Abstract
Purpose:
To investigate the statistical distribution of sphere and cylinder for refractive surgical candidates.
Methods:
Based on population statistics for 393,139,704 eyes in a 2006 national survey and a refractive surgery patient database for 8,246,757 eyes performed in the US until 2006, a statistical analysis was performed. For refractive sphere, a normal distribution is a reasonable measure for virgin eyes. For refractive surgical candidates, however, a Rayleigh distribution better represents the statistics as emmetropic patients do not seek refractive surgery (except for presbyopes). For refractive cylinder, a normal distribution is a reasonable measure for both the virgin eyes and refractive surgical candidates. These statistical models were used to generate random eyes for scientific research or commercial use.
Results:
One million random normal eyes and one million refractive surgical eyes were generated based on the proposed algorithms. Roughly 79% eyes are myopic, 13% are hyperopic and 7% are mixed astigmatic. The statistics of the simulated eyes follow the statistics assumed. This allows a “real-world” testing, or a “simulated clinical trial” for the development of new treatment algorithms, or verification and validation of to-be-released treatment software products, greatly speeds up the process and improves the efficiency for product development in a commercial environment. It may also be used in the vision research when a large number of eyes are required for statistical analysis.
Conclusions:
The refractions of human eyes (sphere and cylinder) follow the proposed statistical distribution based on a large national optometric database. Simulation based on the statistics provides a useful means for vision research and commercial use.
Keywords: 678 refractive surgery •
676 refraction •
473 computational modeling