Abstract
Purpose :
Screening for vision disorders in young children has proven difficult, especially for lay screeners in early education settings. This study evaluated the efficacy of a novel visual developmental (VD) and risk assessment (RA) questionnaire (Q) for this purpose. Age specific questions were developed for children by 6-month age intervals. The VDQ varied by age across the Q’s and the RAQ were uniform for all ages.
Methods :
223 children between ages 3-36 months were recruited from early education centers in Boston and Springfield MA. Parents/primary caregivers completed an age appropriate Q. Masked examiners performed comprehensive eye exam with cycloplegic retinoscopy on all children with completed Q’s. Item analysis and kappa analysis were performed on individual questions in each Q to compare levels agreement with the results of the gold standard eye exam.
Results :
Among the VDQs, kappa coefficients showed moderate agreement for one question in the 13-18 (0.523 (0.074-0.971)) and 25-30 (0.456 (0.130-0.783)) month cohorts. All five RAQs had low kappa coefficients implying poor agreement with the eye exam. However, the observed agreement of two specific RAQ’s were 79.4% and 74.4% indicating a higher than chance agreement with the results of the eye exam despite having low kappa coefficients (0.167 (0.011-0.324) and 0.156 (0.008-0.306) respectively). In addition, the odds ratio calculation for these two RAQs (0.243 (0.100-0.590, p<0.002) and 0.329 (0.152-0.7136, p<0.005) respectively) was also statistically significant. One of these questions evaluated if there was a family history of strabismus and the other evaluated if there was a family history of refractive correction during early childhood.
Conclusions :
Item analyses identified questions that showed better agreement with the results of the gold standard eye exam than others, implying that a subset of questions and the Q in general is a promising tool for vision screening in this difficult to screen population. In addition, when pooling the responses of all questions across age groups, the Q showed moderate sensitivity and specificity in detection of amblyogenic risk factors (69.2% and 65.8% respectively) with an area under the curve of 0.703. Large-scale studies are needed to evaluate the validity of this novel and promising method of detecting vision problems in young children.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.