Treating Comorbid Depression During Care Transitions with Relational Agents (Massachusetts)

Depression is common among individuals with chronic illness such as congestive heart failure (CHF) or chronic pulmonary disease (COPD) and is a strong predictor of poor clinical outcomes and high rates of avoidable 30-day readmissions. However, while depression is a treatable condition, few people with depression receive effective treatment. The project will develop an Internet-accessed relational agent system to deliver depression treatment to patients with CHF or COPD and comorbid depression immediately following hospital discharge to improve the safety of care transitions.

This project builds on work funded by AHRQ: Testing the Re-Engineered Hospital Discharge (Project RED) (HS 015905) and Project RED-Depression (HS 019700). For Project RED, investigators at Boston Medical Center delivered ten mutually reinforcing components of safe hospital discharge, reducing the rate of 30-day readmission by one-third versus usual care. Project RED-Depression (RED-D) is currently testing a 12 week post-discharge mental health intervention to reduce 30-day readmissions. It involves weekly telephone visits with a mental health care manager providing self-management support, navigation, and cognitive behavior therapy (CBT) for medical inpatients with depression symptoms at discharge.

The current project will translate the RED-D intervention for delivery by a computer-driven relational agent (RA) platform to enhance scalability and provide immediate access to the RED-D service. While CBT is an effective treatment for depression, access to care is a significant barrier to treatment. Computerized CBT – self-directed CBT using a text-based computer program – is more accessible and similar in efficacy to face-to-face CBT, but health literacy barriers limit its effectiveness. An alternative to computerized and face-to-face CBT is CBT delivered by novel health technology known as the relational agent (RA). The RA is a human animation program that interacts with patients and integrates best practices from provider-patient communication theory, emulating the face-to-face conversational behavior of an empathic provider emphasizing nonverbal communicative behavior such as gaze, posture, and gestures. The investigators will expand the RA capabilities to include the RED-D, a CBT intervention modified with self-management education (RA-CBT) in an interactive format tailored for CHF or COPD patients with depression.

The specific aims of the project are to:

• Script and program a relational agent system designed to deliver a post-discharge, 5 module modified CBT intervention for depression and physical symptom management (RA-CBT) for CHF and COPD patients with comorbid depressive symptoms at the time of acute hospitalization. 
• Conduct a randomized, controlled pilot study of the post-discharge RA-CBT system to treat comorbid depressive symptoms with 100 CHF or COPD patients with a positive depression screen during an index hospitalization to determine the RA system’s feasibility and acceptability and impact on depression. 
• Compare trends in 30-day hospital reutilization (readmissions and emergency department visits) between pilot intervention participants and control arm participants with CHF or COPD and comorbid depressive symptoms during the index hospitalization. 

If successful, this new approach to depression treatment would expand access and scalability for post-discharge mental health support in the care transition.