Search

Study explores potential of AI assistant in conducting cataract follow-up checks

A group of 225 patients received a phone call from an AI assistant, with the conversation supervised in real-time by an ophthalmologist

Surgeons wearing blue gowns and face masks operate on a patient in an operating theatre
Pixabay/Sasin Tipchai

Researchers from Oxford University Hospitals NHS Foundation Trust and Imperial College Healthcare NHS Trust have highlighted the potential of using AI assistants within the cataract care pathway.

Writing in eClinicalMedicine, the authors outlined how 202 routine cataract surgery patients received an autonomous telemedicine call around three weeks after surgery.

The artificial intelligence (AI) assistant, named Dora R1, was supervised in real-time by an ophthalmologist asking patients a series of questions to determine whether patients required further management.

The outcome of the call was to either recommend that the patient was discharged, or recommend that the patient received a further review.

Patients who were discharged from the hospital service would typically be advised to see their optometrist for a sight test six weeks after surgery.

The ophthalmologist supervising the call made an independent decision about whether the patient should be discharged or receive a review.

The researchers found a high level of agreement between the decisions of the supervising ophthalmologist and Dora R1.

“Only four of the 64 patients recommended for review by the supervising clinician were recommended discharge by Dora, none of whom were deemed to require further clinical review on callback,” the authors highlighted.

Interviews with 20 participants involved in the study found that patients were accepting of the technology in routine circumstances.

However, there were concerns about “the lack of a human element” in cases with complications.

“It will be important to ensure that introducing automation into healthcare does not dehumanise care by removing patient-clinician interaction, but rather increases clinicians’ interactions with patients who have concerns,” the authors emphasised.