Meta-analysis finds SIVH prediction models in preterm infants show moderate discrimination
This systematic review and meta-analysis evaluated the performance of existing prediction models for severe intraventricular hemorrhage (SIVH) in preterm infants. The analysis included 16 prediction models, with 7 models providing sufficient data for meta-analysis of discrimination performance. The pooled area under the curve (AUC) was 0.805 (95% confidence interval: 0.756–0.853), indicating moderate discrimination ability.
All included studies exhibited high risk of bias, primarily in the analysis domain. Methodological shortcomings were frequent, including inadequate handling of missing data (93.75% of studies), use of univariable analysis for predictor selection (62.50%), insufficient calibration assessment (68.75%), and non-robust internal validation (50.00%). These limitations significantly affect the reliability and generalizability of the findings.
Safety and tolerability data were not reported in the meta-analysis. The review did not specify funding sources or conflicts of interest. The authors note that current SIVH prediction models show promise but require substantial methodological improvements before clinical application.
For practice, this evidence suggests that while prediction models for SIVH in preterm infants demonstrate moderate discrimination, their clinical applicability remains uncertain due to methodological limitations. Clinicians should interpret these models cautiously and await more rigorously developed and validated tools before incorporating them into routine care.
View Original Abstract ↓
Standardized formal resuscitation training likely reduces neonatal mortality in low- and middle-income countries compared to no training or basic training.
A Simple Training Program Is Saving Newborn Lives in Poor Countries
Related in Pediatrics
From Other Specialties
