Machine learning risk stratification for high-risk infant follow-up of term and late preterm infants.

BACKGROUND: Term and late preterm infants are not routinely referred to high-risk infant follow-up programs at neonatal intensive care unit (NICU) discharge. We aimed to identify NICU factors associated with abnormal developmental screening and develop a risk-stratification model using machine learning for high-risk infant follow-up enrollment. METHODS: We performed a retrospective cohort study identifying abnormal developmental screening prior to 6 years of age in infants born ≥34 weeks gestation admitted to a level IV NICU. Five machine learning models using NICU predictors were developed by classification and regression tree (CART), random forest, gradient boosting TreeNet, multivariate adaptive regression splines (MARS), and regularized logistic regression analysis. Performance metrics included sensitivity, specificity, accuracy, precision, and area under the receiver operating curve (AUC). RESULTS: Within this cohort, 87% (1183/1355) received developmental screening, and 47% had abnormal results. Common NICU predictors across all models were oral (PO) feeding, follow-up appointments, and medications prescribed at NICU discharge. Each model resulted in an AUC > 0.7, specificity >70%, and sensitivity >60%. CONCLUSION: Stratification of developmental risk in term and late preterm infants is possible utilizing machine learning. Applying machine learning algorithms allows for targeted expansion of high-risk infant follow-up criteria. IMPACT: This study addresses the gap in knowledge of developmental outcomes of infants ≥34 weeks gestation requiring neonatal intensive care. Machine learning methodology can be used to stratify early childhood developmental risk for these term and late preterm infants. Applying the classification and regression tree (CART) algorithm described in the study allows for targeted expansion of high-risk infant follow-up enrollment to include those term and late preterm infants who may benefit most.

Antenatal Breast Milk Expression Survey of Individuals Whose Pregnancy Was Complicated by Diabetes: Exploring Knowledge, Perceptions, Experiences, and Milk Volume Expressed.

OBJECTIVE: Antenatal breast milk expression (ABE) offers a host of benefits, including reduced formula consumption, support for breastfeeding success, and increased maternal satisfaction. Despite these advantages, experience with ABE differs significantly, often leading to anxiety over perceived inadequate milk supply and eventual breastfeeding cessation. This study comprehensively evaluates the knowledge, attitudes, and real-world experiences of individuals with gestational or pregestational diabetes concerning ABE, with a focus on total milk volume expressed prior to birth. STUDY DESIGN: Utilizing a convenience sampling method, we surveyed individuals with gestational or pregestational diabetes from three health care facilities who were trained in ABE. Knowledge and perceptions were gauged through presurvey statements, while postsurvey statements were employed to measure experiences, both using a 5-point Likert scale. In parallel, a retrospective study assessed both maternal and infant outcomes among the same participant pool. Statistical comparisons between individuals with and without reservations were made using the Wilcoxon signed rank sum, Mann-Whitney, chi-square, and Fisher's exact tests. RESULTS: Of the 138 participants, 75% completed both survey segments, and 61% expressed reservations about ABE. Both groups were demographically similar and showed comparable newborn outcomes. However, individuals with reservations experienced heightened pain during ABE, reported lesser lactation support, and were less willing to repeat the process compared to those individuals without reservations. The median total ABE volume was significantly lower by 14 mL among those with reservations (7 vs. 21 mL, p = 0.009). Although both groups demonstrated improved attitudes toward the utility of ABE for individuals with gestational or pregestational diabetes, no significant shift occurred in the perception of ABE difficulty. CONCLUSION: Our results indicate that individuals with gestational or pregestational diabetes who have reservations about ABE face unique challenges and tend to express lower milk volumes. This underlines the need for specialized interventions and ongoing research to address antenatal lactation support and alleviate ABE-related concerns among individuals with gestational or pregestational diabetes. KEY POINTS: · Reservations of ABE were associated with reduced milk volumes.. · Regardless of reservations, ABE was felt to be beneficial.. · Our results underscore the need for more ABE education for those with reservations..

Prediction of short- and long-term outcomes using pre-operative ventricular size in infants with post-hemorrhagic ventricular dilation.

PURPOSE: Post-hemorrhagic ventricular dilation (PHVD) leads to developmental delays in premature infants, yet the optimal timing of neurosurgical interventions is unknown. Neuroimaging modalities have emerged to delineate injury and follow the progression of PHVD. Fronto-temporal horn ratio (FTHR) is used as a marker of ventricular dilation and can be a standardized tool to direct the timing of neurosurgical intervention. Our study determined a pre-operative FTHR measurement threshold to predict short- and long-term outcomes. METHODS: This is a retrospective cohort study of premature infants with severe intraventricular hemorrhage (IVH) who developed PHVD requiring neurosurgical intervention and were treated in a level IV NICU between 2012 and 2019. Receiver operating characteristic (ROC) curve and area under the curve (AUC) analyses were performed to evaluate the accuracy of pre-operative FTHR for predicting developmental delay. In-hospital outcomes and developmental assessments were analyzed. RESULTS: We reviewed 121 charts of infants with IVH and identified 43 infants with PHVD who required neurosurgical intervention. We found FTHR measurements were an excellent predictor of cognitive and motor delay with an AUC of 0.89 and 0.88, respectively. An average pre-operative FTHR of ≥ 0.67 was also associated with worse lung and feeding outcomes. There was excellent inter-observer reliability of individual components of FTHR measurements. CONCLUSIONS: Early intervention for PHVD is ideal but not always practical. Identification of ventricular size thresholds associated with better outcomes is needed to direct timing of neurosurgical intervention.

Variability in neurosurgical management and associated comorbidities and complications among preterm patients with posthemorrhagic hydrocephalus in the United States.

OBJECTIVE: Posthemorrhagic hydrocephalus (PHH) remains a major morbidity of premature birth resulting from intraventricular hemorrhage (IVH). National consensus guidelines for the timing of surgical interventions are lacking, which leads to considerable variations in management among neonatal intensive care units (NICUs). Early intervention (EI) has been shown to improve outcomes, but the authors hypothesized that the timing from IVH to intervention affects the comorbidities and complications associated with PHH management. The authors used a large national inpatient care data set to characterize comorbidities and complications associated with PHH management in premature infants. METHODS: The authors used hospital discharge data from the 2006-2019 Healthcare Cost and Utilization Project (HCUP) Kids' Inpatient Database (KID) to conduct a retrospective cohort study of premature pediatric patients (weight < 1500 g) with PHH. The predictor variable was the timing of the PHH intervention (EI ≤ 28 days vs late intervention [LI] > 28 days). Hospital stay data included hospital region, gestational age, birth weight (BW), length of stay (LOS), PHH treatment procedures, comorbidities, surgical complications, and death. Statistical analysis included chi-square and Wilcoxon rank-sum tests, Cox proportional hazards regression, logistic regression, and a generalized linear model with Poisson and gamma distributions. Analysis was adjusted for demographic characteristics, comorbidities, and death. RESULTS: Of the 1853 patients diagnosed with PHH, 488 (26%) had documented timing of surgical interventions during their hospital stay. More patients had LI than EI (75%). The patients in the LI group of patients had younger gestational age and lower BW. There were significant regional differences in the timing of treatment: hospitals in the West performed EI, whereas hospitals in the South performed LI, even after adjustment for gestational age and BW. The LI group was associated with longer median LOS and more total hospital charges compared with the EI group. More temporary CSF diversion procedures occurred in the EI group, whereas more permanent CSF-diverting shunts were placed in the LI group. Shunt/device replacement and complications did not differ between the two groups. The LI group had 2.5-fold higher odds of sepsis (p < 0.001) and almost 2-fold higher odds of retinopathy of prematurity (p < 0.05) than the EI group. CONCLUSIONS: The timing of PHH interventions differs by region in the United States, whereas the association of potential benefits with treatment timing suggests the importance of national consensus guidelines. Development of these guidelines can be informed by data regarding treatment timing and patient outcomes available in large national data sets, which provide insights into comorbidities and complications of PHH interventions.

Quality improvement sustainability to decrease utilization drift for therapeutic hypothermia in the NICU.

OBJECTIVES: Therapeutic hypothermia (TH) is now standard of care for the neuroprotection of patients with moderate to severe hypoxic-ischemic encephalopathy (HIE). TH misuse results in increased medical complication rates and high health care resource utilization. Quality improvement (QI) methodology can address drift from clinical guidelines. Assessment of sustainability of any intervention over time is an integral part of the QI methodology. METHODS: Our prior QI intervention improved medical documentation using an electronic medical record-smart phrase (EMR-SP) and demonstrated special cause variation. This study serves as Epoch 3 and investigates the sustainability of our QI methods to decrease TH misuse. RESULTS: A total of 64 patients met the diagnostic criteria for HIE. Over the study period, 50 patients were treated with TH, and 33 cases (66%) used TH appropriately. The number of appropriate TH cases between cases of misuse increased to an average of 9 in Epoch 3 from 1.9 in Epoch 2. Of the 50 cases, 34 (68%) had EMR-SP documentation included. Length of stay and TH complication rates did not vary between cases of TH misuse and appropriate TH use. CONCLUSIONS: Our study confirmed a sustained decrease in TH misuse, despite inconsistent use of EMR-SP. We speculate that culture change involving increased awareness of guidelines through education may have contributed more to a lasting change.

Neonatal hyperbilirubinemia: Assessing variation in knowledge and practice.

INTRODUCTION: Neonatal hyperbilirubinemia (NH) is commonly diagnosed and managed by pediatricians in various clinical settings. The 2004 American Academy of Pediatrics (AAP) Clinical Practice Guideline on NH is widely cited, but literature examining variation across pediatric specialties is limited. This study aimed to assess baseline knowledge and practice habits regarding NH among pediatric providers across various specialties immediately prior to the release of the 2022 NH clinical practice guideline. METHODS: A non-probability, convenience, self-selected sampling survey was electronically distributed to 311 subjects across five specialties within one pediatric healthcare institution. The survey included eight multiple choice knowledge-based questions with confidence assessments and five management-based questions assessing respondent agreement on a 5-point scale. To compare groups, the Kruskal-Wallis and Mann-Whitney tests were used for continuous variables, and the chi-square and Fisher's exact tests were used for categorical variables. RESULTS: The overall survey response rate is 46%. There were significant differences between specialties' knowledge regarding NH (p<0.05). There were also significant differences between specialties' confidence ratings, independent of choosing the correct response (p<0.05). For select management-based questions, there were also significant differences between specialties (p<0.05). A majority of respondents (56%) indicated phototherapy treatment thresholds should remain the same in updated management guidelines. CONCLUSIONS: Significant variations in knowledge and management of NH were identified among pediatric specialties. This suggests dissemination of new guidelines must be cognizant of different constraints impacting knowledge and practice across specialties.

Trends in the early care of infants with myelomeningocele in the United States 2012-2018.

PURPOSE: The early care of children with spina bifida has changed with the increasing availability of fetal surgery and evidence that fetal repair improves the long-term outcomes of children with myelomeningocele. We sought to determine current trends in the prevalence and early care of children with myelomeningocele using a national administrative database. METHODS: This is a retrospective, cross-sectional cohort study of infants with spina bifida admitted within the first 28 days of life using the 2012-2018 Healthcare Cost and Utilization Project National Inpatient Database. Patients with spina bifida were identified by ICD code and stratified into a cohort with a coded neonatal repair of the defect and those without a coded repair. This database had no identifier specific for fetal surgery, but it is likely that a substantial number of infants without a coded repair had fetal surgery. RESULTS: We identified 5,090 patients with a coded repair and 5,715 without a coded repair. The overall prevalence of spina bifida was 3.94 per 10,000 live births. The percentage of patients without neonatal repair increased during the study period compared to those with repair (p = 0.0002). The cohort without neonatal repair had a higher risk of death (p < 0.001), prematurity (p < 0.001), and low birth weight (p < 0.001). More shunts were placed in patients who underwent neonatal repair (p < 0.001). Patients without neonatal repair were less likely to have public insurance (p = 0.0052) and more likely to reside in zip codes within the highest income quartile (p = 0.0002). CONCLUSIONS: The prevalence of spina bifida from 2012 to 2018 was 3.94 per 10,000 live births, with an increasing number of patients without neonatal repair of the defect, suggesting increased utilization of fetal surgery. Patients without neonatal repair had a higher risk of death, prematurity, and low birth weight but were more likely to have commercial insurance and reside in high-income zip codes.

Addressing drivers of healthcare utilization for neonatal opioid withdrawal syndrome.

OBJECTIVE: Aim to reduce healthcare utilization (HU) for infants at risk of neonatal opioid withdrawal syndrome (NOWS) by 30% in 1 year and sustain for 2 years. STUDY DESIGN: Baseline data from three Level I & II newborn nurseries from January 2016 to June 2018 informed PDSA cycles from August 2018 to December 2021. Shewhart process control charts evaluated length of stay (LOS), pharmacologic treatment (PT) rates, direct cost (DC), process, and balancing measures for special cause variation (SCV). RESULTS: Two hundred and seventeen infants showed downward SCV in LOS (12.6 to 4.4 days), PT (53% to 17%) and DC ($12593.82 to $5219.17). Onset of the COVID-19 pandemic coincided with reversible SCV. DC varied by provider specialty. CONCLUSION: Transition from MFNASS to ESC led to decrease in healthcare utilization for infants at risk of NOWS. QI methodology identified persistent drivers of variability, including the COVID-19 pandemic and provider specialty.

The Burden of Neonatal Abstinence Syndrome, Opioids, and COVID-19 in Wisconsin.

INTRODUCTION: Wisconsin experienced overlapping and accelerating epidemics of opioid use and COVID-19 after March 2020. We hypothesized that Wisconsin neonatal abstinence syndrome rates increased after March 2020 alongside other markers of opioid burden. METHODS: Retrospective cohort analysis examined deidentified Wisconsin census, birth certificate, death certificate, hospital discharge, Prescription Drug Monitoring Program, emergency medical service run, and COVID-19 diagnosis records spanning January 1, 2019, through December 31, 2021. January 2019 through March 2020 was considered before the onset of COVID-19 (pre); April 2020 through December 2021 was considered post-onset of COVID-19 (post). Wisconsin Department of Health Services guidelines defined 5 Wisconsin regions. Rates pre- to post-onset were compared with P values < 0.05 considered statistically significant. RESULTS: Of 1362 patients, 83.3% completed a COVID-19 vaccination series. Younger patients had increased odds of not completing a COVID-19 vaccination series (mean [SD] 46.7 [14.7] vs 54.3 [15.8]; OR 1.03; 95% CI, 1.02-1.04; P < 0.001). Those who identified as non-White (1.88; 95% CI, 1.16-3.04; P = 0.010) or current smoker (1.85, 95% CI, 1.85-2.79; P = 0.004) had increased odds of not completing a COVID-19 vaccination series. Those who resided in rural ZIP codes (1.81; 95% CI, 1.35-2.43; P < 0.001), had not received a 2019-2020 influenza vaccine (5.13; 95% CI, 3.79-6.96; P < 0.001), or had lower comorbidity scores (2.95; 95% CI, 1.98-4.41; P < 0.001) had higher odds of not completing a COVID-19 vaccination series. CONCLUSIONS: Opioid-associated morbidity and mortality increased in Wisconsin during the study period, including among females age 15 to 44 years. Despite increased opioid burden, neonatal abstinence syndrome incidence decreased in the Southeastern Region. Ongoing neonatal abstinence syndrome and opioid analysis may benefit from region-based contextualization.

Code Blue Events in the Neonatal and Pediatric Intensive Care Units at a Tertiary Care Children's Hospital.

OBJECTIVE: The authors aim to compare all code blue events, regardless of the need for chest compressions, in the neonatal intensive care unit (NICU) versus the pediatric intensive care unit (PICU). We hypothesize that code events in the two units differ, reflecting different disease processes. STUDY DESIGN: This is a retrospective analysis of 107 code events using the code narrator, which is an electronic medical record of real-time code documentation, from April 2018 to March 2019. Events were divided into two groups, NICU and PICU. Neonatal resuscitation program algorithm was used for NICU events and a pediatric advanced life-support algorithm was used for PICU events. Events and outcomes were compared using univariate analysis. The Mann-Whitney test and linear regressions were done to compare the total code duration, time from the start of code to airway insertion, and time from airway insertion to end of code event. RESULTS: In the PICU, there were almost four times more code blue events per month and more likely to involve patients with seizures and no chronic condition. NICU events more often involved ventilated patients and those under 2 months of age. The median code duration for NICU events was 2.5 times shorter than for PICU events (11.5 vs. 29 minutes), even when adjusted for patient characteristics. Survival to discharge was not different in the two groups. CONCLUSION: Our study suggests that NICU code events as compared with PICU code events are more likely to be driven by airway problems, involve patients <2 months of age, and resolve quickly once airway is taken care of. This supports the use of a ventilation-focused neonatal resuscitation program for patients in the NICU. KEY POINTS: · Code blue events are four times more common in PICU.. · NICU code events are 2.5 times shorter in duration compared with PICU events.. · NICU code events are more likely to be attributed to a problem with an airway..

Neonatal Birth Trauma: Analysis of Yearly Trends, Risk Factors, and Outcomes.

OBJECTIVE: To evaluate the trends, proportions, risk factors, resource utilization, and outcomes of neonatal birth trauma in the US. STUDY DESIGN: This cross-sectional study of in-hospital births used the Nationwide Inpatient Sample for 2006-2014. We divided the cases by type of birth trauma: scalp injuries and major birth trauma. Linear regression for yearly trends and logistic regression were used for risk factors and outcomes. A generalized linear model was used, with a Poisson distribution for the length of stay and a gamma distribution for total spending charges. RESULTS: A total of 982 033 weighted records with neonatal birth trauma were found. The prevalence rate increased by 23% from (from 25.3 to 31.1 per 1000 hospital births). Scalp injuries composed 80% of all birth traumas and increased yearly from 19.87 to 26.46 per 1000 hospital births. Major birth trauma decreased from 5.44 to 4.67 per 1000 hospital births due to decreased clavicular fractures, brachial plexus injuries, and intracranial hemorrhage. There were significant differences in demographics and risk factors between the 2 groups. Compared with scalp injuries, major birth trauma was associated with higher odds of hypoxic-ischemic encephalopathy, seizures, need for mechanical ventilation, meconium aspiration, and sepsis. Length of stay was increased by 56%, and total charges were almost doubled for major birth trauma. CONCLUSIONS: Neonatal birth trauma increased over the study period secondary to scalp injuries. Major birth trauma constitutes a significant health burden. Scalp injuries are also associated with increased morbidity and might be markers of brain injury in some cases.

Improved Maternal and Infant Outcomes with Serial, Self-Reported Early Prenatal Substance Use Screening.

INTRODUCTION: Most screening tools identifying women with substance use are not validated, used once in pregnancy, and are not reflective of continued substance use. We hypothesized that serial early prenatal substance screening leads to decreased substance use by the end of pregnancy and improved outcomes. METHODS: This is a retrospective cohort study of mothers and their infants between 1/2015 and 12/2017. A self-reported substance screening tool was administered on the first prenatal visit and subsequent visits until delivery. For analysis, mothers were divided into three groups based on the trimester of their first screen and adjusted for demographics and risk factors. RESULTS: Early first trimester screening resulted in 52% of mothers having ≥ 3 screens throughout pregnancy vs. 6% of mothers with late third trimester screens (p < 0.001). Compared to third trimester screening, there was a five-fold decrease of any substance use at second trimester, a seven-fold decrease at first trimester, and a nine-fold decrease for marijuana at first trimester. Compared to third trimester screening, there was a significant five-fold increase of negative maternal urine drug screen, 3 ½ -fold increase in well newborn diagnosis, and a five-fold increase of no infant morphine treatment at first trimester. DISCUSSION: We identified improved maternal and infant outcomes with serial early prenatal substance use screening. Early maternal substance use identification is crucial for immediate referral for prevention and treatment, and for social and community services. Further research is needed on universal serial early prenatal screenings.

Correlating Scores but Contrasting Outcomes for Eat Sleep Console Versus Modified Finnegan.

OBJECTIVES: The Modified Finnegan Neonatal Abstinence Scoring System (M-FNASS) and the newer Eat, Sleep, and Console (ESC) model guide the clinical management of neonatal opioid withdrawal syndrome (NOWS). In this study, we evaluate how the M-FNASS and ESC model directly compare in inpatient practice. We hypothesized that ESC scores would correlate with M-FNASS scores, whereas ESC management would reduce health care use for infants with NOWS. METHODS: In this retrospective cohort study, we compared management of infants with NOWS admitted to nursery settings. Epoch 1 was managed by using an M-FNASS algorithm. Epoch 2 was scored simultaneously with the M-FNASS and ESC model and managed by using the ESC approach. In the statistical analysis, we compared M-FNASS and ESC scores and outcomes between epochs. RESULTS: A total of 158 infants provided 2101 scoring instances for analysis. Demographic characteristics were similar between epochs. ESC scores significantly correlated with overall M-FNASS scores and specific M-FNASS domains. Receiver operating characteristic (ROC) curve analysis revealed that an ESC score containing at least 1 "no" was best predicted by an M-FNASS cutoff value of 7.5 (sensitivity 0.84; specificity 0.70; area under the curve = 0.842). Length of stay (median 9.5 vs 5 days; P = .0002) and initiation (53% vs. 33%; P = .018) and duration of pharmacologic treatment (median 11 vs 7 days; P = .0042), as well as length of stay for infants who were pharmacologically treated (median 15 vs 10 days; P = .0002), were significantly reduced with ESC-based management after adjustment for covariates. CONCLUSIONS: The ESC approach meaningfully correlates with the M-FNASS to detect NOWS. Management with the ESC approach continues to be associated with reduced health care use when compared with an M-FNASS approach, implying that the ESC approach may facilitate higher-value inpatient care.

Quality improvement for reducing utilization drift in hypoxic-ischemic encephalopathy management.

OBJECTIVES: Therapeutic hypothermia is an effective neuroprotective intervention for infants with moderate or severe hypoxic-ischemic encephalopathy (HIE). With the introduction of new medical therapy comes a learning curve with regards to its proper implementation and understanding of eligibility guidelines. We hypothesized that variation in patient selection and lack of adherence to established protocols contributed to the utilization drift away from the original eligibility guidelines. METHODS: A retrospective cohort study was conducted including infants who received therapeutic hypothermia in the neonatal intensive care unit (NICU) for HIE to determine utilization drift. We then used QI methodology to address gaps in medical documentation that may lead to the conclusion that therapeutic hypothermia was inappropriately applied. RESULTS: We identified 54% of infants who received therapeutic hypothermia who did not meet the clinical, physiologic, and neurologic examination criteria for this intervention based on provider admission and discharge documentation within the electronic medical record (EMR). Review of the charts identified incomplete documentation in 71% of cases and led to the following interventions: 1) implementation of EMR smartphrases; 2) engagement of key stakeholders and education of faculty, residents, and neonatal nurse practitioners; and 3) performance measurement and sharing of data. We were able to improve both adherence to the therapeutic hypothermia guidelines and achieve 100% documentation of the modified Sarnat score. CONCLUSIONS: Incomplete documentation can lead to the assumption that therapeutic hypothermia was inappropriately applied when reviewing a patient's EMR. However, in actual clinical practice physicians follow the clinical guidelines but are not documenting their medical decision making completely. QI methodology addresses this gap in documentation, which will help determine the true utilization drift of therapeutic hypothermia in future studies.

Social disparities negatively impact neonatal follow-up clinic attendance of premature infants discharged from the neonatal intensive care unit.

OBJECTIVE: Neonatal neurodevelopmental follow-up clinic provides continued surveillance and assessment of high-risk premature infants. We hypothesized that attrition is associated with race and social factors. STUDY DESIGN: We performed a retrospective cohort study of neonates born at 26-32 weeks gestation who were admitted to a level IV neonatal intensive care unit. Maternal and neonatal characteristics and follow-up attendance were collected. Statistical analysis was performed with significance set at p value < 0.05. RESULTS: In total, 237 neonates met study criteria. There was a 62% loss to follow-up over 2 years. Factors associated with loss to follow-up included older gestational age, African American race, and maternal cigarette smoking. Protective factors included older maternal age, a neonatal diagnosis of bronchopulmonary dysplasia, and longer hospital length of stay. CONCLUSIONS: Social disparities negatively impact neonatal follow-up clinic attendance. Efforts to identify and target high-risk populations must be started during initial hospitalization before infants are lost to follow-up.

Outcome of neonates with gastroschisis at different gestational ages using a national database.

BACKGROUND/PURPOSE: The optimal time for delivery of neonates with a prenatal diagnosis of gastroschisis (GS) is controversial. We compared the outcomes for GS at three different gestational ages (GAs), 33-34 weeks, 35-36 weeks, and ≥ 37 weeks. METHODS: We analyze hospital discharge data of neonates with GS using the 2006, 2009 and 2012 Healthcare Cost and Utilization Project Kids' Inpatient Database (HCUPKIDS). Multivariable analysis was used to compare the association between GS outcomes and the three GAs. RESULTS: Significantly higher number of 33-34 week infants had coexisting morbidities like respiratory distress syndrome, bronchopulmonary dysplasia, small bowel atresia, stenosis, or stricture, large bowel atresia and/or stenosis, malrotation, and atrial septal defect. In multivariable logistic regression, 33-34 week infants had higher NEC (p value = 0.002, 95% CI1.64-10.32), small bowel resection (0.024, 1.12-5.25) and pRBCs transfusion (0.024, 1.05-2.11). No differences were found between 35-36 weeks and ≥37 weeks gest infants for NEC, malabsorption, small bowel resection, TPN cholestasis, sepsis, CLABSI, number of pRBCs transfusion, length of stay and total charges. CONCLUSION: We did not show benefit for delivering early and in the absence of data, delivery at ≥37 weeks was noninferior to 35-36 weeks. We suggest that waiting for spontaneous onset of labor may be a better approach to balance the effects of prematurity and possible ongoing in utero bowel damage/stillbirth. LEVELS OF EVIDENCE: Level 3 (Retrospective comparative study).