American pediatricians, armed with a new reimbursement code for photoscreening young children (99174), are very interested in the validity of potential instruments.  As such, we were interested in the recent report that compared Plusoptix photoscreener to orthoptic exam in early primary school children, reporting 100% specificity but relatively low sensitivity: “the sensitivity of the PVS to detect amblyopia-associated factors is 44%, with 95% CI from 28 to 62%. Even a sensitivity of 62% would be less than desirable for a standalone screening test”[1]  We are concerned the methodology and terminology for “sensitivity” differs from what many pediatricians might expect. From Table 3, it appears the PlusOptix was able to detect most refractive errors but underestimated cycloplegic hyperopia and strabismus.  We ask that the authors report their results by using pre-defined pediatric ophthalmology risk levels (ref 10) and clarify the limited number of their subject who actually completed the cycloplegic examination which others consider the gold standard[2]. Plusoptix allows users to pre-define age-based referral cut offs for various  amblyopia risk factors.  We wonder why the authors chose their PlusOptix cutoffs that differed from an attempt to calibrate the instrument in reference 21?[3] Furthermore we wonder if changing the cutoff for hyperopia may increase the accuracy of the device. It may be worthwhile to re-analyze this original cohort to see if the sensitivity would improve. It would also be worth noting how the gold standard orthoptic vision screening compares to a dilated cyclopleged examination by a pediatric ophthalmologist, as comparing the Plusoptix to a comprehensive ophthalmology examination may change the reliability. Orthoptists are ideally trained to provide an accurate pediatric vision screening on children, however in certain countries, such as the United States of America, with a population of approximately 305 million people and only approximately 300 orthoptists more efficient methods of pediatric vision screenings must be discovered and refined. In recent population studies the prevalence of childhood amblyopia, primarily comprised of refractive error and manifest esotropia,  is about 2.5%[4].  If an acceptable screening technique has false positive rate less than 1/3, then the ideal referral rate should be about 4%, much closer to the PlusOptix compared to 12.5% for orthoptic screening.  The author’s “gold standard” orthoptic screening has merit, but probably has low predictive value for amblyopia screening.

Robert W. Arnold, MD
Noelle Matta CO, CRC, COT, Orthoptist
Members of the Vision Screening Committee of AAPOS  

References

1. Dahlmann-Noor A, Vrotsou K, Kostakis V, et al. Vision screening in Children by Plusoptix Vision Screener compared with gold standard orthoptic assessment. Br J Ophthalmol 2008;92:e-pub.
2. Donahue S, Arnold R, Ruben JB. Preschool vision screening: What should we be detecting and how should we report it?  Uniform guidelines for reporting results from studies of preschool vision screening. J AAPOS 2003;7:314-6.
3. Clausen MM, Arnold RW. Pediatric Eye/Vision Screening: Referral Criteria for the PediaVision PlusOptix S04 Photoscreener Compared to Visual Acuity & Digital Photoscreening: “Kindergarten Computer Photoscreening”. Binoc Vis and Strabismus Quart 2007;22:83-9.
4. MEPEDS. Prevalence of amblyopia and strabismus in African American and Hispanic children ages 6 to 72 months the multi-ethnic pediatric eye disease study. Ophthalmology 2008;115:1229-36 e1.