Free Papers

Influence of the Invariant Refraction Assumption in the Comparison of Intraocular Lens Power calculation formulae

Free Paper Details

First Author: J.Fernández Pérez SPAIN

Co Author(s):    M. Rodriguez-Vallejo   J. Martínez   N. Burguera   D. Piñero           

Abstract Details

Purpose:

To assess the differences in the comparison between SRK-T and Haigis formulae as a function of using them for deciding the intraocular lens (IOL) implanted power versus the retrospectively estimated with the formula non-used for selecting the IOL power.

Setting:

Retrospective crossed design

Methods:

Median Absolute Prediction Error (MedAE) was calculated for 27 and 20 right eyes for which the SRK-T and Haigis were used for deciding the implanted monofocal IOL power, respectively. Same eyes were retrospectively calculated for the formula non-used for taking the surgery decision. Standards were followed for selecting the sample and constant optimization was conducted before the analysis. The cumulated percentage of eyes for each level of prediction error (PE) was also calculated.

Results:

No significant differences were found in the MedAE [Interquartile-range] between SRK-T and Haigis, 0.18 [0.34] D and 0.23 [0.41] D, respectively, in independent samples comparison for which each one of the formulae were used for selecting IOL power. MedAE increased for the formula non-used for selecting the power, 0.30 [0.37] D for SRK-T in the Haigis implanted group and 0.32 [0.50] D for Haigis in the SRK-T implanted group. Significant differences in favor of the formula used for selecting the IOL power was found in the paired samples comparison. Similar results were obtained for the cumulated percentage calculation.

Conclusions:

Current standards for formulae comparison describe paired design studies on which PE is retrospectively calculated for several formulae non-used for selecting the implanted IOL power. These studies should include the formula used for deciding the final implanted IOL power in the methods section because better results might be obtained for that formula. Studies with independent samples and formula randomization would lead to more reliable results avoiding the invariant refraction assumption limitation.

Financial Disclosure:

-