Development and validation of clinical scores for visual outcomes after cataract surgery

To develop and validate a clinical score to predict visual acuity (VA) and functional changes after phacoemulsification on the basis of readily obtainable preoperative history data and patient assessment.

Prospective follow-up study.

A sample of 5512 patients on waiting lists for phacoemulsification at 17 hospitals in Spain.

Data were obtained at the baseline examination from the 5512 patients. The patients were divided randomly into 2 subgroups: derivation (n = 3285; 60%) and validation (n = 2227; 40%). The preoperative predictors of postoperative gains in VA and visual function index 14 (VF-14) were determined by multivariate logistic regression analysis and implemented using a prediction score.

Probability of postoperative improvement in VA and VF-14 scores. The cutoff points were established for each outcome on the basis of the minimal clinically important difference values.

The predictive variables for VA gain were the baseline VA, patient age, ocular comorbidity, and surgical complexity. Regarding the VF-14, the predictive factors were the preoperative VF-14, the eye with the better VA, and the surgical complexity. In the multivariate logistic model in the derivation sample, the final VA and VF-14 scores ranged from 0 to 44 and from 0 and 24, respectively. Receiver operating characteristic curves were developed in the derivation and validation samples, and no statistical significance was found when their areas under the curve were compared. Areas under the curve ranged from 65% to 80%. Both scores had a positive predictive value from 74% to 85%.

Newly developed and validated clinical prediction scores may assist physicians and patients in decision making about the expected outcomes and benefits of cataract surgery.