Purpose: To analyze which factors influence the mean absolute prediction error after cataract surgery.
Setting: Forty-nine eye departments participating in the Swedish National Cataract Register (NCR).
Methods: In this prospective multicenter comparative nonrandomized study, ophthalmology departments in Sweden voluntarily reported to the NCR the outcomes of every cataract extraction performed during the month of March from 2000 through 2005. During this period, 49 of 55 ophthalmology departments in Sweden reported data. Perioperative data describing visual acuity, sex, age, other eye diseases, target postoperative refraction, and achieved postoperative refraction were included in the study. Factors that could be related to the mean absolute prediction error were evaluated. Of the cataract extractions, 23244 were eligible for analyses. Statistical analyses were performed using multiple regression and logistic regression.
Results: During the study years, the mean absolute prediction error decreased significantly (P < .0001). The mean absolute prediction error was larger in women than in men (P < .001) and in patients with low preoperative visual acuity (P < .0001). Glaucoma in the surgical eye was also significantly related to a large deviation from the target refraction. The precision in reaching the target refraction was not better in second-eye surgery than in first-eye surgery. Some clinics had a significantly smaller mean absolute prediction error than other clinics (P < .0001).
Conclusions: Some preoperative factors were related to less success in reaching the planned postoperative refraction. Taking this into consideration, extra care is warranted in cases with these preoperative risk factors.