The development of a screener for Cerebral Visual Impairment.

Abstract

This study explored the secondary use of Luxembourg's school monitoring tool for a large-scale screening of Cerebral Visual Impairment (CVI)-related difficulties. 44 items, with and without time constraint, were developed, and pretested among 959 children. All children subsequently participated in an individual evaluation of higher-level visual processing (HLVP) measures related with CVI. A clinical outcome was attributed post hoc with 32 children being classified as having CVI-related difficulties. To explore the predictive power of the CVI items included in the monitoring, item responses were matched to the results of the individual HLVP assessment. Of all items, the untimed item targeting the combined functions of surface and rotation significantly distinguished group performances (<.05). To improve condition discrimination, different item combinations were tested. Sensitivity and specificity metrics were computed resulting in ranges of 37.5% - 81.3% and 27% - 88.8% respectively. The item combination with the highest sensitivity (81.3%) was retained considering a viable trade-off between sensitivity and specificity metrics. These results support the secondary use of an existing large-scale monitoring tool to screen for CVI-related difficulties in the beginning of elementary school, provided that additional sources of information are progressively implemented to strengthen the tool's predictive power.