Journal of Perinatology (2014) 34, 582–586 & 2014 Nature America, Inc. All rights reserved 0743-8346/14 www.nature.com/jp

SPECIAL FEATURE

The Children’s Hospitals Neonatal Database: an overview of patient complexity, outcomes and variation in care K Murthy1, FD Dykes2, MA Padula3, EK Pallotto4, KM Reber5, DJ Durand6, BL Short7, JM Asselin6, I Zaniletti8 and JR Evans3 The Children’s Hospitals Neonatal Consortium is a multicenter collaboration of leaders from 27 regional neonatal intensive care units (NICUs) who partnered with the Children’s Hospital Association to develop the Children’s Hospitals Neonatal Database (CHND), launched in 2010. The purpose of this report is to provide a first summary of the population of infants cared for in these NICUs, including representative diagnoses and short-term outcomes, as well as to characterize the participating NICUs and institutions. During the first 2½ years of data collection, 40910 infants were eligible. Few were born inside these hospitals (2.8%) and the median gestational age at birth was 36 weeks. Surgical intervention (32%) was common; however, mortality (5.6%) was infrequent. Initial queries into diagnosis-specific inter-center variation in care practices and short-term outcomes, including length of stay, showed striking differences. The CHND provides a contemporary, national benchmark of short-term outcomes for infants with uncommon neonatal illnesses. These data will be valuable in counseling families and for conducting observational studies, clinical trials and collaborative quality improvement initiatives. Journal of Perinatology (2014) 34, 582–586; doi:10.1038/jp.2014.26; published online 6 March 2014 Keywords: neonatal intensive care unit (NICU); outcomes; levels of neonatal care; Children’s Hospitals Neonatal Consortium (CHNC); practice variation; regional NICU

INTRODUCTION There is a vast body of literature addressing the common morbidities and outcomes of preterm infants. However, benchmarking data on populations of infants with uncommon diseases have been much more infrequent, and most that do exist are typically limited by small numbers and prolonged accrual. In 2007 the Children’s Hospitals Neonatal Consortium (CHNC) formed a multicenter collaboration of clinicians from referralbased regional neonatal intensive care units1 (NICUs) to address this unmet need. In partnership with the Children’s Hospital Association (CHA), the Children’s Hospitals Neonatal Database (CHND) was launched in 2010. The CHND captures data on each admitted infant in the participating NICUs in order to characterize the diseases in this referral population of infants, with the primary goal of improving their care and outcomes. The purposes of this report are to describe the patient characteristics of the infants admitted to CHND NICUs, to present an overview of the unit characteristics of the participating NICUs and to provide a representative selection of diagnoses, interventions, short-term outcomes and inter-center variations in care for this population. METHODS Data fields, sources and participating NICUs The CHND captures core clinical data on all infants admitted to 27 regional children’s hospitals’ NICUs.1,2 Additional data are collected for infants

who have the following diseases or conditions: severe chronic lung disease, gastrointestinal anomalies, necrotizing enterocolitis, congenital diaphragmatic hernia and hypoxic-ischemic encephalopathy. During database development, data fields and definitions were aligned with those existing in other relevant data sets, to permit future comparisons and collaborations. The refining of data fields is an ongoing process, and CHA and CHND jointly design and field-test enhancements to the CHND in order to increase the accuracy and precision of the ascertained data. Each participating NICU met specific criteria: (a) level IIIc designation based on published standards1,2 at the time of initiation of data collection; (b) 4400 admissions annually; (c) X25 inpatient beds; and (d) 450% of admitted infants born outside the participating hospital. While three sites did not precisely fit these criteria, all were level IIIc NICUs and their neonatology services closely matched the 24 other participating NICUs. Each institution received institutional review board oversight before participating in the CHND. Not all hospitals began data collection simultaneously; however, from each hospital’s start date in 2010, all subsequent admissions were captured. Chart abstractors at each site were provided prospective training including review of clinical definitions, participation in web-based seminar tutorials and case-based practice. Both initial and semi-annual measurements of inter-rater agreement scores were calculated at each site; over 90% intra-site concordance in abstraction was required for participation in the CHND. Prospectively, identified and defined variables were abstracted retrospectively from each infant’s medical record. Abstracted elements were entered into a web-based data collection tool, which functions with built-in validations to minimize the risk of key stroke errors and/or invalid entries. Once a patient record was opened, patient data was securely transferred from each site to a centralized server.

1 Ann & Robert H Lurie Children’s Hospital of Chicago, Department of Pediatrics, Feinberg School of Medicine, Northwestern University Chicago, Chicago, IL, USA; 2Children’s Healthcare of Atlanta at Egleston and the Department of Pediatrics, Emory University School of Medicine, Atlanta, GA, USA; 3Division of Neonatology, Department of Pediatrics, Children’s Hospital of Philadelphia, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, USA; 4Children’s Mercy Hospital and Clinics and the Department of Pediatrics, University of Missouri School of Medicine, Kansas City, MO, USA; 5Nationwide Children’s Hospital and the Department of Pediatrics, Ohio State University School of Medicine, Columbus, OH, USA; 6Department of Pediatrics, Children’s Hospital Oakland & Research Center, Oakland, CA, USA; 7Children’s National Medical Center and the Department of Pediatrics, George Washington University School of Medicine, Washington DC, USA and 8Children’s Hospital Association, Overland Park, KS, USA. Correspondence: Dr JR Evans, Department of Pediatrics, Division of Neonatology, Children’s Hospital of Philadelphia, Perelman School of Medicine at the University of Pennsylvania, Room 2NW26 CHOP Main, 34th and Civic Center Blvd, Philadelphia, PA 19104, USA. E-mail: [email protected] Received 4 October 2013; revised 10 December 2013; accepted 13 January 2014; published online 6 March 2014

Children’s Hospitals Neonatal Database K Murthy et al

583 The CHND was queried on 15 March 2013 for this analysis. Accounting for varying dates for sites’ initiation of data collection, the study period was B30 months. In addition to patient data, characteristics of each participating NICU and hospital are collected from each physician site sponsor annually. The survey ascertains the capacity of each NICU, details of organization of patient transport, the available pediatric and surgical subspecialty services, as well as characteristics of co-located obstetric services. Survey results reported here were obtained in the first quarter of 2013 for status in 2012. Census data and annual admissions were obtained from the CHND; calendar year 2011 was used for this information to ensure that all records would be complete and closed for the most accurate reflection of annual census.

Description of population characteristics and participating NICUs The primary aim of this report is to provide representative information on the CHND by characterizing both the infants admitted to participating NICUs as well as the characteristics of the NICUs and their associated services. Mortality and length of stay (LOS), both measured to CHND hospital discharge, are reported for infants with selected neonatal surgical and medical diagnoses. Data on ultimate disposition for some infants transferred to other institutions were incomplete, thus these infants’ LOS and mortality are reported only until the time of transfer. For all infants in the data set, demographic characteristics are described and include categorical descriptions of gestational age at birth, birth weight, referral source and the primary reason for referral. These reasons for referral were prospectively classified into 17 categories, and the five most frequent reasons are reported. For infants born o32 weeks’ gestation or o1.5 kg at birth, selected prematurity-associated morbidities are reported if they were present at the time of referral to the CHND NICU.

Table 1.

Characteristics of the participating NICUs/hospitals

Staffing Median number of neonatologists, n (range) Neonatal fellowship program, n (%) Median number of enrolled fellows, n (range) Pediatric residency program, n (%) Pediatric surgical training program, n (%) Unit characteristics Median NICU beds, n (range) Median patients admitted per center per year, n (IQR) Average daily census per center (median±s.d.) Offered services, n (%) In NICU magnetic resonance imaging In-house labor and delivery/obstetrics Adjacent hospital with obstetrics service Interventional neuroradiology Cardiac surgery Extracorporeal membrane oxygenation Accepts infants from home General pediatric surgery, otolaryngology, urology, neurosurgery and interventional radiology

17 21 8 23 22

(6–70) (78) (0–18) (85) (82)

50 (18, 108) 636 (416, 845) 36.5 (29.6, 51) 5 8 15 22 24 25 26 27

(19) (30) (56) (81) (89) (93) (96) (100)

Abbreviations: IQR, interquartile range; NICU, neonatal intensive care unit. Out of twenty-seven hospitals twenty-five (92%) provide transplantation services including liver,18 kidney,18 cardiac,16 lung8 and bowel.6

Data analysis Mortality For infants who died, the frequency of limitations in medical care in place before death is reported. These limitations were typically either withdrawal of life-sustaining medical intervention(s) or ‘do not resuscitate’ (DNR) orders. Also, the occurrence of cardiopulmonary resuscitation within 6 hours of death is described. The rationale behind the selection of these measures was to begin to define a cohort in which death was anticipated.

Status at discharge The complexity of infants who were discharged directly to home from the NICU is described; reported measures include the proportions of infants prescribed technology-based care or enteral tube feedings at the time of discharge. The proportion of infants receiving any breast milk at the time of hospital discharge is also reported.

Surgical intervention The total number of infants undergoing surgery, their surgical procedures, whether surgical evaluation was the primary indication for referral, and the frequency of multiple surgical interventions per infant is reported. Also, postoperative physiologic measures are reported and defined as hypothermia and hyperthermia (temperature o36 1C and 438 1C, respectively); hypoglycemia and hyperglycemia (bedside or serum glucose o50 and 4200 mg/dl, respectively); acidosis and alkalosis (pHo7.25 and 47.5, respectively); and hypocarbia and hypercarbia (PCO2o30 and 470 mm Hg, respectively).

Variation in care Inter-center practice variation is highlighted for one medical therapy, one surgical practice and two short-term outcomes. The four metrics selected were: (a) the rate of fundoplication amongst all infants who received a gastrostomy tube; (b) the rate of postoperative hypothermia defined as temperature o36 1C in the first 30 min postoperatively amongst infants undergoing non-cardiac surgery; (c) LOS for infants with gastroschisis born X34 weeks’ gestation and without bowel atresia; and (d) the duration of mechanical ventilation expressed as a proportion of hospital days amongst these same infants with gastroschisis. & 2014 Nature America, Inc.

Analyses were performed centrally using SAS v 9.3 (Cary, NC, USA). The Ann and Robert H Lurie Children’s Hospital of Chicago Institutional Review Board provided oversight for this project (# 2011–14673).

RESULTS Twenty-seven regional NICUs contributed to CHND during the study period. All participating NICUs provided pediatric surgical services and almost all accepted infants admitted from home. Few had in-house obstetric services (Table 1). There were 43 070 patient admissions, or episodes of care, amongst the 40 910 infants cared for in these NICUs during the study period. Over half the infants were born X37 weeks’ gestation and were referred primarily for surgical, pulmonary or neurological evaluation/management. Very few were inborn. Among those born o32 weeks’ gestation or o1.5 kg, a significant proportion had prematurity-associated morbidities at the time of referral (Table 2). Relative to the national birth cohort, many of the infants who were referred had uncommon diagnoses or complications. Among infants with the selected conditions, the relationship of non-survival versus survival to hospital LOS varied by diagnosis (Table 3). Overall, there were 27 199 surgical procedures completed in 20 148 surgical timeframes; timeframes were defined as instances during which infants received one or more surgical interventions. During the study period, a median of 789 surgical time frames occurred per center (interquartile range 381, 916). Among all episodes of care, surgical procedures occurred in 32%. There were 9858 (24.1%) admissions for which infants were referred primarily for surgical evaluation and management; 7098 (72%) of these received a surgical procedure during their CHND hospitalization. Conversely, 5993 (19.3%) of the 31 052 infants whose primary reason for referral was for a nonsurgical evaluation ultimately had X1 surgical procedure during their hospitalization. The counts of a sample of surgical interventions are described (Table 4). Journal of Perinatology (2014), 582 – 586

Children’s Hospitals Neonatal Database K Murthy et al

584 Table 2.

Patient characteristics at the time of referral/admission

Variable Unique infants in the study period (n)

Survival and LOS for selected conditions in CHND

Diagnosis description

Total no

% Survival

40 910

Gestational age and birth weight o32 weeks oro1.5 kg, n (%) Median chronological age at admission (days, (IQR)) Median post-menstrual age at admission (weeks, (IQR)) X37 weeks’ gestation, n (%)

22 760 (55.6)

Referral sources, n (%) Other hospital Level II NICU Level III NICU Within hospital delivery unit (inborn) Within hospital other (OR, ICU and ward) Outpatient setting (home, ED or office)

32 470 (79.4) 6569 (16.1) 13 933 (34.1) 1147 (2.8) 6535 (16) 758 (1.9)

8595 (21) 2 (0,12) 38 (35, 40)

Primary reason for referral, n (%) Surgical evaluation/management Respiratory evaluation/management Neurologic evaluation/management Cardiac evaluation/management Preterm birth without other co-morbidities Other

11 823 (28.9)

Morbidities present on admission,* n (%) IVH PDA BPD Medical NEC Intestinal perforation Hydrocephalus ROPXstage 3 Surgical NEC cPVL

1650 (19.2) 1543 (18) 1071 (12.5) 421 (4.9) 413 (4.8) 399 (4.6) 281 (3.3) 235 (2.7) 104 (1.2)

9858 (24.1) 8849 (21.6) 3850 (9.4) 3514 (8.6) 3016 (7.4)

Abbreviations: BPD, bronchopulmonary dysplasia; cPVL, cystic periventricular leukomalacia; ED, emergency department; ICU, intensive care unit; IQR, interquartile range; IVH, intra-ventricular hemorrhage; NEC, necrotizing enterocolitis; NICU, neonatal intensive care unit; OR, operating room; PDA, patent ductus arteriosus; ROP, retinopathy of prematurity. *For infants born o32 weeks’ gestation or o1.5 kg.

Postoperatively, infants frequently had physiologic instability. The frequency of postoperative hypothermia or hyperthermia was 15.9 and 1.6%, of hypoglycemia and hyperglycemia was 0.7 and 10.1%, of acidosis and alkalosis was 14.3 and 3.4%, and of hypocarbia and hypercarbia was 4.5 and 5.5%, respectively. At discharge from the NICU, 29 856 (73%) infants went home or to foster care, 4852 (11.9%) were transferred to another unit in the hospital, 4087 (10%) were transferred to another institution and 2115 (5.2%) had died. After being transferred to another unit in the hospital, an additional 178 (0.4%) infants died before hospital discharge, and thus, the frequency of mortality during the CHND hospitalization was 5.6%. Although mortality for infants transferred to other institutions was rare (16 patients) for the 20 sites that collected this information, follow-up was not universally ascertained for these transferred infants. Of the 2293 infants who died before CHND hospital discharge, 48.2% were born X34 weeks gestation, and the vast majority of deaths were expected. Active withdrawal of intensive care (72.4%) and DNR (52.6%) orders were frequently in place at the Journal of Perinatology (2014), 582 – 586

Table 3.

sBPD 32 wks GA HIE Gastroschisis Surgical NEC Trisomy 21 Imperforate anus Myelomeningocele CDH Omphalocele CPAM Midgut volvulus SCT Total colonic Hirschprung’s Epidermolysis bullosa Cerebral AVM Hemochromatosis Kernicterus ACD

Median LOS (d) Nonsurvivors

Survivors

1073 999 879 739 712 581 542 501 261 201 183 64 40

89 81 98 64 91 93 99 68 79 98 97 94 95

90 4 10 2 16 6 29 16 16 1 2 32 25

112 15 34 97 19 11 11 37 20 3 20 11 16

33 38 23 10 9

82 68 65 80 0

48 7 26 10 14

9 14 20 15 —

Abbreviations: ACD, alveolar capillary dysplasia; AVM, arterio-venous malformation; CDH, congenital diaphragmatic hernia; CHND, Children’s Hospitals Neonatal Database; CPAM, congenital pulmonary airway malfromation; HIE, hypoxic ischemic encephalopathy; LOS, length of stay; NEC, necrotizing enterocolitis; SCT, sacrococcygeal teratoma; sBPD, severe bronchopulmonary dysplasia; wks GA, weeks in gestational age at birth.

Table 4.

Counts of selected surgeries and/or interventions

Gastrostomy tube insertion Ductus arteriosus ligation Gastroschisis final closure Fundoplication Ileostomy Ostomy takedown or reanastomosis Colostomy TEF±EA repair Tracheostomy Retinal laser surgery CDH repair Duodenal atresia, stenosis and web repair Large bowel resection Jejunal, ileal and colonic atresia repair Jejunostomy Omphalocele closure

2188 1153 874 776 743 633 547 531 527 497 437 362 271 234 224 170

Abbreviations: CDH, congenital diaphragmatic hernia; TEF±EA, tracheoesophageal fistula±esophageal atresia.

time of death. Only 18.1% of infants who died had neither withdrawal of intensive care nor a DNR order in place. Of the 33 519 infants discharged directly home from the CHND hospital (from the NICU or from another unit), 4962 (14.8%) received a type of respiratory support, cardio-respiratory monitoring and/or supplemental FiO2 at the time of discharge. Non-oral enteral feedings were prescribed in 3823 (11.4%) at discharge, and the majority of infants were receiving some breast milk (n ¼ 19 316; 57.6%). Inter-center variation in intervention (for example, fundoplication), outcomes (for example, postoperative hypothermia) and diseasespecific resource utilization (for example, in patients with gastroschisis) is striking (Figures 1a–d, Po0.01 for all) & 2014 Nature America, Inc.

Children’s Hospitals Neonatal Database K Murthy et al

80 60 40 20

Median hospital LOS (days)

0

60 50 40 30 20 10 0

% Hospital days on ventilator

% Fundoplication

100

% Post-operative temperature

585 60 50 40 30 20 10 0

35 30 25 20 15 10 5 0

Figure 1. (a–d) Selected inter-center variations across participating CHND sites. Each data point represents a single site’s value. X-axes for all panels are the CHND site, and individual hospital values for each outcome are ordered in the same sequence on each panel. (a) Percent of fundoplication among infants who received a gastrostomy tube; represented 2029 gastrostomies. (b) Percent of infants with postoperative hypothermia (o36 1C) after non-cardiac surgery; represented 18 680 surgeries. (c) Median hospital LOS among infants with gastroschisis born X34 weeks’ gestation and without bowel atresia; represented 833 infants. (d) Median ventilation days as a percent of total hospital days in infants with gastroschisis born X34 weeks’ gestation and without bowel atresia (39 patients had unknown duration of ventilation); represented 794 infants. LOS, length of stay.

DISCUSSION The CHND was designed to allow quality benchmarking for these infants who are typically referred to regional NICUs for management requiring pediatric subspecialty services. In addition to each hospital having full access to its own data, a multidisciplinary collaborative structure is in place to analyze multicenter data, report results and develop quality improvement initiatives congruent with the membership’s priorities. This report demonstrates that CHND is a resource to develop these goals and to inquire about disease-specific care and outcomes.3,4 Owing to the size of the network, rapid accrual of patients is feasible for these analyses. Excellent, longitudinal, worldwide benchmarking data have previously been available for infants cared for primarily in NICUs with large associated obstetric services.5–9 These databases primarily focus on preterm infants and their relatively small number of common associated morbidities, many of which can be used as outcome measures. This report highlights that the CHND patient population differs substantially from those previously well characterized. In the CHND, the majority of patients are born X37 weeks’ gestation, there is a high prevalence of uncommon diagnoses and surgical care is remarkably frequent. In addition, the population of preterm infants in the CHND differs from most systematically studied preterm populations because the timing of referral of preterm infants to CHND NICUs is usually well after birth, and frequently after a significant proportion has acquired severe prematurity-associated morbidities,6 which indeed are often existing at the time of or even the very reason for referral. Thus, these traditional outcome measures are inappropriate for benchmarking the care and outcomes in large regional NICUs in the United States and perhaps serve more appropriately as to highlight the severity of illness in the very preterm infants who were referred. However, for these NICUs, the rarity of the diagnoses managed as well as the historical absence of a large, multicenter collaboration has precluded both the development of appropriate benchmarks for expected outcomes, as well as disease- or intervention-specific quality improvement efforts. & 2014 Nature America, Inc.

Although there has been no registry reflecting care and outcomes for the complex patient population admitted to regional NICUs in the United States, there is a registry for infants admitted to regional NICUs in Canada.8,10 Although premature infants predominate in this registry, a linked data set11 focuses on patients with two surgical conditions admitted to these NICUs. In addition, there are several US-based registries focusing on specific neonatal diseases12,13 as well as specific interventions in critical infants and children. We anticipate interesting comparisons of care and outcomes with these data sets. Given the complexity and severity of diseases captured in the CHND, the overall mortality was surprisingly low and unexpected death was, in fact, rare. However, infants who ultimately died frequently had long LOS and high resource utilization. In addition, survivors in this cohort often experienced high resource utilization both during and after their hospitalization, often had ongoing medical issues even at the time of discharge and are expected to be at high risk for adverse early childhood outcomes, underscoring the importance of a better understanding of their illnesses and the need to develop mechanisms to improve their care. One area of great interest for the potential of improving outcomes is understanding variability in practice across sites. In this report, although only a few selected measures are shown, inter-center variation in care, resource utilization and outcomes were striking. Although we report unadjusted analyses, it remains likely that much of the observed variability is not due to differences in case-mix or illness severity but due to systematic practice differences across sites. Although we can only speculate on which factors may drive the practice differences, they have occurred in the historical context of limited multicenter collaborations amongst clinicians in regional NICUs and few standards of care for many of the reported diseases, including use of therapies with unproven risk/benefit data and significant costs. Prior studies have highlighted inter-center variations of NICU care, their associations with outcomes14–17 as well as improved outcomes after adopting evidence-based disease-specific care standards.18 Thus, the findings highlighted in this report suggest substantial Journal of Perinatology (2014), 582 – 586

Children’s Hospitals Neonatal Database K Murthy et al

586 opportunities within regional NICUs1 to improve outcomes and to optimize resource utilization by identification and implementation of best practices. There are inherent limitations to all data sets, including CHND. Data were captured only on infants referred to participating NICUs, and thus, these results may well not be generalizable to infants who were not referred. In addition, limited data were obtained regarding medical information before referral or after CHND hospital discharge for the infants transferred to other institutions. Although short-term, inpatient outcomes are quantified, post-discharge outcomes including neuro-developmental sequelae remain unmeasured. In summary, this report describes a large contemporary sample of infants with selected, uncommon illnesses and their associated short-term outcomes who are referred to regional children’s hospitals’ NICUs. We intend to focus future efforts of this collaborative on (a) describing disease-specific characteristics, interventions and outcomes in neonatal populations that have not previously been well characterized; (b) defining inter-site variations in care and outcomes that will provide foundations for clinical trials and ‘best practice’ implementation; (c) predicting disease-specific outcomes to aid in counseling families and provide information about expected resource utilization for various populations of infants referred to large regional NICUs; and (d) continuing quality improvement activities that address these infants’ unique and complex issues in health-care delivery (for example, perioperative care). Mechanisms to capture early childhood outcomes for selected populations are being explored.

CONFLICT OF INTEREST JMA received a portion of salary support from CHA in 2011–2013 for the development and maintenance of the database that was analyzed for this study; JRE received a stipend from CHA in 2012; KM received a portion of salary support from CHA in 2012. The remaining authors declare no conflict of interest.

ACKNOWLEDGEMENTS The CHNC (http://www.thechnc.org) partnered with Children’s Hospital Association, Inc. (Washington, DC and Overland Park, KS) in order to design, launch and maintain the CHND. Inquiries and information can be obtained from Kate Conrad, Vice President ([email protected]) We are indebted to the following institutions that serve the infants and their families, and these institutions also have invested in and continue to participate in the Children’s Hospitals Neonatal Database (CHND). Site sponsors during the study period are indicated in brackets: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

Children’s Healthcare of Atlanta, Atlanta, GA (Francine Dykes, Anthony Piazza) Children’s Healthcare of Atlanta at Scottish Rite (Gregory Sysyn) Children’s Hospital of Alabama, Birmingham, AL (Carl Coghill) Le Bonheur Children’s Hospital, Memphis, TN (Ramasubbareddy Dhanireddy) Children’s Hospital Boston, Boston, MA (Anne Hansen) Ann & Robert H. Lurie Children’s Hospital of Chicago, Chicago, IL (Karna Murthy) Nationwide Children’s Hospital, Columbus, OH (Kristina Reber) Children’s Medical Center, Dallas, TX (Rashmin Savani, Luc Brion) Children’s Hospital Colorado, Aurora, CO (Theresa Grover) Children’s Hospital of Michigan, Detroit, MI (Girija Natarajan) Cook Children’s Health Care System, Fort Worth, TX (Jonathan Nedrelow, Annie Chi) Texas Children’s Hospital, Houston, TX (Yvette Johnson) Children’s Mercy Hospitals & Clinics, Kansas City, MO (Eugenia Pallotto) Arkansas Children’s Hospital, Little Rock, AR (Becky Rodgers) Children’s Hospital Los Angeles, Los Angeles, CA (Lisa Kelly*, Steven Chin) Children’s Hospital & Research Center Oakland, Oakland, CA (David Durand, Jeanette Asselin) The Children’s Hospital of Philadelphia, Philadelphia, PA (Jacquelyn Evans, Michael Padula) Children’s Hospital of Pittsburgh of UPMC, Pittsburgh, PA (Beverly Brozanski) St. Louis Children’s Hospital, St Louis, MO (Joan Rosenbaum, Tasmin Najaf) All Children’s Hospital, St. Petersburg, FL (Victor McKay)

Journal of Perinatology (2014), 582 – 586

21. 22. 23. 24. 25. 26. 27.

Rady Children’s Hospital, San Diego, CA (Mark Speziale) Children’s National Medical Center, Washington, DC (Billie Short) AI DuPont Hospital for Children, Wilmington, DE (Kevin Sullivan) Primary Children’s Medical Center, Salt Lake City, UT (Donald Null) Children’s Hospital of Wisconsin, Milwaukee, WI (Michael Uhing) Children’s Hospital of Omaha (Lynne Willett, John Grebe) Florida Hospital for Children (Rajan Wadhawan)

*deceased

DISCLAIMER Individual disclosures are listed below. CHA had no input in the design of the project, the aims or the decision whether to submit this manuscript for publication. Staff from CHA (IZ and Cary Thurm) led the statistical analyses presented in this report.

REFERENCES 1 Committee on Fetus & Newborn, American Academy of Pediatrics. Levels of neonatal care. Pediatrics 2012; 130(3): 587–597. 2 Stark AR. Committee on Fetus and Newborn, American Academy of Pediatrics. Levels of neonatal care. Pediatrics 2004; 114(5): 1341–1347. 3 Padula MA, Grover TR, Brozanski B, Zaniletti I, Nelin LD, Asselin JM et al. Therapeutic interventions and short-term outcomes for infants with severe bronchopulmonary dysplasia born ato32 weeks’ Gestation. J Perinatol 2013; 33(11): 877–881. 4 Natarajan G, Johnson YR, Brozanski B, Farrow KN, Zaniletti I, Padula M et al. Postnatal weight gain in preterm infants with severe bronchopulmonary dysplasia. Am J Perinatol. e-pub ahead of print May 2013; doi:10.1055/s-00331345264. 5 Stoll BJ, Hansen NI, Bell EF, Shankaran S, Laptook AR, Walsh MC et al. Neonatal outcomes of extremely preterm infants from the NICHD neonatal research network. Pediatrics 2010; 126(3): 443–456. 6 Horbar JD, Carpenter JH, Badger GJ, Kenny MJ, Soll RF, Morrow KA et al. Mortality and neonatal morbidity among infants 501 to 1500 grams from 2000 to 2009. Pediatrics 2012; 129(6): 1019–1026. 7 Lahra MM, Beeby PJ, Jeffrey HE. Intrauterine inflammation, neonatal sepsis, and chronic lung disease: a 13-year hospital cohort study. Pediatrics 2009; 123: 1314–1319. 8 Isayama T, Lee SK, Mori R, Kusuda S, Fujimura M, Ye XY et al. Comparison of mortality and morbidity of very low birth weight infants between Canada and Japan. Pediatrics 2012; 130(4): e957–e965. 9 Spitzer AR, Ellsbury DL, Handler D, Clark RH. The pediatrix babysteps data warehouse and the pediatrix qualitysteps improvement project system--tools for ‘‘meaningful use’’ in continuous quality improvement. Clin Perinatol 2010; 37(1): 49–70. 10 Jones HP, Karuri S, Cronin CMG, Ohlsson A, Peliowski A, Synnes A et al. Actuarial survival of a large Canadian cohort of preterm infants. BMC Pediatr 2005; 5: 40. 11 Skarsgard ED, Claydon J, Bouchard S, Kim PCW, Lee SK, Laberge J-M et al. Canadian pediatric surgical network: a population-based pediatric surgery network and database for analyzing surgical birth defects. The first 100 cases of gastroschisis. J Ped Surg 2008; 43(1): 30–34. 12 Pfister RH, Bingham P, Edwards EM, Horbar JD, Kenny MJ, Inder T et al. The Vermont Oxford Neonatal Encephalopathy Registry: rationale, methods, and initial results. BMC Pediatr 2012; 12: 84. 13 Doyle NM, Lally KP. The CDH study group and advances in the clinical care of the patient with congenital diaphragmatic hernia. Semin Perinatol 2004; 28(3): 174–184. 14 Baird R, Eeson G, Safavi A, Puligandla P, Laberge J-M, Skarsgard ED et al. Institutional practice and outcome variation in the management of congenital diaphragmatic hernia and gastroschisis in Canada: a report from the Canadian pediatric surgery network. J Ped Surg 2011; 46(5): 801–807. 15 Lagatta J, Clark R, Spitzer A. Clinical predictors and institutional variation in home oxygen use in preterm infants. J Pediatr 2012; 160(2): 232–238. 16 Aliaga S, Boggess K, Ivester TS, Price WA. Influence of neonatal practice variation on outcomes of late preterm birth. Am J Perinatol. e-pub ahead of print September 2013; doi:10.1055/s-0033-1356484. 17 Alleman BW, Bell EF, Li L, Dagle JM, Smith PB, Ambalavanan N et al. Individual and center-level factors affecting mortality among extremely low birth weight infants. Pediatrics 2013; 132(1): e175–e184. 18 Tracy ET, Mears SE, Smith PB, Danko ME, Diesen DL, Fisher KA et al. Protocolized approach to the management of congenital diaphragmatic hernia: benefits of reducing variability in care. J Ped Surg 2010; 45(6): 1343–1348.

& 2014 Nature America, Inc.