A continuous-scale measure of child development for population-based epidemiological surveys: a preliminary study using Item Response Theory for the Denver Test
Source: Paediatric & Perinatal Epidemiology, Volume 21, Number 2, March 2007 , pp. 138-153(16)
Abstract:Summary Drachler ML, Marshall T, de Carvalho Leite JC. A continuous-scale measure of child development for population-based epidemiological surveys: a preliminary study using Item Response Theory for the Denver Test. Paediatric and Perinatal Epidemiology 2007; 21: 138–153.
A method for translating research data from the Denver Test into individual scores of developmental status measured in a continuous scale is presented. It was devised using the Denver Developmental Screening Test (DDST) but can be used for Denver II. The DDST was applied in a community-based survey of 3389 under-5-year-olds in Porto Alegre, Brazil. The items of success were standardised by logistic regression on log chronological age. Each child’s ability age was then estimated by maximum likelihood as the age in this reference population corresponding to the child’s success and failures in the test. The score of developmental status is the natural logarithm of this ability age divided by chronological age and thus measures the delay or advance in the child’s ability age compared with chronological age. This method estimates development status using both difficulty and discriminating power of each item in the reference population, an advantage over scores based on total number of items correctly performed or failed, which depend on difficulty only. The score corresponds with maternal opinion of child developmental status and with the 3-category scale of the DDST. It shows good construct validity, indicated by symmetrical and homogeneous variability from 3 months upwards, and reasonable results in describing gender differences in development by age, the mean score increasing with socio-economic conditions and diminishing among low-birthweight children. If a standardised measure of development status (z-scores) is required, this can be obtained by dividing the score by its standard deviation. Concurrent and discriminant validity of the score must be examined in further studies.
Document Type: Research Article
Affiliations: 1: Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, UK 2: School of Allied Health Professions, Institute of Health, University of East Anglia, Norwich, and
Publication date: March 1, 2007