Prediction of Postpartum Haemorrhage After Labour Induction: An Internally Validated 10-Hour Risk-Stratification Threshold

Background/Objectives: Postpartum haemorrhage (PPH) remains a leading cause of maternal morbidity, and the relationship between labour-induction duration and haemorrhagic risk has not been translated into a practical intrapartum risk-stratification framework. We aimed to derive a clinically interpretable induction-duration threshold for PPH risk stratification and to explore an internally validated parsimonious clinical decision-support model. Methods: In this retrospective cohort of 1128 induced singleton labours at ≥37 weeks at a Turkish tertiary centre, laboratory-defined PPH was operationalised as a haemoglobin drop ≥ 2 g/dL. Multivariable logistic regression identified independent predictors; receiver operating characteristic (ROC) analysis with 2000-replicate bootstrap internal validation derived a duration threshold; calibration, decision curve analysis, and a probability-scaled nomogram were additionally evaluated. Results: PPH occurred in 143 patients (12.7%). Four predictors were independently associated with PPH: induction duration (adjusted odds ratio 1.243 per hour; 95% CI 1.191–1.298; p < 0.001), parity, emergency caesarean, and maternal age. Induction duration achieved an apparent area under the curve of 0.773 (optimism-corrected 0.773); a Youden-optimal threshold of 10.1 h yielded 86.7% (95% CI 80.2–91.3%) sensitivity and 96.8% (95% CI 95.0–97.9%) negative predictive value. The model showed favourable calibration and positive net clinical benefit on decision curve analysis. Conclusions: Induction duration was independently associated with PPH; the internally validated 10.1 h threshold therefore represents a hypothesis-generating risk-stratification benchmark—reflecting an association rather than a demonstrated causal effect—that requires prospective external validation in independent populations before clinical application. Whether this association is causal or reflects an underlying myometrial phenotype requires prospective study.