Abstract
Purpose :
There are advantages to objective autorefraction measurements. However, all instruments rely on algorithms to compute the refractive path of light. Frequently, the subjective and objective refractions differ. This is particularly true for subjects with higher order aberrations, such as those with keratoconus, corneal transplants, post-refractive surgery, cataract, trauma, etc. Yet, these are the subjects with the highest clinical need for objective data. To use objective data effectively, we have developed a metric to predict the probability that the objective measurement will yield the best acuity.
Methods :
A clinical study with 145 eyes was conducted, where each eye was measured with a wavefront aberrometer as part of the normal patient workup. Subjective refractions were also collected. The study included a mix of eye conditions, including normal, keratoconus, and post-surgical corneas. Three different algorithms for calculating the refraction were compared to the manifest refraction. The difference from manifest was plotted vs. RMS wavefront error (WFE). This data was used to derive the number of measurements within a specified tolerance (0.25D or 0.5D) of the manifest refraction. The confidence index is the fraction within this tolerance.
Results :
With increasing aberration, the difference between the subjective and objective refraction increased. The average difference increased linearly with a corresponding increase in variance. The linear increase in was used to calculate the fraction that was within 0.25D and 0.5D respectively. The different refraction algorithms each had different characteristics.
Conclusions :
The ability to identify objective measurements that match the manifest refraction is extremely useful to the instrument operator. From confidence indicator, for aberration less than 0.25 um RMS all methods match to manifest within 0.25D. For 0.5 um RMS WFE, the methods match within 0.5D. Thus, an operator can know whether to trust the objective measurement or to spend additional time measuring subjectively.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.