Loading ...

TOPLINE:

A voice-based generative AI assistant aimed at educating patients with wet age-related macular degeneration about eye injections was found to be helpful to all participants; most rated the AI assistant’s information as understandable, and 81% reported increased confidence in AI’s role in healthcare. Retinal specialists and practice managers who reviewed the recordings also found the assistant to be accurate, empathetic, and potentially useful for patient education.

METHODOLOGY:

• The qualitative focus-group study in the US evaluated Polaris, a voice-based generative AI aide designed to support clinical objectives such as wellness check-ins, assessments of medication adherence and diet, appointment confirmations, and the review of procedural logistics and tested its effectiveness in educating patients about intravitreal therapy for wet age-related macular degeneration.
• A total of 26 patients (6 male, 20 female) aged 24-83 years participated, of whom two reported age-related macular degeneration and four reported diabetic macular edema; participants role-played as patients newly diagnosed with wet age-related macular degeneration and engaged in open, unscripted conversations with the AI assistant.
• Pre- and postinteraction surveys evaluated patients’ understanding, usability, and confidence in AI-assisted care.
• Nineteen retinal specialists and five practice managers reviewed audio recordings and provided feedback through structured surveys and discussions.

TAKEAWAY:

• Before interaction with the AI assistant, 19 (73%) patients believed AI would improve healthcare; however, after the sessions, all (100%) found the technology helpful, 25 (96%) rated the AI as organized and understandable, 22 (85%) considered it useful for patients newly diagnosed with age-related macular degeneration, and 21 (81%) reported increased confidence in AI’s role in healthcare.
• A total of 16 (85%) healthcare professionals and four (80%) practice managers believed AI would improve healthcare before exposure; after reviewing patient interactions, all agreed the AI was accurate, helpful, and appropriate for patient education.
• Also, 12 (63%) healthcare professionals and four (80%) practice managers reported increased confidence in AI’s potential, with healthcare professionals praising its empathetic tone and conversational realism and noting its potential uses in pre- and postdiagnosis education, adherence monitoring, and patient triage.

IN PRACTICE:

“A fully generative, voice-based AI assistant was well received by participants. This approach may support patient education but requires evaluation in real-world settings,” the researchers of the study reported.

SOURCE:

The study was led by Allon Jacobs, MS, F. Hoffmann-La Roche, Basel, Switzerland. It was published online May 14 in JAMA Ophthalmology.

LIMITATIONS:

Limitations include the small sample size of 26 patients and the use of a hypothetical scenario where participants role-played as patients newly diagnosed with wet age-related macular degeneration rather than actual patients receiving care.

DISCLOSURES:

F. Hoffmann-La Roche provided funding for the study. Some authors reported employment at and stock ownership in F. Hoffmann-La Roche, employment at and stock ownership in Hippocratic AI, and employment with Genentech, with some reporting additional support from these organizations. Several authors are named inventors on patents and patent applications held by Hippocratic AI and may have had a financial interest in the commercialization of the technology evaluated in the study.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.