PRESCHOOL VISION SCREENING IN PEDIATRIC PRACTICE: A STUDY FROM THE PEDIATRIC RESEARCH IN OFFICE SETTINGS (PROS) NETWORK

Richard C. Wasserman, MD, MPH [*,***]; Candace A. Croft, PhD [*]; and Sarah E. Brotherton, PhD [*]

------------------------------------------------------------------------------

ABSTRACT. In this cross-sectional study, the vision-screening process is described for 8417 children aged 3 to 5 seen for health supervision in a group of 102 pediatric practices in 23 states and Puerto Rico. Three hundred forty children who failed screening (63% of those who failed) were followed up 2 months after initial screening. The sample was 52% male, 86% white, 9% black, 3% Hispanic, and 1% Asian. Vision screening was attempted on 66% of children overall. Pediatricians' reasons for not screening were "not routine" (44%), "too young" (40%), and "screening done previously" (17%). Younger children were less likely to be screened than older children (39% of those aged 3), and Hispanics were less likely to be screened than other ethnic groups (P <.001). Thirty-three percent of children received no screening for latent strabismus. Two months later, 50% of parents whose child had failed a vision test were unaware of this fact on questionnaire follow-up. Eighty-five percent of children referred to an eye specialist had made or kept an appointment. It is concluded that pediatricians need to increase vision screening among younger preschool children and communicate more effectively to parents the results of screening failure. Pediatrics 1992;89:834-838; vision, visual acuity, strabismus, binocularity, amblyopia, screening, preschool children, health supervision.

------------------------------------------------------------------------------

ABBREVIATIONS. AAP, American Academy of Pediatrics; PROS, Pediatric Researchin Office Settings.

------------------------------------------------------------------------------

Undetected vision problems are common in preschool children, with an estimated prevalence of 5% to 10%. [1] Vision problems in preschoolers can lead to a variety of adverse consequences. The most serious of these is amblyopia, a loss of visual acuity and binocular vision that becomes irreversible after age 5. The technology to detect and treat vision problems in this age group is widely available. Earlier treatment of patients who either have or are at risk for amblyopia is widely accepted as leading to improved outcomes. [2,3] Since the early identification and referral of vision problems is the key to preventing serious consequences, a variety of expert groups, including the US Preventive Services Task Force [4] and the American Academy of Pediatrics (AAP) Committee on Practice and Ambulatory Medicine, [5] recommend that all preschool children be screened for visual acuity and strabismus by age 4 as a part of regular preventive care.

Published reports indicate that half of all children with amblyopia are diagnosed after age 5, when therapy is no longer effective. [6] This may be partially due to underscreening. Two studies have estimated that only 25% of the nation's preschoolers receive any kind of vision testing. [7,8] A more recent report suggests, however, that as many as one third of children who receive a late diagnosis of amblyopia have actually been screened as preschoolers. [9] This failure of prevention could indicate a problem either with screening test accuracy (eg, a high false-negative rate) or with postscreening referral and treatment.

Very little is known about the performance of vision screening among pediatricians, who provide approximately half of all preventive care delivered to preschool children. [10] Identification of difficulties and pitfalls in the pediatric vision-screening process could lead to more effective early identification and treatment of preschool vision problems. To expand knowledge of current vision-screening practices, the AAP's Pediatric Research in Office Settings (PROS) Network undertook a study to (1) describe the vision-screening process for children aged 3 through 5 years being seen for health supervision visits in a group of pediatric practices, (2) identify factors associated with attempted and successful screening, and (3) determine what happens to children who fail vision screening in pediatric practice.

METHODS

The PROS Network was established in 1986 to study primary care pediatric problems in practice settings. Practices were recruited by volunteer pediatricians (chapter coordinators; see Appendix) from AAP chapters around the United States. In 1988, the PROS Network consisted of 356 self-selected pediatricians in 102 practices situated in 23 states and the Commonwealth of Puerto Rico. The practices that participated in the study are listed in the Appendix.

To assess the degree to which PROS practitioners corresponded to the AAP membership, comparisons were made between PROS pediatricians and groups of randomly selected AAP members who responded to the AAP Research Department's Periodic Surveys of Fellows in 1988 and 1989. [11] The Periodic Surveys included a variety of questions about members' demographic and practice characteristics, as well as specific questions about vision-screening practices. The same questions were asked of the participating PROS practitioners. PROS pediatricians were younger (by an average of 1 year) and more likely to practice in groups (55% vs 47%) than the practitioners responding to the Periodic Survey. In each group, 90% reported performing routine vision screening for at least some health supervision visits.

The study time frame was a 12-week period from May to July of 1988. All children aged 3 through 5 who were being seen for scheduled health supervision visits were to be enrolled in the study, regardless of whether vision screening was to be a part of that health supervision visit. Vision screening was defined as the testing of visual acuity and/or binocular (stereoscopic) vision performed apart from the physical examination.

A vision-screening attempt was defined as 10 or more seconds spent trying to get the child to cooperate with vision screening. Criteria for pass and fail of vision-screening tests at different ages were taken from the AAP Committee on Practice and Ambulatory Medicine guidelines. [5] A result was judged uninterpretable if the screener made an attempt to screen a child but was unable to reliably assign a pass or fail result to the child's performance (ie, child uncooperative or results ambiguous).

At the visit, data were collected on a questionnaire filled out by parents and on a data card completed by pediatricians and their office staffs. Two months later, a follow-up questionnaire was mailed to the parents of children who had failed the vision screen, failed the cover/uncover test on physical examination, or had been referred to an eye specialist.

The parent questionnaire elicited demographic information, the child's past medical history, and the family history of vision problems. The pediatric office data card elicited history and physical examination data relevant to vision; whether, how, and why vision screening was performed (or not performed); the results obtained from screening; and the disposition. The follow-up questionnaire addressed the parents' understanding of the vision-screening results and actions that had been taken by the parents. Parent questionnaires in Spanish were available for those parents who wanted them.

Chapter coordinators in each state trained the practitioners and office staffs for each practice using a study protocol and training scenarios developed by the investigators. Data collected at each practice site were mailed weekly to the PROS Network office at the AAP. Subjects were assigned unique identification numbers by each practice to preserve subject anonymity while allowing for follow-up. Subjects meeting the criteria for follow-up were identified for each practice by the PROS Network office. Follow-up questionnaires for those subjects (including postage-paid return envelopes) were sent to each practice, which in turn mailed them to the subjects' families for completion.

To estimate the completeness of subject enrollment, practices were asked to keep a log of all 3- through 5-year-old patients seen for health supervision during the study time frame (whether or not they were entered as subjects). Analysis of practice logs revealed that, on average, 95.9% of eligible subjects were enrolled per practice.

Data analyses (1) described the study sample and the types of vision screening that were performed, (2) characterized the children on whom screening had been attempted and successfully completed, (3) identified factors associated with screening, and (4) identified what had happened to children who had failed vision screening at the health supervision visit. Descriptive statistics, Chi^2, t tests, and Pearson's correlation coefficients were calculated using SPSS-PC. All statistical tests of significance were two-tailed.

RESULTS

Sample

Pediatricians completed data cards on a total of 8417 children. The parents of 7754 of these children (92%) completed the initial study questionnaire, and these children comprise the study sample. The sample is described in Table 1. A disproportionate number of Hispanic parents failed to complete the study questionnaire, and therefore Hispanic children were underrepresented in the study sample.

Reliability

Data cards on a random 10% of subjects were sent back to pediatricians several months after the initial data collection to assess reliability on items that might be determined retrospectively from medical records. Results revealed item agreement percentages in the 83% to 97% range for demographic variables, vision-related history, whether or not vision screening had been attempted, and screening results.

Testing Patterns

According to the parents, 63% of the children had never had vision testing and 12% had already seen an eye specialist (ophthalmologist or optometrist). Vision screening was carried out on 66% of the study sample, including only 64% of those who had never been tested previously. Thirty-four percent of the sample were screened for visual acuity only; 31% were screened for both acuity and binocular vision. The most commonly identified reasons given by pediatricians when vision screening was not attempted (more than one response allowed) were "not routine" (44%), "child too young" (40%), "screening already done" (17%), and "screening will be done elsewhere" (9%).

Table 2 shows the proportion of children on whom screening was attempted by the age of the child. Notably, vision screening was not attempted on more than 60% of the 3-year-old children. Pediatricians' mst common reasons for not attempting vision screening in 3-year-old children were "child too young" (58%) and "not routine" (52%). In general, the younger the child, the less likely that vision screening had been attempted. Table 3 shows attempts at vision screening according to the child's ethnic background. Screening was attempted less frequently for Hispanics than for any other ethnic group. None of the other between-group ethnic differences seen in Table 3 were statistically significant. Analyses of the Hispanic group alone revealed that the lower rate of screening attempts was not explained by an excess of younger children in this group.

Many of the children who had been screened for visual acuity were not screened for binocular vision. We therefore performed analyses to identify whether the cover/uncover test, the physical examination maneuver that detects latent strabismus and therefore also addresses binocular vision, had been performed on these children. In fact, 33% of the study sa mple had neither thecover/uncover test nor binocular vision screening performed.

There were no associations between characteristics of the pediatric practice (solo vs. group) and screening patterns. Data on source of payment for the visit were not collected, and therefore this could not be investigatedas a factor in screening.

Interpretability: Age and Test Factors

Table 4 shows the percentages of children in each group whose vision-screening tests yielded interpretable results for acuity and binocular vision. Not unexpectedly, 3-year-old children were considerably less likely to have interpretable results than those 4 or 5 years old. Even among 3-year-old children, however, a full 70% of results were interpretable.

Table 5 shows the types of screening methods that pediatricians used in screening for acuity and binocular vision. To see whether particular screening methods were more likely to yield interpretable results among the youngest (therefore hardest to screen) children, we examined the interpretability of results according to the screening method for 3-year-old children alone. Results of this analysis for acuity screening can be seen in Table 6. Three-year-old children screened with hand-held cards (eg, Allen Cards) had a higher rate of interpretable results than those tested with screening machines or wall charts. This was not the case among 4- and 5-year-old children. In a similar analysis of methods of binocular vision screening, a trend toward higher success rates with stereograms than with screening machines was seen among 3-year-old children, but the numbers were too small for statistical significance.

Abnormalities: Detection and Disposition

Overall, 8% of those screened failed acuity testing, and 5% failed binocular vision testing. An additional 1% of the sample failed only the cover/uncover test on the physical examination. These figures are consistent with previous reports on failure rates for vision-screening tests. [12] Disposition for those who failed a vision-screening test was as follows: 21% were newly referred to an eye specialist, 57% were told to return to the pediatrician's office for repeated screening (typically in 12 months), 5% were referred back to their eye specialist, and 15% were told to recheck as needed. The duration of time specified for patients who were told to return to the office for repeated screening did not vary by the age of the patient (ie, 4- and 5-year-old children who failed screening were no less likely to be told to return in 12 months than 3-year-old children).

The follow-up questionnaire was mailed to the 538 families whose child had failed one or both vision-screening tests, had failed the cover/uncover test only, or had been referred to an eye specialist. Of these, 340 (63%) responded. Respondents had higher levels of maternal education, were more likely to be white, and were less likely to have a 3-year-old child than nonrespondents.

Among those children who, according to the office data card, had failed a vision test, 50% of the parents claimed to be unaware of the fact. Nine percent were unaware or unsure that the test had been done, and 41% thought that the results were normal. Another 9% of parents were aware of the result and had plans to make a future follow-up appointment, and 33% had already made or kept a follow-up appointment. Notably, of those families who had been referred to an eye specialist, 85% had either made or kept an appointment.

Parents' lack of knowledge of vision test results was not associated with their ethnicity or level of education.

DISCUSSION

The description of the preschool vision-screening process that emerges from this study suggests several possible reasons why amblyopia is diagnosed too late for effective therapy in a substantial number of school-age children. These reflect problems of underscreening by pediatricians as well as difficulties among those who are actually screened.

Underscreening is clearly a problem. In more than one third of the sample, and in nearly two thirds of the 3-year-old children, no vision screening was attempted. Some of the children cited as "too young" may, in fact, have been uncooperative. For more than half of the 3-year-old children for whom no attempt was made, however, the reason cited was "not routine." And although some children had been screened previously and others would probably be screened elsewhere, the results clearly indicate a missed opportunity for prevention in the comprehensive care setting. We are especially concerned about the failure to screen younger children, since they are the ones most likely to benefit both from screening and from interventions to prevent amblyopia. Also with respect to preventing amblyopia, it is distressing that one third of children in all age groups had received no testing to detect either latent strabismus (cover/uncover test on physical examination) or problems with binocular vision. The lower rates of screening found among Hispanic children are disturbing as well. This finding may reflect language difficulties between the preschoolers and those administering the tests.

The results offer some insight as to why pediatricians may attempt vision screening less frequently with younger children. Tests performed on 3-year-old children were substantially less likely to yield interpretable results than those done on older children. The good news is that, even among 3-year-old children, 70% of screening tests were interpretable. Although it is unlikely that attempts at screening all 3-year-old children would yield equally high rates of interpretability, the results suggest that attempts at screening this younger age group are probably worthwhile. In addition, among 3-year-old children, hand-held cards (such as Allen Cards) appeared more likely to yield interpretable results than other methods of acuity testing. This result needs to be confirmed prospectively in a controlled trial.

Several factors related to this study's design should be considered in interpreting the results on screening patterns. First, PROS practices are not representative of pediatric practices in general, and they may actually constitute a group of practitioners more adherent to official recommendations than other pediatricians. Second, although PROS practitioners were told specifically not to alter their usual vision-screening practices for the purposes of the study, we cannot guarantee that alterations did not occur. Third, we relied entirely on practitioner self-report for data about completeness of enrollment. In assessing the potential biases introduced by these design factors, it is likely that the study actually overestimates the frequency of vision screening in pediatric practice and, therefore, underestimates te magnitude of the problem.

In addition to the problem of underscreening, the results suggest that all may not be well among those who are being screened. Our study design does not permit an answer to the question of screening test accuracy. Screening test inaccuracy might explain why some children who have had their vision screened might still receive a late diagnosis of amblyopia. [9] The study results do, however, suggest another possible explanation for this phenomenon. Difficulties are evident in the plans made for follow-up and with the communication of results to parents among those children who fail vision screening.

The follow-up strategies used for some of the patients in this study may be inadequate. Fifty-seven percent of those who failed vision testing were told to return to the office for repeated screening. Although this is a very sensible approach, the strategy of waiting 12 months to rescreen (recommended for 75% of this group) is probably inadvisable for anyone older than age 3, and yet it was uniformly employed for all age groups. Rechecking on an as needed basis, recommended for 15% of those who failed, makes little sense from a preventive standpoint.

Communication of results about vision-screening failure and the need for follow-up appears less than ideal. Half of parents whose child had failed a vision test claimed to be unaware of the results. The reason for this miscommunication is not clear, but this 50% figure is not inconsistent with the literature on the communication of diagnostic information and therapeutic instructions to patients. [13] Perhaps the parents simply did not understand, or pediatricians, wishing to minimize parental anxiety about an abnormal result, failed to emphasize the results. In any case, lack of information about an abnormal test precludes effective follow-up. There is, however, reason for encouragement. Among patients who understood that their child had failed vision screening and who had been referred to an eye specialist, 85% claimed to have made or kept an appointment within 2 months of the visit. This suggest that parents are able to follow through when pediatricians successfully communicate the seriousness of the problem.

Screening is basic to health supervision. Among all screening efforts recommended for children younger than age 6, vision screening stands out as a worthwhile procedure because of its demonstrated effectiveness in preventing a permanently disabling condition. Results of the PROS vision-screening study indicate room for improvement in pediatric vision-screening practices.

We recommend that pediatricians inform themselves with the latest information about vision-screening techniques [14] and screen their preschool patients for vision problems more completely, making special efforts to

include the younger children who are most likely to benefit from earlier intervention and to perform tests that address binocular vision and strabismus. Pediatricians need to find ways to communicate the results of vision-screening failure more effectively to parents, perhaps by employing written recommendations. They need to ensure the follow-up of patients with uninterpretable or abnormal results, possibly by scheduling a subsequent visit for vision testing in their offices at the time of vision-screening failure. All of these steps will help to ensure the identification of children at risk for amblyopia in sufficient time to treat them.

Finally, researchers must develop new vision-screening technologies for practitioners to employ. New techniques such as photorefraction [15] may improve our accuracy in detecting young children at risk for amblyopia. These new technologies must then be field tested in office settings and compared with present methods of screening to identify the best way to perform preschool vision screening in pediatric practice.

ACKNOWLEDGMENTS

This work was supported by grant MCJ-173400 from the US Maternal and Child Health Bureau. We thank the PROS practitioners and their office staffs for their efforts in data collection; the PROS chapter coordinators for contributions to the design of the study and data collection materials and for regional oversight of data collection; Evan Charney, MD, and Gretchen Fleming, PhD, for their roles in the establishment of the PROS Network and the management of this project; Karen O'Connor for the Periodic Survey; Cynthia Hasemeier for data analysis; and Katherine Kaufer Christoffel, MD, MPH, and the members of the PROS Steering Committee for manuscript review.

APPENDIX

The PROS chapter coordinators at the time of the study were as follows: Arizona, Burris Duncan, MD; California, Robert Pantell, MD; Colorado, Jules Amer, MD; Connecticut, Frederick Berrien, MD; Florida, Lorna Katz, MD; Georgia, Edward Gotlieb, MD; Illinois, Katherine Kaufer Christoffel, MD; Indiana, Virginia Wagner, MD; Maryland, John Straus, MD; Michigan, Sheldon Brenner, DO; New Jersey, Harris Lilienfeld, MD; New York II, Thomas Marr, MD; North Carolina, George Prince, MD; North Dakota, Gerald Atwood, MD; Ohio, Richard Simpson, MD; Pennsylvania, Kenneth Rogers, MD; Puerto Rico, Carlos Bourdony, MD; Rhode Island, Anthony Alario, MD; South Carolina, Howard Trent III, MD; Tennessee, Charles Fish, MD; Uniformed Services West, Jose Pascual, MD; Utah/Idaho, Gordon Glade, MD; Vermont, Richard Wasserman, MD; Virginia, Harry Gewanter, MD; Wyoming, Gary Lang, MD.

The pediatric practices or individual practitioners who participated in this study are listed here by AAP chapter: Arizona, Kachina Pediatrics, Ltd (Tucson), Mesa Pediatrics Prof Assoc (Mesa); California, Harvey, Ball, & Sharron (Palo Alto), Valley Pediatric Medical Group, Inc (Encino); Colorado, Arvada Pediatric Associates (Arvada); Connecticut, Dr Gerald Calnen (Enfield), Children's Medical Group (Bloomfield), Guilford Pediatrics (Guilford), Dr. Barry Keller (Danbury), St. Francis Pediatric Primary Care Center (Hartford); Florida, Lorne Katz, MD (Coral Springs), Altamonte Pediatric Associates (Altamonte Springs); Georgia, The Pediatric Center (Stone Mountain), Northside Pediatrics and Adolescent Medicine, PC (Atlanta), Kravtin & Levy, MDs, PC (Columbus), Children's Medicine, PC (Lilburn); Illinois, DuPage Pediatrics (Darlen), Emalee G. Flaherty, MD (Chicago), Kamala A. Ghaey, MD (Chicago), Pediatrics (Arlington Heights), Children's Memorial Primary Care (Chicago); Idaho, Pediatric Center (Twin Falls), Pocatello Children & Adolescent Clinic (Pocatello); Indiana Patricia Becker, MD (St Vincent Pediatric Clinic, Indianapolis), Jonathan Evans, MD (MetroHealth, Indianapolis), Claudia J. Somes, MD (Indianapolis), MetroHealth Lafayette Square (Indianapolis); Maryland, Green-spring Pediatric Assoc (Baltimore), Daniel Levy, MD (Owings Mills), Clinical Associates, PA (Towson), Valley Pediatric Assoc, PA (Owings Mills); Michigan, General Pediatric Services (Detroit), Pedicenter Associates, PC (Detroit), Children's Health Care of Port Huron, PC (Port Huron); New Jersey, Delaware Valley Pediatric Associates, PA (Lawrenceville), Cinnaminson Pediatric Associates, PA (Cinnaminson), Curtis, Factor, Tokar, Zablocki, Rubin (Highland Park), Bina Kapila, MD (East Orange), Richard Lander, MD, PA (Livingston), Barry Kessler, MD, PA (Cardiff); New York II, S. M. Copperman (Merrick); North Carolina, Eastover Pediatrics, PA (Charlotte), Michael J. Grode (Charlotte), Valley Pediatrics, PA (Fayetteville), Goldsboro Pediatrics, PA (Goldsboro), Fleming, Edwards, Goldman, Carr, & Lehan (Raleigh), Gastonia Children's Clinic (Gastonia); North Dakota, Fargo Clinic-Pediatrics (Fargo), Dakota Clinic, Ltd (Fargo), Medical Arts Clinic (Minot), Grand Forks Clinic (Grand Forks); Ohio, Bryan Medical Group (Bryan); Pennsylvania, Northland Pediatric Association (Pittsburgh), James G. Pitcavage, MD (Sewickley), Rod Hartman, MD (Pittsburgh), Leonard D. Leibowitz, MD (Monroeville), Mindy F. Rosenblum, MD (Bala Cynwyd), Episcopal Pediatric Clinic (Philadelphia), Reading Pediatrics, Inc (Wyomissing), David Sands, MD (Kingston), Pediatric and Adolescent Medicine, PC (Pittsburgh), Paul David Francis, MD (Danville), P. J. Bruno, MD (Sunbury); Puerto Rico, Felisa Santiago, MD (Paradise Hills); Rhode Island, Barrington Pediatric Associates (East Providence), Marvin Wasser, MD (Cranston); South Carolina, Anderson Pediatric Group (Anderson), Children's Clinic, PA (Spartanburg), Greenwood Pediatrics (Greenwood), J. Rutledge Lawson, MD (Greenville), Mount Pleasant Pediatrics (Mount Pleasant), Palmetto Pediatric and Adolescent Clinic (Columbia), Pediatrics (Rock Hill), Sandhills Pediatric Clinic (Columbia); Tennessee, Johnson City Pediatrics, PC (Johnson City), Jackson Clinic (Jackson); Uniform Services West, Department of Pediatrics, Wilford Hall USAF Medical Center (San Antonio); Utah, Gordon Glade, MD (American Fork), David Nuthall, MD (American Fork), Granger Medical Clinic (West Valley City), Salt Lake Clinic ( Sandy); Vermont, Sarayu Balu, MD (Morrisville), H. Taylor Yates, Jr (St Albans), Essex Pediatrics (Essex Junction), Newport Pediatrics (Newport), Practitioners of Pediatric Medicine (South Burlington), Uiversity Pediatrics (Burlington); Virginia, Danville Pediatric Assoc, Inc (Danville), Pediatric Associates of Richmond, Inc (Richmond), Physicians to Children, Inc (Roanoke), Lewis Gale Children's Clinic (Salem), Virginia Pediatrics & Adolescent Center (Springfield), Division of Ambulatory and Emergency Care (Richmond), James River Pediatrics (Midlothian); Wyoming, Gary Lang, MD (Lander), Jackson Pediatrics (Jackson).

REFERENCES

1. Brown MS. Vision screening of preschool children: how to check on visual acuity and heterophoria as part of a routine physical examination. Clin Pediatr (Phila). 1975;14:968-973

2. Simons K, Reinecke RD. A reconsideration of amblyopia screening and stereopsis. Am J Ophthalmol. 1974;78:707-714

3. Lang J; Cibis GW, trans. Strabismus. Thorofare, NJ: SLACK Inc; 1984:141

4. US Preventive Services Task Force. Guide to Clinical Preventive Services: An Assessment of the Effectiveness of 169 Interventions. Baltimore, MD: Williams & Wilkins; 1989

5. Committee on Practice and Ambulatory Medicine, American Academy of Pediatrics. Vision screening and eye examination in children. Pediatrics. 1986;77:918-919

6. Pollard ZF, Manley DM. Long term results in treatment of unilateral high myop

Copyright (c) American Academy of Pediatrics 1992