Comparison of autorefractor with focometer in patients with refractive errors attending Lagos University Teaching Hospital, Lagos, Nigeria - A cross sectional survey

Background: Uncorrected refractive error is the most common cause of visual impairment globally. Yet, there is paucity of refractionists in rural areas of most developing countries. Thus, there is a need for a cost effective but accurate method of refraction that could be used by rural health workers with minimal training. To compare refractive error measurements of autorefractor with that of focometer with a view to determining the accuracy and reliability of focometer. Methods: This was a comparative cross-sectional study conducted among patients with refractive errors attending the Guinness Eye Centre Clinic, Lagos University Teaching Hospital, Lagos, Nigeria. Consecutively consenting patients who met the eligibility criteria were recruited until the sample size was attained. All participants had a standardized protocol examination including visual acuity assessment and ocular examination. Refractive error was measured using the autorefractor, focometer and subjective refraction in both eyes of each participant. Comparison was done based on the means of variables of autorefractor, subjective refraction and focometer measurements using the paired-sample t-tests, Pearson's correlation and linear regression. Agreement between the measurements was investigated using the Bland-Altman analysis and reliability of the repeated measurements tested with Cronbach's alpha. The analysis was considered statistically significant when the P < 0.05. Results: Four hundred eyes of 200 patients were analyzed in this study. The mean age of respondents was 45.1 ± 16.3yrs and the male:female ratio was 1: 2.1. There was a statistically significant difference between the mean spherical (P < 0.001) and cylindrical (P < 0.001) readings of the focometer and autorefractor. However, the mean difference between the spherical equivalent of focometer and that of the autorefractor was not statistically significant (P = 0.66). Pearson correlation coefficient was high for the compared methods of refraction as both the bivariate linear regression between the autorefractor and focometer, and that between the subjective refraction and focometer showed good linearity. Bland-Altman plot showed good agreement between the mean focometer measurements with both the autorefractor (mean difference = +0.02 ± 0.85 DS; mean difference ± 1.96 standard deviation [SD] = 1.69 to − 1.65 DS) and subjective refractive (mean difference = +0.06 ± 0.72 DS; mean difference ± 1.96 SD = 1.49 to − 1.36 DS) measurements. Cronbach's alpha showed good reliability of focometer and autorefractor repeated measurements. Conclusion: This study showed a good correlation and agreement between focometer and autorefractor. Hence, focometer could be used for refraction in low resource settings where locals could be trained in its use.