Drawing Characteristics Associated with Attention-Deficit/Hyperactivity Disorder in Children: Findings from the...

This study aimed to examine drawing characteristics in the Synthetic House-Tree-Person (S-HTP) drawing test among children with attention-deficit/hyperactivity disorder (ADHD) and to explore their associations with ADHD symptom severity.

A total of 60 children aged 7-11 years (30 with clinically diagnosed ADHD and 30 healthy controls) completed the S-HTP drawing test and the SNAP-IV scale. Drawing characteristics were independently coded by two trained raters using predefined criteria. Group differences were analyzed using chi-square and rank-sum tests. Logistic regression analyses were performed to evaluate associations with ADHD status, and ROC analyses were used to assess discrimination. FDR correction and bootstrap internal validation were additionally applied.

Six drawing characteristics differed significantly between groups in the initial analyses, of which five remained significant after FDR correction. Seven drawing characteristics showed exploratory associations with SNAP-IV scores. After adjustment for age, gender, family structure, and Raven's Standard Progressive Matrices scores, three drawing characteristics remained associated with ADHD status: A3 (tracing in shaky lines; OR = 4.31, 95% CI: 1.10-16.94), A4 (lines jagged and not joined; OR = 5.45, 95% CI: 1.45-20.49), and A8 (more than seven different colors; OR = 0.20, 95% CI: 0.05-0.72). Exploratory ROC analyses indicated that combining these characteristics improved discriminatory performance. The combined model achieved an AUC of 0.835. Bootstrap internal validation yielded an optimism-corrected AUC of 0.812, indicating reasonable model stability within the present sample.

Certain S-HTP drawing characteristics were associated with ADHD status and symptom severity in this sample. In particular, A3, A4, and A8 were identified as drawing characteristics of potential interest for future research. However, given the exploratory design, relatively small sample size, and lack of external validation, further studies are required before the clinical utility of these findings can be established.