AI as a Communication Facilitator: Shared Decision-Making Inspired Strategies for Bipolar Disorder Diagnosis and Treatment

TL;DR

This paper reviews how AI communication facilitators can improve bipolar disorder diagnosis and treatment by supporting shared decision-making, addressing the disorder's complexity and subjective symptom assessment.

Contribution

It introduces a novel AI communication facilitator model that embodies shared decision-making to enhance bipolar disorder diagnosis and treatment processes.

Findings

AI can support accurate bipolar diagnosis by capturing patient heterogeneity.

Shared decision-making improves treatment adherence and patient satisfaction.

AI bridges gaps between mental healthcare and human-AI interaction.

Abstract

This literature review involves the use of AI communication facilitators to detect mood disorders such as bipolar disorder, a psychiatric condition in which patients experience drastic mood shifts. Due to the ill-defined nature of the disorder, it is difficult for even a psychiatrist alone to be confident with their diagnosis. Changes in mental and mood state are often highly subjective and difficult to pinpoint through short-term surveys and psychiatric consultations. For many patients, diagnosis and treatment based on trial-and-error is unavoidable. A timely and thorough diagnosis and treatment plan is associated with the need for an equal involvement of both the patient and the psychiatrist throughout the process. This conclusion is reached through a detailed assessment of current interventions for (i) the ill-defined nature of the disorder, and (ii) the trial-and-error requirement for medication and diagnosis. As a result, I propose the implementation of an AI communication facilitator that can aid in the appropriate diagnosis and treatment of bipolar disorder by embodying the shared decision-making model. I propose that the model can be broken down into specific critical decision points with considerations made for each party involved in the process, aligning with the service blueprint model. I conclude by emphasizing the importance of AI in bipolar disorder diagnosis and treatment due to its ability to embrace patient heterogeneity, and bridge the gap between mental healthcare and human-AI interaction.