THE ROLE OF AI-BASED CONVERSATIONAL ASSISTANT DORA IN CATARACT SURGERY PATHWAY: A ONE-MONTH ANALYSIS

Healthcare systems globally, including the UK, face a workforce crisis, necessitating innovative solutions. Artificial intelligence (AI) technologies, such as Dora, an automated conversational assistant, delivering clinical consultations over the telephone, offer the potential to support healthcare staff, optimise resources, and enhance patient accessibility. This study aims to evaluate the utilisation of Dora for post operative cataract follow up.

All patients having a postoperative call with Dora at Frimley Park Hospital, UK in January 2023 were included.

All patients having routine cataract surgery who could have a conversation in English were listed for a call with Dora, as part of the standard care pathway. The Dora call asks patients about key post operative symptoms (for example about visual recovery) and also gives patients the opportunity to ask questions. If any significant concerns are identified by Dora the patient receives a callback from a clinician. If no concerns are identified (“pass”), the patient  is discharged to the community or continues with 2nd eye surgery. This evaluation assessed the outcomes of the calls with Dora, and retrospective review of the clinical notes identified the outcomes after the Dora call up to 3 months post operatively.

Dora placed a total of 118 calls. 102 calls were answered with a 93.6% completion rate. The mean patient age was 77 years. 54 (52.9%) patients were called after their first cataract surgery, 43 (42.1%) were called after their second cataract surgery and 5 (4.9%) were called after their bilateral cataract surgery. 51 (50%) patients had no concerns identified by Dora (“pass”), while 44 (43.1%) patients had possible concerns. A total of 7 (6.9%) calls remained incomplete either because the timing wasn't suitable, the patients ended the calls, or due to technical issues.

Of those patients “passed” by Dora, one presented to eye casualty 15 days after the call with new floaters, and was diagnosed with a posterior vitreous detachment. No patients “passed” by Dora had an unexpected management change within 3 months. For those patients Dora identified potential concerns, when they were contacted by a clinician, 7/44(15.9%) had F2F review arranged, 5/44 (1.1%) were listed for 2nd eye surgery. The three most frequently reported issues during these follow-up communications were blurred vision, watery eyes and floaters for which the clinician gave telephone advice.

The use of AI-based conversational assistant Dora within the cataract surgery pathway shows promise in optimising outpatient follow-up care. With the potential to automate a ‘high-volume, low complexity' healthcare pathway, like cataract surgery, Dora provides an opportunity to alleviate workforce pressures and free up outpatient clinic space. The results of this one-month analysis indicate the feasibility, safety and effectiveness of utilising AI in post-operative patient care, marking a step forward in enhancing healthcare delivery.