Predictive model for the cure rate of NICU admitted preterm neonates: the mixture cure model

Mohammad Aghaali, Parvaneh Sadeghi-Moghaddam, Soheila Khodakarim


Background: Predicting the survival of preterm neonates admitted to Neonatal Intensive Care Unit (NICU) care is important for counseling parents, informing care, and planning services. This study was conducted to establish a predictive model for evaluating the survival of NICU admitted preterm neonates by the cure model.

Materials and methods: 1,200 preterm neonates admitted to the NICU of a tertiary referral hospital in Qom, Iran, were enrolled to build a model and to train relevant parameters for prediction. The time to discharge from hospital and potential predictive factors were collected for analysis. A mixture cure model with the Stata version 14.2 was applied to predict the cure rate. Established factors and significant variables in the simple models were included in the multiple model.

Results: The cure rate of all patients by birth weight were: > 2,500 g, 87.31%; 1,500-2,500 g, 84.68%; 1,000-1,499 g, 75.31%; and < 1,000 g, 46.39%. The significant predictors in the final model include congenital abnormality, resuscitation, Apgar score, invasive procedure, pneumothorax, ventilation and birth weight. The result of the area under the curves for the final model on the validation data was 0.89 (95% CI, 0.823-0.951). It showed that predictive validity was satisfactory.

Conclusion: By using the cure model survival analysis, we identified significant predictors of the cure rate of NICU admitted preterm neonates. The model showed a satisfactory predictive validity, which prompted one to make an individual prediction.


survival analysis; infant; premature; intensive care units; neonatal; mixture cure model; Apgar score; pneumothorax

Full Text: PDF Number of abstract views: 848 Number of PDF views/downloads: 566


N.B. All JPNIM articles are accessible in Open Access. You can access the page containing the full PDF article just by clicking on the “Full Text: PDF” link at the bottom of the abstract page. On the full article page, if the article doesn’t load properly in the PDF view window, please wait a few seconds or click on the “Download this PDF file” link.

Technical advice. If you are using Firefox and you are experiencing problems, please set the browser preferences as follows: Firefox > Preferences > Advanced > General > Accessibility > deselect "Warn me when web sites try to redirect or reload the page".

Privacy Policy. English text: Privacy Policy; Italian text: Privacy Policy.

Cookie Policy. English text: Cookie Policy; Italian text: Cookie Policy.