Predicting death and survival without major morbidity for extremely preterm infants using information on hospital admission: a multicenter cohort study

Introduction

Extremely preterm infants (EPIs; ≤28 weeks’ gestation) represent a group of neonates facing the highest risk of severe morbidity and mortality, associated with adverse long-term outcomes and prolonged hospitalization with addition medical resources (1). Accurate prediction during the early stage after birth can assist clinicians and parents in making decisions when confronted with potential severe adverse outcomes. While numerous predictive models have been developed, they mainly rely on therapeutic measures and laboratory values, limiting their applicability primarily to high-resource settings (2-9). Additionally, despite improvements, survival rates of EPIs in low- or middle-income countries, especially rates of survival without major morbidity, still lag behind those in high-income countries (10). This gap indicates that the existing prognostic models may not be entirely suitable for routine use in countries such as China. The recent increase in the number of EPIs in China underscores the absence of well-established, locally validated predictive models for outcomes (10,11).

Therefore, this study aimed to develop and validate locally relevant prediction models for mortality and survival without major morbidity among EPIs in China, using variables available within the first hour of neonatal intensive care unit (NICU) admission. The new-derived model is expected to serve as an open-source tool for professionals and families to estimate potential outcomes for EPIs, aiding the postnatal counselling and complex decision making for these vulnerable infants. Our work will provide a benchmark for quality improvement initiatives and ongoing studies, to further improve implementation in clinical practice in China. We present this article in accordance with the TRIPOD reporting checklist (12) (available at https://tp.amegroups.com/article/view/10.21037/tp-2025-33/rc).

Methods

Data source and study population

Development cohort

Our study used data from the Reduction of Infection in Neonatal Intensive Care Units using the Evidence-based Practice for Improving Quality (REIN-EPIQ) study for model development. The REIN-EPIQ study (NCT02600195) enrolled all infants <34 weeks’ gestation who were admitted to 25 participating tertiary hospitals within 7 days of birth between May 1, 2015 and April 30, 2018 in China. All 25 hospitals were national or provincial neonatal referral centers or regional referral centers in metropolitan cities. The detailed information regarding the validity and completeness of the REIN-EPIQ data can be viewed in previous reports (13-15). The current study included all infants born at 24⁺⁰ to 28⁺⁶ weeks’ gestation admitted to NICU within 24 hours of birth who were provided with complete treatment in the REIN-EPIQ dataset. Infants with major congenital malformations or with incomplete records were excluded.

Validation cohort

The study used the Chinese Neonatal Network (CHNN) cohort as the external validation cohort. The CHNN is a national non-profit neonatal network of Chinese tertiary NICUs established to facilitate collaborative researches and to improve neonatal outcomes and quality of care (10). A standardized perinatal-neonatal database in the CHNN has been established and maintained to investigate care practices and outcomes among participating hospitals from January 1, 2019 (16). By 2021, the CHNN encompassed 79 NICUs in China. All preterm infants with gestational age of <32⁺⁰ weeks or birth weight <1,500 g admitted to CHNN participating NICUs have been enrolled in CHNN cohort. Similar to the development cohort, the current study included all EPIs of 24⁺⁰ to 28⁺⁶ weeks admitted to CHNN NICUs within 24 hours of birth with complete treatment between January 1, 2019 and December 31, 2021. Infants with major congenital malformations or with incomplete records were also excluded.

Ethical approval

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the ethics review board of Children’s Hospital of Fudan University (No. 2018-296), which was recognized by all participating centers. Waiver of consent was granted at all sites owing to the use of deidentified patient data.

Study outcomes

The study aimed to develop two separate predictive models for mortality and survival without major morbidity, respectively. Mortality was defined as death before discharge. Survival without major morbidity was considered when the infant discharged without severe neurological injury including grade 3 or 4 intraventricular hemorrhage (IVH) or periventricular leukomalacia (PVL), severe retinopathy of prematurity (ROP) (stage 3 or higher), stage 2 or 3 necrotizing enterocolitis (NEC) or bronchopulmonary dysplasia (BPD).

Candidate variables

Potential predictors were selected by literature review and expert opinions. The predictors should be of known association with neonatal outcomes, and could be commonly obtained on NICU admission in China, and were frequently included in previous neonatal outcome prediction models or illness severity scores (8,17,18). A total of seven potential predictors were identified, including gestational age, birth weight, sex, inborn or outborn status, receipt of antenatal steroids, 5-min Apgar score, and invasive ventilation on NICU admission. Gestational age and birth weight were treated as continuous variables in the model building for more accurate prediction.

Definitions

Gestational age was determined using the hierarchy of best obstetric estimate based on prenatal ultrasonography, menstrual history, obstetric examination, or all three factors. If the obstetric estimate was not available or was different from the postnatal estimate of gestation by more than 2 weeks, the gestational age was estimated using the Ballard score (19). Complete week of gestational age was used during modeling. Inborn status was defined as infants born in perinatal centers and admitted in the NICU of the same hospital. Receipt of antenatal steroids was defined as any exposure of any type of antenatal steroids. Outborn status was defined as infants born in other centers and subsequent transported to the NICU. Invasive ventilation was defined as requiring mechanical ventilation within the first hour of NICU admission.

IVH was defined according to the criteria of Papile from the worst findings on head ultrasound during the NICU stay (20). PVL was defined as the presence of periventricular cysts on cranial ultrasound or cranial magnetic resonance imaging (MRI) scans. ROP was classified according to international classification (21). NEC was defined according to Bell’s criteria (22). BPD was defined as need for any form of respiratory support (oxygen or positive pressure support) at 36 weeks corrected gestational age or at the time of discharge (23).

Statistical analysis

Descriptive statistics were listed as frequencies with percentage, and means with standard deviation (SD) or medians with interquartile range (IQR). Chi-squared test was used for categorical variables, and Wilcoxon rank-sum test was used for continuous variables to compare characteristics. A P value of <0.05 denoted statistical significance.

Logistic regression analysis was used for model development. For screening predictors, candidate variables were retained by univariate analysis of P value <0.1. To further build the predictive models, multivariable logistic regression was used with backward stepwise selection with a P value greater than 0.05 for removal of variables.

For validation, the models were applied to the CHNN database to evaluate the discrimination ability. The area under the curve (AUC) in the receiver operating characteristic (ROC) curve, equivalent to C-index, was used to evaluate the discrimination performance of the models. Model calibration was assessed by comparing the observed and predicted outcomes using Hosmer-Lemeshow goodness-of-fit test dividing into 10 groups in the validation dataset. Given the sample size in the validation dataset, and recommended by TRIPOD, we also used calibration plots with bootstraps of 1,000 resamples for calibration. To test the discrimination ability of the current model for mortality, we compared it with references including Transport Risk Index of Physiologic Stability (TRIPS) score, gestational age, and birth weight using Delong’s test (24). The TRIPS score was commonly used in China as a disease severity score on NICU admission (6). Analyses were performed using Stata/SE version 15.0 (StataCorp) and R version 4.1.2. Finally, we translated our model as a web application for clinical use.

Results

Participant characteristics

A total of 3,030 EPIs were included in the REIN-EPIQ database. Of these, 131 infants with major congenital anomalies and 461 infants with incomplete records were excluded (279 infants for 5-min Apgar score, 148 infants for receipt of antenatal steroids, 2 infants for invasive ventilation of NICU admission, and 32 infants for survival without major morbidity). Among the final 2,438 infants eligible in the development cohort, the median gestational age was 28 weeks (IQR, 27–28 weeks) and the mean birth weight was 1,091 g (SD, 211 g), of which 58.5% were male. In the validation cohort from the CHNN database, 5,045 infants were included, exhibiting a lower gestational age {median [IQR], 27 [27–28] weeks, P=0.004} and smaller birth weight {mean [SD], 1,061 [205] g, P<0.001} compared to infants in the development cohort (a flow chart of participant selection is shown in Figure 1). Characteristics of all the participants are summarized in Table 1 and Table S1.

Figure 1 Flow study participants. (A) The development cohort (REIN-EPIQ, 2015–2018); (B) the validation cohort (CHNN, 2019–2021). Total participants indicate neonates of 24–28 weeks’ gestation. *, antenatal steroids (n=148), 5-min Apgar score (n=279), invasive ventilation of NICU admission (n=2), survival without major morbidity (n=32). ^#, antenatal steroids (n=287), 5-min Apgar score (n=188), invasive ventilation of NICU admission (n=121), sex (n=1). CHNN, Chinese Neonatal Network; NICU, neonatal intensive care unit; REIN-EPIQ, Reduction of Infection in Neonatal Intensive Care Units using the Evidence-based Practice for Improving Quality.

Among the 2,438 infants in the development cohort, 17.7% (431/2,438) died, and 52.5% (1,281/2,438) survived without major morbidity. Among the 5,045 infants in the validation cohort, 9.2% (463/5,045) died, and 59.1% (2,981/5,045) survived without major morbidity.

Model derivation

In the univariate logistic regression analysis, lower gestational age, lower birth wright, invasive ventilation of NICU admission, non-receipt of antenatal steroids, and lower 5-min Apgar score were identified as potential predictors of mortality, and larger gestational age, larger birth wright, non-invasive ventilation of NICU admission, female and larger 5-min Apgar score were potential predictors of survival without major morbidity (Table 2).

In the multivariable logistic regression analysis based on the above factors, gestational age, birth weight, invasive ventilation of NICU admission, receipt of antenatal steroids, and 5-min Apgar score were identified as independent predictors of mortality. Gestational age, birth wright, invasive ventilation of NICU admission, sex, and 5-min Apgar score were independent predictors of survival without major morbidity. Table 3 presents coefficients and 95% confidence intervals (CIs) of the selected predictors in the final models. The models were translated into a publicly available web application (http://60.205.178.199:3838/shiny-app/).

Model performance

The AUC of the model for mortality was 0.77 (95% CI: 0.75–0.79), and the Hosmer-Lemeshow goodness-of-fit statistics indicated a good fit (P=0.19) between observed and predicted outcomes. For survival without major morbidity, the AUC was 0.72 (95% CI: 0.70–0.74), and the Hosmer-Lemeshow test demonstrated that the model was a good fit (P=0.40). The calibration curves of both models were close to ideal diagonal lines (Figure 2).

Figure 2 Calibration curves in the development and validation cohorts. (A) Mortality model in the development cohort; (B) mortality model in the validation cohort; (C) survival without major morbidity model in the development cohort; (D) survival without major morbidity model in the validation cohort.

Model validation

In the external validation cohort, the CHNN Neonatal Outcome Model resulted in an AUC of 0.76 (95% CI: 0.73–0.78) for mortality, and 0.70 (95% CI: 0.68–0.71) for survival without major morbidity (Table 4). The calibration curves of the models were close to ideal diagonal lines (Figure 2).

Model comparison

Using the gestational age to predict mortality yielded an AUC of 0.69 (95% CI: 0.67–0.71) in the validation cohort. Using birth weight alone had an AUC of 0.71 (95% CI: 0.68–0.74), and TRIPS resulted in an AUC of 0.64 (95% CI: 0.61–0.66) (Figure 3). Delong’s statistical test comparing the AUCs between the CHNN Neonatal Outcome Model and gestational age or birth weight or TRIPS was significant (P<0.001 for CHNN Neonatal Outcome Model vs. gestational age or birth weight or TRIPS).

Figure 3 Comparisons of discrimination ability for mortality in the validation cohort. AUC, area under the curve; BW, birth weight; GA, gestational age; TRIPS, Transport Risk Index of Physiologic Stability.

Discussion

This study, utilizing the two largest prospective databases of EPIs from China, has developed and validated models to predict mortality and survival without major morbidity among neonates of 24⁺⁰ to 28⁺⁶ weeks’ gestational age at the time of admission to the NICU. The models demonstrated moderate discrimination among EPIs admitted to Chinese NICUs, and have been further translated as a web application for ease of use.

The performance of the model is comparable to other EPI mortality prediction models from high-income countries, such as the BAG (birth weight, 5-min Apgar score, gestational age) model and the National Institute of Child Health and Human Development (NICHD) calculator (17,25). Several other neonatal mortality models yielded higher AUCs (>0.80) than our model, but these models targeted infants of larger gestational ages or birth weights with less site variation in care practices and outcomes (4,6,8,18). Our model shows better prediction performance than TRIPS score, which has been commonly used in current clinical practice in China. By calculating at the bedside within a few minutes after NICU admission, this model may be used as a simple, user-friendly and relatively more accurate outcome prediction tool in healthcare settings similar to China.

The quality of survival for EPIs is as important as survival, because major morbidities of EPIs are significantly associated with adverse long-term neurodevelopmental outcomes (26). Therefore, our study not only developed the model to predict mortality, but also attempted to assess the intact survival of these vulnerable infants. However, the occurrence of major morbidity was influenced by many factors during long-term hospitalization, so the discrimination ability of the survival without major morbidity model was not very ideal and needed further improvement.

Factors in our model including gestational age, birth weight, sex, respiratory support, antenatal corticosteroid use, and 5-min Apgar score, are all well-known predictors of survival in preterm infants (27). Gestational age, birth weight and sex are the most important factors and can be easily accessible in clinical practice. Despite improvements over past years, the antenatal corticosteroid use rate (about 70%) in our cohort remained much lower than in high-income countries (28), therefore receipt of antenatal steroids was still a significant predictor of neonatal outcomes in China. We used 5-min Apgar score and respiratory support on NICU admission to quantify illness severity. Although utilizing 5-minute Apgar score alone is of concerns due to subjectivity and reliability, adding other predictors such as gestational age would improve the discriminatory ability for mortality (29). Five-min Apgar score was also available for all EPIs in China. Score for Neonatal Acute Physiology (SNAP), or the Clinical Risk Index for Babies (CRIB) were not included as predictors, because they required prolonged data collection of therapeutic measures and laboratory values over 12 hours, which might not be easily available. Invasive ventilation of NICU admission was also included. While the respiratory support mode might be influenced by treatment practice within the unit, it could also, to some extent, serve as an indicator for the severity of the disease. Therefore, the simplicity and availability of all the predictors allowed our models to display an estimated probability of mortality and survival without major morbidity shortly after NICU admission for almost all EPIs. Furthermore, it is important to emphasize that the predictive models serve solely as a reference to enable parents and caregivers to make informed choices. Using this tool for decision-making in the context of healthcare provision may not be appropriate.

The strengths of our study include the unitality of the two largest cohort of EPIs in China, the ability for external validation, assessment of both mortality and survival without major morbidity, and the use of simple predictors that are available immediately on admission. There are some limitations in this study. The data in the REIN-EPIQ and the CHNN dataset were both from tertiary NICUs in China, and therefore our study may not be fully representative of the general population or diverse geographic cohorts. However, this is unlikely to impact the generalizability of our findings since the majority of these smallest infants are typically born or transported to hospitals with the highest level of neonatal care. It is also worth noting that half of the participants were at 28 weeks’ gestation, and 90% were at 26–28 weeks’ gestation. Therefore, with the anticipated increase in the number of EPIs in the future, it becomes imperative to conduct further validation among infants of <26 weeks’ gestation. Hospital variation was also considered as a potential predictor since post-admission treatments of different NICUs could have impacts on the outcomes, and therefore exclusions of treatment information as predictors could influence the predictive performance.

Conclusions

Our study has developed a multivariable model to predict short-term neonatal outcomes including mortality and survival without major morbidity for EPIs, utilizing clinical information available within one hour of NICU admission. Future directions involve additional external validations in a larger cohort to evaluate effectiveness in practical use and impacts on prognosis, as well as periodic modifications with improvements in neonatal outcomes over time.

Acknowledgments

The authors gratefully acknowledge the support of the Canadian Neonatal Network (CNN) and all the data abstractors from the REIN-EPIQ study and the CHNN. Group Information of the Chinese Neonatal Network: Children’s Hospital of Fudan University; The Third Affiliated Hospital of Zhengzhou University; Guangzhou Women and Children’s Medical Center; Tianjin Obstetrics & Gynecology Hospital; Gansu Provincial Maternity and Child Care Hospital; Northwest Women’s and Children’s Hospital; Shenzhen Maternity and Child Health Care Hospital; Guizhou Women and Children’s Hospital; Suzhou Municipal Hospital Affiliated to Nanjing Medical University; Shengjing Hospital of China Medical University; Children’s Hospital of Shanxi; Quanzhou Women and Children’s Hospital; Fujian Women and Children’s Medical Center; Children’s Hospital of Nanjing Medical University; Hunan Children’s Hospital; Qingdao Women and Children’s Hospital; Nanjing Maternity and Child Health Care Hospital; The First Bethune Hospital of Jilin University; The First Affiliated Hospital of Anhui Medical University; Women and Children’s Hospital of Guangxi Zhuang Autonomous Region; The First Affiliated Hospital of Xinjiang Medical University; Foshan Women and Children’s Hospital; The Affiliated Hospital of Qingdao University; Henan Children’s Hospital; Children’s Hospital of Shanghai; Chongqing Health Care Center for Women and Children; Children’s Hospital of Chongqing Medical University; Wuxi Maternity and Child Healthcare Hospital; Children’s Hospital of Soochow University; People’s Hospital of Xinjiang Uygur Autonomous Region; Yuying Children’s Hospital Affiliated to Wenzhou Medical University; Shanghai First Maternity and Infant Hospital; Anhui Provincial Hospital; Xinhua Hospital Affiliated to Shanghai Jiaotong University School of Medicine; Qilu Children’s Hospital of Shandong University; The First Affiliated Hospital of Zhengzhou University; General Hospital of Ningxia Medical University; Hebei Children’s Hospital; Hainan Women and Children’s Hospital; The Second Xiangya Hospital of Central South University; Ningbo Women & Children Hospital; Xiamen Children’s Hospital; Shanxi Provincial People’s Hospital; The Affiliated Hospital of Southwest Medical University; Shanghai Children’s Medical Center affiliated to Shanghai Jiaotong University School of Medicine; First Affiliated Hospital of Kunming Medical University; Changzhou Maternal and Children Health Care Hospital; Shenzhen Children’s Hospital; Jiangxi Provincial Children’s Hospital; Xiamen Maternity and Child Health Care Hospital; Zhuhai Center for Maternal and Child Health Care; Guangdong Women and Children’s Hospital; Wuhan Children’s Hospital; Beijing Children’s Hospital of Capital Medical University; Maternal and Children Hospital of Shaoxing; The First People’s Hospital of Yunnan Province; Dehong People’s Hospital of Yunnan Province; First Affiliated Hospital of Xian Jiaotong University; Inner Mongolia Maternal and Child Health Care Hospital; Dalian Municipal Women and Children’s Medical Center; Lianyungang Maternal and Children Health Hospital; Children’s Hospital Affiliated to Capital Institute of Pediatrics; Anhui Children’s Hospital; Fuzhou Children’s Hospital of Fujian Province; Kunming Children’s Hospital; Shenzhen Hospital of Hongkong University; Peking Union Medical College Hospital; Obstetrics & Gynecology Hospital of Fudan University; The Affiliated Hospital of Guizhou Medical University; Qinghai Women and Children Hospital; The International Peace Maternity & Child Health Hospital of China Welfare Institute; Children’s Hospital of Zhejiang University; Inner Mongolia People’s Hospital; Xiamen Humanity Hospital; Shanghai General Hospital; The First People’s Hospital of Yinchuan; The Third Hospital of Nanchang.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tp.amegroups.com/article/view/10.21037/tp-2025-33/rc

Data Sharing Statement: Available at https://tp.amegroups.com/article/view/10.21037/tp-2025-33/dss

Peer Review File: Available at https://tp.amegroups.com/article/view/10.21037/tp-2025-33/prf

Funding: This study was funded by the Shanghai Science and Technology Commission’s Scientific and Technological Innovation Action Plan (No. 21Y21900800) and the Canadian Institutes of Health Research (No. CTP87518).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tp.amegroups.com/article/view/10.21037/tp-2025-33/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the ethics review board of Children’s Hospital of Fudan University (No. 2018-296), which was recognized by all participating centers. Waiver of consent was granted at all sites owing to the use of deidentified patient data.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Zhu Z, Yuan L, Wang J, et al. Mortality and Morbidity of Infants Born Extremely Preterm at Tertiary Medical Centers in China From 2010 to 2019. JAMA Netw Open 2021;4:e219382. [Crossref] [PubMed]
2. Richardson DK, Gray JE, McCormick MC, et al. Score for Neonatal Acute Physiology: a physiologic severity index for neonatal intensive care. Pediatrics 1993;91:617-23. [Crossref] [PubMed]
3. Richardson DK, Corcoran JD, Escobar GJ, et al. SNAP-II and SNAPPE-II: Simplified newborn illness severity and mortality risk scores. J Pediatr 2001;138:92-100. [Crossref] [PubMed]
4. The CRIB (clinical risk index for babies) score: a tool for assessing initial neonatal risk and comparing performance of neonatal intensive care units. The International Neonatal Network. Lancet 1993;342:193-8. Erratum in: Lancet 1993;342:626. [Crossref] [PubMed]
5. Parry G, Tucker J, Tarnow-Mordi W, et al. CRIB II: an update of the clinical risk index for babies score. Lancet 2003;361:1789-91. [Crossref] [PubMed]
6. Lee SK, Zupancic JA, Pendray M, et al. Transport risk index of physiologic stability: a practical system for assessing infant transport care. J Pediatr 2001;139:220-6. [Crossref] [PubMed]
7. Lee SK, Aziz K, Dunn M, et al. Transport Risk Index of Physiologic Stability, version II (TRIPS-II): a simple and practical neonatal illness severity score. Am J Perinatol 2013;30:395-400. [PubMed]
8. Ge WJ, Mirea L, Yang J, et al. Prediction of neonatal outcomes in extremely preterm neonates. Pediatrics 2013;132:e876-85. [Crossref] [PubMed]
9. Shah PS, Ye XY, Synnes A, et al. Prediction of survival without morbidity for infants born at under 33 weeks gestational age: a user-friendly graphical tool. Arch Dis Child Fetal Neonatal Ed 2012;97:F110-5. [Crossref] [PubMed]
10. Cao Y, Jiang S, Sun J, et al. Assessment of Neonatal Intensive Care Unit Practices, Morbidity, and Mortality Among Very Preterm Infants in China. JAMA Netw Open 2021;4:e2118904. [Crossref] [PubMed]
11. Deng K, Liang J, Mu Y, et al. Preterm births in China between 2012 and 2018: an observational study of more than 9 million women. Lancet Glob Health 2021;9:e1226-41. [Crossref] [PubMed]
12. Collins GS, Reitsma JB, Altman DG, et al. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement. BMJ 2015;350:g7594. [Crossref] [PubMed]
13. Jiang S, Yan W, Li S, et al. Mortality and Morbidity in Infants <34 Weeks’ Gestation in 25 NICUs in China: A Prospective Cohort Study. Front Pediatr 2020;8:33. [Crossref] [PubMed]
14. Jiang S, Zhang L, Yan W, et al. Antibiotic Use in Neonatal Intensive Care Units in China: A Multicenter Cohort Study. J Pediatr 2021;239:136-142.e4. [Crossref] [PubMed]
15. Yu W, Zhang L, Li S, et al. Early Antibiotic Use and Neonatal Outcomes Among Preterm Infants Without Infections. Pediatrics 2023;151:e2022059427. [Crossref] [PubMed]
16. Sun J, Cao Y, Hei M, et al. Data Quality Improvement and Internal Data Audit of the Chinese Neonatal Network Data Collection System. Front Pediatr 2021;9:711200. [Crossref] [PubMed]
17. Rysavy MA, Horbar JD, Bell EF, et al. Assessment of an Updated Neonatal Research Network Extremely Preterm Birth Outcome Model in the Vermont Oxford Network. JAMA Pediatr 2020;174:e196294. [Crossref] [PubMed]
18. Medvedev MM, Brotherton H, Gai A, et al. Development and validation of a simplified score to predict neonatal mortality risk among neonates weighing 2000 g or less (NMR-2000): an analysis using data from the UK and The Gambia. Lancet Child Adolesc Health 2020;4:299-311. [Crossref] [PubMed]
19. Ballard JL, Novak KK, Driver M. A simplified score for assessment of fetal maturation of newly born infants. J Pediatr 1979;95:769-74. [Crossref] [PubMed]
20. Papile LA, Burstein J, Burstein R, et al. Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm. J Pediatr 1978;92:529-34. [Crossref] [PubMed]
21. The International Classification of Retinopathy of Prematurity revisited. Arch Ophthalmol 2005;123:991-9. [Crossref] [PubMed]
22. Bell MJ, Ternberg JL, Feigin RD, et al. Neonatal necrotizing enterocolitis. Therapeutic decisions based upon clinical staging. Ann Surg 1978;187:1-7. [Crossref] [PubMed]
23. Shennan AT, Dunn MS, Ohlsson A, et al. Abnormal pulmonary outcomes in premature infants: prediction from oxygen requirement in the neonatal period. Pediatrics 1988;82:527-32. [Crossref] [PubMed]
24. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44:837-45. [Crossref] [PubMed]
25. Moreira A, Benvenuto D, Fox-Good C, et al. Development and Validation of a Mortality Prediction Model in Extremely Low Gestational Age Neonates. Neonatology 2022;119:418-27. [Crossref] [PubMed]
26. Schmidt B, Asztalos EV, Roberts RS, et al. Impact of bronchopulmonary dysplasia, brain injury, and severe retinopathy on the outcome of extremely low-birth-weight infants at 18 months: results from the trial of indomethacin prophylaxis in preterms. JAMA 2003;289:1124-9. [Crossref] [PubMed]
27. Medlock S, Ravelli AC, Tamminga P, et al. Prediction of mortality in very premature infants: a systematic review of prediction models. PLoS One 2011;6:e23441. [Crossref] [PubMed]
28. Stoll BJ, Hansen NI, Bell EF, et al. Trends in Care Practices, Morbidity, and Mortality of Extremely Preterm Neonates, 1993-2012. JAMA 2015;314:1039-51. [Crossref] [PubMed]
29. Shah PS, Norman M, Rusconi F, et al. Five-minute Apgar score and outcomes in neonates of 24-28 weeks’ gestation. Arch Dis Child Fetal Neonatal Ed 2022;107:437-46. [Crossref] [PubMed]