Automated Collection Of Patient-Reported Outcome Measures (Proms) For Cataract Surgery Using Artificial Intelligence: A Multicenter Evaluation In Uk Nhs Trusts

Patient-reported outcome measures (PROMs) are essential for assessing the impact of cataract surgery on vision-related quality of life (QoL). Many current assessments rely only on objective visual acuity, which does not capture patient experience. By enabling routine pre and post-op collection of PROMs, we can evaluate outcomes at both individual, surgeon and service levels. We implemented an artificial intelligence, telephone-based clinical assistant, ‘Dora,’ (Ufonia Ltd, Oxford, UK) to automate PROMs collection for cataract surgery patients. This study aims to analyse pre- and post-operative PROMs data collected through Dora and assess the feasibility and effectiveness of automated telephone calls in routine UK clinical practice.

Four NHS Trusts in the UK used Dora as the standard of care to collect the Cat PROM5 via automated telephone calls. Patients received pre-operative calls within one week prior to cataract surgery and post-operative calls at 12 weeks post-op. Any patient able to have a conversation in English on the telephone was eligible. Data from these conversations were analysed to determine changes in vision-related QoL. Data collection was performed under a multi-site audit. 

Patients who received at least one PROM call between May 2023- May 2024 were included. The catPROM5, a 5-item PROM developed for UK Cataract patients and the National Ophthalmology Database (NOD) was used. CatPROM5 scores were converted using Rasch calibration, which involves transforming raw scores into a logit scale, and then mapping them onto a standardised 0-100 scale.  

Demographic and clinical data were retrieved from electronic medical records for those with paired pre-post op calls. Statistical analysis assessed differences in PROMs scores based on comorbidities and post-operative complications.

9543 patients had a PROM call with Dora, with an overall completion of 68%. 583 (6%) were listed for both the pre and post operative PROM call. 316 completed both calls, (mean age 76 years, SD 9.3; 59% female). Completion ranged from 37%- 77% across the sites. The mean pre-operative Cat-PROM5 score was 47.2 (SD 14.8), increasing to 64.5 (SD 18.7) post-operatively, with a mean improvement of 17.3 (SD 19.1). 80.4% of patients experienced improved vision-related QoL, while 13.6% reported worsening. Patients with comorbidities and post-operative complications had significantly worse outcomes (p<0.05 and p<0.001, respectively). 

Automated PROMs calls achieved a 68% response rate, higher than other means such as electronic collection (22%) in the UK National Ophthalmology Database pilot. Routine PROMs collection via Dora offers a scalable, efficient way to assess vision-related quality of life in cataract surgery patients. However, operational challenges led to many patients not being listed for calls both pre- and post-operatively, which is being addressed in current deployment . 

When available, PROMs data enhances patient counseling, can identify post-operative concerns, and support service evaluation. Automated PROMs collection overcomes historical barriers, providing key insights for clinical decisions and healthcare commissioning.