testing AI-Delivered Dialectical Behavioral Therapy Skills and Coaching in Older Patients With Serious Illness

Overview

Lay / General Audience Summary Many older adults living with serious illnesses experience high levels of emotional distress, including anxiety, sadness, fear, and feeling overwhelmed. These feelings often worsen as illness progresses and can make physical symptoms harder to manage, reduce quality of life, and increase the use of medications. Unfortunately, many people with serious illness cannot access mental health care because there are not enough providers, services are hard to reach, appointments are physically exhausting, or care is not available in their language. This study will test a new way of delivering emotional support that does not require in-person therapy. The program teaches practical coping skills that help people manage strong emotions and stress. These skills are delivered through short videos and daily supportive check-ins using an automated phone or text system that people can use from home. The system is available in many languages and can be used on basic mobile phones, making it accessible to a wide range of patients. We will enroll older adults with serious illnesses such as cancer, heart disease, lung disease, or kidney disease who are experiencing emotional distress. Participants will be randomly assigned to receive either their usual medical care alone or usual care plus the new skills program. We will examine whether people find the program easy to use, helpful, and acceptable, and whether it improves their confidence in managing difficult emotions and reduces distress. The goal of this project is to develop a low-burden, widely accessible way to support emotional well-being for people living with serious illness. If successful, this approach could expand access to mental health support for patients who currently have few options.

Detailed description

Project Abstract

Psychological distress is prevalent among older adults with serious illness, affecting approximately 30-50% of patients. Psychological distress is associated with greater symptom burden, poorer quality of life, higher health care utilization, and potentially inappropriate medication use. Despite its clinical significance, access to evidence-based psychological care remains limited due to workforce shortages, Medicare access constraints, transportation barriers, cost, and poor scalability of traditional in-person psychotherapy. As a result, psychological outcomes in serious illness care remain suboptimal.

Dialectical Behavior Therapy (DBT) is a principle-based, skills-focused intervention with strong transdiagnostic efficacy across psychiatric and medical populations. DBT's emphasis on strengthening self-efficacy to manage negative emotions, a core mechanism underlying diverse forms of psychological distress, makes it particularly well suited to serious illness populations characterized by multimorbidity and fluctuating symptomatology. However, DBT has limited research in serious illness, largely due to its reliance on frequent sessions with trained providers.

This project will evaluate the feasibility, acceptability, and preliminary efficacy of a video and AI-chatbot-delivered DBT skills and coaching (AI-DBT) for older adults with serious illness and elevated psychological distress. AI-DBT was developed and refined through prior funded studies and delivers brief DBT skills videos combined with coaching via a multilingual, constrained-logic conversational agent accessible by phone or text. In a pilot randomized controlled trial (n=80), participants with serious illness and psychological distress will be randomized to usual care or usual care plus AI-DBT.

Aim 1 will assess feasibility and acceptability using recruitment, retention, engagement metrics, participant-reported satisfaction, and qualitative interviews. Aim 2 will examine preliminary efficacy on the mechanistic target of self-efficacy to manage emotions and exploratory clinical outcomes including psychological distress, anxiety, depression, and quality of life.This study will generate critical early-stage data to support a future R01 and advance scalable, mechanism-driven psychosocial care for people living with serious illness.

Interventions

• Behavioral AI DBT
  The skills component of the intervention consists of 14 skills-based videos, available in English and Spanish, that are each under 5 minutes. These videos were adapted in prior research programs to be more suitable to older adults and adults with medical comorbidities. Two videos will be sent to participants per week via text message. In addition to the videos, Angel, the constrained-AI DBT skills chatbot, will check in with participants every day to collect daily-diary data and inquire if a par

Primary outcome measures

• Acceptability [Time frame: 8 weeks]
• Feasibility [Time frame: 8 weeks]
• Preliminary efficacy [Time frame: 8 weeks]

Secondary outcome measures (3)

• depressive symptoms [Time frame: 8 weeks]
• anxiety [Time frame: 8 weeks]
• psychological distress [Time frame: 8 weeks]

Eligibility criteria

Inclusion Criteria:1) Aged 60+, 2) diagnosis of serious, life limiting illness including stage 4 breast, prostate, or colon cancer; stage 3b or 4 lung or blood cancer, New York Heart Association Stage 3 or 4 CHF, or COPD with hypercapnia74 or CKD stage 4 or ESRD; 3) psychological distress rated 4 or higher on the distress thermometer;75-77 4.) has access to a mobile phone and 5.) speaks English or Spanish fluently.

\-

Exclusion criteria

\-

Criteria are shown verbatim from the registry (in English). Final eligibility is always assessed by the study center.

Healthy volunteers: No

Study design

Allocation

Randomized

Model

Parallel assignment

Masking

Double blind

Primary purpose

Treatment

Study locations

Center list to be confirmed — check the primary protocol.

Identifiers

NCT: NCT07748312 · Pro2026000988 · 1U54AG093230-01

Primary sources (government registries)

• ClinicalTrials.gov (US NIH)

View this study on ClinicalTrials.gov ↗