Deviation from predicted postoperative refractive outcome (DPPOR): a desktop study investigating the association between patients' eyes in a real-world data set

Session Details

Session Title: Presented Poster Session: Cataract Surgery Special Cases

Venue: Poster Village: Pod 1

First Author: : A.Haq UK

Co Author(s): :    S. Perera                       

Abstract Details

Purpose:

NICE advises to “consider using 50% of the first-eye prediction error in observed refractive outcome to guide calculations for the intraocular lens power for second-eye cataract surgery” (NICE NG77 1.3.9), making the assumption that the prediction error for the first and second eye are associated, rather than independent variables. The 50% correction factor is based on a number of studies, all graded by NICE as poor evidence level. Our study looked to investigate whether: 1) A patient’s eyes were deviating from predicted in a similar way. 2) The 50% recommended correction factor gives desirable results in our practice.

Setting:

This exploratory desktop study was carried out using one surgeon’s data in a private practice setting in the United Kingdom.

Methods:

• Retrospective data collection • N=134 patients who underwent uncomplicated sequential cataract surgery, performed by the same surgeon. • Biometry readings taken on a daily calibrated IOLMaster 700, by trained operator. • Lens choice with Hoffer Q formula for AL <22 and SRKT for all others.  The DPPOR was calculated and compared between left eyes (LE) and right eyes (RE) for each patient. Correlation was assessed and a Bland-Altman plot constructed looking at the level of agreement in DPPOR between the patient’s eyes. Test descriptive statistics were produced for hypothetical correction factor usage from 10-100% in 10% increments.

Results:

• LE and RE DPPOR moderately positively correlated R = 0.57. • Bland-Altman Plot: - Mean discrepancy between LE and RE DPPOR  = 0.07 - 95% limits of agreement = -1.02 to +1.16. - Variability constant, no increase or decreased bias at extreme values. - Even distribution of positive and negative values, no tendency to under or overcorrect in either eye. • Scenario Analysis: - 50% correction factor reduces average deviation from aimed refractive outcome cf. 10% factor from 0.34D to 0.29D, with range reducing from 2.73D to 2.075D - 100% correction factor range = 1.62D with 0.32D average deviation from aimed outcome.

Conclusions:

LE and RE behaved similarly in terms of DPPOR. However, the limits of agreement are wide considering the clinical quality benchmark for 85% of cases to be within +/- 1D of aimed refractive outcome. A mid-range correction factor has the smallest average DPPOR. Range decreases with increasing correction factor. A 50% correction factor would therefore make outcomes more accurate, on average, but for patients at the extremes of range outcomes would be worse than without correction. We did not find sufficient evidence of superiority of outcomes to confidently adopt a 50% correction factor. Prospective study is required to further investigate.

Financial Disclosure:

None