Agreement Between Retinoscopy, Autorefractometry and Subjective Refraction for Determining Refractive Errors in Congolese Children.

Objective: To assess the agreement between retinoscopy and autorefractometry and between subjective refraction and both retinoscopy and autorefractometry in Congolese children.

Patients and methods: Fifty-four children (6-17 years old) were enrolled consecutively in this cross-sectional study. Refraction was evaluated before and after cycloplegia (1% cyclopentolate) with retinoscopy and autorefractometry. Readings were compared (paired t-test) and agreement assessed with Bland-Altman plots. Subjective refraction was compared with the two methods to determine which one provides better reference estimates for subjective refraction.

Results: Under cycloplegia, the spherical power was comparable between retinoscopy and autorefractometry (1.12 ± 1.37 D vs 1.22 ± 1.06D, P = 0.70), cylinder power was significantly more myopic on retinoscopy than autorefractometry (0.80 ± 1.10D vs -0.62 ± 0.66, P = 0.019), and SE was greater on autorefractometry than retinoscopy (0.91 ± 1.10D vs 0.72 ± 1.00D, P = 0.014). Retinoscopy and autorefractometry overestimated the power of spherical (P = 0.022 and 0.002, respectively) and cylindrical components (all P < 0.001). There was an agreement between retinoscopy and autorefractometry in measuring spherical (bias: 0.09 ± 0.16D; limit of agreement, LoA: -0.40 to 0.22D) and cylindrical power (bias: -0.18 ± 0.20D; LoA: -0.57 to 0.21D). Subjective refraction agreed with cycloplegic retinoscopy for determining SE power (bias: 0.11D; LoA: -0.51 to 0.73D).

Conclusion: Retinoscopy and autorefractometry can be used interchangeably in children for determining the power of spherical and cylindrical components. Cycloplegic retinoscopy is better than autorefractometry to obtain SE reference values for subjective refraction in children.