Answers are grounded in this paper and cite their sources.
This retrospective model-development and same-centre temporal-validation study evaluated staged prediction of paediatric dental non-cooperation in 1120 unique children. The development period comprised 577 training and 248 internal-selection children; 295 later children formed a non-overlapping temporal-validation dataset. Five information scenarios represented basic triage (S1), child behavioural vulnerability (S2), caregiver-family context (S3), full pre-treatment information (S4), and a focused peri-visit set (S5). Non-cooperation occurred in 35.4% (397/1120). Temporal ROC AUCs for S1–S5 were 0.700 (95% CI 0.637–0.764), 0.824 (0.774–0.869), 0.856 (0.807–0.897), 0.905 (0.867–0.937), and 0.857 (0.812–0.897), respectively. S4 achieved PR-AUC 0.864 and Brier score 0.119; S5 had lower paired temporal discrimination than S4 (ΔROC AUC − 0.049, 95% CI − 0.080 to − 0.018). S4 had approximately seven training events per encoded feature, indicating limited information relative to model complexity. The study quantified predictive performance and decision-analytic value. Multicentre external validation and prospective implementation studies are required to establish robustness and clinical effectiveness before routine use.
No takes yet. Share an insight, caveat, or question.
Yu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: