Abstract
Purpose: :
Effective screening can help prevent amblyopia development in children. The Vision in Preschoolers study (VIP) recently showed that the Welsh–Allyn SureSight (WASS) can be an effective screening instrument. However, in a recent field validation study, the WASS had a high over–referral rate for astigmatism (11.5%) and a low positive predictive value (30%) when using the referral criteria proposed by VIP to have 90% specificity (Hudson, ARVO, 2005). Our aim was to refine the cylinder cutoff value in order to increase the positive predictive value of the WASS.
Methods: :
As part of an ongoing vision screening program, The Tennessee Lions Program has screened children for amblyogenic factors including astigmatism, anisometropia, hyperopia and high myopia. Using the published Phase 1 VIP criteria for 90% specificity, cylinder values greater than 1.5 would result in referral for possible astigmatism. Referred children were examined by local Ophthalmologists and the results were used as the gold standard. The data analysis looked to determine the PPV with differing cyl values ranging from 1.5–2.3 by intervals of .1 to determine the most effective cylinder value for screening purposes.
Results: :
The program screened 3258 children ages 1– to 5– years–old with the WASS. Eighty–nine percent (2890/3258) of the children passed the initial screening and nine percent (298/3258) were referred for MD/OD formal examination by the WASS for possible astigmatism. Of the 298 children referred to an MD/OD with suspected cylinder, 142 (48%) have thus far received a comprehensive examination and cycloplegic refraction. Forty–seven of these children had astigmatism ≥ 1.5 in one or both eyes. The positive predictive value for a WASS C–value ≥1.5 was 41%. The PPV increased incrementally for each additional .1 unit increase in referral criteria up to approximately 2.1 D, which had a PPV of 60%. Except for two patients with astigmatism of 2.5D and 2.25D, no child with astigmatism > 2.0 would have been missed by adjusting the referral criteria to 2.1D. There was a small but significant relationship between the reliability value of the WASS reading and the PPV.
Conclusions: :
Over–referrals for suspected astigmatism can be decreased with appropriate alterations in the astigmatism referral value, and the PPV can be 60% when the cutoff value for referral is increased to 2.1D. Very few children with substantial magnitude astigmatism are missed by not referring those children having WASS cylinder readouts of 1.5–2.0. A modification of the WASS with multiple sets of referral criteria based upon desired sensitivity, specificity and predictive values would be welcome.
Keywords: screening for ambylopia and strabismus • amblyopia • clinical (human) or epidemiologic studies: biostatistics/epidemiology methodology