Abstract
Purpose:
To propose a predictive model of myopia onset based on fulfilling a percentage of risk criteria out of several standard optometric tests.
Methods:
A retrospective study on patient files of a single private practice optometrist in the south-west of France (Gaillac). Patients between 6 and 19 years with initial subjective spherical equivalent (SE) of [0.00; +0.75] D in both eyes were selected. 73 patients in which SE remained within [0.00; +0.75] D for at least 2 years were classified as stable emmetropes (E); 52 patients in which SE became strictly negative within less than 2 years were classified as pre-myopes (PM). Values of standard optometric tests were compared between PM and E using Student’s t-test. For values showing statistically significant differences (p<0.05), risk criteria were determined as thresholds with Youden’s index J > 0.40. For monocular tests, right eye values were used.
Results:
23 tests showed significant differences between PM and E, out of which 8 reached J > 0.40: subjective binocular refraction ≤ 0.25 D (p < 0.001, J = 0.44), first perceived blur at distance ≤ 0.50 D (p < 0.001, J = 0.43), distance NRA ≤ +1.00 (p < 0.001, J = 0.45), distance retinoscopy ≤ 0.00 D (p < 0.001, J = 0.42), quantitative direct ophthalmoscopy ≤ -0.50 D (p < 0.001, J = 0.46), monocular subjective refraction ≤ 0.00 D (p < 0.001, J = 0.41), distance binocular red-green duochrome ≤ +0.25 D (p<0.001, J = 0.43) and pupil diameter measured in low luminance ≥ 5.5 mm (p = 0.049, J = 0.45). ROC curve analysis showed that fulfilling at least 60% of those risk criteria identified PMs with a sensitivity of 0.83 and a specificity of 0.74, which was higher than for each test taken separately.
Conclusions:
Using a percentage of fulfilled risk criteria among 8 standard optometric tests allowed a better detection of the onset of myopia than through criteria taken separately.
Keywords: 605 myopia •
464 clinical (human) or epidemiologic studies: risk factor assessment