Early Experience Of Using Virtual Assistants To Conduct Routine Cataract Surgery Pre-Operative Assessments

To evaluate early experience with an AI-powered virtual assistant in conducting pre-operative assessments (POA) for patients within a public high-volume cataract service (circa N=51,000 procedures during 2024). 

Optegra Eye Health Care, UK.

In July 2024, a new virtual assistant (‘Iris’) was integrated into the POA pathway for cataract surgery patients. Iris mirrors human-led POAs in terms of workflow and patients are invited to use a telephone or web chat virtual assistant for the POA. Based on responses, patient suitability for surgery or the need for further review by staff is determined.  Iris’s performance was compared to traditional clinician led-assessments. After successful implementation, areas for refinement were identified and introduced.  We compare early Iris performance data, before and after refinements were applied.

Since its introduction 62% of POAs are delivered by Iris; engagement with Iris-led POA is 15% greater than human-led POAs, due to round-the-clock, multi-language (web chat) availability; 30% of next appointment bookings are successfully completed by Iris at first contact, and patient waiting time to POA has reduced (from 6 to 2 days). Compared to human-led POAs, rejection rates were similar (~7%), although Iris had somewhat higher clinical review rates.  Continuing refinement of the virtual assistant process has led to a 50% reduction in the rate of dropped calls at the identification stage, a reduction in medication recording/spelling errors and patient satisfaction scores over 90%.

AI-powered virtual assistants demonstrate strong potential in enhancing the efficiency of pre-operative cataract surgery assessments. Their integration into ophthalmic care pathways could alleviate workload pressures on clinicians and improve patient access to timely interventions. Optimising virtual assistants is key to fully harnessing the advantages. Further research is warranted to evaluate AI-human collaboration for broader clinical applications to allow healthcare providers to meet growing patient demand while maintaining high-quality care.