- Email: [email protected]
Validation of photoscreening technology in the general pediatrics office: a prospective study
Validation of photoscreening technology in the general pediatrics office: a prospective study Jana Bregman, MD,a,b and Sean P. Donahue, MD, PhDa BACKGROUND
Photoscreening instruments have been widely validated in pediatric ophthalmology clinics and field studies; however, validation by general pediatricians is lacking. We performed the first prospective, multisite evaluation of a commercially available photoscreener in the medical home.
METHODS
Eleven practices in Middle Tennessee recruited over 3,100 children between 12 months and 5 years to be screened at well-child examinations. Participants were those who received a “refer” result; controls received a “pass.” Referred children received a comprehensive eye examination with cycloplegic retinoscopy. A subset of control children underwent eye examinations in an attempt to determine sensitivity and specificity. The overall referral rate was 10%. Amblyopia risk factors (ARFs) were confirmed in 47% of referred children, with positive predictive values (PPVs) of 77.8% for suspected hyperopia, 60% for myopia, 50% for anisometropia, and 44.8% for astigmatism by the 2013 guidelines of the American Association of Pediatric Ophthalmology and Strabismus Vision Screening Committee. Using the 2003 guidelines, the overall PPV was 60.3%; PPVs were determined for suspected hyperopia (77.8%), myopia (60%), anisometropia (67.6%), and astigmatism (61.2%). Of referred children who received follow-up, 18 (13.2%) had amblyopia. PPVs for children #36 months (n 5 79) did not differ from those 37-72 months (n 5 57). No child who passed screening and had a follow-up examination had any ARFs.
RESULTS
CONCLUSIONS
Our results replicate those of previously published field studies and support recent United States Preventive Services Task Force and American Academy of Pediatrics position statements. They provide prospective evidence that photoscreening is an effective tool for children aged 12-72 months. ( J AAPOS 2016;20:153-158)
A
mblyopia remains the most common cause of monocular vision loss in children and young adults.1 Efforts to detect amblyopia risk factors (ARFs) and initiate timely treatment are widely viewed as integral to pediatric primary care.2,3 The estimated prevalence of ARFs in the pediatric population is 15%-20%.1-4 There is ample evidence that there is a direct relationship between age and depth of amblyopia in the absence of appropriate treatment,4,5 moreover,
Author affiliations: aVanderbilt Eye Institute, Nashville, Tennessee; bVanderbilt University School of Medicine, Nashville, Tennessee Funding support included a grant from Welch Allyn to the Cumberland Pediatric Foundation and an unrestricted grant from Research to Prevent Blindness, New York, NY. Financial disclosure: Sean Donahue was a prior consultant for Welch Allyn, IScreen, PediaVision, Plus Optix, Diopsys, and Gobiquity. Conflict of interest: Sean Donahue is the lead author of the forthcoming AAP position statement on preschool vision screening. Presented as a paper at the 41st Annual Meeting of the American Association for Pediatric Ophthalmology and Strabismus, New Orleans, Louisiana, March 25-29, 2015. Submitted October 8, 2015. Revision accepted January 4, 2016. Correspondence: Jana Bregman, MD, Vanderbilt Eye Institute, 2311 Pierce Avenue, Nashville, TN 37232 (email: [email protected]). Published by Elsevier Inc. on behalf of the American Association for Pediatric Ophthalmology and Strabismus. 1091-8531/$36.00 http://dx.doi.org/10.1016/j.jaapos.2016.01.004
Journal of AAPOS
there is a known inverse relationship between age and the effectiveness of amblyopia treatment.4,6 Therefore, the detection of ARFs and early initiation of therapy during the critical window of visual development is imperative for preventing future visual complications, such as a doubled lifetime risk of bilateral visual impairment with untreated amblyopia.1,4 Since the first automated device for preschool vision screening became commercially available approximately 2 decades ago, photoscreening instruments have become easier to use without compromising accuracy and precision for detecting ARFs.7 In addition, automated vision screening poses several potential advantages over traditional screening techniques, including shorter time to screen per child and ability to screen less cooperative subjects.8 Photoscreening is now recognized by both the US Preventive Services Task Force (USPSTF) and the American Academy of Pediatrics (AAP) as a valid screening modality in the preschoolers.2,3 Although automated screening instruments have undergone extensive validation in pediatric ophthalmology clinics and field studies,9-12 validation in the medical home is lacking. In addition, there are fewer studies that specifically address the efficacy of vision screening among children \36 months of age.13-15 Additional research
153
154
Bregman and Donahue
Volume 20 Number 2 / April 2016
Table 1A. Manufacturer criteria for the Spot v.2.1.4a Age-adjusted criteria for amblyopia risk factors, D Age, mos
Anisometropia
Astigmatism
Myopia
Hyperopia
Anisocoria
6-12 12-36 36-72 72-240 240-1200
$1.5 $1 $1 $1 $1
$2.25 $2 $1.75 $1.5 $1.5
$ 2 $ 2 $ 1.25 $1.0 $ 0.75
$3.5 $3 $2.5 $2.5 $1.5
$1 $1 $1 $1 $1
Media opacity Any Any Any Any Any
D, diopter. a Referral criteria according to the preprogrammed Spot software v.2.14. A patient’s vision screen results in a Refer if any of the above criteria are met for either or both eyes and a Pass if none of the above criteria are met.
addressing these key issues is needed for this technology, which has already been approved by the current USPSTF and AAP guidelines, to become more widely adopted as standard practice in general pediatric clinics. Key controversies surrounding automated vision screening lie in determining the appropriate age range for screening and optimizing referral criteria to reduce over-referrals while capturing children at risk of developing amblyopia.7 The purpose of this study is to describe the first large-scale, prospective, multi-site evaluation of a commercially available photoscreening device in the medical home, and to begin to address key controversies associated with this technology by evaluating its use in the younger pediatric population, and identifying elements of the referral process that need to be refined.
Subjects and Methods Approval from the Institutional Review Board of Vanderbilt University was obtained for this study, which conformed to the requirements of the US Health Insurance Portability and Accountability Act of 1996. Informed consent was obtained for all participants. Eleven large pediatric practices were recruited to perform photoscreening of healthy preschool children during well-child examinations (see e-Supplement 1 at jaapos.org for a complete listing). Screenings were performed by nurses at the various practices who had completed formal training led by a company representative of the photoscreening device prior to beginning screening. Inclusion criteria consisted of children between 12 and 72 months who were to be seen for a well-child examination. Children were screened using the Spot Vision Screener v2.1.4 (Welch Allyn, Syracuse, NY). The Spot Vision Screener is a handheld, digital photoscreener with touch screen and electronic data storage and transfer. It uses corneal and retinal reflections to screen for ARFs. It estimates refractive error in a range of 7.5 D (spherical equivalent) and compares the estimated value with preprogrammed, age-adjusted criteria for 7 amblyopia risk factors (media opacity, strabismus, anisocoria, anisometropia, hyperopia, astigmatism, and myopia); calculates estimates of spherical and cylindrical refractive error, axis, and gaze vector; and generates a “pass” or “refer” result according to manufacturer specifications (Table 1A, 1B). If the device detects an amount of refractive error
Table 1B. Manufacturer criteria for Spot v.2.1.4 Component of gaze deviation, prism diopters Age, mos.
Vertical
Nasal
Temporal
Asymmetry
6-12 12-36 36-72 72-240 240-1200
$8 $8 $8 $8 $8
$5 $5 $5 $5 $5
$8 $8 $8 $8 $8
$8 $8 $8 $8 $8
that is greater than the ability of the instrument to quantitate, it automatically generates a refer output. The device was held at eye-level, 3 feet away from the seated child in a dimly lit room, to complete the data capture. Possible reasons for a refer if estimates exceeded preprogrammed, age-adjusted values, included the following: media opacity, strabismus, anisocoria, anisometropia, hyperopia, astigmatism, and myopia. Suspected reasons for referral and the affected eye(s) were listed. Physicians whose patients received a referal for their examination were notified during the patient visit that referral to an ophthalmologist or optometrist was warranted. Referred children who completed follow-up received a gold standard pediatric ophthalmic examination performed by a community ophthalmologist or optometrist using a specified data sheet; the examination included age appropriate assessment of visual acuity, a cover test for the detection of strabismus, and cyclopegic retinoscopy. To determine the primary diagnosis from the gold standard eye examination, amblyopia risk factors were ranked in the following order: (1) media opacity, (2) manifest strabismus, (3) anisometropia, (4) hyperopia, (5) astigmatism, and (6) myopia. The revised 2013 American Association of Pediatric Ophthalmology and Strabismus Vision Screening Committee (AAPOS-VSC) guidelines were used to determine whether an ARF was present (Table 2A).7 The 2013 guidelines are age-based and of higher magnitude than the 2003 criteria. Because the 2013 guidelines are more stringent and require higher levels of refractive error, the data were also evaluated according to the 2003 AAPOS-VSC guidelines so that the data could be compared with previously published field studies that relied on the older criteria (Table 2B).16 All patients who received a pass were pooled, and a randomly selected subset was recruited for a free, comprehensive, gold standard pediatric ophthalmic examination in an effort to calculate the sensitivity and specificity of the study. An attempt to contact these individuals was made by telephone using the
Journal of AAPOS
Volume 20 Number 2 / April 2016
Bregman and Donahue
Table 2A. 2013 Amblyopia risk factor (ARF) criteria of the American Association of Pediatric Ophthalmology and Strabismus Vision Screening Committeea Age-adjusted criteria for amblyopia risk factors Age, mos 12-30 31-48 .48 All ages
Anisometropia, D
Hyperopia, D
.2.5 .4.5 .2.0 .4.0 .1.5 .3.5 Manifest strabismus of .8 PD Media opacity of .1 mm
Astigmatism, D
Myopia, D
.2.0 .2.0 .1.5
. 3.5 . 3.0 . 1.5
D, diopters; PD, prism diopters. Updated criteria for automated preschool vision screening based on randomized controlled trials of amblyopia treatment and field studies of screening technology, with the goal of attaining high specificity for ARF detection in young children and high sensitivity for ARF detection in older children.
a
contact number provided on the consent form. A study subject was excluded for further recruitment if contact failed on three separate occasions.
Results During an 8-month period, 3,134 children aged 1272 months at 11 pediatric practices in Middle Tennessee were recruited and screened. Recruitment was restricted, so that no single practice screened more than 20% (n 5 627) of the total study population. Of the 3,134 children screened, 306 (9.8%) were referred for a formal eye examination, and 2,828 (90.2%) children had normal screening results. Table 3 shows the referral rate by practice and the relative distribution of the number of children screened per practice. The overall referral rate of 10% was fairly consistent across all practices; the only outlier was Southern Hills, with a referral rate of 24%. Southern Hills is located in a predominately Hispanic population, which has a higher prevalence of astigmatism and therefore an expectedly higher referral rate. The most common reasons for referral were as follows: suspected astigmatism (n 5 189 [61.8%]), suspected anisometropia (n 5 53 [17.3%]), and strabismus (n 5 39 [12.7%]). Referrals for suspected hyperopia (12 [3.9%]), myopia (8 [2.6%]), anisocoria (5 [1.6%]), and media opacity (0) were all low. See Table 4. Of the 306 referred children, 136 (44.4%) completed a gold standard examination, including cycloplegic retinoscopy. Demographics for this population are provided in Table 5. The age distribution of this subgroup (58%, 12-36 months; 42%, 37-72 months) reflected the entire referred population (61%, 12-36 months; 39%, 37-72 months). The primary reasons for referral for this subgroup do not differ significantly from the distribution of the total referred population (Table 4). The most common reason for referral in examined children was suspected astigmatism (n 5 67 [49.3%]), and a large percentage were due to suspected anisometropia (n 5 34 [25%]) or strabismus (n 5 18 [13.2%]). The numbers of children
Journal of AAPOS
155
Table 2B. 2003 AAPOS VSC amblyopia risk factor criteriaa ARF identified Anisometropia (spherical or cylindrical) Strabismus Hyperopia Myopia Media opacity Astigmatism Ptosis Visual acuity
Criteria, not age adjusted .1.5 D Any manifest strabismus .3.5 D in any meridian Magnitude .3.00 D in any meridian Any media opacity .1 mm in size .1.5 D at 90 or 180 .1.0 in an oblique axis (.10 eccentric to 90 or 180 ) #1 mm margin reflexb distance Per age-appropriate standards
D, diopter. a Original 2003 criteria for automated preschool vision screening that are referred to by the vast majority of currently published field studies of automated technology. b A standard objective measurement of ptosis equal to the distance from the corneal light reflex to the margin. Table 3. Referral rate by practice and the relative distribution of screenings by practice Practicea
No. children screened (% referred)
Nurture Tullahoma CCN Rivergate Hermitage Mt. Juliet CCE Lebanon Brentwood Capstone Southern Hills Lebanon
470 (11) 508 (7) 199 (8) 62 (10) 280 (10) 605 (8) 166 (12) 558 (11) 114 (12) 91 (24) 81 (7)
a
11 pediatric practices located in Middle Tennessee. See e-Supplement 1 at jaapos.org for formal list.
examined after referral for suspected hyperopia (n 5 9 [6.6%]), myopia (n 5 5 [3.7%]), anisocoria (n 5 3 [2.2%]) and media opacity (n 5 0) were all low. When the 2013 AAPOS-VSC guidelines were used as the gold standard, ARFs were detected in 64 of 136 patients for an overall positive predictive value (PPV) of 47%. The PPV specified by primary reason for referral is listed in Table 6. The PPV for suspected hyperopia was 77.8%; for myopia, 60%. The PPV for suspected anisometropia was 50%; for astigmatism, 44.8%. The PPV for suspected strabismus was only 33.3% and the PPV for suspected anisocoria was 33.3% (only 1 of 3 children examined for this reason had anisocoria confirmed on examination). In order to address previously published field studies, which refer to the 2003 AAPOS guidelines, these data were also analyzed using the older guidelines as gold standard. This analysis resulted in 82 patients with amblyopia risk factors (PPV of 60.3%). The PPV specified by primary reason for referral is listed in Table 6. PPVs were calculated for suspected hyperopia (77.8%), anisometropia (67.6%), astigmatism (61.2%), and myopia (60%). The PPVs for suspected strabismus and anisocoria remained unchanged.
156
Bregman and Donahue
Volume 20 Number 2 / April 2016
Table 4. Primary reason for Spot 2.1.4 referrala No. Primary reason for referral
All children referred N 5 306 (10% total study population)
Percent of population (with follow-up)
Astigmatism Anisometropia Strabismus Hyperopia Myopia Anisocoria Media opacity
189 53 39 12 8 5 0 Children with completed follow-up N 5 136 (44% referred population)
61.8 17.3 12.7 3.9 2.6 1.6 0
Astigmatism Anisometropia Strabismus Hyperopia Myopia Anisocoria Media opacity
67 34 18 9 5 3 0
49.3b 25.0b 13.2b 6.6b 3.7b 2.2b 0b
a Approximately 44% of children referred completed a follow-up appointment. Reasons for referral were comparable between the total referred population and the referred population with follow-up. b Percentage refers to fraction of referred population with follow-up rather than total referred population.
Table 5. Patient characteristics for referred population with completed follow-up
Characteristics Age group, mos 12-24 25-36 37-48 49-60 61-72 Sex Female Male Ethnicity White Hispanic African American Asian Middle Eastern Unknown Insurance Status Commercial Government Unknown
All patients referred by Spot 2.1.4 for a suspected amblyopia risk factor, no. (%) (n 5 136) 46 (33.8) 33 (24.3) 28 (20.6) 20 (14.7) 9 (6.6) 70 (51.5) 66 (48.5) 88 (64.7) 20 (14.7) 11 (8.1) 9 (6.6) 3 (2.2) 5 (3.7) 63 (46.3) 62 (45.6) 11 (8.1)
The data were also analyzed by patient age. Of the 136, 79 (58.1%) who were referred and completed follow-up were #36 months of age. The USPSTF guidelines state that evidence of photoscreening effectiveness in this age group is inconclusive.2 According to the 2013 AAPOSVSC criteria, the overall PPV for this younger age group was less than that in the older age group (38.0% vs
59.6%); however, according to the 2003 guidelines, the overall PPVs were comparable (60.8% vs 59.6%; Table 6). Of the 136 children who were referred and completed follow-up, 18 (13.2%) had amblyopia (a difference of 2 lines of best spectacle-corrected visual acuity or, for preverbal children, unmaintained fixation using fixation preference testing). Amblyopia was detected in 6 children 12-36 months old and 12 children 37-72 months old. Sixteen cases were due to anisometropic amblyopia; 2 cases, to strabismic amblyopia. Assuming that amblyopia was evenly distributed in the referred population who did not receive follow-up examinations, we estimated that the Spot referred 33 children with amblyopia. This equates to 1.0% of the screened population and is on the lower end of the range of estimates of amblyopia in the general pediatric population.17,18,19 To determine the sensitivity of the Spot vision screener in this study, parents of approximately 300 children who passed their screening examination were contacted to schedule a free follow-up, gold standard pediatric ophthalmic examination with cycloplegic retinoscopy. About 75 appointments were scheduled successfully; however, only 12 of the scheduled children appeared for an examination. All 12 of these children had normal ophthalmic examination results with no child who received a pass diagnosed as having amblyopia or an ARF. We considered contacting additional children to increase the sample of passed children, but decided that, with an examination rate of only 4% of contacted children, there was a potential for introducing significant selection bias in the study.
Discussion The present study demonstrates that the Spot v.2.1.4 photoscreener is an effective tool for detecting ARFs when used in general pediatrics clinic. Our overall referral rate (10%) and PPV (60.3%) replicate the outcome measurements from previously published field studies and key validation studies. In addition, over 60% of children referred were found to have ARFs and 13% had amblyopia detected on examination, both of which are significantly more than in the general population, which is thought to be 15%20% for ARFs and 1%-2% for amblyopia. The USPSTF and AAP endorse automated vision screening in the 3- to 5-year-olds. Both identify photoscreening as a valid modality with potential for use in the 6- to 36-month age range when other vision screening techniques are not age-appropriate.2,3 The former recommends “screening for all children at least once between the ages of 3 and 5 years” and lists photoscreening as one modality that is “feasible” for testing in the primary care setting. The AAP echoes the USPSTF recommendation for routine screening but further specifies that screening should be performed at an early age, at regular intervals, and “within the medical home.” The USPSTF provides a class “B” rating for vision screening in children aged 3-5 years and an “I”
Journal of AAPOS
Volume 20 Number 2 / April 2016
Bregman and Donahue
Table 6. Comparison of positive predictive value (PPV) by primary reason for referral according to 2013 and 2003 AAPOS VSC guidelines and by age groupa PPV: no. true positives (%) Primary reason for referral
No.
All age groups (n 5 136) Overall PPV 136 Astigmatism 67 Anisometropia 34 Strabismus 18 Hyperopia 9 Myopia 5 Anisocoria 3 Media opacity 0 Children aged 12-36 months (n 5 79) Overall PPV 79 Astigmatism 37 Anisometropia 19 Strabismus 12 Hyperopia 5 Myopia 3 Anisocoria 3 Media opacity 0 Children aged 37-72 months (n 5 57) Overall PPV 57 Astigmatism 30 Anisometropia 15 Strabismus 6 Hyperopia 4 Myopia 2 Anisocoria 0 Media opacity 0
By 2013 guidelines
By 2003 guidelines
64 (47.0) 30 (44.8) 17 (50) 6 (33.3) 7 (77.8) 3 (60.0) 1 (33.3) –
82 (60.3) 41 (61.2) 23 (67.6) 6 (33.3) 7 (77.8) 3 (60.0) 1 (33.3) –
30 (38.0) 15 (40.5) 7 (36.8) 3 (25.0) 4 (80.0) 1 (33.3) 1 (33.3) –
48 (60.8) 26 (70.3) 13 (68.4) 3 (25.0) 4 (80.0) 1 (33.3) 1 (33.3) –
34 (59.6) 16 (53.3) 10 (66.7) 3 (50.0) 3 (75.0) 2 (100.0) – –
34 (59.6) 16 (53.3) 10 (66.7) 3 (50.0) 3 (75.0) 2 (100.0) – –
a
Calculation of PPVs included only those study participants who had completed follow-up (44.4% of total referred population). PPVs according to the 2013 criteria resulted in lower numbers than when referencing the less stringent 2003 criteria.
rating for vision screening in children under 3 years in whom they feel the data are inconclusive. In our study population, according to the 2003 AAPOSVSC criteria, the PPV for children aged 12-36 months was comparable to that of the older age group (60.8% vs 59.6%). Evaluating the data with the more restrictive 2013 criteria resulted in a lower PPV for the younger age group (38.0% vs 59.6%), because these guidelines require higher magnitudes of refractive error to be considered ARFs at younger ages. Nevertheless, the advantage of early detection is that prompt treatment in this age group helps prevent amblyopia from becoming entrenched.4,6 Our results, along with those of previous studies, support the use of photoscreening for the younger age group12-15,20 and should help make the inconclusive rating obsolete. It has been previously demonstrated that at baseline, automated screening outperforms traditional methods in terms of sensitivity and specificity.8 Efforts to enhance both sensitivity and specificity of photoscreening are ongoing with new technology and software updates. High specificity is important for reducing health care costs and parental anxiety as well as increasing confidence in the screening devices by physicians and other users. Sensitivity
Journal of AAPOS
157
is not as crucial with a single screening, because vision screening should be viewed as a continuous process throughout childhood, with multiple opportunities to detect pathology. Consequences of false positives include unnecessary follow-up appointments. Our false positive rate was 53% according to 2013 guidelines and 40% according to 2003 guidelines, with the largest proportion of false positives due to suspected strabismus (PPV of 33.3% for both). This is not surprising, because the photoscreener can easily “mistake” head tilting or distraction for gaze deviation. These results highlight the need for redefining the current algorithm used to detect strabismus as well as optimizing sensitivity and specificity to enhance the cost effectiveness and utility of photoscreening.12 A significant limitation to our study was the poor response rate for children who passed and were offered a free, comprehensive eye examination. There are a number of possible explanations for the poor response; however, it is unlikely that a large study can be performed in the future without financial compensation to encourage parents to have their children with “normal” vision examined. Having a study coordinator in charge of scheduling appointments and ensuring attendance, as in the Iowa MTI,21 Alaska Blind Child Discovery,22 and Tennessee Lions Outreach studies, would almost certainly have improved our follow-up rates. Regardless, because the number of follow-ups was small, sensitivity could not be adequately assessed. We felt that repeated attempts to recruit children who passed the photoscreening would bias the results by including children possibly more likely to have pathology. Our compliance rate for follow-up among children who received a refer was 44%. Without optimal compliance, there is a potential for selection bias; however, the distribution of pathology among children with completed followup did not differ significantly from that of all children referred (Table 4). This finding is therefore reassuring that our sample of children with completed follow-up is representative of the total referred population. An additional limitation of our study was the use of doctors with variable levels of training (ophthalmologists and optometrists) to complete the follow-up examinations. We chose to include multiple community physicians to maximize appointment slots and optimize follow-up rates. This aspect of the study design could be improved in follow-up studies by having all examinations completed by a single ophthalmologist to increase the accuracy of the examination. Lastly, we were unable to draw statistically strong conclusions regarding some parameters because of the small number of children referred for suspected myopia, hyperopia, anisocoria, or media opacity. Although we were not surprised by the lower numbers of referred children with less common pathology, the data and our conclusions appropriately reflect the ARFs that are more prevalent in the general pediatric population. Photoscreening is a single cost-effective test, can be administered quickly by nonmedical personnel, can be
158
Bregman and Donahue
used in many environments, and does not require prolonged patient cooperation. Prior to this study, photoscreeners had not been validated in the general pediatrics clinic and therefore may not have been widely incorporated as standard practice at well-child examinations. Our results support ongoing efforts to introduce and standardize the use of automated screening practices in the medical home. Additional, prospective, studies should clarify circumstances when intervention should be undertaken in very young children.
Acknowledgments The authors thank Sandy Owings, COA. and Janet Cates. References 1. Gunton K. Advances in amblyopia: what we have we learned from PEDIG trials? Pediatrics 2013;131:540-47. 2. US Preventive Services Task Force. Vision screening for children 1 to 5 years of age: US preventive services task force recommendation statement. Pediatrics 2011;127:340-46. 3. American Academy of Pediatrics Section on Ophthalmology, Committee on Practice and Ambulatory Medicine, American Academy of Ophthalmology, American Association for Pediatric Ophthalmology and Strabismus, American Association of Certified Orthoptists. Instrument-based pediatric vision screening policy statement. Pediatrics 2012;130:983-6. 4. Donahue SP. Relationship between anisometropia, patient age, and the development of amblyopia. Am J Ophthalmol 2006;142:132-40. 5. Arnold RW. Amblyopia risk factor prevalence. J Pediatr Ophthalmol Strabismus 2013;50:213-17. 6. Holmes JM, Lazar EL, Melia BM, et al., Pediatric Eye Disease Investigator Group. Effect of age on response to amblyopia treatment in children. Arch Ophthalmol 2011;129:1451-7. 7. Donahue SP, Arthur B, Neely DE, Arnold RW, Silbert D, Ruben JB. AAPOS Vision Screening Committee. Guidelines for automated preschool vision screening: a 10-year, evidence-based update. J AAPOS 2013;17:4-8. 8. Salcido AA, Bradley J, Donahue SP. Predictive value of photoscreening and traditional screening of preschool children. J AAPOS 2005;9: 114-20.
Volume 20 Number 2 / April 2016 9. Matta NS, Singman EL, Silbert DI. Performance of the Plusoptix vision screener for the detection of amblyopia risk factors in children. J AAPOS 2008;12:490-92. 10. Peterseim MM, Papa CE, Wilson ME, et al. The effectiveness of the Spot Vision Screener in detecting amblyopia risk factors. J AAPOS 2014;18:539-42. 11. Garry GA, Donahue SP. Validation of Spot screening device for amblyopia risk factors. J AAPOS 2014;18:476-80. 12. Arnold RW, Tulip D, McArthur E, et al. Predictive value from pediatrician Plusoptix screening: Impact of refraction and binocular alignment. Binoc Vis Strabolog Q Simms Romano 2012;27:227-32. 13. Lowry EA, Wang W, Nyong’o O. Objective vision screening in 3year-old children at a multispeciality practice. J AAPOS 2015;19: 16-20. 14. Halegoua J, Schwartz RH. Vision photoscreening of infants and young children in a primary care pediatric office: can it identify asymptomatic treatable amblyopic risk factors? Clin Pediatr 2015; 54:33-9. 15. Longmuir SQ, Boese EA, Pfeifer W, et al. Practical community photoscreening in very young children. Pediatrics 2013;131:764-9. 16. Donahue SP, Arnold RW, Ruben JB, AAPOS Vision Screening Committee. Preschool vision screening: what should we be detecting and how should we report it? Uniform guidelines for reporting results of preschool vision screening studies. J AAPOS 2003;7:314-16. 17. Cotter SA, Varma R, Tarczy-Hornoch K, et al. Risk factors associated with childhood strabismus. Ophthalmology 2011;118:2251-61. 18. Multi-ethnic Pediatric Eye Disease Study Group. Prevalence of amblyopia and strabismus in African American and Hispanic children ages 6 to 72 months. Ophthalmology 2008;115:1229-36. 19. Cotter S, Cyert L, Miller J, Quinn GE, National Expert Panel to the National Center for Children’s Vision and Eye Health. Vision screening for children 36 to \72 months: recommended practices. Optom Vis Sci 2015;92:6-16. 20. McKean-Cowdin R, Cotter SA, Tarczy-Hornoch K, Multi-Ethnic Pediatric Eye Disease Study Group. Prevalence of amblyopia or strabismus in Asian and Non-Hispanic white preschool children. Ophthalmology 2013;120:2117-24. 21. Longmuir SQ, Pfeifer W, Leon A, Olson RJ, Short L, Scott WE. Nine-year results of a volunteer lay network photoscreening program of 147,809 children using a photoscreener in Iowa. Ophthalmology 2010;117:1869-75. 22. Arnold RW, Donahue SP. The yield and challenges of charitable state-wide photoscreening. Binocul Vis Strabismus Q 2006;21: 93-100.
Journal of AAPOS
Photoscreening instruments have been widely validated in pediatric ophthalmology clinics and field studies; however, validation by general pediatricians is lacking. We performed the first prospective, multisite evaluation of a commercially available photoscreener in the medical home.
METHODS
Eleven practices in Middle Tennessee recruited over 3,100 children between 12 months and 5 years to be screened at well-child examinations. Participants were those who received a “refer” result; controls received a “pass.” Referred children received a comprehensive eye examination with cycloplegic retinoscopy. A subset of control children underwent eye examinations in an attempt to determine sensitivity and specificity. The overall referral rate was 10%. Amblyopia risk factors (ARFs) were confirmed in 47% of referred children, with positive predictive values (PPVs) of 77.8% for suspected hyperopia, 60% for myopia, 50% for anisometropia, and 44.8% for astigmatism by the 2013 guidelines of the American Association of Pediatric Ophthalmology and Strabismus Vision Screening Committee. Using the 2003 guidelines, the overall PPV was 60.3%; PPVs were determined for suspected hyperopia (77.8%), myopia (60%), anisometropia (67.6%), and astigmatism (61.2%). Of referred children who received follow-up, 18 (13.2%) had amblyopia. PPVs for children #36 months (n 5 79) did not differ from those 37-72 months (n 5 57). No child who passed screening and had a follow-up examination had any ARFs.
RESULTS
CONCLUSIONS
Our results replicate those of previously published field studies and support recent United States Preventive Services Task Force and American Academy of Pediatrics position statements. They provide prospective evidence that photoscreening is an effective tool for children aged 12-72 months. ( J AAPOS 2016;20:153-158)
A
mblyopia remains the most common cause of monocular vision loss in children and young adults.1 Efforts to detect amblyopia risk factors (ARFs) and initiate timely treatment are widely viewed as integral to pediatric primary care.2,3 The estimated prevalence of ARFs in the pediatric population is 15%-20%.1-4 There is ample evidence that there is a direct relationship between age and depth of amblyopia in the absence of appropriate treatment,4,5 moreover,
Author affiliations: aVanderbilt Eye Institute, Nashville, Tennessee; bVanderbilt University School of Medicine, Nashville, Tennessee Funding support included a grant from Welch Allyn to the Cumberland Pediatric Foundation and an unrestricted grant from Research to Prevent Blindness, New York, NY. Financial disclosure: Sean Donahue was a prior consultant for Welch Allyn, IScreen, PediaVision, Plus Optix, Diopsys, and Gobiquity. Conflict of interest: Sean Donahue is the lead author of the forthcoming AAP position statement on preschool vision screening. Presented as a paper at the 41st Annual Meeting of the American Association for Pediatric Ophthalmology and Strabismus, New Orleans, Louisiana, March 25-29, 2015. Submitted October 8, 2015. Revision accepted January 4, 2016. Correspondence: Jana Bregman, MD, Vanderbilt Eye Institute, 2311 Pierce Avenue, Nashville, TN 37232 (email: [email protected]). Published by Elsevier Inc. on behalf of the American Association for Pediatric Ophthalmology and Strabismus. 1091-8531/$36.00 http://dx.doi.org/10.1016/j.jaapos.2016.01.004
Journal of AAPOS
there is a known inverse relationship between age and the effectiveness of amblyopia treatment.4,6 Therefore, the detection of ARFs and early initiation of therapy during the critical window of visual development is imperative for preventing future visual complications, such as a doubled lifetime risk of bilateral visual impairment with untreated amblyopia.1,4 Since the first automated device for preschool vision screening became commercially available approximately 2 decades ago, photoscreening instruments have become easier to use without compromising accuracy and precision for detecting ARFs.7 In addition, automated vision screening poses several potential advantages over traditional screening techniques, including shorter time to screen per child and ability to screen less cooperative subjects.8 Photoscreening is now recognized by both the US Preventive Services Task Force (USPSTF) and the American Academy of Pediatrics (AAP) as a valid screening modality in the preschoolers.2,3 Although automated screening instruments have undergone extensive validation in pediatric ophthalmology clinics and field studies,9-12 validation in the medical home is lacking. In addition, there are fewer studies that specifically address the efficacy of vision screening among children \36 months of age.13-15 Additional research
153
154
Bregman and Donahue
Volume 20 Number 2 / April 2016
Table 1A. Manufacturer criteria for the Spot v.2.1.4a Age-adjusted criteria for amblyopia risk factors, D Age, mos
Anisometropia
Astigmatism
Myopia
Hyperopia
Anisocoria
6-12 12-36 36-72 72-240 240-1200
$1.5 $1 $1 $1 $1
$2.25 $2 $1.75 $1.5 $1.5
$ 2 $ 2 $ 1.25 $1.0 $ 0.75
$3.5 $3 $2.5 $2.5 $1.5
$1 $1 $1 $1 $1
Media opacity Any Any Any Any Any
D, diopter. a Referral criteria according to the preprogrammed Spot software v.2.14. A patient’s vision screen results in a Refer if any of the above criteria are met for either or both eyes and a Pass if none of the above criteria are met.
addressing these key issues is needed for this technology, which has already been approved by the current USPSTF and AAP guidelines, to become more widely adopted as standard practice in general pediatric clinics. Key controversies surrounding automated vision screening lie in determining the appropriate age range for screening and optimizing referral criteria to reduce over-referrals while capturing children at risk of developing amblyopia.7 The purpose of this study is to describe the first large-scale, prospective, multi-site evaluation of a commercially available photoscreening device in the medical home, and to begin to address key controversies associated with this technology by evaluating its use in the younger pediatric population, and identifying elements of the referral process that need to be refined.
Subjects and Methods Approval from the Institutional Review Board of Vanderbilt University was obtained for this study, which conformed to the requirements of the US Health Insurance Portability and Accountability Act of 1996. Informed consent was obtained for all participants. Eleven large pediatric practices were recruited to perform photoscreening of healthy preschool children during well-child examinations (see e-Supplement 1 at jaapos.org for a complete listing). Screenings were performed by nurses at the various practices who had completed formal training led by a company representative of the photoscreening device prior to beginning screening. Inclusion criteria consisted of children between 12 and 72 months who were to be seen for a well-child examination. Children were screened using the Spot Vision Screener v2.1.4 (Welch Allyn, Syracuse, NY). The Spot Vision Screener is a handheld, digital photoscreener with touch screen and electronic data storage and transfer. It uses corneal and retinal reflections to screen for ARFs. It estimates refractive error in a range of 7.5 D (spherical equivalent) and compares the estimated value with preprogrammed, age-adjusted criteria for 7 amblyopia risk factors (media opacity, strabismus, anisocoria, anisometropia, hyperopia, astigmatism, and myopia); calculates estimates of spherical and cylindrical refractive error, axis, and gaze vector; and generates a “pass” or “refer” result according to manufacturer specifications (Table 1A, 1B). If the device detects an amount of refractive error
Table 1B. Manufacturer criteria for Spot v.2.1.4 Component of gaze deviation, prism diopters Age, mos.
Vertical
Nasal
Temporal
Asymmetry
6-12 12-36 36-72 72-240 240-1200
$8 $8 $8 $8 $8
$5 $5 $5 $5 $5
$8 $8 $8 $8 $8
$8 $8 $8 $8 $8
that is greater than the ability of the instrument to quantitate, it automatically generates a refer output. The device was held at eye-level, 3 feet away from the seated child in a dimly lit room, to complete the data capture. Possible reasons for a refer if estimates exceeded preprogrammed, age-adjusted values, included the following: media opacity, strabismus, anisocoria, anisometropia, hyperopia, astigmatism, and myopia. Suspected reasons for referral and the affected eye(s) were listed. Physicians whose patients received a referal for their examination were notified during the patient visit that referral to an ophthalmologist or optometrist was warranted. Referred children who completed follow-up received a gold standard pediatric ophthalmic examination performed by a community ophthalmologist or optometrist using a specified data sheet; the examination included age appropriate assessment of visual acuity, a cover test for the detection of strabismus, and cyclopegic retinoscopy. To determine the primary diagnosis from the gold standard eye examination, amblyopia risk factors were ranked in the following order: (1) media opacity, (2) manifest strabismus, (3) anisometropia, (4) hyperopia, (5) astigmatism, and (6) myopia. The revised 2013 American Association of Pediatric Ophthalmology and Strabismus Vision Screening Committee (AAPOS-VSC) guidelines were used to determine whether an ARF was present (Table 2A).7 The 2013 guidelines are age-based and of higher magnitude than the 2003 criteria. Because the 2013 guidelines are more stringent and require higher levels of refractive error, the data were also evaluated according to the 2003 AAPOS-VSC guidelines so that the data could be compared with previously published field studies that relied on the older criteria (Table 2B).16 All patients who received a pass were pooled, and a randomly selected subset was recruited for a free, comprehensive, gold standard pediatric ophthalmic examination in an effort to calculate the sensitivity and specificity of the study. An attempt to contact these individuals was made by telephone using the
Journal of AAPOS
Volume 20 Number 2 / April 2016
Bregman and Donahue
Table 2A. 2013 Amblyopia risk factor (ARF) criteria of the American Association of Pediatric Ophthalmology and Strabismus Vision Screening Committeea Age-adjusted criteria for amblyopia risk factors Age, mos 12-30 31-48 .48 All ages
Anisometropia, D
Hyperopia, D
.2.5 .4.5 .2.0 .4.0 .1.5 .3.5 Manifest strabismus of .8 PD Media opacity of .1 mm
Astigmatism, D
Myopia, D
.2.0 .2.0 .1.5
. 3.5 . 3.0 . 1.5
D, diopters; PD, prism diopters. Updated criteria for automated preschool vision screening based on randomized controlled trials of amblyopia treatment and field studies of screening technology, with the goal of attaining high specificity for ARF detection in young children and high sensitivity for ARF detection in older children.
a
contact number provided on the consent form. A study subject was excluded for further recruitment if contact failed on three separate occasions.
Results During an 8-month period, 3,134 children aged 1272 months at 11 pediatric practices in Middle Tennessee were recruited and screened. Recruitment was restricted, so that no single practice screened more than 20% (n 5 627) of the total study population. Of the 3,134 children screened, 306 (9.8%) were referred for a formal eye examination, and 2,828 (90.2%) children had normal screening results. Table 3 shows the referral rate by practice and the relative distribution of the number of children screened per practice. The overall referral rate of 10% was fairly consistent across all practices; the only outlier was Southern Hills, with a referral rate of 24%. Southern Hills is located in a predominately Hispanic population, which has a higher prevalence of astigmatism and therefore an expectedly higher referral rate. The most common reasons for referral were as follows: suspected astigmatism (n 5 189 [61.8%]), suspected anisometropia (n 5 53 [17.3%]), and strabismus (n 5 39 [12.7%]). Referrals for suspected hyperopia (12 [3.9%]), myopia (8 [2.6%]), anisocoria (5 [1.6%]), and media opacity (0) were all low. See Table 4. Of the 306 referred children, 136 (44.4%) completed a gold standard examination, including cycloplegic retinoscopy. Demographics for this population are provided in Table 5. The age distribution of this subgroup (58%, 12-36 months; 42%, 37-72 months) reflected the entire referred population (61%, 12-36 months; 39%, 37-72 months). The primary reasons for referral for this subgroup do not differ significantly from the distribution of the total referred population (Table 4). The most common reason for referral in examined children was suspected astigmatism (n 5 67 [49.3%]), and a large percentage were due to suspected anisometropia (n 5 34 [25%]) or strabismus (n 5 18 [13.2%]). The numbers of children
Journal of AAPOS
155
Table 2B. 2003 AAPOS VSC amblyopia risk factor criteriaa ARF identified Anisometropia (spherical or cylindrical) Strabismus Hyperopia Myopia Media opacity Astigmatism Ptosis Visual acuity
Criteria, not age adjusted .1.5 D Any manifest strabismus .3.5 D in any meridian Magnitude .3.00 D in any meridian Any media opacity .1 mm in size .1.5 D at 90 or 180 .1.0 in an oblique axis (.10 eccentric to 90 or 180 ) #1 mm margin reflexb distance Per age-appropriate standards
D, diopter. a Original 2003 criteria for automated preschool vision screening that are referred to by the vast majority of currently published field studies of automated technology. b A standard objective measurement of ptosis equal to the distance from the corneal light reflex to the margin. Table 3. Referral rate by practice and the relative distribution of screenings by practice Practicea
No. children screened (% referred)
Nurture Tullahoma CCN Rivergate Hermitage Mt. Juliet CCE Lebanon Brentwood Capstone Southern Hills Lebanon
470 (11) 508 (7) 199 (8) 62 (10) 280 (10) 605 (8) 166 (12) 558 (11) 114 (12) 91 (24) 81 (7)
a
11 pediatric practices located in Middle Tennessee. See e-Supplement 1 at jaapos.org for formal list.
examined after referral for suspected hyperopia (n 5 9 [6.6%]), myopia (n 5 5 [3.7%]), anisocoria (n 5 3 [2.2%]) and media opacity (n 5 0) were all low. When the 2013 AAPOS-VSC guidelines were used as the gold standard, ARFs were detected in 64 of 136 patients for an overall positive predictive value (PPV) of 47%. The PPV specified by primary reason for referral is listed in Table 6. The PPV for suspected hyperopia was 77.8%; for myopia, 60%. The PPV for suspected anisometropia was 50%; for astigmatism, 44.8%. The PPV for suspected strabismus was only 33.3% and the PPV for suspected anisocoria was 33.3% (only 1 of 3 children examined for this reason had anisocoria confirmed on examination). In order to address previously published field studies, which refer to the 2003 AAPOS guidelines, these data were also analyzed using the older guidelines as gold standard. This analysis resulted in 82 patients with amblyopia risk factors (PPV of 60.3%). The PPV specified by primary reason for referral is listed in Table 6. PPVs were calculated for suspected hyperopia (77.8%), anisometropia (67.6%), astigmatism (61.2%), and myopia (60%). The PPVs for suspected strabismus and anisocoria remained unchanged.
156
Bregman and Donahue
Volume 20 Number 2 / April 2016
Table 4. Primary reason for Spot 2.1.4 referrala No. Primary reason for referral
All children referred N 5 306 (10% total study population)
Percent of population (with follow-up)
Astigmatism Anisometropia Strabismus Hyperopia Myopia Anisocoria Media opacity
189 53 39 12 8 5 0 Children with completed follow-up N 5 136 (44% referred population)
61.8 17.3 12.7 3.9 2.6 1.6 0
Astigmatism Anisometropia Strabismus Hyperopia Myopia Anisocoria Media opacity
67 34 18 9 5 3 0
49.3b 25.0b 13.2b 6.6b 3.7b 2.2b 0b
a Approximately 44% of children referred completed a follow-up appointment. Reasons for referral were comparable between the total referred population and the referred population with follow-up. b Percentage refers to fraction of referred population with follow-up rather than total referred population.
Table 5. Patient characteristics for referred population with completed follow-up
Characteristics Age group, mos 12-24 25-36 37-48 49-60 61-72 Sex Female Male Ethnicity White Hispanic African American Asian Middle Eastern Unknown Insurance Status Commercial Government Unknown
All patients referred by Spot 2.1.4 for a suspected amblyopia risk factor, no. (%) (n 5 136) 46 (33.8) 33 (24.3) 28 (20.6) 20 (14.7) 9 (6.6) 70 (51.5) 66 (48.5) 88 (64.7) 20 (14.7) 11 (8.1) 9 (6.6) 3 (2.2) 5 (3.7) 63 (46.3) 62 (45.6) 11 (8.1)
The data were also analyzed by patient age. Of the 136, 79 (58.1%) who were referred and completed follow-up were #36 months of age. The USPSTF guidelines state that evidence of photoscreening effectiveness in this age group is inconclusive.2 According to the 2013 AAPOSVSC criteria, the overall PPV for this younger age group was less than that in the older age group (38.0% vs
59.6%); however, according to the 2003 guidelines, the overall PPVs were comparable (60.8% vs 59.6%; Table 6). Of the 136 children who were referred and completed follow-up, 18 (13.2%) had amblyopia (a difference of 2 lines of best spectacle-corrected visual acuity or, for preverbal children, unmaintained fixation using fixation preference testing). Amblyopia was detected in 6 children 12-36 months old and 12 children 37-72 months old. Sixteen cases were due to anisometropic amblyopia; 2 cases, to strabismic amblyopia. Assuming that amblyopia was evenly distributed in the referred population who did not receive follow-up examinations, we estimated that the Spot referred 33 children with amblyopia. This equates to 1.0% of the screened population and is on the lower end of the range of estimates of amblyopia in the general pediatric population.17,18,19 To determine the sensitivity of the Spot vision screener in this study, parents of approximately 300 children who passed their screening examination were contacted to schedule a free follow-up, gold standard pediatric ophthalmic examination with cycloplegic retinoscopy. About 75 appointments were scheduled successfully; however, only 12 of the scheduled children appeared for an examination. All 12 of these children had normal ophthalmic examination results with no child who received a pass diagnosed as having amblyopia or an ARF. We considered contacting additional children to increase the sample of passed children, but decided that, with an examination rate of only 4% of contacted children, there was a potential for introducing significant selection bias in the study.
Discussion The present study demonstrates that the Spot v.2.1.4 photoscreener is an effective tool for detecting ARFs when used in general pediatrics clinic. Our overall referral rate (10%) and PPV (60.3%) replicate the outcome measurements from previously published field studies and key validation studies. In addition, over 60% of children referred were found to have ARFs and 13% had amblyopia detected on examination, both of which are significantly more than in the general population, which is thought to be 15%20% for ARFs and 1%-2% for amblyopia. The USPSTF and AAP endorse automated vision screening in the 3- to 5-year-olds. Both identify photoscreening as a valid modality with potential for use in the 6- to 36-month age range when other vision screening techniques are not age-appropriate.2,3 The former recommends “screening for all children at least once between the ages of 3 and 5 years” and lists photoscreening as one modality that is “feasible” for testing in the primary care setting. The AAP echoes the USPSTF recommendation for routine screening but further specifies that screening should be performed at an early age, at regular intervals, and “within the medical home.” The USPSTF provides a class “B” rating for vision screening in children aged 3-5 years and an “I”
Journal of AAPOS
Volume 20 Number 2 / April 2016
Bregman and Donahue
Table 6. Comparison of positive predictive value (PPV) by primary reason for referral according to 2013 and 2003 AAPOS VSC guidelines and by age groupa PPV: no. true positives (%) Primary reason for referral
No.
All age groups (n 5 136) Overall PPV 136 Astigmatism 67 Anisometropia 34 Strabismus 18 Hyperopia 9 Myopia 5 Anisocoria 3 Media opacity 0 Children aged 12-36 months (n 5 79) Overall PPV 79 Astigmatism 37 Anisometropia 19 Strabismus 12 Hyperopia 5 Myopia 3 Anisocoria 3 Media opacity 0 Children aged 37-72 months (n 5 57) Overall PPV 57 Astigmatism 30 Anisometropia 15 Strabismus 6 Hyperopia 4 Myopia 2 Anisocoria 0 Media opacity 0
By 2013 guidelines
By 2003 guidelines
64 (47.0) 30 (44.8) 17 (50) 6 (33.3) 7 (77.8) 3 (60.0) 1 (33.3) –
82 (60.3) 41 (61.2) 23 (67.6) 6 (33.3) 7 (77.8) 3 (60.0) 1 (33.3) –
30 (38.0) 15 (40.5) 7 (36.8) 3 (25.0) 4 (80.0) 1 (33.3) 1 (33.3) –
48 (60.8) 26 (70.3) 13 (68.4) 3 (25.0) 4 (80.0) 1 (33.3) 1 (33.3) –
34 (59.6) 16 (53.3) 10 (66.7) 3 (50.0) 3 (75.0) 2 (100.0) – –
34 (59.6) 16 (53.3) 10 (66.7) 3 (50.0) 3 (75.0) 2 (100.0) – –
a
Calculation of PPVs included only those study participants who had completed follow-up (44.4% of total referred population). PPVs according to the 2013 criteria resulted in lower numbers than when referencing the less stringent 2003 criteria.
rating for vision screening in children under 3 years in whom they feel the data are inconclusive. In our study population, according to the 2003 AAPOSVSC criteria, the PPV for children aged 12-36 months was comparable to that of the older age group (60.8% vs 59.6%). Evaluating the data with the more restrictive 2013 criteria resulted in a lower PPV for the younger age group (38.0% vs 59.6%), because these guidelines require higher magnitudes of refractive error to be considered ARFs at younger ages. Nevertheless, the advantage of early detection is that prompt treatment in this age group helps prevent amblyopia from becoming entrenched.4,6 Our results, along with those of previous studies, support the use of photoscreening for the younger age group12-15,20 and should help make the inconclusive rating obsolete. It has been previously demonstrated that at baseline, automated screening outperforms traditional methods in terms of sensitivity and specificity.8 Efforts to enhance both sensitivity and specificity of photoscreening are ongoing with new technology and software updates. High specificity is important for reducing health care costs and parental anxiety as well as increasing confidence in the screening devices by physicians and other users. Sensitivity
Journal of AAPOS
157
is not as crucial with a single screening, because vision screening should be viewed as a continuous process throughout childhood, with multiple opportunities to detect pathology. Consequences of false positives include unnecessary follow-up appointments. Our false positive rate was 53% according to 2013 guidelines and 40% according to 2003 guidelines, with the largest proportion of false positives due to suspected strabismus (PPV of 33.3% for both). This is not surprising, because the photoscreener can easily “mistake” head tilting or distraction for gaze deviation. These results highlight the need for redefining the current algorithm used to detect strabismus as well as optimizing sensitivity and specificity to enhance the cost effectiveness and utility of photoscreening.12 A significant limitation to our study was the poor response rate for children who passed and were offered a free, comprehensive eye examination. There are a number of possible explanations for the poor response; however, it is unlikely that a large study can be performed in the future without financial compensation to encourage parents to have their children with “normal” vision examined. Having a study coordinator in charge of scheduling appointments and ensuring attendance, as in the Iowa MTI,21 Alaska Blind Child Discovery,22 and Tennessee Lions Outreach studies, would almost certainly have improved our follow-up rates. Regardless, because the number of follow-ups was small, sensitivity could not be adequately assessed. We felt that repeated attempts to recruit children who passed the photoscreening would bias the results by including children possibly more likely to have pathology. Our compliance rate for follow-up among children who received a refer was 44%. Without optimal compliance, there is a potential for selection bias; however, the distribution of pathology among children with completed followup did not differ significantly from that of all children referred (Table 4). This finding is therefore reassuring that our sample of children with completed follow-up is representative of the total referred population. An additional limitation of our study was the use of doctors with variable levels of training (ophthalmologists and optometrists) to complete the follow-up examinations. We chose to include multiple community physicians to maximize appointment slots and optimize follow-up rates. This aspect of the study design could be improved in follow-up studies by having all examinations completed by a single ophthalmologist to increase the accuracy of the examination. Lastly, we were unable to draw statistically strong conclusions regarding some parameters because of the small number of children referred for suspected myopia, hyperopia, anisocoria, or media opacity. Although we were not surprised by the lower numbers of referred children with less common pathology, the data and our conclusions appropriately reflect the ARFs that are more prevalent in the general pediatric population. Photoscreening is a single cost-effective test, can be administered quickly by nonmedical personnel, can be
158
Bregman and Donahue
used in many environments, and does not require prolonged patient cooperation. Prior to this study, photoscreeners had not been validated in the general pediatrics clinic and therefore may not have been widely incorporated as standard practice at well-child examinations. Our results support ongoing efforts to introduce and standardize the use of automated screening practices in the medical home. Additional, prospective, studies should clarify circumstances when intervention should be undertaken in very young children.
Acknowledgments The authors thank Sandy Owings, COA. and Janet Cates. References 1. Gunton K. Advances in amblyopia: what we have we learned from PEDIG trials? Pediatrics 2013;131:540-47. 2. US Preventive Services Task Force. Vision screening for children 1 to 5 years of age: US preventive services task force recommendation statement. Pediatrics 2011;127:340-46. 3. American Academy of Pediatrics Section on Ophthalmology, Committee on Practice and Ambulatory Medicine, American Academy of Ophthalmology, American Association for Pediatric Ophthalmology and Strabismus, American Association of Certified Orthoptists. Instrument-based pediatric vision screening policy statement. Pediatrics 2012;130:983-6. 4. Donahue SP. Relationship between anisometropia, patient age, and the development of amblyopia. Am J Ophthalmol 2006;142:132-40. 5. Arnold RW. Amblyopia risk factor prevalence. J Pediatr Ophthalmol Strabismus 2013;50:213-17. 6. Holmes JM, Lazar EL, Melia BM, et al., Pediatric Eye Disease Investigator Group. Effect of age on response to amblyopia treatment in children. Arch Ophthalmol 2011;129:1451-7. 7. Donahue SP, Arthur B, Neely DE, Arnold RW, Silbert D, Ruben JB. AAPOS Vision Screening Committee. Guidelines for automated preschool vision screening: a 10-year, evidence-based update. J AAPOS 2013;17:4-8. 8. Salcido AA, Bradley J, Donahue SP. Predictive value of photoscreening and traditional screening of preschool children. J AAPOS 2005;9: 114-20.
Volume 20 Number 2 / April 2016 9. Matta NS, Singman EL, Silbert DI. Performance of the Plusoptix vision screener for the detection of amblyopia risk factors in children. J AAPOS 2008;12:490-92. 10. Peterseim MM, Papa CE, Wilson ME, et al. The effectiveness of the Spot Vision Screener in detecting amblyopia risk factors. J AAPOS 2014;18:539-42. 11. Garry GA, Donahue SP. Validation of Spot screening device for amblyopia risk factors. J AAPOS 2014;18:476-80. 12. Arnold RW, Tulip D, McArthur E, et al. Predictive value from pediatrician Plusoptix screening: Impact of refraction and binocular alignment. Binoc Vis Strabolog Q Simms Romano 2012;27:227-32. 13. Lowry EA, Wang W, Nyong’o O. Objective vision screening in 3year-old children at a multispeciality practice. J AAPOS 2015;19: 16-20. 14. Halegoua J, Schwartz RH. Vision photoscreening of infants and young children in a primary care pediatric office: can it identify asymptomatic treatable amblyopic risk factors? Clin Pediatr 2015; 54:33-9. 15. Longmuir SQ, Boese EA, Pfeifer W, et al. Practical community photoscreening in very young children. Pediatrics 2013;131:764-9. 16. Donahue SP, Arnold RW, Ruben JB, AAPOS Vision Screening Committee. Preschool vision screening: what should we be detecting and how should we report it? Uniform guidelines for reporting results of preschool vision screening studies. J AAPOS 2003;7:314-16. 17. Cotter SA, Varma R, Tarczy-Hornoch K, et al. Risk factors associated with childhood strabismus. Ophthalmology 2011;118:2251-61. 18. Multi-ethnic Pediatric Eye Disease Study Group. Prevalence of amblyopia and strabismus in African American and Hispanic children ages 6 to 72 months. Ophthalmology 2008;115:1229-36. 19. Cotter S, Cyert L, Miller J, Quinn GE, National Expert Panel to the National Center for Children’s Vision and Eye Health. Vision screening for children 36 to \72 months: recommended practices. Optom Vis Sci 2015;92:6-16. 20. McKean-Cowdin R, Cotter SA, Tarczy-Hornoch K, Multi-Ethnic Pediatric Eye Disease Study Group. Prevalence of amblyopia or strabismus in Asian and Non-Hispanic white preschool children. Ophthalmology 2013;120:2117-24. 21. Longmuir SQ, Pfeifer W, Leon A, Olson RJ, Short L, Scott WE. Nine-year results of a volunteer lay network photoscreening program of 147,809 children using a photoscreener in Iowa. Ophthalmology 2010;117:1869-75. 22. Arnold RW, Donahue SP. The yield and challenges of charitable state-wide photoscreening. Binocul Vis Strabismus Q 2006;21: 93-100.
Journal of AAPOS