2 minute read

AI-Enabled Postoperative Follow-up Calls

Next Article
Industry News

Industry News

New AI system performs automated follow-up calls for cataract surgery patients. Roibeárd O’hÉineacháin reports

Anatural language artificial intelligence (AI) assistant “Dora” (Ufonia, Oxford, UK)— which performs automated followup safety checks over the telephone for cataract patients—met with high levels of satisfaction among participants in a mixed-method cohort study, said Sarah Khavandi MBBS, BSc(Hons).

“Dora is seen as a highly acceptable method of routine follow-up post-cataract surgery by patients. Whilst many enjoy human interaction, patients appreciate that automation saves time and money for the National Health Service and find automated telephone followup simple and easy to use,” Dr Khavandi said.

She noted that Dora is the first CE-marked (now UKCA) AI technology-driven clinical assistant capable of delivering cataract surgery follow-up calls. Clinicians simply provide Dora with a patient list, and it then automatically contacts patients by telephone, without any additional need for training of patients or clinicians. Dora uses speech transcription, natural language understanding, speech generation, and a machine-learning conversation model to enable contextual conversations.

In their study, Dr Khavandi and her associates used Dora to call 184 patients with planned telephone follow-up calls after uncomplicated cataract surgery from June to September 2021. The patients were between 41 and 98 years old and had a mean age of 76 years. They received calls three to four weeks postoperatively, and a human ophthalmologist supervised all calls in real time.

ASKS AND ANSWERS QUESTIONS When patients received the automated call from Dora, they first confirmed they were the patient, providing their name and date of birth. They then answered a series of questions regarding symptoms, such as pain, eye redness, change in vision, flashing lights, and floaters. Dora then inquired if they wished to go ahead with the second eye surgery and responded to patients’ queries and frequently asked questions.

As part of the conversation, patients gave a Net Promoter Score (NPS) in answer to the question, “On a scale of 1 to 10, how likely would you be to recommend this automated service to a friend or colleague?” The median NPS response was 9 out of 10.

A randomly selected cohort of 21 patients also underwent a remote semi-structured interview to assess their opinions about Dora’s usability, acceptability, appropriateness, and level of satisfaction. Emerging themes from interview data include convenience, ease of use, and the preference of some patients to speak to a clinician for human interaction.

The patients gave responses to the validated Telephone Usability Questionnaire (TUQ). On a scale of one to five, the patients gave overall satisfaction a mean score of four. Simplicity, timesaving, and ease of use scored the highest with a median of five, while “speaking to Dora feels the same as speaking to a clinician” scored a median of three.

“Patient views are an integral part of improving the design and development of such innovation. With high rates of patient acceptability, the use of automated AI assistants such as Dora has the potential for a transformative, system-wide increase in efficiency of high volume, low complexity care,” Dr Khavandi said.

Dr Khavandi presented at the ESCRS Virtual Winter Meeting 2022.

Sarah Khavandi MBBS, BSc(Hons) is a Clinical Education Fellow at Imperial College School of Medicine, London, UK, and is an AI Research Clinician at Ufonia, Oxford, UK. sk@ufonia.co

This article is from: