Recent US Preventive Services Task Force recommendations on vision screening reported insufficient data to recommend vision screening in children <3 years of age. The Iowa photoscreening program, KidSight, has screened children from 6 months of age and older since 2000. We report our experience with vision screening in these children and compare the results of the photoscreens in children younger than 3 years with those of children of preschool age and older.
A retrospective review of results from the Iowa KidSight database using the MTI PhotoScreener containing results of children screened between May 1, 2000, and April 30, 2011.
During the 11 years of the study, 210 695 photoscreens on children were performed at 13 750 sites. In the <3-year age group, the unreadable rate was 13.0%, the referral rate was 3.3%, and the overall positive-predictive value was 86.6%. In the 3- to 6-year-old children, the unreadable rate was 4.1%, the referral rate was 4.7%, and the overall positive-predictive value was 89.4%.
No statistically significant difference was found in screening children from 1 to 3 years old compared with screening children >3 years old. These results confirm that early screening, before amblyopia is more pronounced, can reliably detect amblyogenic risk factors in children younger than 3 years of age, and we recommend initiation of photoscreening in children aged 1 year and older.
Comments
Photoscreening for risk factors vs. disease
I write with concerns about the photoscreening study by Longmuir and colleagues. First, while the referral rate was comparable to that of another large photoscreening study,(1) more patients had unreadable measurements (5.9%) than were referred (4.4%). This high unreadable rate reduces the effectiveness of the screening program and adds additional "noise" to calculations of program accuracy. Second, I am concerned that they did not report the number of referred children who were actually examined. Third, the important question of sensitivity is impossible to answer, since none of the 200,000 children who passed photoscreening underwent eye examinations. Lacking information on sensitivity, the low referral rate cannot be placed in context.
It is possible to infer an expected referral rate from the published prevalences of strabismus (2%) and amblyopia (3%). We can estimate that of the 1030 children referred for strabismus, 920 (1030 x 89.4% study PPV) actually had strabismus, while the expected referral would have been 4200 (2% x 210,000). That is, only 21% of study children with strabismus (920/4200) may have been referred. This is not surprising, since only 28% of children with strabismus have refractive errors considered amblyogenic.(2) We expect that 6300 in the population (3% x 210,000) had amblyopia, but we cannot estimate how many of the 7917 referred for refractive risk factors actually had amblyopia, because visual acuity was not reported in this study, and it has been estimated that only 1 of 8 children with risk factors will develop amblyopia.(3) Therefore this study cannot be used to measure either the sensitivity or the specificity of photoscreening for detecting either strabismus or amblyopia.
In conclusion, it appears that this photoscreening program had very low sensitivity for detecting strabismus, and that neither the sensitivity nor specificity for amblyopia detection could be determined. The goal of vision screening is early detection and treatment of strabismus and amblyopia to prevent vision loss, and other instrument-based technology, which detects binocular alignment rather than refractive risk factors, has shown very high (>95%) sensitivity and specificity for detecting both conditions.(4) For proper comparison of all technologies, studies of vision screening should determine both positive and negative predictive value for detecting both strabismus and amblyopia, not just refractive risk factors.(5)
REFERENCES
1. Donahue SP, Baker JD, Scott WE, Rychwalski P, Neely DE, Tong P, Bergsma D, Lenahan D, Rush D, Heinlein K, Walkenbach R, Johnson TM. Lions Clubs International Foundation Core Four Photoscreening: results from 17 programs and 400,000 preschool children. J AAPOS. 2006 Feb;10(1):44-8.
2. Cotter SA, Varma R, Tarczy-Hornoch K, et al. Risk factors associated with childhood strabismus: the multi-ethnic pediatric eye disease and Baltimore pediatric eye disease studies. Ophthalmology 2011;118:2251-61.
3. Arnold RW. The high prevalence of AAPOS amblyopia risk factors [Abstract]. J AAPOS 2011;15:e2.
4. Loudon SJ, Rook CA, Nassif DS, Piskun NV, Hunter DG. Rapid, High- Accuracy Detection of Strabismus and Amblyopia Using the Pediatric Vision Scanner. Invest Ophthalmol Vis Sci 2011;52:5043-8.
5. Hunter DG. Targeting treatable disease - not just risk factors - in pediatric vision screening. J AAPOS 2013;17:2-3.
Conflict of Interest:
Dr. Hunter holds equity in REBIScan, Inc, a company that he founded to commercialize new vision screening technology. He is also co-inventor on two patents related to vision screening, both of which have been licensed by REBIScan. Dr. Hunter has received no direct income or royalties related to these activities, but the Children's Hospital Ophthalmology Foundation has been reimbursed for a portion of his time spent consulting for REBIScan.
Re: Parental Compliance in Preschool Vision Screening Children
Dear Authors,
I was enthusiastic to see your publication titled "Practical Community Photoscreening in Very Young Children" in which your impressive number of 210,695 photo vision screening results over an 11 year period are retrospectively reviewed against a gold standard Ophthalmologist cycloplegic examination.
Several years ago we ran a pilot preschool vision screening program in Edmonton, Alberta, Canada. In our program, a trained nurse went out to local preschools and carried out a standardized vision screening protocol with HOTV matching, Titmus stereopsis, Krimsky, ocular motility, pupils, Bruckner test(unpublished). We had excellent compliance to testing in both the >3 and <3 year old age groups, and similar positive predictive values to comparable published screening programs. The major difficulty we found arose from an unexpected source: despite obtaining written consent from all parents prior to screening, and the nurse calling and arranging follow ups for children who failed screening, only approximately 50% of children referred actually attended for their gold standard eye exam.
Parental compliance following through with recommendations when a child fails a screening test significantly reduces the impact of mass screening interventions, but often goes unreported. It would be interesting to know what percentage of the children who were referred for failed the photo screening were subsequently lost to follow up without the results of the gold standard exam in your study group.
Sincerely,
Brad Wakeman BSc, OC(C) Department of Ophthalmology, University of Alberta Faculty of Medicine
Conflict of Interest:
None declared