Non-cycloplegic screening for refractive errors in children with the hand-held autorefractor Retinomax: final results and comparison with non-cycloplegic photoscreening.

AIMS: To establish the results of refractive screening of preschool children with the hand-held autorefractor Retinomax under non-cycloplegic conditions, and to compare these results with those of photoscreening.
METHODS: Among 1218 children undergoing non-cycloplegic refractive screening, 302 (25%) were also refracted under cycloplegia using the same refractor and were used as controls. Our criteria for a positive screening test were based on the spherical or cylinder values and were: myopia over 3D, astigmatism > or = 2D, spherical or cylindrical anisometropia > or = 1.5D, and hyperopia > or = 1.5D. Absolute myopia over 3D, absolute astigmatism > or = 2D, absolute anisometropia > or = 1.5D and absolute hyperopia > 3.5D were considered as true positives. The sensitivity, specificity, and positive and negative predictive values (PPV and NPV) were calculated within the group of controls for each refractive anomaly. On the basis of Bayes' theorem, these figures were then corrected to yield the true screening results that would be expected in a population without verification and selection bias. To determine the usefulness of this screening technique, the likelihood ratios for positive test results (+LR) were also calculated. The results of this screening in terms of sensitivity, specificity and predictive values were then compared with those of photoscreening.
RESULTS: The basic results of screening with the hand-held autorefractor were as follows: -EHyperopia: sensitivity 46%, specificity 97%, PPV 55%, NPV 96%, +LR 15; -EAstigmatism: sensitivity 37%, specificity 99%, PPV 69%, NPV 96%, +LR 37; -EAnisometropia: sensitivity 66%, specificity 93%, PPV 19%, NPV 99%, +LR 9; -EMyopia: sensitivity 87%, specificity 99%, PPV 33%, NPV 100%, +LR 87. The comparison with photoscreening revealed a similar performance when screening for hyperopia, but the hand-held autorefractor yielded much better figures when screening for astigmatism. In the case of myopia and anisometropia, the lack of consistent information concerning photoscreening invalidates any comparison.
CONCLUSION: The hand-held autorefractor Retinomax appears to have potential as a screening device. Our experience with the non-cycloplegic screening of preschool children for refractive anomalies indicates definite usefulness and reasonable accuracy of the Retinomax for detecting myopia, astigmatism and hyperopia. The weak point of this screening technique is the diagnosis of anisometropia, with only moderate utility and poor accuracy.