Abstract
Purpose: :
To compare results of screening preschool children using two different approaches to administration of visual acuity (VA) screening with Lea Symbols.
Methods: :
As part of the Vision In Preschoolers Study, two different formats for presentation of Lea Symbols were used: 1) a linear array of Lea Symbols (1 per acuity level) surrounded by a crowding rectangle, presented on cards at 10 feet; and 2) single Lea Symbols surrounded by four crowding bars, presented on a disk at 5 feet. Both tests included an identical pretest to establish testability. Screeners were either Licensed Eye Professionals (LEPs), Nurse Screeners (NS), or Lay Screeners (LS). All screeners were trained and certified within VIP. Overall, more than 4,500 children age 3 to 5 years were screened with at least one of the 5 approaches of test format and screener. Accuracy of the screening test results was assessed by comparison to results of comprehensive eye examinations conducted by examiners masked to screening results. Sensitivity to detect children with ≥1 targeted condition at 90% specificity was used as the primary outcome measure; sensitivity to detect the most serious conditions (Group 1) was also assessed. Tests of significance were performed by chi–square tests of two independent proportions (different children tested) and two correlated proportions (the same children tested twice).
Results: :
: Child testability was 99% for both test formats by all screeners. With specificity set at 90%, sensitivity was significantly better for the single symbol format than for the linear symbols format for detection of one or more VIP targeted conditions for LS (61% vs. 37%; p<0.0001) but not for NS (52% vs. 49%; p=0.53). Further, with specificity set at 90%, sensitivity for the Group 1 conditions was significantly better for the single symbol format than for linear symbols format for both LS (80% vs. 50%; p<0.0001) and NS (82% vs. 60%; p=0.0009). LS and NS administration of the single symbol format yielded sensitivity similar to that achieved by LEP screeners (61% for ≥ 1 condition, 77% for Group 1 conditions) using the linear symbols format.
Conclusions: :
When single crowded symbols are used with standardized testing procedures at 5 feet instead of at 10 feet, VA test sensitivity improves substantially for lay screeners and achieves sensitivity equivalent to that achieved by LEPs with linear format.
Keywords: visual acuity • screening for ambylopia and strabismus • clinical (human) or epidemiologic studies: systems/equipment/techniques