Association Between Neonatal Intensive Care Unit Type and Quality of Care in Moderate and Late Preterm Infants

Elizabeth G Salazar; Sara C Handley; Lucy T Greenberg; Erika M Edwards; Scott A Lorch

doi:10.1001/jamapediatrics.2022.5213

. 2023 Jan 17;177(3):278–285. doi: 10.1001/jamapediatrics.2022.5213

Association Between Neonatal Intensive Care Unit Type and Quality of Care in Moderate and Late Preterm Infants

Elizabeth G Salazar ^1,^2,^✉, Sara C Handley ^1,^2,³, Lucy T Greenberg ^4,⁵, Erika M Edwards ^4,^5,⁶, Scott A Lorch ^1,^2,³

¹Division of Neonatology, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania

²Leonard Davis Institute of Health Economics, Philadelphia, Pennsylvania

³Perelman School of Medicine at the University of Pennsylvania, Philadelphia

⁴Vermont Oxford Network, Burlington, Vermont

⁵Department of Mathematics and Statistics, The University of Vermont, Burlington

⁶Department of Pediatrics, The Robert Larner, MD College of Medicine, The University of Vermont, Burlington

Accepted for Publication: October 18, 2022.

Published Online: January 17, 2023. doi:10.1001/jamapediatrics.2022.5213

^✉

Corresponding Author: Elizabeth G. Salazar, MD, Division of Neonatology, The Children’s Hospital of Philadelphia, 3401 Civic Center Blvd, 2NW, Philadelphia, PA 19103 (salazare@chop.edu).

Author Contributions: Drs Edwards and Greenberg had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Salazar and Handley served as co–first authors and contributed equally to the work.

Concept and design: Salazar, Handley, Edwards, Lorch.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Salazar.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Greenberg, Edwards.

Obtained funding: Salazar.

Supervision: Edwards, Lorch.

Conflict of Interest Disclosures: Dr Handley reported receiving grants from the American Academy of Pediatrics Neonatal Resuscitation Program outside the submitted work. Dr Edwards reported receiving grants from the Vermont Oxford Network during the conduct of the study. No other disclosures were reported.

Funding/Support: This study was supported by grant T32 HL09854 from the National Institutes of Health (Dr Salazar).

Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 2.

Additional Contributions: The authors thank their colleagues who submit data to the Vermont Oxford Network on behalf of infants and their families.

Additional Information: The centers contributing data to this study are listed in eTable 9 in Supplement 1.

^✉

Corresponding author.

PMCID: PMC9857785 PMID: 36648939

Key Points

Question

Does moderate and late preterm and extremely and very preterm infant care quality vary by neonatal intensive care unit (NICU) type?

Findings

In this cohort study that analyzed 376 219 moderate and late preterm and 57 595 extremely and very preterm infants, moderate and late preterm infants had lower composite quality scores at NICUs with more subspecialty capabilities. No difference in composite quality scores by NICU type for extremely and very preterm infants was found.

Meaning

Moderate and late preterm infants may receive higher-quality care at NICUs with fewer complex services available.

This cohort study uses national data to evaluate the association between neonatal intensive care unit (NICU) type and care quality in moderate and late preterm infants.

Abstract

Importance

A higher level of care improves outcomes in extremely and very preterm infants, yet the impact of neonatal intensive care unit (NICU) level on moderate and late preterm (MLP) care quality is unknown.

Objective

To examine the association between NICU type and care quality in MLP (30-36 weeks’ gestation) and extremely and very preterm (25-29 weeks’ gestation) infants.

Design, Setting, and Participants

This cohort study was a prospective analysis of 433 814 premature infants born in 465 US hospitals between January 1, 2016, and December 31, 2020, without anomalies and who survived more than 12 hours and were transferred no more than once. Data were from the Vermont Oxford Network all NICU admissions database.

Exposures

NICU types were defined as units with ventilation restrictions without surgery (type A with restrictions, similar to American Academy of Pediatrics [AAP] level 2 NICUs), without surgery (type A) and with surgery not requiring cardiac bypass (type B, similar to AAP level 3 NICUs), and with all surgery (type C, similar to AAP level 4 NICUs).

Main Outcomes and Measures

The primary outcome was gestational age (GA)–specific composite quality measures using Baby-Measure of Neonatal Intensive Care Outcomes Research (Baby-MONITOR) for extremely and very preterm infants and an adapted MLP quality measure for MLP infants. Secondary outcomes were individual component measures of each scale. Composite scores were standardized observed minus expected scores, adjusted for patient characteristics, averaged, and expressed with a mean of 0 and SD of 1. Between May 2021 and October 2022, Kruskal-Wallis tests were used to compare scores by NICU type.

Results

Among the 376 219 MLP (204 181 [54.3%] male, 172 038 [45.7%] female; mean [SD] GA, 34.2 [1.7] weeks) and 57 595 extremely and very preterm (30 173 [52.4%] male, 27 422 [47.6%] female; mean [SD] GA, 27.7 [1.4] weeks) infants included, 6.6% received care in type A NICUs with restrictions, 29.3% in type A NICUs without restrictions, 39.7% in type B NICUs, and 24.4% in type C NICUs. The MLP infants had lower MLP-QM scores in type C NICUs (median [IQR]: type A with restrictions, 0.4 [−0.1 to 0.8]; type A, 0.4 [−0.4 to 0.9]; type B, 0.1 [−0.7 to 0.7]; type C, −0.7 [−1.6 to 0.4]; P < .001). No significant differences were found in extremely and very preterm Baby-MONITOR scores by NICU type. In type C NICUs, MLP infants had lower scores in no extreme length of stay and change-in-weight z score.

Conclusions and Relevance

In this cohort study, composite quality scores were lower for MLP infants in type C NICUs, whereas extremely and very preterm composite quality scores were similar across NICU types. Policies facilitating care for MLP infants at NICUs with less complex subspecialty services may improve care quality delivered to this prevalent, at-risk population.

Introduction

Comprising more than 80% of all preterm births in the US, moderate and late preterm (MLP) infants are a prevalent population in the neonatal intensive care unit (NICU).^1,2,3 The risk of morbidity and mortality in MLP infants, defined in this study as infants born at 30 to 36 weeks’ gestation, is unique yet understudied.² Moderate and late preterm infants have lower risks than extremely and very preterm infants but higher risks than full-term infants.^4,5,6 The American Academy of Pediatrics (AAP) guidelines on neonatal levels of care state that well newborn nurseries (level 1) should be capable of caring for infants born between 35 and 37 weeks’ gestation, and special care nurseries (level 2) should be capable of caring for infants born at 32 weeks’ gestation or greater.⁷ Thus, MLP infants may be born at hospitals with varied neonatal capabilities.

Understanding where MLP infants receive optimal care is an opportunity to improve outcomes for this population. Infants born at 23 to 37 weeks’ gestation have lower mortality at higher-level NICUs; however, the association of NICU level with MLP infant care quality and outcomes specifically was not studied.^{8,9,10,11,12,13,14} The objective of this study was to evaluate the association of NICU type with the quality of care delivered to the MLP population who require NICU admission compared with extremely and very preterm infants. We hypothesized that type A and B NICUs may provide higher-quality care to the MLP population.

Methods

Population

In this cohort study, we used data from January 1, 2016, to December 31, 2020, from the all NICU admissions database maintained by the Vermont Oxford Network (VON), a voluntary worldwide community dedicated to improving the quality, safety, and value of neonatal care, to conduct a prospective analysis of care quality measures by NICU type for infants born in US hospitals (eTable 9 in Supplement 1).¹⁵ The database includes (1) very low-birth-weight (≤1500 g) infants who are admitted anywhere in a hospital or die anywhere in a hospital within 28 days of birth and (2) infants weighing more than 1500 g who within 28 days of birth either died anywhere in the hospital or were admitted to a NICU, where a NICU is defined as a location where continuous positive airway pressure or intermittent mechanical ventilation can be administered to infants (not including the delivery room or another location where respiratory interventions are provided briefly for stabilization).¹⁶ Vermont Oxford Network data are manually abstracted at the unit level using a prespecified data dictionary, evaluated using error checking software, and manually corrected using previously described methods.¹⁷ Units contribute data to the all NICU admissions database voluntarily.

Exclusion criteria were congenital anomalies (n = 24 825) (eTable 1 in Supplement 1), deaths in the delivery room or within 12 hours of NICU admission (n = 1213), 1 or more transfers (n = 798), and implausible values for birth weight (n = 1168), defined as birth weight less than 201 g or more than 4 SDs from the mean for gestational age (GA) and sex. Infants were divided into 2 groups by GA: MLP infants born at 30 to 36 weeks’ gestation, the primary focus of our study, and extremely and very preterm infants born at 25 to 29 weeks’ gestation, as a reference group and national validation of previous studies. Although maternal race and ethnicity data are collected by the VON, this study was not designed or a hypothesis defined with regard to how race and ethnicity mediate the association between NICU type and quality of care.

The institutional review board of The University of Vermont determined that the use of the VON research repository for this analysis was not human participants research and, thus, exempt from informed consent. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.

Outcome Variables

Our primary outcomes were GA-specific quality measures. The outcome for extremely and very preterm infants was infant care quality as assessed by the Baby-Measure of Neonatal Intensive Care Outcomes Research (Baby-MONITOR), an established measure used for very low-birth-weight infants.^18,19 For MLP infants, the outcome was an MLP quality measure (MLP-QM) based on Baby-MONITOR and adapted to the MLP population.²⁰ Secondary outcomes were the individual components of the GA-specific quality measures.²⁰ The Baby-MONITOR measure consists of infant-level process measures (no admission hypothermia, no health care–associated infection, discharge on any human milk, antenatal steroid exposure, timely retinal examination), and outcome measures (survival to hospital discharge, no nonsurgically induced pneumothorax, no chronic lung disease, greater than median growth velocity).¹⁹ The MLP-QM consists of infant-level process measures (no admission hypothermia, no health care–associated infection, discharge on any human milk) and outcome measures (survival to hospital discharge, no nonsurgically induced pneumothorax, no oxygen at 28 days of life or at discharge, change-in-weight z score, no extreme length of stay). The development of Baby-MONITOR and the MLP-QM adaptation have been described in prior work.^19,20,21

Quality scores were calculated, adjusted, and standardized as previously described.^21,22,23,24 The Baby-MONITOR and MLP-QM scores were adjusted for relevant infant characteristics (GA, sex, 5-minute Apgar score, receipt of prenatal care, inborn or outborn status, small for GA, multiple gestation, and cesarean delivery) as in prior work^{21,22,25,26,27,28,29,30,31,32} (eTable 2 in Supplement 1). Oxygen use at 28 days of life and chronic lung disease were adjusted for center elevation.³³ Extreme length of stay was defined as total hospital stay greater than the 95th percentile for the predicted value and included adjustment for birth weight, ventilation status, respiratory distress syndrome, surgery (other than for retinopathy of prematurity [ROP]), 1-minute Apgar score, small for GA, reason for transfer, cesarean delivery, inborn or outborn status, sex, and prenatal care. Measures were calculated to appropriately attribute events for infants transferred between hospitals by excluding infants if transferred out by day 3 from the growth, infection, and human milk measures or if admitted after day 3 for the oxygen measures. The standardized scores for each component of the composite measure were equally weighted, averaged, and rescaled to derive the composite quality score. The composite score is a modified z score with an SD of 1. Higher scores indicate higher care quality. Analysis was performed at the admission level, and infants could contribute to the score at multiple hospitals. Further details regarding the quality score methodology used in this analysis are described in prior work.^19,21

Center Definitions and Characteristics

Using member surveys, VON classifies NICUs into 4 types based on their ventilation restrictions and surgical capabilities. Type A NICUs with restrictions on assisted ventilation are those that do not perform neonatal surgery (omphalocele repair, ventriculoperitoneal shunt, tracheoesophageal fistula and/or reanastomosis, meningomyelocele repair, patent ductus arteriosus ligation, cardiac catheterization, or cardiac surgery requiring bypass) and transfer infants to another center for assisted ventilation based on infant characteristics, such as GA or duration of ventilation. Type A NICUs without ventilation restrictions are those that do not perform neonatal surgery. Type B NICUs have no ventilation restrictions and perform neonatal surgery, except cardiac surgery requiring bypass. Type C NICUs have no ventilation restrictions and perform all surgeries.³⁴ While the AAP classification system includes GA criteria preventing direct comparison, level 2 units have similar resource availability as type A NICUs with ventilation restrictions, level 3 units are similar to type A NICUs without ventilation restrictions and type B NICUs, and level 4 units are type C NICUs.⁷ The VON survey includes the following hospital characteristics: ownership, obstetric level of care, the number of intensive and intermediate care beds, the annual number of hospital births, and available respiratory services. The VON data were used to derive the number of inborn admissions (all preterm [<37 weeks’ gestation], MLP [30-36 weeks’ gestation], and extremely and very preterm [<30 weeks’ gestation]) and the proportion of NICU admissions of extremely and very preterm infants (the number of infants born at 25-29 weeks’ gestation divided by the total number of admissions).

Statistical Analysis

We performed the data analysis between May 2021 and October 2022. For each hospital, we derived scores for Baby-MONITOR and each of its components based on extremely and very preterm infants and scores for MLP-QM and each of its components based on MLP infants. To receive scores, hospitals needed data on at least 1 infant for each measure. We examined the distributions of composite quality scores and individual components by unit type. Kruskal-Wallis tests were used to compare differences in quality score and hospital characteristics by unit type. Given heterogeneity in MLP infant outcomes, analyses were stratified by GA subgroup (30-31, 32-34, and 35-36 weeks). We performed linear regression models, unadjusted and adjusted for hospital ownership, level of obstetric service, log-transformed preterm volume (25-36 weeks’ gestation), and proportion born at 25 to 29 weeks’ gestation, to assess the association between unit type and quality score in the MLP and extremely and very preterm populations. Given a strong correlation with log-transformed preterm volume, number of NICU beds, total births, and total admissions were not included. The threshold for significance was set at a 2-sided P ≤ .01. R, version 4.021 software (R Foundation for Statistical Computing) was used for the statistical analysis.

Results

Our cohort included 376 219 MLP (54.3% males, 45.7% females) and 57 595 extremely and very preterm infants (52.4% males, 47.6% females) cared for at 465 US hospitals born between January 1, 2016, and December 31, 2020. Mean (SD) GA for MLP infants was 34.2 (1.7) weeks and for extremely and very term infants, 27.7 (1.4) weeks. Table 1 compares MLP and extremely and very preterm infant characteristics. The MLP infants had lower rates of cesarean delivery (60.6% vs 73.3% for extremely and very preterm infants) and respiratory support after resuscitation (59.0% vs 99.4%). For both MLP and extremely and very preterm infants, a higher percentage received ventilation after initial resuscitation in type C NICUs compared with infants of the same GA at type A units (MLP type A with restrictions, 6.6%; MLP type C, 15.5%; extremely and very preterm type A with restrictions, 52.7%; extremely and very preterm type C, 70.9%). In the cohort, 6.6% received care in type A NICUs with restrictions, 29.3% in type A NICUs without resrictions, 39.7% in type B NICUs, and 24.4% in type C NICUs.

Table 1. Characteristics of Moderate and Late Preterm (MLP) and Extremely and Very Preterm Infant Cohorts by Type of Neonatal Intensive Care Unit (NICU)^a.

Variable	No. (%)
	MLP (30-36 wk GA) infants					Extremely and very preterm (25-29 wk GA) infants
	All NICUs	Type A with restrictions	Type A	Type B	Type C	All NICUs	Type A with restrictions	Type A	Type B	Type C
No. of infants	376 219 (100.0)	27 338 (7.3)	113 228 (30.1)	149 206 (39.7)	86 447 (23.0)	57 595 (100.0)	1392 (2.4)	13 878 (24.1)	23 018 (40.0)	19 307 (33.5)
Birth weight, g, mean (SD)	2225 (575)	2289 (545)	2244 (570)	2227 (573)	2177 (588)	1036 (275)	1137 (252)	1058 (274)	1033 (277)	1018 (274)
GA, wk, mean (SD)	34.2 (1.7)	34.5 (1.6)	34.3 (1.7)	34.2 (1.7)	34.1 (1.8)	27.7 (1.4)	28.3 (1.3)	27.9 (1.4)	27.7 (1.4)	27.6 (1.4)
SGA	41 900 (11.1)	2916 (10.7)	12 615 (11.1)	16 081 (10.8)	10 288 (11.9)	4976 (8.6)	61 (4.4)	1079 (7.8)	1996 (8.7)	1840 (9.5)
Sex
Female	172 038 (45.7)	429 (45.5)	51 987 (45.9)	68 063 (45.6)	39 559 (45.8)	27 422 (47.6)	712 (51.1)	6659 (48.0)	11044 (48.0)	9007 (46.6)
Male	204 181 (54.3)	14 909 (54.5)	61 241 (54.1)	81 143 (54.4)	46 888 (54.2)	30 173 (52.4)	680 (48.9)	7219 (52.0)	11 974 (52.0)	10 300 (53.4)
Apgar score at 5 min, mean (SD)	8.3 (1.2)	8.4 (1.1)	8.4 (1.1)	8.3 (1.2)	8.1 (1.4)	7.1 (1.8)	7.3 (1.8)	7.2 (1.8)	7.2 (1.8)	6.9 (1.9)
Cesarean birth	227 884 (60.6)	15 671 (57.3)	68 308 (60.3)	89 961 (60.3)	53 944 (62.4)	42 210 (73.3)	953 (68.5)	10 192 (73.4)	16 976 (73.8)	14 089 (73.0)
Multiple gestation	90 892 (24.2)	6071 (22.2)	27 161 (24.0)	36 245 (24.3)	21 415 (24.8)	13 637 (23.7)	297 (21.3)	3183 (22.9)	5418 (23.5)	4739 (24.5)
Any prenatal care	363 520 (96.9)	26 366 (96.6)	109 439 (96.9)	144 557 (97.1)	83 158 (96.6)	54 991 (95.8)	1313 (94.5)	13 248 (95.8)	22 103 (96.2)	18 327 (95.5)
Inborn admissions	334 704 (89.0)	25 538 (93.4)	106 509 (94.1)	134 701 (90.3)	67 956 (78.6)	48 987 (85.1)	1217 (87.4)	12 766 (92.0)	20 520 (89.1)	14 484 (75.0)
Ventilation after initial resuscitation (conventional or high frequency)	43 751 (11.6)	1794 (6.6)	10 573 (9.3)	18 108 (12.1)	13 352 (15.5)	39 245 (68.2)	733 (52.7)	8827 (63.6)	16 010 (69.6)	13 679 (70.9)
Any respiratory support after initial resuscitation	221 640 (59.0)	13 914 (51.0)	64 720 (57.2)	90 289 (60.5)	52 803 (61.2)	57 170 (99.4)	1381 (99.4)	13 763 (99.3)	22 904 (99.6)	19 127 (99.3)

Open in a new tab

Abbreviations: GA, gestational age; SGA, small for gestational age.

^{^a}

N = 433 814. NICU levels were defined as those with ventilation restrictions and without surgery (type A with restrictions), without surgery (type A), with surgery except surgery requiring cardiac bypass (type B), and with all surgery (type C). Characteristics are at the patient level, while the analysis was at the admission level.

Table 2 compares hospital characteristics by unit type. Of the 465 units, 74 (16%) were type A with restrictions, 176 (38%) were type A, 151 (32%) were type B, and 64 (14%) were type C. Compared with type C units, type A units were more likely to be investor owned (type A, 26%; type C, 8%), were less likely to have level 3 obstetric services (type A, 66%; type C, 78%), and had fewer annual inborn admissions and preterm admissions.

Table 2. Hospital and Neonatal Intensive Care Unit (NICU) Characteristics by Unit Type.

Variable	NICU type (N = 465)
Variable	Type A with restrictions (n = 74)	Type A (n = 176)	Type B (n = 151)	Type C (n = 64)
Hospital ownership, No. (%)
Government	2 (3)	8 (5)	12 (8)	6 (9)
Nongovernment, not for profit	55 (74)	122 (69)	111 (74)	50 (78)
Investor owned, for profit	16 (22)	45 (26)	28 (19)	5 (8)
No. of hospital births, median (IQR)	1551 (988-2162)	1973 (1441-3007)	2696 (1897-3750)	2673 (1635-4206)
Level of obstetric service, No. (%)
1	3 (4)	1 (1)	0	0
2	57 (77)	59 (34)	10 (7)	1 (2)
3	12 (16)	116 (66)	138 (91)	50 (78)
No obstetric service	2 (3)	0	3 (2)	13 (20)
No. of beds, median (IQR)
Intensive care	8 (3-12)	16 (12-23)	26 (18-40)	54 (40-73)
Intermediate care	4 (0-8)	3 (0-11)	8 (0-18)	11 (0-29)
Annual No. of admissions, median (IQR)
Inborn	154 (110-233)	252 (191-366)	425 (279-618)	560 (298-824)
Inborn preterm (<37 wk)	77 (52-95)	128 (93-186)	204 (131-299)	308 (94-404)
25-36 wk	80 (59-114)	140 (98-203)	219 (145-321)	348 (232-485)
30-36 wk	76 (57-108)	126 (90-181)	191 (126-279)	297 (194-390)
25-29 wk	3 (2-6)	15 (9-23)	30 (17-46)	65 (40-97)
Percentage born at 25-29 wk gestation, median (IQR)^a	4 (3-6)	10 (8-13)	13 (11-15)	18 (16-22)
Available respiratory services, No. (%)
Nasal CPAP	71 (96)	174 (99)	151 (100)	64 (100)
Conventional assisted ventilation	71 (96)	175 (99)	151 (100)	64 (100)
High-frequency ventilation	14 (19)	152 (86)	149 (99)	64 (100)

Open in a new tab

Abbreviation: CPAP, continuous positive airway pressure.

^{^a}

Percentage born at 25 to 29 weeks’ gestation calculated using the number of infants born at 25 to 29 weeks divided by number of infants born at 25 to 36 weeks times 100.

The MLP infants had lower MLP-QM composite scores in type C NICUs compared with all other types (median [IQR]: type A with restrictions, 0.4 [−0.1 to 0.8]; type A, 0.4 [−0.4 to 0.9]; type B, 0.1 [−0.7 to 0.7]; type C, −0.7 [−1.6 to 04]; P < .001) (Figure 1A). Extremely and very preterm infants had no significant difference in Baby-MONITOR composite scores by unit type (Figure 1B) (median: type A with restrictions, −0.1 [IQR, −0.7 to 0.3]; type A, −0.1 [IQR, −0.8 to 0.5]; type B, 0.1 [IQR, −0.7 to 0.8]; type C, 0.1 [IQR, −0.8 to 1.0]; P = .35).

Figure 1. — Higher scores indicate higher care quality. The horizontal bar inside the boxes indicates the median, and the lower and upper ends of the boxes are the first and third quartiles. The whiskers indicate the minimum and maximum values. Composite quality measure score distributions were compared using Kruskal-Wallis test). Baby-MONITOR indicates Baby-Measure of Neonatal Intensive Care Outcomes Research; MLP-QM, moderate and late preterm quality measure.

^aP < .001.

^bP = .35.

In unadjusted and adjusted regression analyses, type C NICUs were associated with lower MLP-QM scores compared with type A with restrictions, type A, and type B units (eTable 3 in Supplement 1). A higher percentage of infants born at 25 to 29 weeks’ gestation was associated with decreased MLP-QM scores (eTable 3 in Supplement 1). In unadjusted and adjusted regression analyses in extremely and very preterm infants, no association was seen with NICU type and Baby-MONITOR scores (eTable 4 in Supplement 1). Log-transformed preterm volume was associated with improved Baby-MONITOR scores (eTable 4 in Supplement 1).

When stratified by GA, MLP-QM scores were lower in type C NICUs and consistent across all GA categories of 30 to 31 weeks (median: type A with restrictions, 0.4 [IQR, 0.0-0.6]; type A, 0.3 [IQR, −0.4 to 0.8]; type B, −0.1 [IQR, −0.8 to 0.6]; type C, −0.3 [IQR, −0.7 to 0.7]; P = .01), 32 to 34 weeks (median: type A with restrictions, 0.3 [IQR, −0.2 to 0.6]; type A, 0.4 [IQR, −0.1 to 0.8]; type B, 0.0 [IQR, −0.7 to 0.5]; type C, −0.8 [IQR, −1.4 to 0.6]; P < .001), and 35 to 36 weeks (median: type A with restrictions, 0.3 [IQR, −0.0 to 0.9]; type A, 0.4 [IQR, −0.2 to 0.8]; type B, 0.2 [IQR, −0.4 to 0.7]; type C, −0.9 [IQR, −1.7 to 0.0]; P < .001) (Figure 2) (eTable 5 in Supplement 1). For infants born at 30 to 31 and 32 to 34 weeks’ gestation, differences were driven by change-in-weight z score. For infants born at 35 to 36 weeks’ gestation, differences were also seen in no hypothermia, no infection, no extreme length of stay, and no oxygen at 28 days (eFigure and eTable 6 in Supplement 1).

Figure 2. — Higher scores indicate higher care quality. The horizontal bar inside the boxes indicates the median, and the lower and upper ends of the boxes are the first and third quartiles. The whiskers indicate minimum and maximum values. Composite quality measure score distributions were compared using Kruskal-Wallis test.

^aP = .01.

^bP < .001.

Figure 3 depicts quality score components by unit type for MLP and extremely and very preterm infants. The MLP infants in type C NICUs had lower scores in no extreme length of stay (median: type A with restrictions, 0.5 [IQR, 0.2-0.7]; type A, 0.4 [IQR, −0.2 to 0.7]; type B, 0.1 [IQR, −0.5 to 0.6]; type C, 0.0 [IQR, −0.7 to 0.4]; P < .001) and change-in-weight z score (median: type A with restrictions, 0.1 [IQR, −0.4 to 0.6]; type A, 0.2 [IQR, −0.4 to 0.7]; type B, 0.0 [IQR, −0.7 to 1.0]; type C, −0.3 [IQR, −1.1 to 0.3]; P = .005) (Figure 3A) (eTable 7 in Supplement 1). Extremely and very preterm infants in type C NICUs had higher scores in antenatal steroids (median: type A with restrictions, −0.8 [IQR, −1.5 to 0.0]; type A, −0.2 [IQR, −1.0 to 0.1]; type B, −0.1 [IQR, −0.7 to 0.6]; type C, 0.7 [IQR, 0.2-1.1]; P < .001) and timely ROP examination (median: type A with restrictions, −0.2 [IQR, −0.8 to 0.1]; type A, 0.0 [IQR, −0.6 to 0.3]; type B, 0.3 [IQR, −0.2 to 0.6]; type C, 0.4 [IQR, −0.1 to 0.8]; P < .001) and lower scores in growth velocity compared with other unit types (median: type A with restrictions, 0.0 [IQR, −0.3 to 0.4]; type A, 0.2 [IQR, −0.2 to 0.6]; type B, 0.0 [IQR, −0.7 to 0.6]; type C, −0.4 [IQR, −1.2 to 0.5]; P = .001) (Figure 3B) (eTable 8 in Supplement 1).

Discussion

This cohort study describes the novel finding of higher quality of care for MLP infants at type A and B NICUs after adjusting for patient characteristics.^{8,9,10,11,12,13} Improved care quality for MLP infants at units with less complex surgical services appears to be primarily driven by improved scores in no extreme length of stay and change-in-weight z score. Consistent with research in a California cohort, extremely and very preterm care quality did not vary by unit type.¹⁸ Although improved quality scores may reflect healthier infants in type A and B units, this selection bias is reduced in the risk-adjusted analysis, and stratified analyses support the importance of risk-appropriate care across the GA continuum.

Type A and B units had higher composite quality scores than type C units for MLP infants in this study. Potential explanations for this finding are lower illness severity or improved MLP-specific care in type A units. We addressed variation in patient acuity by NICU type by excluding infants with anomalies, those who died within 12 hours after birth, and those who had multiple transfers. Additionally, we adjusted for infant and maternal risk factors associated with increased postnatal illness severity and inborn and outborn status in quality score creation. Alternatively, type A and B units may have structures and processes that address the MLP infant’s unique risk profile, such as feeding issues, hyperbilirubinemia, respiratory distress syndrome, and sepsis,³⁵ that lead to more efficient care and shorter length of stay.³⁶ Staff experience with MLP infant care, such as feeding cues, may also lead to decreased length of stay.³⁶ Such hypotheses are consistent with our findings that decreases in extreme length of stay and better growth significantly contributed to better overall composite quality scores in type A and B units. Additionally, these units may have more refined experience with noninvasive respiratory and surfactant administration in the MLP population, interventions that would decrease the need for transfer to a type C unit. The hospital-level regression analysis suggests that resources associated with type C NICUs, such as more complex surgical capabilities or extremely and very preterm infant volume, may be less relevant for disease processes common to MLP infants. Instead, resources such as hospital ownership and patient acuity, as measured by the percentage of extremely and very preterm infants, may be more relevant. These findings highlight the need to identify structures and processes specific to MLP care in type A and B units and consider translating these practices to type C units to improve MLP care quality.

Consistent with prior work in California,¹⁸ this study did not find a significant difference in care quality delivered by unit type for extremely and very preterm infants. Potential explanations for this finding include effective regionalization or equal weighting of the outcomes in Baby-MONITOR given strong prior evidence supporting the association between improved mortality for this population at higher level centers.^9,18 We found that type C NICUs had higher scores in process measures, such as antenatal steroids and timely ROP examination, and lower scores in outcomes measures, such as growth velocity, differences that may reflect selection bias inadequately addressed by risk adjustment.

The finding of improved quality of care at units with less complex surgical capabilities for MLP infants has broad implications for the organization and structure of neonatal care delivery. This study supports risk-appropriate care, a key tenet of regionalization, specifically for lower risk infants. Providing risk-appropriate care for lower risk infants at type A and B units may facilitate better care for high- and low-risk infants admitted to the NICU by allowing additional space and resources at units for higher risk, ill infants. Leveraging regionalization with the MLP population in mind and studying ways to optimize the delivery of risk-appropriate care may improve outcomes.

Limitations

This study has several limitations. First, although the quality scores were adjusted for patient characteristics, adjustment for illness acuity may be inadequate. To mitigate this potential bias, we excluded infants with anomalies, those who died within the first 12 hours after birth, and those who had multiple transfers and still found differences among NICU types. The persistent differences among NICU types when stratified by GA also support adequate control for illness acuity, though additional differences in infants born at 35 to 36 weeks’ gestation may be associated with an indication for NICU admission. Second, this study used a recently published adapted measure for MLP care quality.¹⁹ Notably, novel components in the MLP-QM, including no extreme length of stay and change-in-weight z score, revealed significant variation by NICU type. While the MLP-QM composite score allowed for efficient assessment of MLP care quality by unit type, future evaluation of the association of unit type and MLP care quality with a formally validated MLP measure (eg, via a Delphi process) will ensure inclusion of key components of MLP care, such as neuroimaging for more premature infants. Third, transfers may lead to lower quality scores for type A and B units; however, quality scores were adjusted for inborn and outborn status. Fourth, this study used a convenience sample of hospitals contributing to the VON all NICU admissions database. While there may be a bias in centers that opt to contribute to this database, we included more than 450 centers of varying types across the US. Fifth, although our adjusted models accounted for some hospital characteristics, future analyses could consider additional factors, such as percentage of Medicaid-insured patients, nursing staff experience, and additional clinical data to account for patient illness severity.

Despite these limitations, this study uses a large, national data set of NICU admissions from multiple centers and reveals the novel finding of improved quality of care for MLP infants at type A and B units. This work highlights the importance of risk-appropriate care for lower risk infants and ongoing need to assess MLP infant care quality. Further research to identify the processes at units with fewer cardiac and surgical capabilities that contribute to improved outcomes in the MLP population is warranted.

Conclusions

In a national cohort of premature infants, type A and B NICUs were associated with higher care quality in MLP infants, with significant differences in no extreme length of stay and change-in-weight z score by unit type. There was no association between unit type and extremely and very preterm infant quality of care. Further identification of processes specific to units with less subspecialized care that lead to improved MLP infant care quality may facilitate dissemination of these processes to other types of NICUs.

Supplement 1.

eFigure. Component Quality Scores by Neonatal Intensive Care Unit Type Withing Gestational Age Strata for Moderate and Late Preterm Infants

eTable 1. Congenital Anomalies as Defined by the Vermont Oxford Network

eTable 2. Risk Adjustment Component Covariates for Quality Measures

eTable 3. Linear Regression of Moderate and Late Preterm Quality Measure Scores by Neonatal Intensive Care Unit (NICU) Type While Controlling for Hospital Characteristics

eTable 4. Linear Regression of Baby-Measure of Neonatal Intensive Care Outcomes Research Scores by Neonatal Intensive Care Unit (NICU) Type While Controlling for Hospital Characteristics

eTable 5. Moderate and Late Preterm Quality Measure (MLP-QM) Scores by Neonatal Intensive Care Unit Type Within Gestational Age Strata

eTable 6. Component Quality Scores by Neonatal Intensive Care Unit Type Within Gestational Age (GA) Strata for Moderate and Late Preterm Infants

eTable 7. Component Quality Scores of Moderate and Late Preterm (MLP) Infants by Neonatal Intensive Care Unit Type

eTable 8. Component Quality Scores of Extremely and Very Preterm Infants by Neonatal Intensive Care Unit Type

eTable 9. Vermont Oxford Network Contributing Centers

Click here for additional data file.^{(505.3KB, pdf)}

Supplement 2.

Data Sharing Statement

Click here for additional data file.^{(14.8KB, pdf)}

References

1.Martin JA, Hamilton BE, Osterman MJK, Driscoll AK. Births: final data for 2019. Natl Vital Stat Rep. 2021;70(2):1-51. [PubMed] [Google Scholar]
2.Escobar GJ, McCormick MC, Zupancic JAF, et al. Unstudied infants: outcomes of moderately premature infants in the neonatal intensive care unit. Arch Dis Child Fetal Neonatal Ed. 2006;91(4):F238-F244. doi: 10.1136/adc.2005.087031 [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Hibbard JU, Wilkins I, Sun L, et al. ; Consortium on Safe Labor . Respiratory morbidity in late preterm births. JAMA. 2010;304(4):419-425. doi: 10.1001/jama.2010.1015 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Teune MJ, Bakhuizen S, Gyamfi Bannerman C, et al. A systematic review of severe morbidity in infants born late preterm. Am J Obstet Gynecol. 2011;205(4):374.e1-374.e9. doi: 10.1016/j.ajog.2011.07.015 [DOI] [PubMed] [Google Scholar]
5.Stewart DL, Barfield WD; Committee on Fetus and Newborn . Updates on an at-risk population: late-preterm and early-term infants. Pediatrics. 2019;144(5):e20192760. doi: 10.1542/peds.2019-2760 [DOI] [PubMed] [Google Scholar]
6.Tomashek KM, Shapiro-Mendoza CK, Davidoff MJ, Petrini JR. Differences in mortality between late-preterm and term singleton infants in the United States, 1995-2002. J Pediatr. 2007;151(5):450-456, 456.e1. doi: 10.1016/j.jpeds.2007.05.002 [DOI] [PubMed] [Google Scholar]
7.American Academy of Pediatrics Committee on Fetus and Newborn . Levels of neonatal care. Pediatrics. 2012;130(3):587-597. doi: 10.1542/peds.2012-1999 [DOI] [PubMed] [Google Scholar]
8.Chung JH, Phibbs CS, Boscardin WJ, et al. Examining the effect of hospital-level factors on mortality of very low birth weight infants using multilevel modeling. J Perinatol. 2011;31(12):770-775. doi: 10.1038/jp.2011.29 [DOI] [PubMed] [Google Scholar]
9.Phibbs CS, Baker LC, Caughey AB, Danielsen B, Schmitt SK, Phibbs RH. Level and volume of neonatal intensive care and mortality in very-low-birth-weight infants. N Engl J Med. 2007;356(21):2165-2175. doi: 10.1056/NEJMsa065029 [DOI] [PubMed] [Google Scholar]
10.Lasswell SM, Barfield WD, Rochat RW, Blackmon L. Perinatal regionalization for very low-birth-weight and very preterm infants: a meta-analysis. JAMA. 2010;304(9):992-1000. doi: 10.1001/jama.2010.1226 [DOI] [PubMed] [Google Scholar]
11.Cifuentes J, Bronstein J, Phibbs CS, Phibbs RH, Schmitt SK, Carlo WA. Mortality in low birth weight infants according to level of neonatal care at hospital of birth. Pediatrics. 2002;109(5):745-751. doi: 10.1542/peds.109.5.745 [DOI] [PubMed] [Google Scholar]
12.Marlow N, Bennett C, Draper ES, Hennessy EM, Morgan AS, Costeloe KL. Perinatal outcomes for extremely preterm babies in relation to place of birth in England: the EPICure 2 study. Arch Dis Child Fetal Neonatal Ed. 2014;99(3):F181-F188. doi: 10.1136/archdischild-2013-305555 [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Ismail AQT, Boyle EM, Pillay T; OptiPrem Study Group . The impact of level of neonatal care provision on outcomes for preterm babies born between 27 and 31 weeks of gestation, or with a birth weight between 1000 and 1500 g: a review of the literature. BMJ Paediatr Open. 2020;4(1):e000583. doi: 10.1136/bmjpo-2019-000583 [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Lorch SA, Baiocchi M, Ahlberg CE, Small DS. The differential impact of delivery hospital on the outcomes of premature infants. Pediatrics. 2012;130(2):270-278. doi: 10.1542/peds.2011-2820 [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Horbar JD, Edwards EM, Greenberg LT, et al. Variation in performance of neonatal intensive care units in the United States. JAMA Pediatr. 2017;171(3):e164396. doi: 10.1001/jamapediatrics.2016.4396 [DOI] [PubMed] [Google Scholar]
16.Vermont Oxford Network . Manual of Operations: Part 1: Guidelines for Database Participation. Vermont Oxford Network; 2016. [Google Scholar]
17.Vermont Oxford Network . Manual of Operations, Part 1 and 2. Vermont Oxford Network; 2020. [Google Scholar]
18.Profit J, Gould JB, Bennett M, et al. The association of level of care with NICU quality. Pediatrics. 2016;137(3):e20144210. doi: 10.1542/peds.2014-4210 [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Profit J, Gould JB, Zupancic JAF, et al. Formal selection of measures for a composite index of NICU quality of care: Baby-MONITOR. J Perinatol. 2011;31(11):702-710. doi: 10.1038/jp.2011.12 [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Salazar EG, Handley SC, Greenberg LT, Edwards EM, Lorch SA. Measuring quality of care in moderate and late preterm infants. J Perinatol. 2022;42(10):1294-1300. doi: 10.1038/s41372-022-01377-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Profit J, Kowalkowski MA, Zupancic JAF, et al. Baby-MONITOR: a composite indicator of NICU quality. Pediatrics. 2014;134(1):74-82. doi: 10.1542/peds.2013-3552 [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Horbar JD, Edwards EM, Greenberg LT, et al. Racial segregation and inequality in the neonatal intensive care unit for very low-birth-weight and very preterm infants. JAMA Pediatr. 2019;173(5):455-461. doi: 10.1001/jamapediatrics.2019.0241 [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Draper DGM. Statistical analysis of performance indicators in UK higher education. J R Stat Soc. 2004;167(3):449-474. doi: 10.1111/j.1467-985X.2004.apm12.x [DOI] [Google Scholar]
24.Helkey DDD. dghrank: Construct Draper-Gittoes-Helkey Institution Rankings; 2018. Accessed December 12, 2018. https://github.com/dhelkey/dghrank
25.Phibbs CS, Bronstein JM, Buxton E, Phibbs RH. The effects of patient volume and level of care at the hospital of birth on neonatal mortality. JAMA. 1996;276(13):1054-1059. doi: 10.1001/jama.1996.03540130052029 [DOI] [PubMed] [Google Scholar]
26.Ancel PY, Goffinet F, Kuhn P, et al. ; EPIPAGE-2 Writing Group . Survival and morbidity of preterm children born at 22 through 34 weeks’ gestation in France in 2011: results of the EPIPAGE-2 cohort study. JAMA Pediatr. 2015;169(3):230-238. doi: 10.1001/jamapediatrics.2014.3351 [DOI] [PubMed] [Google Scholar]
27.Clausson B, Gardosi J, Francis A, Cnattingius S. Perinatal outcome in SGA births defined by customised versus population-based birthweight standards. BJOG. 2001;108(8):830-834. [DOI] [PubMed] [Google Scholar]
28.Tyson JE, Parikh NA, Langer J, Green C, Higgins RD; National Institute of Child Health and Human Development Neonatal Research Network . Intensive care for extreme prematurity–moving beyond gestational age. N Engl J Med. 2008;358(16):1672-1681. doi: 10.1056/NEJMoa073059 [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Thavarajah H, Flatley C, Kumar S. The relationship between the five minute Apgar score, mode of birth and neonatal outcomes. J Matern Fetal Neonatal Med. 2018;31(10):1335-1341. doi: 10.1080/14767058.2017.1315666 [DOI] [PubMed] [Google Scholar]
30.Korb D, Goffinet F, Seco A, Chevret S, Deneux-Tharaux C; EPIMOMS Study Group . Risk of severe maternal morbidity associated with cesarean delivery and the role of maternal age: a population-based propensity score analysis. CMAJ. 2019;191(13):E352-E360. doi: 10.1503/cmaj.181067 [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Luke B, Keith LG. The contribution of singletons, twins and triplets to low birth weight, infant mortality and handicap in the United States. J Reprod Med. 1992;37(8):661-666. [PubMed] [Google Scholar]
32.Chen XK, Wen SW, Yang Q, Walker MC. Adequacy of prenatal care and neonatal mortality in infants born to mothers with and without antenatal high-risk conditions. Aust N Z J Obstet Gynaecol. 2007;47(2):122-127. doi: 10.1111/j.1479-828X.2007.00697.x [DOI] [PubMed] [Google Scholar]
33.Britton JR. Altitude, oxygen and the definition of bronchopulmonary dysplasia. J Perinatol. 2012;32(11):880-885. doi: 10.1038/jp.2012.6 [DOI] [PubMed] [Google Scholar]
34.Edwards EM, Horbar JD. Variation in use by NICU types in the United States. Pediatrics. 2018;142(5):e20180457. doi: 10.1542/peds.2018-0457 [DOI] [PubMed] [Google Scholar]
35.Trembath AN, Payne AH, Colaizy TT, Bell EF, Walsh MC. The problems of moderate preterm infants. Semin Perinatol. 2016;40(6):370-373. doi: 10.1053/j.semperi.2016.05.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Whyte RK. Neonatal management and safe discharge of late and moderate preterm infants. Semin Fetal Neonatal Med. 2012;17(3):153-158. doi: 10.1016/j.siny.2012.02.004 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials