A Predictive Model for Personalized Fitting of Peripheral Defocus Spectacles Based on Ocular Biometric Parameters.

Area of Science:

• Ophthalmology
• Optometry
• Biomedical Engineering

Background:

• Myopia is a growing public health concern, particularly in children.
• Peripheral defocus spectacles are a common intervention for myopia control.
• Personalized spectacle selection can optimize treatment efficacy.

Purpose of the Study:

• To develop predictive models for personalized recommendation of peripheral defocus spectacle designs.
• To base these recommendations on individual ocular biometric parameters.

Main Methods:

• A retrospective cohort study of 2824 myopic children using five different peripheral defocus spectacle designs.
• Development of multiple linear regression models using ocular biometric data.
• Model validation using training (80%) and validation (20%) datasets.

Main Results:

• No statistically significant difference in axial elongation control among the five designs.
• High predictive accuracy (R-squared 0.88-0.97) and low error (RMSE 0.04-0.07 mm) for the developed models.
• Good agreement between predicted and observed values shown by calibration plots.

Conclusions:

• Predictive models based on ocular biometric parameters can guide personalized fitting of peripheral defocus spectacles.
• While average efficacy was similar across designs, personalization enhances clinical decision-making.