# Practical Algorithm Evaluating Preoperative Risk Factors for Posterior Capsule Rupture During Phacoemulsification

## TL;DR

This study develops an algorithm to assess preoperative risk factors for complications during cataract surgery, aiming to improve surgical safety.

## Contribution

A practical algorithm was developed and validated for preoperative risk assessment in cataract surgery using statistical data.

## Key findings

- The algorithm included age, sex, laterality, and other significant factors to predict posterior capsule rupture risk.
- The model had a 5.9% misclassification error rate and an area under the ROC curve of 0.62 in prospective testing.
- External validation is recommended to assess the algorithm's applicability in other departments.

## Abstract

Purpose: The purpose of the study is to create a practical and efficient tool for the preoperative assessment of cataract surgery based on statistical data.

Methods: A two-phase study was conducted in a tertiary teaching ophthalmology department, including a retrospective cohort for 2014-2015 and a prospective cohort for 2017-2018. In the first phase of the study, all preoperative files of cataract patients (excluding trauma, uveitis related and pediatric cataracts) were gathered and analysed for 2014-2015. An algorithm was created based on their preoperative assessment and then tested in a prospective cohort for 2017-2018, following the same inclusion criteria.

Results: The selection of predictors among the 1792 patients in the retrospective cohort was based on univariate and multivariate logistic regression analysis. The model with the lowest Akaike Information Criterion was formulated including three factors regardless of their p-value (age, sex, laterality) and the statistically significant factors, mature cataract, pseudoexfoliation, phacodonesis, diabetes, glaucoma, monocularity and resident surgeon at different rates of influence. The algorithm was tested in the prospective cohort (2017-2018) in 2057 cataract patients. The overall misclassification error rate was 5.9%, and the area below the ROC curve was 0.62 (CI 0.57-0.67).

Conclusions: The model created can assess patients and preoperatively evaluate their perioperative risk of complications while planning surgery with greater safety. Every population under study has unique characteristics, and safer assumptions can be made when particularities have been identified and taken into account. External validation would provide more information on its applicability in other teaching ophthalmology departments.