Background/Objectives: Individual responses to recombinant human growth hormone (rhGH) therapy in children with idiopathic short stature (ISS) vary substantially, limiting pretreatment decision-making. This study aimed to develop and externally validate an interpretable machine learning model for predicting suboptimal 12-month growth response to rhGH therapy. Methods: In this retrospective dual-center study, 901 children from Center 1 were used for model development and internal testing, and 51 children from Center 2 formed an independent external validation cohort. Routinely collected baseline demographic, laboratory, hormonal, radiographic, and family-history variables were used to develop multiple machine learning models. A soft-voting ensemble classifier was constructed and interpreted using SHapley Additive exPlanations (SHAP). The primary outcome was suboptimal growth response, defined as failure to achieve a height gain of at least 0.5 standard deviation score after 12 months of treatment. Results: The optimized ensemble model showed strong discrimination in the internal test set, with an area under the receiver operating characteristic curve of 0.927, and maintained robust performance in the external validation cohort, with an AUC of 0.897. SHAP analysis identified luteinizing hormone, body mass index, TW3 RUS bone age, and insulin-like growth factor 1 as the leading contributors to predicted suboptimal-response risk. Conclusions: An interpretable ensemble machine learning model based on routinely available pretreatment data can predict suboptimal short-term rhGH response in children with ISS and may support individualized risk stratification in pediatric endocrine practice. Clinical trial registration was not required because this was a retrospective analysis.

Background To investigate the value of the SNAP-IV scale combined with IVA-CPT objective attention/executive function indicators in identifying the risk of cognitive ability in children with ADHD, and to evaluate its incremental predictive ability over traditional subjective scales, providing a basis for clinical screening and stratified interventions. Methods A total of 248 children with ADHD (191 boys, 57 girls), with a mean age of 8.98 ± 1.68 years, were enrolled from the ADHD outpatient clinic at the Children's Health Center of Capital center for Children's Health, Capital Medical University. All participants completed the SNAP-IV scale, WISC and IVA-CPT Attention Test. Cognitive high-risk was defined by a Wechsler Full Scale IQ < 85. A logistic regression model (Model 1) was constructed including only the SNAP-IV total score, age, sex, parental ADHD status, and whether the child was an only child. Model 2 was constructed by adding IVA-CPT composite scores for attention quotient, response control quotient, and auditory/visual attention and control quotients. The area under the curve (AUC) of the ROC curve and DCA curve was compared between the two models. Results The AUC of Model 1 was 0.612, with a sensitivity of 0.804, and specificity of 0.426, indicating that its predictive ability for cognitive high-risk was close to random. After incorporating IVA indicators, the AUC of Model 2 improved to 0.707, with a sensitivity of 0.587, and specificity of 0.787, which was significantly superior to Model 1 (p = 0.034). DCA analysis further demonstrated that Model 2 offered more favorable guidance for decision-making at moderate-risk thresholds.Additionally, no significant correlation was found between the SNAP-IV score and any IVA dimension score, indicating that subjective symptom scales and objective attention function indicators assess different aspects of cognitive performance. Conclusion The SNAP-IV scale alone is insufficient for effectively identifying the cognitive high-risk individuals among children with ADHD. However, the addition of IVA-CPT attention and response control indicators significantly improves predictive performance. This study supports the use of a combined screening strategy of "subjective behavioral scales + objective cognitive tests" in ADHD assessments, which can help identify cognitive high-risk children early and facilitate early intervention. Trial registration Not applicable.