JMIR Publications

ABSTRACT

Background:

Communication is at the center of every healthcare profession, rendering the training of communication skills in these fields most important. Technological advances, such as artificial intelligence (AI) and particularly machine learning (ML), may support this cause and might provide students with an always-available, low-threshold communication training possibility.

Objective:

This scoping review aims to summarize the current status of the utilization of AI or ML in the acquisition of communication skills within academic healthcare professions.

Methods:

We conducted a comprehensive literature search across the databases PubMed, Scopus, Cochrane Library, Web of Science Core Collection and Cumulative Index to Nursing and Allied Health (CINAHL) to identify articles that deal with the application of AI or ML to the training of communication for undergraduate students of healthcare professions. Additionally, articles that incorporate potential AI- or ML-based ways of improvements for communication training for caretaker-patient-encounters were included. Using an inductive approach, the included studies were organized into three distinct categorical themes. Study characteristics, ways of AI or ML applications and main outcomes of included studies were evaluated and compared. Furthermore, hindering factors of the application of AI/ML in healthcare communication training were depicted.

Results:

339 titles and abstracts were identified, of which we examined 23 papers for full-text review. 13 studies met the inclusion criteria. The three distinct themes that were agreed upon comprised the specific utilization of AI/ML, the combination of AI/ML and virtual reality and the combination of AI/ML and virtual patients – each within the frame of academic healthcare communication training. The implementation of AI/ML could roughly be distinguished between two themes: the usage of virtual patients and the analysis of communication and feedback-provision. Reported barriers in the application of AI/ML in communication training revolved around the lack of authenticity and restricted natural flow of language of the utilized AI/ML-based virtual patient systems. Furthermore, the usage of educational AI/ML-based systems in healthcare communication training is currently limited to only a few cases, topics and clinical domains.

Conclusions:

The utilization of AI and ML in health professional communication trainings clearly is a growing and promising field with a potential to render trainings more cost-effective, less time-consuming and to serve learners as an individualized, always-available practice method. In most cases, however, the depicted applications and technical solutions are limited in matters of access or targeted scenarios as well as regarding natural conversation flows and authenticity. These issues still stand in the way of any widespread implementation ambitions.