Value of the Goodenough Drawing Test as a research tool to detect developmental delay in South African preschool children

Abstract

There is a need for simple, cost-effective research tools to detect developmental delay in preschool children in low- and middle-income countries where insufficient resources are often a barrier to detection and management. The Goodenough Draw-a-Person test is freely available, easily administered, and requires limited language ability and equipment; it is thus potentially useful in resource-constrained settings. We aimed to determine the diagnostic accuracy of the Draw-a-Person test to identify developmental delay in 5-year-old preschool children using the Griffiths Mental Developmental Scales-Extended Revised eye-hand coordination subquotient as the gold standard. This was a cross-sectional analysis of drawings by South African preschool children from low-income families, whose Griffiths Mental Developmental Scales-Extended Revised assessments included a human figure drawing. Draw-a-Person test quotients were estimated independently by a developmental paediatrician and two medical officers to calculate inter-rater agreement. The paediatrician’s scores were used to determine the diagnostic accuracy of the Draw-a-Person test quotient (<85) to predict developmental delay with the eye-hand coordination subquotient (<75). A total of 125 children were included, with a mean age of 60.8 months (range 59–66 months) of which 48.8% were boys. The mean Draw-a-Person test score was 94 (standard deviation 15) with 28 Draw-a-Person test scores below 85. Applying the Draw-a-Person test cut-off of 85, sensitivity of the Draw-a-Person test to the eye-hand coordination subquotient was 80% and specificity 89%. The area under the receiver operator characteristic curve was 0.87 (95% confidence interval [0.78–0.96]). The Goodenough Draw-a-Person test could thus be a useful research tool for detecting fine motor and visuoperceptual delay in South African preschool children.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 5 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.