Abstract
Abstract: :
Purpose: To evaluate which of four commonly administered stereo acuity tests produced results most consistent with that of a "gold standard," and to examine the relationship between stereopsis testing performance and the variables of age, sex, visual acuity and phoria. Methods: Patients, ages 5 through 12, were prospectively examined and measured for near visual acuity, phoria and refractive error using an automated refractometer. Stereo acuity testing was performed in a randomized order by one of two examiners using the Titmus, Randot, TNO and Frisby tests according to the manufacturer's instructions as well as the Randot 1 test, designated our gold standard based on Cooper, et al's 1979 study of stereo acuity in children. Subjects who had anisometropia of greater than one diopter, a tropia or near visual acuity in either eye of 20/40 or worse were excluded from the study. Results: Data from 92 patients were analyzed. Spearman's rho analysis of the data revealed strong correlations between performance on each of the four tests compared with that of the gold standard. Correlation coefficients were: r= .73 for Titmus, r= .69 for Randot, r= .66 for TNO, and r= .59 for Frisby (P<.05). The differences between the four correlation coefficient values were not statistically significant. Spearman's rho analysis of stereopsis testing performance with respect to age for each of the five tests showed a statistically significant (P<.05) inverse correlation. Performance on the stereopsis tests was positively correlated with combined visual acuity (OD + OS / 2) for four of the five tests (P<.05): r= .35 for Randot 1, r= .33 for Frisby, r= .30 for Randot and r= .26 for TNO. A statistically significant correlation between visual acuity and performance on the Titmus test was not observed (r= .15, P= .17). No statistically significant relationship was observed between the presence or quantity of a phoria and stereopsis performance on the five tests (r values ranged from -.10 to +.09). Lastly, Mann-Whitney analysis of stereo acuity regressed against sex for each of the five tests did not reveal a statistically significant relationship (P values ranged from .34 to .97). Conclusion: Although no one stereopsis test outperformed the other three when rigorously analyzed against the gold standard, observationally, the Titmus performed the best and the Frisby performed the worst. Our data also showed that stereopsis testing performance improved (test scores became smaller) with increasing age and that stereopsis generally improved with finer visual acuities. No predictive relationship was observed between either patient sex or phoria and stereopsis testing performance.
Keywords: 329 binocular vision/stereopsis • 386 depth