Purpose: To determine the efficacy of distance stereotesting as a screening device.
Methods: Distance stereoacuity using the global Random Dot and contour Circle test of the Mentor BVAT II-SG computerized testing system was measured for 216 patients, ages 6 to 18 years, before the clinical examination. Patients were classified into pass/fail groups in the areas of refractive error change (REC), ocular deviation (DEV), visual acuity (VA), and all three together (EXAM). Legitimate cutoff scores were obtained when patients were classified as 'pass' as follows: REC if the change was 0.50 D or less in sphere or cylinder relative to the habitual correction or to emmetropia if no habitual correction; DEV if there was no heterophoria or strabismus at distance (criteria of heterophoria of < 6 prism diopters and heterophorias of any magnitude were also tested); VA if the acuity at distance was better than or equal to 20/25 in the poorer eye and better than or equal to 20/20 in the better eye; EXAM if they were pass in REC, DEV, and VA. Optimal pass/fail cutoff values for the stereopsis measurements were determined by finding the maximum chi2 value from contingency tables constructed using pass/fail levels for the screening test at each of the observed levels.
Results: The pass rates for REC, DEV, VA, and EXAM were 45%, 72%, 42%, and 24%, respectively. Patients passed the BVAT at the analytically determined optimal cutoff values of less than or equal to 120 sec arc for global and less than or equal to 30 sec arc for contour stereopsis. The sensitivity and specificity for global stereopsis were 0.90 and 0.40 for REC, 0.89 and 0.30 for DEV, 0.93 and 0.51 for VA, and 0.87 and 0.63 for EXAM. For contour stereopsis, the corresponding values were 0.85 and 0.42, 0.89 and 0.34, 0.91 and 0.53, and 0.84 and 0.62.
Conclusion: Distance stereotesting is highly sensitive to small refractive error changes, heterophorias and strabismus, visual acuities < 20/25, or any of the three. Global stereopsis is only slightly better than contour stereopsis at classifying patients. Distance stereotesting has potential as an effective screening test.