Near retinoscopy as described by Mohindra is a non invasive, non cycloplegic technique for assessment of refractive error (R.E.) easily utilized with the pediatric population. The desirability of such a non invasive technique for the infant population is easily understood. The literature, however, is sparse for reliability and variability studies on an Infant population. An assessment of the variability on test retest by observer and between different observers on a pedriatric population, 0-2 years of age, was initiated. Each of 17 full term, healthy, non strabismic infants (28 eyes) were tested using the near retinoscopy technique by each of five trained clinical observers. Each observer performed the technique five times on each eye. The observer was blind to the neutralizing lens powers. All testing occurred over a 2 month period. As a control condition observers each neutralized 1 0 unknown power lenses 1 o times each on a fixed schematic eye. All experimental data was converted to equivalent spheres.
For the control condition, significant differences existed between observers for all lenses and the true power of each lens at the .05 level. The maximum mean difference from true lens power for any observer was .330 with a minimum difference of .220. Significance occurred due to very small variability by observers and was clinically insignificant. For the experimental condition no significant difference for range variability within observer was found. Mean R.E. ranges (D) (by subject) for observers varied from a minimum of .430 to a maximum of .960. Between observer ranges (by subject) showed significant differences at a .05 level for two of the 10 comparisons. ·Ranges of R.E. by subject for the 5 observers varied from a low of .870 to a high of 6.750. Mean R.E. data, by subject, between observers showed significant differences for one of the 5 examiners (4 of the 10 comparisons) at a .05 level.
The usefulness of near retinoscopy as a clinical technique in an infant population is questioned based on the inter observer differences in range and mean findings, and the magnitude of variability within and between observers. The variability may in large part be due to the interaction variables between clinicians and infants for any given examination.
Files are restricted to Pacific University. Sign in to view.