Feasibility of Mental Health Triage Call Priority Prediction Using Machine Learning
Rajib Rana, Niall Higgins, Kazi Nazmul Haque, John Reilly, Kylie Burke, Kathryn Turner, Terry Stedman
TL;DR
The paper investigates whether mental health call priority can be inferred from caller voice using deep learning. By building a hierarchical audio-based classifier trained on 459 de-identified helpline calls, the authors achieve a robust $92\%$ balanced accuracy in distinguishing high- from low-priority calls, suggesting real-time decision support for triage. They present a prototype application interface for call-takers and discuss ethical, legal, and practical considerations, emphasizing augmentation rather than replacement of clinicians. The work highlights the promise of voice cues as a supplementary signal in mental health triage and outlines future directions for refinement, data expansion, and governance.
Abstract
Ensuring accurate call prioritisation is essential for optimising the efficiency and responsiveness of mental health helplines. Currently, call operators rely entirely on the caller's statements to determine the priority of the calls. It has been shown that entirely subjective assessment can lead to errors. Furthermore, it is a missed opportunity not to utilise the voice properties readily available during the call to aid in the evaluation. Incorrect prioritisation can result in delayed assistance for high-risk individuals, resource misallocation, increased mental health deterioration, loss of trust, and potential legal consequences. It is vital to address these risks to guarantee the reliability and effectiveness of mental health services. This study delves into the potential of using machine learning, a branch of Artificial Intelligence, to estimate call priority from the callers' voices for users of mental health phone helplines. After analysing 459 call records from a mental health helpline, we achieved a balanced accuracy of 92\%, showing promise in aiding the call operators' efficiency in call handling processes and improving customer satisfaction.