1. Astigmatism in a Countywide Preschool Population 1 L. Cowen, 2 W.R. Bobier 1 Dept. Of Statistics and Actuarial Science; 2 School of Optometry, University of Waterloo, Waterloo, Ontario, Canada Introduction Retinomax Preschool Screening Conclusions Retinomax Validation Studies Analysis ' Autorefractive measures of a population of preschool children were taken as part of an annual screening of kindergarten registrants. 6 Vision screening tests were performed during the spring of 1999 (figure 1). ' Non-cycloplegic refractive error measures were taken using the Retinomax Plus (Nikon Co) on 1162 participants of the study. ' It is generally held that astigmatic errors undergo emmetropization during the early years of life. However, large-scale population data is mostly limited to infant and school aged children. 1,2,3 ' There is a general pattern of against-the-rule astigmatism (AGR) in infancy, which typically changes to with-the-rule (WTR) by school age. 3 Studies of preschool children show that prevalence of AGR decreases with age, while WTR increases. 5 ' With the recent development of more ‘child-friendly’ autorefractors, the opportunity for population based measures has increased. ' Large screenings usually preclude the use of cycloplegic agents and thus validation of measures is important.  The sphero-cylinder refractive error measurements were decomposed into three independent components using a Fourier transformation. The components are based on equivalent sphere (M), and two (Jackson Cross) cylinder components (J 0 and J 45 ). For negative-cylinder refractive measure (S, C x  ) we used the transformation as follows: Figure 1. Flow Chart of the Oxford County Vision-Screening Programme. Preschool Child enters vision screening study Fail screening Pass screening Referred to Eye Care Practitioner Retinoscopy taken for validation of Retinomax measures Data sent to Oxford County Public Health Unit Clinical Results collected by Oxford County Board of Health Data sent to Waterloo Data Analysis Objective measure with Nikon Co. Retinomax Plus auto-refractor Discharged from the study ' Validation studies supported the use of the cylinder components and not the equivalent sphere. ' In the preschool study, the equivalent sphere is significantly more myopic for the Retinomax measures (table 1). ' The cylinder components, show smaller biases and less variation in the differences between retinoscopy and Retinomax measures. 'Oxford County preschool children astigmatisms can be broken down into 58% WTR, 30% AGR and 12% oblique. 'No significant differences in cylinder frequencies when age group stratified the population. 'The norm for this population’s cylinder measure was -1.25D. Prescribing patterns of eye-care practitioners show a threshold of prescribing which falls close to this norm (1.50D for 4 to 7 years 8 ). 1. Baldwin WR. in Rosenbloom AA, Morgan MW, editors. Principles and practice of pediatric optometry. Philadelphia: J.B. Lippincott Company; 1990; Ch6, Refractive status of infants and children. p. 104-52. 2. Banks MS. Infant refraction and accommodation. Int Ophthalmol Clin 1980;20(1):205-32. 3. Hirsch MJ. Hirsch MJ, Wick RE, editors. Vision of Children. Philadelphia: Chilton; 1963; The refraction of children. p. 150-2. 4. Abrahamsson M, Fabian G, Sjostrand J. Changes in astigmatism between the ages of 1 and 4 years: A longitudinal study. Br J Ophthalmol 1988;72(2):145-9. 5. Gwiazda J, Scheiman M, Mohindra I, Held R. Astigmatism in children: Changes in axis and amount from birth to six years. Invest Ophthalmol Vis Sci 1984;25:88-92. 6. Robinson B, Bobier WR, Martin E, Bryant L. Measurement of the validity of a preschool vision screening program. Am J Pub.Health 1999;89(2):193-8. 7. Thibos LN, Wheeler W, Horner D. Power vectors: an application of fourier analysis to the description and statistical analysis of refractive error. Optom Vis Sci 1997;74(6):367-76. 8. Miller JM, Harvey EM. Spectacle prescribing recommendations of AAPOS Members. J Pediatr Ophthalmol & Strab 1998;35(8):51-2. Acknowledgements The assistance and advice of the members of the Oxford County Board of Health and of Drs Barbara Robinson, Carloyn Machan and Melanie Campbell at the School of Optometry, University of Waterloo is greatly appreciated. We thank Welch Allyn and NSERC Canada for financial support of this project. Table 1. Differences in practitioner and Retinomax measures of M, J 0 and J 45 for 121 preschool children. Mean, standard deviation and significance level of a paired t-test are shown. ' Preschool children who failed the screening were referred to an eye- care practitioner to receive a retinoscopy. In 23% of the 155 referrals, cycloplegic refractions were conducted. Retinoscopies provided the means for validation of the Retinomax measures. ' A clinical population provided two adult samples (n 1 =167, n 2 =69) where both Retinomax and retinoscopy refractive error measures were taken. ' Retinomax performance was determined by comparison with retinoscopic measures. Retinomax Performance ' Equivalent sphere differences between retinoscopy and the Retinomax in the adult population were smaller and less variable (Table2). ' Differences in the astigmatic components and their variation were similar to those found in the preschool population (Table 1 and 2). Table 2. Differences in practitioner and Retinomax measures of M, J 0 and J 45 for a) 167 adult study and b) 69 adult study. Mean, standard deviation and significance level of a paired t-test are provided. Measurment Validation ' Transposition from Fourier form back to sphero-cylinder form gives an over-correction of (0.93, –0.09 x 137  ) for the mean differences in table 1. ' If added to the mean Retinomax readings, the over-correction would correct them to that of the clinical retinoscopy. Note that cylindrical adjustments are small. ' ANOVA split-plot design showed that population (preschool vs adult) has a significant effect on the equivalent sphere component (p=0.0001). Differences in the cylinder components are coming from the same source in the 2 populations; likely in the fact that the Retinomax provides a more accurate measure of small cylinders than the reported clinical measures. ' Retinomax astigmatism measurments are valid, however equivalent sphere measures are not. Problems with equivalent sphere seem to be consistent with over-accommodation. Stratification ' No significant differences were found between three and four year old age-groups in regards to astigmatism frequencies in the preschool population. ' Astigmatism breakdown into WTR, AGR and oblique shows that most cylinder magnitudes are between –1 and 0 D (figure 4). References Figure 4. WTR, AGR, and oblique astigmatisms for the pooled sample. ' This transformation was used to determine the validity of using Retinomax measures to describe the pattern of refractive errors found in this preschool population. As the Fourier transpositions allows M,J 0 and J 45 to be analysed seperately, simple over-accommodation would affect equivalent sphere and not the astigmatic components. ' A non-parametric 95 th percentile was deemed to be the limit of normal astigmatism. a) b) Figure 2. Frequency distribution of astigmatism for the preschool population. Figure 3. Cumulative density of the cylinder magnitudes showing the 95th percentile to be -1.25D. ' The frequency distribution of all astigmatisms is given in figure 2. WTR astigmatisms dominate this preschool population. Population Norm ' The cumulative density of the cylinder magnitudes was used to determine the 95 th percentile, which was found to be -1.25D (figure 3). Preschool Astigmatism