Using Artificial Intelligence To Evaluate The Results Of Cataract Surgery According To The Ichom Guidelines

The International Consortium for Health Outcomes Measurement (ICHOM) is an international organization that aims to standardize the collection of patient-centered health-related outcomes. In ophthalmology, it recommends a group of tests or “standard sets” to be performed as related to cataract surgery. Our sites have been collecting these data in all patients undergoing cataract surgery since 2022. This study presents the results of patient satisfaction reports analyzed with artificial intelligence (AI) methods.

The study was conducted from 2022 to 2025 in OMIQ clinic, a group of private ophthalmology sites in the area of Barcelona, Spain.

Prospective, longitudinal, multicentric study. Data from all cataract surgeries (demographic, refractive, functional, as well as complications and patient reported outcome measurements/PROM) were collected according to the ICHOM guidelines up to 3 months after surgery. We used a gradient boosting machine (GBM) to determine the features associated with a Net Promoter Score of 10 (the NPS is a 1 to 10 scale that measures patient satisfaction and loyalty). The model was trained to optimize the area under the curve (AUCPR) in the training set (80% of the sample) and evaluated in the test set (the other 20%). Given the large number of variables in relation to the number of cases, conventional biostatistical approaches were not applicable.

We included 2.282 eyes of 1.375 patients; 57.0% were female, and mean age was 71.6 years (standard deviation 10.3). Two-thirds reported an NPS of 10 and one third of 9 or less. The optimal GBM obtained an AUCPR of 0.803 in the test set, with minimal overfitting. According to the analysis of variable importance, the surgeon and the response to certain PROM questions (for example, question 2 in the Catquest-9SF, “Are you satisfied or dissatisfied with your sight at present?”) were the most important factors related to the NPS.

The systematic measurement of the data related to cataract surgery allows establishing the performance of each surgeon and site, foster transparency and ease the transition towards value-based medicine. The application of AI complements conventional biostatistical analyses and may enrich the interpretation of the results.