Supplementary file 1_Perinatal and gestational factors associated with Pavlik harness outcomes in developmental dysplasia of the hip: a retrospective cohort study.doc

Internal validation was performed using bootstrap resampling with 200 iterations.</p>Results<p>Among 792 children (910 hips) included, low Graf classification, higher maternal education level, cephalic intrauterine position, parity > 1, maternal combined vitamin A and D supplementation during pregnancy and regular maternal milk consumption during pregnancy were significantly associated with higher treatment success rates.

The overall prediction model, integrating Graf classification with the five perinatal and gestational factors, achieved a bootstrap-corrected C-index of 0.781, with significantly higher diagnostic accuracy than any single factor or alternative combination (AUC = 0.796, P < 0.05). In the Graf-stratified prediction models, the ROC curves for the five perinatal and gestational factors demonstrated acceptable discriminative ability (Stratum 1: C-index = 0.724; Stratum 2: C-index = 0.707).

However, the model showed no predictive value in Graf type IV hips, where disease severity appeared to be the dominant determinant of treatment failure. Based on these findings, both overall and Graf-stratified nomograms were developed to predict Pavlik harness treatment outcomes in DDH.</p>Conclusion<p>Several perinatal and gestational factors were independently associated with Pavlik harness treatment outcomes in DDH.

The nomograms developed in this study may serve as preliminary decision-support tools for predicting treatment outcomes and facilitating the selection of more appropriate clinical management strategies, particularly for Graf types IIc and D/III. Given the observational study design, these findings should be interpreted as exploratory associations rather than causal determinants, and external validation is warranted before clinical implementation.</p>