Back to Freepaper Session

Multiple regression analysis in myopic LASIK nomogram development: iteration 2

Session Details

Session Title: LASIK

Session Date/Time: Wednesday 09/10/2013 | 08:00-10:30

Paper Time: 08:36

Venue: Auditorium (First Floor)

First Author: : B.Allan UK

Co Author(s): :    A. Ieong   H. Hassan           

Abstract Details

Purpose:

Multivariate modeling can improve myopic wavefront LASIK accuracy by factoring in influences other than the attempted sphere in predicting the achieved ablation. In a second iteration of a multivariate nomogram, we added a correction for astigmatism and dropped factors no longer shown to be statistically significant when refitting the original model to updated results. Here we evaluate outcomes for the second iteration of our nomogram.

Setting:

Moorfields Eye Hospital, London

Methods:

3 month refractive outcome data from consecutive cases of myopic wavefront LASIK (?10D sphere; up to ?4D cyl) were collected prospectively. Stepwise and best subsets methods were used to fit a regression equation to 3 month spherical equivalent (SE) outcomes in the form: achieved SE= a(4mm wavefront SE) + b(manifest SE – 4mm wavefront SE) + c. The calculated % boost derived from the constant a was fully applied in iteration 2. Compensation for the calculated % undercorrection in the achieved surgically induced refractive change in the model data set was added to the new nomogram with the assumption of no coupling with sphere. 3 month outcome data for consecutive cases of myopic wavefront LASIK using the new nomogram (iteration 2) were then compared with similar cases treated using the earlier multiple regression derived nomogram (iteration 1) in which the % overall boost derived from constant a was only partially applied. AMO (Santa Clara, CA) WaveScan 3.67 software and an AMO S4 IR excimer laser were used throughout.

Results:

The results for the new nomogram(iteration 2), with comparator results for the older nomogram (iteration 1 - in brackets) were: n= 248 (299); mean postoperative SE error = +0.07±0.33D (-0.19±0.32D); % eyes ?0.5D from target SE = 91.9 (88.3); % eyes?0.0 (6/6) uncorrected distance visual acuity = 94.6 (84.2)

Conclusions:

Visual results in myopic wavefront LASIK were improved in iteration 2 of the multiple regression analysis derived nomogram in which less significant modifiers were excluded, the calculated % overall boost was fully applied, and a compensation for cylindrical undercorrection was added.

Financial Interest:

... gains financially from product or procedure presented