D3 – Development of RehaBot

 
 

Project team

Project PI – Amanda Rabinowitz

George Collier

 Purpose/aims

The overall aim of this project is to develop and evaluate RehaBot, a conversational agent to support behavioral activation in self-management of depressive symptoms. The design of RehaBot is based on technological advances in conversational user interfaces (a.k.a. chatbots). The conversational, interactive nature of RehaBot will provide an engaging and fully interactive experience, capitalizing on how users naturally interface with their devices, which can be deployed within any platform that supports synchronous messaging (e.g., SMS, Facebook Messenger)—all features that are particularly advantageous for users with cognitive deficits, such as those associated with TBI.  

RehaBot is designed to supplement the clinician-delivered treatments that rely on engagement of goal-directed behaviors between clinical contacts. RehaBot was conceptualized as an adjunct to Behavioral Activation (BA), a well-established treatment for depression, which holds promise for persons with TBI and with which our group has had considerable experience. Successful completion of this project will result in a tool to augment outpatient therapy with an always-available conversational companion—increasing opportunities for supportive therapeutic interactions, skill learning, and generalization of therapeutic gains, without the need for additional clinic visits—a known barrier to treatment for persons with TBI.

 Status

A fully integrated working prototype of RehaBot was developed with feedback from consumer focus groups. User-demonstration trials were conducted to evaluate feasibility and refine the design. We completed and published a pilot usability and feasibility project demonstrating preliminary evidence of RehaBot’s usability, user-satisfaction, and efficacy in The Journal of Head Trauma Rehabilitation.

We demonstrated that participants interacted with RehaBot multiple times per day throughout the course of a 1-2 week trial. Users found RehaBot highly usable and provided positive feedback on their experience in open-ended qualitative interviews. Results also revealed that participants were more successful in completing planned activities when they had access to RehaBot, as compared to treatment as usual (TAU) (Rabinowitz & Collier, 2022).

We have designed an expanded feature list for RehaBot to enhance the product for use as an adjunct to BA and extend its application as an adjunct to a physical activity promotion program for users with chronic moderate to severe TBI. New feature design is described in the table below and was guided by the Capability Opportunity Motivation-Behavior (COM-B) framework by Susan Michie and colleagues (Michie, Van Stralen & West, 2022 Implementation Science).

These updates and improvements are currently in development with a new industry partner, Docbotic, a startup clinician-led tech company that specializes in developing text bots designed to facilitate, augment, and replace in-person therapy.

We successfully obtained grant funding to support a 70-person randomized controlled trial (RCT) to test the updated RehaBot in the context of a physical activity promotion intervention. This research will be conducted as Moss’s local project, funded by the Moss Traumatic Brain Injury Model System NIDILRR award. An IRB protocol for this RCT has been submitted and is currently under review at Thomas Jefferson University.

Key accomplishments

Presentations and Publications:

  1. Rabinowitz, A. R., Collier, G., Vaccaro, M., & Wingfield, R. (2022). Development of RehaBot-A Conversational Agent for Promoting Rewarding Activities in Users with Traumatic Brain Injury. The Journal of Head Trauma Rehabilitation, 10.1097/HTR.0000000000000770.

  2. Rabinowitz, A., & Collier, G. (2021). Development of a conversational agent for promoting increased activity in users with traumatic brain injury. Archives of Physical Medicine and Rehabilitation, 102(10), e83. 

Grant funding:

  1. A NIDLRR TBI Model System (90DPTB0019, Rabinowitz: PI) was awarded in September, 2022 and will support a 70 person RCT of a physical activity promotion program that includes RehaBot (GetUp&Go).

 Challenges and course corrections

As the scope of development expanded, we reached a bottleneck in developer time, and have engaged an outside developer in Docbotic, which is a likeminded startup company with expertise in text-based therapeutic chatbots.