Abstract
OBJECTIVE: To assess the accuracy of the ten questions screen as a measure of childhood disability for epidemiologic studies in populations lacking resources for professional assessment of children's development and functioning. DESIGN: Household survey and screening of children in phase one followed by clinical assessments in phase two. SETTING: Karachi, Pakistan. PARTICIPANTS: A cluster sample of 6365 children, aged 2 to 9 years, screened using the ten questions and a subsample referred for clinical assessments. MAIN RESULTS: Although the sensitivity of the ten questions as a global screen for serious cognitive, motor, and seizure disabilities is high (84-100%), its sensitivity for identifying and distinguishing specific types of disability and for detecting vision, hearing, and mild disabilities, overall, is limited (generally < 80% and as low as 4% for mild vision disability). The predictive value of a positive screening result is also limited-using the ten questions in surveys without clinical confirmation results in overestimation of the prevalence of serious disability by more than 300%. CONCLUSIONS: The ten questions screen is not an assessment tool. Its utility lies in its ability to screen or select a fraction of the population at high risk for serious disability. As a screening tool, it allows scarce diagnostic and other professional resources to be efficiently directed toward those at high risk.