Racial/Ethnic Disparities in NeonatalIntensive Care: A Systematic ReviewKrista Sigurdson, PhD,a,b,c Briana Mitchell, BS,a,c Jessica Liu, PhD,a,c Christine Morton, PhD,d Jeffrey B. Gould, MD, MPH,a,c

Henry C. Lee, MD, MS,a,c Nicole Capdarest-Arest, MA, LIS, AHIP,e Jochen Profit, MD, MPHa,c

abstractCONTEXT: Racial and ethnic disparities in health outcomes of newborns requiring care in theNICU setting have been reported. The contribution of NICU care to disparities in outcomes isunclear.

OBJECTIVE: To conduct a systematic review of the literature documenting racial/ethnicdisparities in quality of care for infants in the NICU setting.

DATA SOURCES: Medline/PubMed, Scopus, Cumulative Index of Nursing and Allied Health, andWeb of Science were searched until March 6, 2018, by using search queries organized aroundthe following key concepts: “neonatal intensive care units,” “racial or ethnic disparities,” and“quality of care.”

STUDY SELECTION: English language articles up to March 6, 2018, that were focused on racial and/or ethnic differences in the quality of NICU care were selected.

DATA EXTRACTION: Two authors independently assessed eligibility, extracted data, and cross-checked results, with disagreements resolved by consensus. Information extracted focused onracial and/or ethnic disparities in quality of care and potential mechanism(s) for disparities.

RESULTS: Initial search yielded 566 records, 470 of which were unique citations. Title andabstract review resulted in 382 records. Appraisal of the full text of the remaining 88 records,along with the addition of 5 citations from expert consult or review of bibliographies, resultedin 41 articles being included.

LIMITATIONS: Quantitative meta-analysis was not possible because of study heterogeneity.

CONCLUSIONS:Overall, this systematic review revealed complex racial and/or ethnic disparities instructure, process, and outcome measures, most often disadvantaging infants of color,especially African American infants. There are some exceptions to this pattern and each areamerits its own analysis and discussion.

aPerinatal Epidemiology and Health Outcomes Research Unit, Division of Neonatology, Department of Pediatrics, School of Medicine, Stanford University and Lucile Packard Children’s Hospital,Palo Alto, California; bDepartment of Social and Behavioral Sciences, School of Nursing, University of California, San Francisco, San Francisco, California; cCalifornia Perinatal Quality CareCollaborative, Palo Alto, California; dCalifornia Maternal Quality Care Collaborative, Palo Alto, California; and eUniversity of California, Davis, Davis, California

Dr Sigurdson conceptualized and designed the study, ran the search strategy, drafted the initial manuscript, and revised the manuscript; Ms Mitchell conceptualizedand designed the study, ran the search strategy, assisted in drafting the initial manuscript, and revised the manuscript; Dr Liu assisted in drafting the initialmanuscript and revised the manuscript; Drs Morton, Gould, and Lee assisted in running the search strategy and revised the manuscript; Ms Capdarest-Arest designedthe search strategy, assisted in drafting the initial manuscript, and revised the manuscript; Dr Profit conceptualized and designed the study and reviewed and revisedthe manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

Ms Mitchell’s current affiliation is Department of Obstetrics, Gynecology and Reproductive Sciences, University of California, San Francisco, San Francisco, CA.

DOI: https://doi.org/10.1542/peds.2018-3114

Accepted for publication Apr 10, 2019

To cite: Sigurdson K, Mitchell B, Liu J, et al. Racial/Ethnic Disparities in Neonatal Intensive Care: A Systematic Review. Pediatrics. 2019;144(2):e20183114

PEDIATRICS Volume 144, number 2, August 2019:e20183114 REVIEW ARTICLE by guest on December 27, 2020www.aappublications.org/newsDownloaded from

The preterm birth rate is 49% higherfor black or African American womenthan for all other women in theUnited States.1 Preterm birth is linkedto increased mortality, makingpreterm birth or very low birthweight (VLBW) the leading cause ofinfant death among black infants.2 Anadditional component to racial and/or ethnic inequity in infant outcomesis differences in NICU quality of carestratified by race and/or ethnicity.3

Disparities in quality have been foundin other areas of health care,including, for instance, cancer care,4

cardiovascular care,5 and pediatriccare.6

The advantage of addressingdisparities in care delivery is thatthey are amenable to improvementthrough quality improvement (QI)methodology, with the potential forspreading solutions at scale withlong-term benefit. To facilitateapproaches for measurement,benchmarking and improvement, wehave conducted a systematic reviewof the literature to summarize racialand/or ethnic disparities in quality ofNICU care delivery. Health care isdelivered within a context of societal,institutional, and interpersonalracism that may contribute todisparate care and outcomes.

METHODS

The authors conducted a systematicreview of the literature in accordancewith the Preferred Reporting Itemsfor Systematic Reviews and Meta-Analyses guidelines.6 The searchstrategies and inclusion and exclusioncriteria were specified in advance anddocumented in a protocol.

We used Donabedian’s7 structure-process-outcome quality of careframework to review the literature onracial and/or ethnic disparity in NICUquality of care delivery. Measuringquality of care is nuanced and thisliterature cannot always isolatemeasures or disparities strictlyrelated to quality of neonatal care,

biological diversity, maternalcomorbidities, or a challengedsocioeconomic environment.Research in this area is observationaland open to interpretation. Indetermining which articles to includeor exclude, we relied on expertassessments of measures of quality8

and the principle that qualitymeasures should be malleable. Wesorted articles by structure-process-outcome according to the dominanttheme in the article and by groupingsimilar articles.

Data Sources

We searched Medline/PubMed,Scopus, Cumulative Index of Nursingand Allied Health, and Web of Scienceuntil March 6, 2018, using searchqueries organized around thefollowing key concepts: “neonatalintensive care units,” “racial or ethnicdisparities,” and “quality of care.” Thefull search strategy for the Medline/PubMed query, which also formed thebasis for the other database queries,can be found in SupplementalInformation. Additional articles werecollected on the basis ofrecommendations from experts in thefield and reviewing bibliographies ofrelevant articles.

Study Selection

According to the eligibility criteriaoutlined in our protocol, we screenedall English language articles retrievedand available in full text throughMarch 6, 2018, using the followinginclusion and exclusion criteria.

Inclusion criteria:

• Studies that reported data on thequality of NICU care inclusive of thefollowing:

○ process-of-care measures (eg,use of appropriateinterventions);

○ outcome measures insofar asthey are used as quality metrics(eg, morbidity or mortality wereused to indicate quality of care);

○ structural measures of care (eg,use of or access to health careservices insofar as this isindicative of quality);

○ patient evaluations of care (eg,patient satisfaction); or

○ direct observations of care; and

○ report data stratified by patientrace and/or ethnicity (eitherwithin or between NICUs,hospitals, or regions).

Exclusion criteria:

• Studies that reported the following:

○ only maternal or post-NICUdischarge measures of quality;

○ only outcomes that may notreflect quality of care (eg,morbidity or mortality rates notat the NICU level);

○ only data from a non-US setting,

○ data not stratified by race and/orethnicity,

○ non–English languagepublication, or

○ or nonhuman studies.

Data Extraction

Articles were divided into structure-process-outcome according to theirdominant theme, with 1 articlesummarized under both structureand process.9 For this review, wecategorize analyses of minorityserving hospitals under the structuredomain. Articles were furthercategorized thematically. Two authorsindependently assessed eligibility,extracted data, and cross-checkedresults, with disagreements resolvedby consensus. Information extractedfocused on racial and/or ethnicdisparities in quality of care andpotential mechanism(s) fordisparities. We did not performadditional bias assessment.

RESULTS

Figure 1 displays our search results.Of an initial pool of 566 potentiallyeligible records, 470 representedunique citations. After review of titles

2 SIGURDSON et al by guest on December 27, 2020www.aappublications.org/newsDownloaded from

and abstracts, we excluded a further382 records. After examination of thefull texts of the remaining 88 records,we retained 36 records for thesystematic review. Five records wereadded based on expert suggestionand review of bibliographies ofrelevant articles for a total of 41selected for inclusion.

Structure

We identified 12 articles thatexamined the role of structure orhealth systems in racial and/or ethnicdisparities in quality of NICU care(Table 1).

Nursing Characteristics

The authors of 2 articles10,11

addressed disparities in NICU nursingcharacteristics and how they relatedto outcomes, both finding that highblack-serving hospitals deliveredlower quality of care. Lake et al10

found that infants with VLBW born inhigh black concentration hospitalshad worse outcomes as well as morenurse understaffing and poorerpractice environments. In addition,Lake et al11 surveyed NICU nurses in

4 US states to understand how nursestaffing and work environment affectsoutcomes. They found that thepatient-to-nurse ratio wassignificantly higher in high black-serving hospitals and NICUs in highblack-serving NICUs missed nearly50% more nursing care than those inlow black-serving NICUs.

Appropriate Setting

The authors of 2 studies looked atdisparities in birth in appropriatesettings,12,13 and the authors of 1study looked at the benefits of beingborn in an appropriate setting,14

often finding that mothers of colorbenefit in both scenarios. In a study ofAlabama VLBW births between 1988and 1990, Bronstein et al12 found thatmothers of color with early prenatalcare were more likely than whitemothers to give birth to infants withVLBW in hospitals with NICUs.Similarly, in a California study, Gouldet al13 found that black women areless likely to give birth to infants withVLBW in inappropriate settings thanwhite and Hispanic women.

Gortmaker et al14 asked whether thebenefit of NICU technology wasequally beneficial to black and whiteinfants with VLBW. They foundsignificant differences in survival bytype of hospital, with infants withVLBW born in hospitals with NICUsmore likely to survive. However, theresearchers found greater survivalrates for black infants compared withwhite infants at the same level ofhospital care. The authors suggestthat this indicates that black infantswith VLBW have better survival thanwhite infants in general, not thatthere is a differential in quality.

Geography

In 2 articles,15,16 researchersinvestigated the relationship betweengeography and racial disparities inquality of care; in both articles, it wasfound that proximity did not makea difference in terms of accessinga high-quality hospital15 or loweringneonatal mortality rate.16 Hebertet al15 found that black women wereless likely to deliver in top-tierhospitals and that top-tier hospitalswere less likely to be in blackwomen’s neighborhoods. Distance,however, did not contribute to theracial disparity in use of top-tierhospitals, in that black and whitewomen often bypassed theirneighborhood hospital. Usinga different approach, Featherstoneet al16 found no significantassociation between travel time todelivery hospital and neonatalmortality. However, they found thatblack infants experienced lower oddsof neonatal death overall.

Minority-Serving Hospitals

Researchers for 2 articles17,18 lookedat the relationship between theproportion of black or white infantswith VLBW in a hospital and itsneonatal mortality rate, finding thathospitals with more black infantswere of lower quality. Morales et al17

found that “minority-serving”hospitals (defined as hospitals where.35% of infants with VLBW that

FIGURE 1Replicable search results. CINAHL, Cumulative Index to Nursing and Allied Health Literature.

PEDIATRICS Volume 144, number 2, August 2019 3 by guest on December 27, 2020www.aappublications.org/newsDownloaded from

TABLE1System

aticReview

Structure

Variable,Author,

andReference

Demographicsof

Study

(IfAvailable)

Outcom

e(s)

ofInterest

Summaryof

MainFindings

RelativeRisk

orMeasuresof

Effect

(IfApplicable)

SuggestedMechanism

Nursing

characteristics

Lake

etal10

8252

infantswith

VLBW

Relationshipbetweennursing

characteristicsandhospital-level

disparities

inVLBW

outcom

es

Higher

infectionrates,higher

ratesof

dischargewithoutbreast

milk,m

ore

nurseunderstaffing,and

poorer

practiceenvironm

ents

inhospitals

with

ahigh

proportionof

black

patients(“high-black”)

compared

with

hospitalswith

alowproportion

(“low-black”);70%

ofblackinfants

with

VLBW

areborn

athigh-black

hospitals,sothesedisparities

stronglyaffectthem

.Black

infantsare

particularlydisadvantagedeven

inhigh-black

hospitalswhenitcomes

tobreastfeeding.Blackinfantsand

white

infantsno

different

within

hospitalsoverall.

aORof

nosocomialinfectionin

high-

blackversus

low:1.64.aORfor

dischargewithoutBM

high

versus

low:1.61.Differences

notsignificant

whencontrolling

fornursing

characteristics.Practice

Environm

ent(scaleof

1–4):high-

black2.95,low

-black

3.16,P

5.004.

Understaffing

(measuredinfraction

ofanurseperinfant

needed

tomeetguidelines):high-black

0.22,

low-black

0.14,P

5.004.

Nursingcharacteristicdifferences

likely

dueto

financialconstraintsandpoor

managem

ent.Nursing

characteristicslikelydriveinfection

anddischargewithoutbreast

milk.

Lake

etal11

Random

samplesurvey

oflicensednurses

in4

largeUS

states:1037

staffnurses

in134

NICUs.

Averagepatient

load,individual

nurses’patient

load,professional

nursingcharacteristics,nurse

workenvironm

ent,andnursing

care

missedon

thelast

shift.

Morecare

activities

aremissedin

hospitalsservingahigher

proportion

ofblackinfants(.

31%)than

inhospitalsthat

serveasm

aller

proportionof

blackinfants(,

11%).

Thisindicatesthatquality

ofcare

may

belower

inhigh-black

hospitals.

Nurses

inhigh-black

NICUsmissed

nearly50%

morecare

activities

than

thosein

low-black

NICUs

(average

1.51

vs1.05

activities

missedpershift,P

5.03).

Disparities

inmissedcare

werelikely

dueto

higher

patient-to-nurse

ratios

inhigh-black

hospitals.H

igh-black

hospitalsweremorelikelyto

have

larger

proportionof

Medicaid

patients,which

canlead

tofinancial

strain

andthereforefewer

staffing

resources.

Appropriate

setting

Bronsteinet

al12

Infantswith

VLBW

;1118

white

and1478

infants

ofcolor.

Deliveryin

hospitalswith

NICUs.

Mothers

ofcolorwith

earlyprenatal

care

werethegroupmostlikelyto

deliver

theirinfantswith

VLBW

inhospitalswith

NICUs.White

wom

enwith

Medicaidweremorelikelyto

betransferredbefore

birththan

white

wom

enwith

privateinsurance.

OR:1.353

(1.119–1.636)

formothers

ofcolorversus

white

mothers.O

R:1.733(1.316–2.283)

formothers

ofcolorwith

Medicaidandearly

prenatalcare

versus

whitemothers

with

noMedicaidandearlyprenatal

care.

Wom

enwith

earlyprenatal

care

are

likelyableto

contacttheircare

providerssooner

ifthey

gointolabor

prem

aturelyor

arepart

ofcare

system

swith

establishedreferral

relationships

forem

ergency

situations.Som

ehospitalsmay

selectivelyretain

privatelyinsured

wom

enforhigh-riskdeliveries,refer

less

Medicaid-insuredwom

ento

subspecialtyregional

centers.

Gouldet

al13

Data

from

1state

(analyzedby

region)

from

1989

to1993

of24

094live-born

infants

with

VLBW

.

Factorsassociated

with

likelihoodof

VLBW

birthat

aLevel1hospital

(hospitalw

ithouta

NICU

or24-hon-

callneonatologist)

10.5%

(24094)

infantswith

VLBW

were

delivered

inLevel1hospitals.

Significant

regional

variation:from

3.1%

to24.3%.The

odds

were

decreasedforAfricanAm

ericansand

SoutheastAsians

andincreasedin

Odds

ofinappropriatedeliverysite

ranged

from

0.37

to2.75

across

California’s

9geographicperinatal

regions.OR

forAfricanAm

ericans:

0.65.O

RforSoutheastAsians:0.54.

Common

toall3racial

ethnicgroups:

risksof

Level1VLBW

delivery

associated

with

(1)less

than

adequate

prenatal

care,(2)

livingin

azipcode

that

isonly50%–75%

urban,and(3)livingat

adistance

to

4 SIGURDSON et al by guest on December 27, 2020www.aappublications.org/newsDownloaded from

TABLE1

Continued

Variable,Author,

andReference

Demographicsof

Study

(IfAvailable)

Outcom

e(s)

ofInterest

Summaryof

MainFindings

RelativeRisk

orMeasuresof

Effect

(IfApplicable)

SuggestedMechanism

Hispanicwom

enas

comparedwith

white

non-Hispanicwom

en.For

all

wom

en,lessthan

adequate

prenatal

care,livingin

a50%–75%

urbanzip

code,and

living.25

milesfrom

the

nearestNICU

significantlyincreased

theodds

ofVLBW

deliveryat

aLevel1

hospital.

ORforwhite

1.00.O

RforHispanic

1.16.

ahospitalwith

aNICU.H

owever,

regional

differences

intheodds

for

inappropriateVLBW

birthremained

afteradjustingforcontributingsocio-

demographicsandgeographical

risk

factors.Hispanicteenage

pregnancieswereat

particular

risk

ofinappropriatebirthsetting.

Gortmaker

etal14

Data

from

4states

for

1978

and1979

used

toestim

atesurvival

curves

forfirst24

hof

life.

Survival

byhospitalsetting(high-

technology

centersversus

urbanor

ruralcenters).

Among750–1000

ginfants:cumulative

probabilityof

survival

atLevelIII

center

was

∼0.63

amongwhite

infantsand∼0.70

amongblack

infants.Am

ong1000–1500

g:∼0.90

amongwhite

infantsand∼95

among

blackinfants(LevelIII).

Blackinfantshadbetter

access

tospecialized

services.InWashington

andTennessee,.50%

ofblack

infantswith

VLBW

wereborn

inspecialized

centers.Forallbirth

weightcategories,survivalduring

thefirst96

hwas

greateram

ong

blackthan

white

infantsat

the

samelevelofhospitalcare.Birthin

aLevelIIIcenter

conferredthe

highestsurvival

rates.

Birthin

aLevelIIIcenter

greatly

improves

infant

outcom

es.

Differentialaccess

tospecialized

LevelIIIcentersbetweenblack

infantsandwhiteinfantsmay

explain

thehigher

observed

survivalratesin

blackinfants.

Geography

Hebert

etal15

VLBW

deliveriesin

New

York

Cityfrom

1996

to2001

tonon-Hispanic

blackandnon-

Hispanicwhite

mothers.

Theroleof

geographicdistributionof

hospitalsin

theracial

disparity

intheuseof

top-tierhospitals(those

inthelowesttertile

ofhospitals

ranked

byratio

ofobserved

toexpected

deaths).

Blackmothers

less

likelyto

deliver

intop-tierhospitals(white5

44%,black

528%;P

,.001).Top-tierhospitals

less

likelyto

bein

blackmothers’

neighborhoods(white

540%,black

533%;P

,.001).Distance,how

ever,

didnotcontribute

tothedisparity

inuseof

top-tierhospitals:m

others

ofboth

racesoftenbypassed

their

neighborhood

hospital(black

562%

bypassed,w

hite

571%;P

,.001).

ORof

relativerisk

blackor

white

ofprobability

that

amotherof

aneonatewith

VLBW

used

atop-tier

hospital:0.6.

Theinfluenceof

geographyon

theuse

oftop-tierhospitalsformothers

ofneonates

with

VLBW

iscomplex.

Otherpersonal

andhospital

characteristics,notjust

distance

orgeography,also

influenced

hospital

usein

NewYork

City.R

esearchers

provideinsights

tothosewho

would

report

risk

adjusted

hospital

mortalityrankings

tofoster

top-tier

selection.However,improvingquality

atpoorlyperformingfacilitiescould

makegreaterimpact

than

encouragingchoice

ofbetter-

performinghospitals.

Featherstone

etal16

Thelinkedbirthand

deathrecordsof

singletoninfantswith

VLBW

born

between

2010

and2012

(n5

2030).

Impact

oftraveltim

efrom

maternal

residenceto

deliveryhospitalon

neonatal

mortalityrate.

Nosignificant

associationbetween

traveltim

eto

deliveryhospitaland

neonatalmortalityafteradjustingfor

confounders.1-wkincrease

inGA

and

non-Hispanicblackmothers

(versus

non-Hispanicwhite

mothers)were

associated

with

lower

odds

ofneonatal

death,whereas

non-

NICU

admission

atbirthwas

associated

with

increasedodds

ofdeath.

1-wkincrease

inGA

associated

with

lower

odds

ofmortality:(OR:

0.61);

non-Hispanicblackmothers

(OR:

0.68)associated

with

lower

death;

non-NICU

admission

atbirth(OR:

5.9)

increasedodds

ofdeath.

Ahigh

proportionof

neonatal

deaths

occurred

within

24hof

birth.

Underlying

causes

ofdeath,

particularlyin

thefirst24

h,arenot

sensitive

toaccess

tocare

butare

morecloselyalignedwith

other

maternal,neonatal,orhospital-level

factors.

PEDIATRICS Volume 144, number 2, August 2019 5 by guest on December 27, 2020www.aappublications.org/newsDownloaded from

TABLE1

Continued

Variable,Author,

andReference

Demographicsof

Study

(IfAvailable)

Outcom

e(s)

ofInterest

Summaryof

MainFindings

RelativeRisk

orMeasuresof

Effect

(IfApplicable)

SuggestedMechanism

Minority-serving

hospitals

Morales

etal17

74050infantswith

VLBW

treatedby

332VON

hospitalsbetween

1995

and2000

Whether

thereisan

association

betweenproportionof

black

infantswith

VLBW

treatedat

ahospitalandneonatal

mortality

forblackandwhite

infantswith

VLBW

.“Minority-serving”definedas

.35%of

infantswith

VLBW

treated

wereblack;othercategories

were

,15%

and15%–35%.

Minority-serving

hospitalshad

significantlyhigher

risk-adjusted

neonatal

mortalityratesthan

,15%.

Differences

werenotexplainedby

either

hospitalor

treatm

ent

variables.

Risk-adjustedneonatal

mortalityrates

in.35%

blackhospitalscompared

with

,15%:w

hiteOR

:1.30,blackOR

:1.29,p

ooledOR

:1.28

Minority-serving

hospitalsmay

beprovidinglower

quality

ofcare

toinfantswith

VLBW

.Results

werenot

explainedby

otherhospital

characteristicslooked

atin

this

study(location,teaching

status,%

admissionscoveredby

Medicaid),

buttherecouldbe

other

characteristicsnotlooked

atin

this

studythat

might

have

aneffect.

Howellet

al18

11781infants500–1499

gin

NewYork

Cityborn

between1996

and

2001.

Risk-adjustedneonatalmortalityrates

foreach

NewYork

Cityhospitaland

racial

andethnicdistribution

amongthosehospitals.

White

infantswith

VLBW

morelikelyto

beborn

inlowestmortalitytertile

ofhospitals(49%

)comparedwith

black

infantswith

VLBW

(29%

).Estim

ated

thatifblackwom

endelivered

insame

hospitalsas

white

wom

en,black

VLBW

mortalityrateswould

bereducedby

6.7per1000

VLBW

births

ordisparity

would

bereducedby

34.5%.

——

Military

vscivilian

care

Kugler

etal19

Blackor

white

singleton

livebirths

inPierce

Countydelivered

between1982

and

1985.

Theeffect

ofsystem

ofcare

(military

care)on

differences

inlowbirth

weightandneonatal

mortality

betweenblackinfantsandwhite

infants.

(Note:notincludingfindings

onlow

birthweightfindings

becausenot

quality

ofcare.)Civilianblackinfants

hadapproximatelytwicetheneonatal

deathratesof

civilianwhite

infants.

Theneonatal

mortalityratesfor

military

blackinfants,however,did

notdiffersignificantlyfrom

either

groupof

white

infants.Disparity

did

notapplyto

birthweight,2500

g.

Civilianblackinfantsversus

civilian

white

infantscruderisk

ratio:2.33;

civilianblackinfantsversus

military

blackinfantscruderisk

ratio:3.08;

military

blackinfantsversus

military

white

infantscruderisk

ratio:1.04.

Military

care

appeared

tobe

associated

with

aprotectiveeffect

forneonatal

mortalityforblackinfants.Thiseffect

was

notdueto

differences

inbirth

weightdistributionor

tothequantity

ofprenatal

care

received.The

effect

was

mostprom

inentfornorm

alweightblackinfants,especiallyfor

thosefrom

lowincomecensus

tracts.

6 SIGURDSON et al by guest on December 27, 2020www.aappublications.org/newsDownloaded from

were treated were black) hadsignificantly higher risk-adjustedneonatal mortality rates thanhospitals with ,15% black infants,and that these differences were notexplained by hospital or treatmentvariables. Similarly, Howell et al18

examined which hospitals black andwhite infants were most likely to beborn in. They found that white infantswere more likely to be born inhospitals in the lowest tertile ofmortality rates (49%) compared withblack infants with VLBW (29%) inNew York City hospitals. Theyestimated that if black women gavebirth in the same hospitals as whitewomen, black VLBW mortality rateswould be significantly reduced.

Military Versus Civilian Care

In looking at the influence of healthcare systems (military hospital versuscivilian hospital) on neonatalmortality, Kugler et al19 found thatmilitary care appeared to beassociated with a protective effect forneonatal mortality for black infants inPierce County, Washington. Theyfound that in civilian care, blackinfants had approximately twice theneonatal death rates of white infants,but that difference disappeared inmilitary care, suggesting a protectiveeffect of military care for blackinfants.

Composite Quality

We identified 2 articles9,20 in whichthe authors investigated multiplequality measures, sometimescombining them into an explicitcomposite measure. Both suggest thatinfants of color are vulnerable to low-quality NICU care. Profit et al9 usedBaby-MONITOR (a compositemeasure of NICU quality) to studydisparities within and between NICUswherein a higher score indicatesbetter quality of care. They foundsignificant variation both betweenand within NICUs. Although non-Hispanic white infants scored higheron measures of process, non-Hispanicblack infants scored higher onTA

BLE1

Continued

Variable,Author,

andReference

Demographicsof

Study

(IfAvailable)

Outcom

e(s)

ofInterest

Summaryof

MainFindings

RelativeRisk

orMeasuresof

Effect

(IfApplicable)

SuggestedMechanism

Composite

quality

Profitet

al9

18616infantswith

VLBW

in134CaliforniaNICUs

betweenJanuary1,

2010,and

December

31,2014.

Baby-MON

ITOR

score(a

compositeof9

processandoutcom

emeasuresof

quality).Foreach

NICU,a

risk-

adjusted

composite

andindividual

component

quality

scoreforeach

race

andethnicity.

Compositequality

scores

ranged

by5.26

standard

units

(range:2

2.30

to2.96).

Non-Hispanicwhite

infantshigher

onmeasuresof

processcomparedwith

non-Hispanicblackinfantsand

Hispanicinfants.Comparedwith

white

infants,non-Hispanicblack

infantsscored

higher

onmeasuresof

outcom

e;Hispanicinfantsscored

lower

on7of

the9Baby-MON

ITOR

subcom

ponents.

Differencebetweenhighestandlowest

performingNICUswas

large(5.2

standard

units).

Someof

thedisparity

createdby

inferior

performance

among

modifiablemeasuresof

process

rather

than

outcom

e,suggesting

acriticalrole

forQI

efforts.

Howellet

al20

7177

infants,GA

24–31

wk

(verypreterm)

Mortalityor

severe

neonatal

morbidity

Morbidity

andmortalityhigher

among

blackandHispanicthan

white

VPTBs.

Risk-adjustedmorbidityandmortality

rate

was

twiceas

high

forVPTBsin

hospitalsinthehighestm

orbidityand

mortalitytertile

than

thoseborn

inthelowesttertile

hospitals.Black

and

HispanicVPTBsweremorelikelyto

occurin

thehighestmorbidity

and

mortalityhospitalsthan

white.M

ost

ofthedisparity

canbe

attributed

todifferences

ininfant

health

risks,but

birthhospitalalso

playsasignificant

role.

Percentage

ofdisparity

explainedby

birthhospital:black-white

540%

(95%

CI,30%

–50%);Hispanic-white

30%

(95%

CI,10%

–49%).

Distance

tothehospital,insurance,

hospitalstructural

characteristics,

patterns

ofracial

segregation,

community

factors,physician

referral,riskperception,patient

choice,access,andthemanagem

ent

ofmedical

emergenciesduring

pregnancymay

allcontribute

toblackandHispanicwom

engiving

birthat

hospitalswith

higher

mortalityandmorbidityrates.

“African

American”and“black”areoftenused

interchangeablyintheliteraturereview

ed.Inthistable,weusethesamelanguage

asthearticlescited.aOR,adjusted

odds

ratio;BM,breastmilk;CI,confidenceinterval;OR,odds

ratio;VON

,Vermont

Oxford

Network;VPTB,verypreterm

birth;—

,not

applicable.

PEDIATRICS Volume 144, number 2, August 2019 7 by guest on December 27, 2020www.aappublications.org/newsDownloaded from

outcome measures. Overall, theauthors found that, “…as [NICU]quality scores rise, whites tend toperform better than AfricanAmericans. However, AfricanAmericans in high-performing NICUsoften fare better than AfricanAmericans in low-performing NICUs.”Howell et al20 examined differencesin neonatal morbidity and mortalityrates in very preterm infants by siteof delivery in New York Cityhospitals. The authors createda composite measure of mortality andmorbidity and found it to be higheramong black and Hispanic than whitevery preterm infants. The risk-adjusted morbidity and mortalityrates were twice as high for theseinfants in hospitals in the highestmorbidity and mortality tertile thanthose born in lowest tertile hospitals.Black and Hispanic very pretermbirths were more likely to occur inthe highest morbidity and mortalityhospitals. The authors20 found thatalthough differences in infant riskfactors accounted for a large portionof the disparity overall, birth hospitalstill accounted for 30% to 40% of theexplained disparity in outcomes forblack or Hispanic infants comparedwith white infants.

Processes

We identified 19 articles in whichdisparities in NICU care processeswere addressed (Table 2).

Breast Milk

Researchers for 8 articles9,21–26,38

addressed racial disparities in NICUbreast milk feeding (BMF), factorsimportant for successful BMF, use ofdonor human milk (DHM) in NICUs,or the effect of single family rooms onbreastfeeding. Many pointed todisparities that disadvantage blackinfants.

Researchers for 3 articles9,21,38

studied disparities in BMF atdischarge. Lee et al21 found thatCalifornia BMF rates were highest forwhite, lower for Hispanic, and lowest

for black infants. After separatingNICUs by quartile percentage of eachrace, Lee et al21 found that BMF rateswere higher for all races andethnicities in hospitals with morewhite mothers. When units werecompared, black infants benefited themost by being cared for in NICUs witha higher proportion of white infants.Lee et al’s21 findings are consistentwith those of Profit et al,9 whoinclude BMF within a composite NICUquality measure, finding that whiteinfants scored higher than otherracial and/or ethnic groups on BMF atdischarge across NICUs. Riley et al38

studied the effect of neighborhoodstructural factors on BMF at NICUdischarge for VLBW infants, findingthat these did not significantly impactBMF, but that maternal race and/orethnicity predicted BMF at dischargesuch that black race was associatedwith reduced likelihood of BMF atdischarge. They also found that unlikewhite mothers, there was a trend ofdecreased BMF at discharge for blackmothers without access to a car.

Researchers for 2 articles exploredfactors important for BMF in theNICU, in particular for black mothers.Using qualitative research, Cricco-Lizza22 found that black non-Hispanicmothers reported receiving limitedbreastfeeding education and supportduring pregnancy, childbirth, NICUstays, postpartum, and recovery inthe community. Fleurant et al23

looked at the association of socialfactors (eg, breastfeeding support,etc) with human milk feeding atdischarge. They found that theSupplemental Nutrition Program forWomen, Infants, and Children (WIC)program enrollment negativelypredicted (and maternal goal settingnear time of discharge positivelypredicted) BMF at discharge formothers of all races and ethnicities.Surprisingly, they found thatperceived breastfeeding support froman infant’s maternal grandmothernegatively predicted BMF atdischarge for black mothers.

Researchers for 2 studiesincorporated the use of DHM asa measure of quality, either seeking toidentify reasons associated withnonconsent for DHM24 or in assessingwhether demographic disparitiesexist in the use of mother’s own milkand DHM.25 Brownell et al24 foundthat nonwhite race was among thepredictors of nonconsent for DHM,and Boundy et al25 found that NICUsin postal codes with low percentagesof black residents had higher rates ofmother’s own milk use and of donormilk use.

Vohr et al26 compared infants caredfor in single family room NICUs tothose cared for in open bay NICUs.They found that, overall, infantsbenefitted from single family roomsthrough higher rates and volume ofbreast milk use and higher cognitiveand language scores, but nonwhiterace was associated with a decreasedeffect of this benefit.

High-risk Infant Follow-up Referrals

Researchers for 2 articles focused ondisparities in referrals for follow-upcare in infants with VLBW cared forin NICUs,27,28 showing disadvantagesfor black infants. We included theseinvestigations because referrals tofollow-up care should be initiated inthe NICU and reflect the highestquality of care. Barfield et al27 foundthat early intervention referrals werelower for infants of black mothers.Similarly, Hintz et al28 found thatblack and Hispanic infants were lesslikely to be referred for high-riskfollow-up than white infants.

Parental Satisfaction and FamilyExperience

Researchers for 2 recent studiesdocumented disparities in parentalsatisfaction or how familiesexperience NICU care,29,30 suggestingthat families of color may bevulnerable to worse care. Martinet al29 surveyed white and blackfamilies on their experiences andsatisfaction with NICU care and found

8 SIGURDSON et al by guest on December 27, 2020www.aappublications.org/newsDownloaded from

TABLE2System

aticReview

Process

Variable,Author,

andReference

Demographicsof

Study(If

Available)

Outcom

e(s)

ofInterest

Summaryof

MainFindings

RelativeRisk

orMeasuresof

Effect

(IfApplicable)

SuggestedMechanism

Breast

milk

Leeet

al21

90%

ofinfantsadmitted

into

Californian

NICUswith

birth

weightof

500–1500

gand

born

ator

transferredto

aCPQCCcenter

within

2dof

birthfrom

2006

to2008.

Anyam

ount

ofBM

Fat

timeof

discharge.Hospitalswere

dividedinto

quartiles

byhospitalpercentage

ofeach

race,and

BMFratesby

race

werecompared.

BMFfoundtobe

higher

forallracialand

ethnicgroups

inhospitalswith

morewhite

mothers.D

ifferences

wereless

pronounced

forHispanicmothers

than

for

blackmothers.African

Americanshadthelowest

ratesof

BMFwhencaredforin

hospitalswith

more

AfricanAm

ericans(43%

)versus

hospitalswith

fewer

AfricanAm

ericans(59%

,P5

.003).On

theother

hand,w

hitemothers

hadthelowestB

MFrates(49%

)in

hospitalswith

thefewestwhite

mothers.

With

risk

adjustment,white

mothers

weremorelikelyto

engage

inBM

Fwhenthere

weremorewhitemothers

atthat

hospital(oddsratio

1.15

for10%

increase

inwhite

mothers,95%

confidenceinterval

1.04–1.28).Blackmothers

wereless

likelyto

engage

inBM

Fwhenmoreblack

mothers

wereat

the

hospital(oddsratio

0.80

for

each

10%

increase

inblack

mothers,95%

confidence

interval

0.67–0.97).

Hospitalsservingmorepatientsof

colorwereless

likelyto

have

prem

atureinfantsfedbreast

milk

forallraces.Targetingsuch

hospitalsforQI

may

help

toimproveBM

Fratesoveralland

reduce

disparities.

Profitet

al9

18616infantswith

VLBW

in134

CaliforniaNICUsbetween

January1,2010,and

December31,2014.

Baby-MON

ITOR

score(a

composite

of9processand

outcom

emeasuresof

quality).Foreach

NICU,

arisk-adjustedcomposite

andindividual

component

quality

scoreforeach

race

andethnicity.

Compositequality

scores

ranged

by5.26

standard

units

(range

22.30

to2.96).Non-Hispanicwhite

infants

higher

onmeasuresof

processcomparedwith

non-

HispanicblackinfantsandHispanicinfants.

Comparedwith

white

infants,non-Hispanicblack

infantsscored

higher

onmeasuresof

outcom

e;Hispanicinfantsscored

lower

on7of

the9Baby-

MON

ITOR

subcom

ponents.

Differencebetweenhighest

andlowestperforming

NICUswas

large(5.2

standard

units).

Someof

thedisparity

createdby

inferior

performance

was

amongmodifiablemeasuresof

processrather

than

outcom

e,suggestingacriticalrole

forQI

efforts.

Rileyet

al38

410VLBW

infantsborn

between

February

2008

and2012

and

admitted

totheLevelIIINICU

atRush

UniversityMedical

Center

inChicago,Illinois.

HMfeedingat

dischargeas

mitigatedby

neighborhood

structural

factors.

InbivariateanalysisHM

feedingat

dischargewas

negativelycorrelated

with

neighborhood

structural

factorsincludingneighborhood

concentrated

disadvantage

(OR:

0.72;C

I:0.60–0.86),neighborhood

violentcrimerate

(OR:

0.93;CI:0.89–0.98),and

negativelycorrelated

with

patient

factorsincluding

blackrace

and/or

ethnicity

(OR:

0.41;C

I:0.24–0.70).

Inmultivariate

analysis,onlymaternalrace

and/or

ethnicity

(OR:

1.04;CI:1.00–1.09),WIC

eligibility,and

length

ofNICU

hospitalizationpredictedHM

feeding

atdischargefortheentirecohort.The

interaction

betweenaccess

toacarandrace

and/or

ethnicity

significantlydifferedbetweenblackandwhite

orAsianmothers,although

thepredictedprobabilityof

HMfeedingat

dischargewas

notsignificantly

affected

byaccess

toacarforanyracial

and/or

ethnicsubgroup.

Inmultivariate

analysis,

maternalrace

and/or

ethnicity

(OR:

1.04;C

I:1.00-

1.09)predictedHM

feeding

atdischargeforentire

cohort.

Socioeconomicstatus

affected

breastfeedingratesinallracial/

ethnicgroups,b

utdisproportionatelyaffected

blackmothers

toagreater

degree

than

white

orHispanic

mothers.

Cricco-Lizza

22130blacknon-Hispanicmothers

enrolledin

theNewYork

SpecialSupplemental

Audiotaped

interviewsandfield

noteswereanalyzed

for

mothers’descriptions

of

Foundlim

itedbreastfeedingeducationandsupport

during

pregnancy,childbirthstay

inNICU,

postpartum

,and

recovery

inthecommunity.They

N/A

Lack

oftrustingrelationships

(neglectingtheaffective

elem

ents

ofhealth

care).

PEDIATRICS Volume 144, number 2, August 2019 9 by guest on December 27, 2020www.aappublications.org/newsDownloaded from

TABLE2

Continued

Variable,Author,

andReference

Demographicsof

Study(If

Available)

Outcom

e(s)

ofInterest

Summaryof

MainFindings

RelativeRisk

orMeasuresof

Effect

(IfApplicable)

SuggestedMechanism

Nutrition

Program

forWIC

weregeneralinform

ants.

From

thisgroup,11

key

inform

ants

hadclosefollow-

upduring

pregnancyandthe

firstpostpartum

year.

infant-feedingeducationand

supportfrom

nurses

and

physicians.

also

expressedtrustor

distrust

concerns

and

varyingdegreesof

anxietyabouthowthey

were

treatedby

nurses

andphysicians.

Fleurant

etal23

362raciallydiversemothers

ofinfantswith

VLBW

.HM

feedingat

discharge.

Forall362mothers,W

ICnegativelypredictedHM

feedingat

dischargeandmaternalgoalneartim

eof

dischargepositivelypredictedHM

feedingat

discharge.Perceivedbreastfeedingsupportfrom

infant’smaternalgrandm

othernegativelypredicted

HMfeedingat

discharge.

Forallmothers:W

ICeligibility

(OR:

0.34;0.15–0.75;P

5.008),breastfeedingsupport

from

mothers’mother(OR:

0.45;0.26–0.79;P

5.005),

goal

ofanyHM

near

discharge(OR:

8.38;

3.42–20.53;P#

.001).

Stratified

byrace

and/or

ethnicity:“Forblack

mothers,support

from

the

mother’s

mother(OR0.27

[95%

CI0.11–0.68],P5

.006).

Goalof

HMat

dischargepredicting

HMfeedingconsistent

with

otherresearch.Surprised

byfindingthat

higher

maternal

grandm

othersupportpredicts

lower

BM.Suggestionthat

this

relatesto

maybe

theperception

ofsupport(ratherthan

actual

support)andthat

mothers

coparentingwith

theirow

nmothers

aremorelikelyto

experience

parentingstress.

Brow

nell

etal24

Mothers

ofinfantseligibleto

receivedonormilk

(#32

weeks’gestationor

#1800

g)born

betweenAugust

2010

and2015.

Odds

ofnonconsent.

Ofthe486mother-infant

dyadsfrom

thefirst5yofthe

donormilk

program,nonwhite

race

(aOR:1.69;95%

CI:1.04–2.76)andincreasing

GA(aOR:1.11;95%

CI:

1.03–1.21)independently

predictednonconsent.Each

year

theprogram

existed,therewas

a48%

reductionin

odds

ofnonconsent

(aOR:0.52;95%

CI:

0.43–0.62).Themostcommon

reason

givenfor

nonconsent

was

‘‘it’s

someone

else’smilk.’’

Nonw

hite

race

(aOR:1.69;95%

CI:1.04–2.76);increasing

GA(aOR:1.11;95%

CI:

1.03–1.21)independently

predictednonconsent.

Program

durationwas

associated

with

reducednonconsent

rates

andmay

reflectincreased

exposure

toinform

ationand

acceptance

ofdonormilk

use

amongNICU

staffandparents.

Despite

overallimprovem

ents

inconsentrates,race-specific

disparities

inratesof

nonconsent

fordonormilk

persistedafter5yof

thisdonor

milk

program.

Boundy

etal25

Data

from

CDC’s2015

Maternity

Practices

inInfant

Nutrition

andCare

survey,linkedtothe

2011–2015

USCensus

Bureau’sAm

erican

Community

Survey.

Theuseof

mother’s

ownmilk

anddonormilk

inhospitals

with

NICUsby

thepercentage

ofnon-Hispanicblack

residentsin

thehospital

postal

code

area,

categorizedas

beingabove

orbelowthenational

average(12.3%

).

Inpostal

codeswith

.12.3%blackresidents,48.9%of

hospitalsreported

usingmothers’ow

nmilk

in$75%ofinfantsintheNICU,and

38.0%reported

not

usingdonormilk,com

paredwith

63.8%and29.6%of

hospitals,respectively,in

postal

codeswith

#12.3%

blackresidents.

N/A

Differences

foundmay

berelated

tovariations

inhealth

care

personnelsupport,hospital

policiesandpractices,m

others’

know

ledgeandaccess

toinform

ation,andcommunity-

levelsupportforbreastfeeding.

Donormilk

usemight

also

beaffected

byhospitalp

roximity

tomilk

banks,stateregulations,

andhospitalpoliciesrelatedto

theprovisionof

donormilk

and

insurancereimbursem

ent.

10 SIGURDSON et al by guest on December 27, 2020www.aappublications.org/newsDownloaded from

TABLE2

Continued

Variable,Author,

andReference

Demographicsof

Study(If

Available)

Outcom

e(s)

ofInterest

Summaryof

MainFindings

RelativeRisk

orMeasuresof

Effect

(IfApplicable)

SuggestedMechanism

Vohr

etal26

Infantsweighing#1250

gcared

forin

anopen-bay

NICU

(January

2007–August

2009)

(n5

394)

versus

thosein

aSFRNICU

(January

2010–December2011)(n

5297).

Human

milk

provisionat1,4wk

anddischarge,and4wk

volume(m

L/kg/d).Also,

18–24

moBayley

III.

Infantscaredforin

theSFRNICU

hadhigher

Bayley

IIIcognitive

andlanguage

scores,higherratesof

human

milk

provisionat

1and4wk,andhigher

human

milk

volumeat

4wk.TheSFRNICU

was

associated

with

a2.55-point

increase

inBayley

cognitive

scores

and3.70-point

increase

inlanguage

scores.Every

10mL/kg

perdincrease

ofhuman

milk

at4wkwas

associated

with

increasesin

Bayley

cognitive,language,andmotor

scores

(0.29,0.34,

and0.24,respectively).M

edicaidwas

associated

with

decreasedcognitive

andlanguage

scores,and

lowmaternaleducationandnonw

hite

race

with

decreasedlanguage

scores.

Nonw

hite

race

with

decreased

language

scores

(25.8).

Authorsdo

notcommenton

race

orethnicity

butnote

that

low

maternaleducation,poverty,and

insurancestatus

likelymean

less

abilityto

spendtim

ein

the

NICU

andprovidebreast

milk,

thus

leadingto

lower

Bayley

IIIcognitive

scores.

High-riskinfant

follow-up

referrals

Barfield

etal27

1233

infantswith

VLBW

(,1200

g)in

Massachusetts

born

January1998–June

2003

Ratesof

referral

andtim

eto

referral

byrace

and/or

ethnicity.

Blackinfantshadlowestreferral

rates,especiallyat

0–12

mo.12%

blackinfantsnotreferred

atall,

comparedwith

6.8%

overall.

aHRof

referral

ofblacknon-

Hispanicinfantscompared

with

white

non-Hispanic

infantswas

0.85.

Possiblethat

minority

families

are

less

confident

intheirabilityto

access

specialty

health

care

services,suchas

EI.Also

mentions

barriers

toaccess,

physicianmistrust,concerns

and/or

misunderstandings

regardingtreatm

ent,and

disparities

ininsurancestatus

(which

was

included

and

adjusted

forwhencalculated

aHR,

butmentionedin

discussion).

Hintzet

al28

CPQCCinfants,1500

g,2010–2011

High-riskinfant

follow-up

referral

rates.

Lower

odds

ofreferral

forinfantswith

maternalrace

Hispanicor

AfricanAm

erican

versus

white.

aORcomparedwith

white:

AfricanAm

erican:0.58,

Hispanic:0.65.

Previous

research

haspointedto

issues

ofnoncom

plianceand

lackingsupportsystem

s;in

unadjusted

analyses,hospitals

servinghighestproportionof

AfricanAm

erican

infantshave

lower

referral

ratesoverall.

PEDIATRICS Volume 144, number 2, August 2019 11 by guest on December 27, 2020www.aappublications.org/newsDownloaded from

TABLE2

Continued

Variable,Author,

andReference

Demographicsof

Study(If

Available)

Outcom

e(s)

ofInterest

Summaryof

MainFindings

RelativeRisk

orMeasuresof

Effect

(IfApplicable)

SuggestedMechanism

Parental

satisfaction

andfamily

experience

Martin

etal29

Self-reportsfrom

non-Hispanic

blackandnon-Hispanicwhite

participants

who

hadan

infant

born

ataGA

#35

wk

orbirthwt,2000

g,presentingwithin

2moafter

NICU

dischargeto

any1of

30Children’sHospitalof

Philadelphiaprimarycare

centersbetweenJanuary1,

2010,and

January1,2013.

Parental

reported

trust,

communicationstyle,

expectations

ofthehealth

care

system

,and

satisfaction

with

theirphysicianandthe

NICU

course.Collected

parental

data

included

sex,

age,em

ployment,education,

homeow

nership,ethnicity,

race,m

aritalstatus,

householdresident

inform

ation,andincome.

Although

morecommentingparentswerewhite

(62%

)than

black(38%

),blackparent

comments

were

morenegative(58%

negativevs

42%

positive)

comparedwith

white

parent

comments

(33%

negativevs

67%

positive).The

nurse-parent

relationshipanddistinct

positiveandnegative

nursingbehaviorsareimportantfactorsaffecting

parental

satisfactionwith

NICU

care.B

lack

parents

weremostdissatisfied

with

nursingsupport,

wantingcompassionate

andrespectfu

lcommunicationandnurses

that

wereattentiveto

theirchildren.White

parentsweremostdissatisfied

with

inconsistent

nursingcare

andlack

ofinform

ativeexchanges,wantingeducationabout

theirchild’sshort-andlong-term

needs.Both

groups

describedachaotic

NICU

environm

entwith

high

nursingturnover.

N/A

Blackpatientsareless

likelyto

report

satisfactionwith

health

care,p

ossiblystem

mingfrom

differences

inprovider

communicationstyles,clinician’s

attitudes,m

edical

mistrust,and

perceivedracism

.Provider

uncertainty,biases,and

stereotyping

may

also

contributeto

unequaltreatment.

Sigurdson

etal30

Attendeesat

the2016

Verm

ont

Oxford

NetworkQuality

Congress

composedof

providers(nurses,

physicians,and

otherclinical

specialists)andNICU

family

advocates.

Open-ended

survey

ofaccounts

ofdisparatecare.

Studygathered

324accounts

ofdisparatecare,

majority

describedperceivedworse

care

offamilies,

notstrictlyinfants.Accounts

describedworse

care

basedon

intersectingfactorsof,eg,language,

culture

orethnicity,race,etc.Respondents

describedfamilies

receivingneglectfu

lcare,

judgmentalcare,and

system

icbarriers

tocare.

N/A

Adverseconsequences

ofdisparate

care

forinfantsmediated

throughadverseinteractions

with

families

moreso

than

throughdifferences

incare

toinfant.Socialinequalitiesshape

thecontentandtone

ofhealth

care

encounters

such

that

families

receiveworse

care

basedon

intersectingfactors.

Shared

parent-

provider

decision-

making

VanMcCrary

etal31

Twocasesin

aNewYork

City

tertiary

care

unit.

Nonm

edical

barriers

that

impede

decision-makingin

theNICU.

ManyNICUsarenotequipped

todealwith

complicated

culturalandlinguistic

barriers

that

preventthem

from

communicatingeffectivelywith

theirpatients.

N/A

Forboth

families

ofcolorin

these

case

studies,areinforcingcycleof

disapprovalhad

form

ed.Each

family

made,or

hadpreviously

made,decisionsthattheNICU

staff

didnotagreewith,sothestaff

implicitlyresented

thefamily.In

both

cases,care

decisionswere

influenced

bysocialandcultural

factorsthat

weremisunderstood

byNICU

medicalstaff.

12 SIGURDSON et al by guest on December 27, 2020www.aappublications.org/newsDownloaded from

TABLE2

Continued

Variable,Author,

andReference

Demographicsof

Study(If

Available)

Outcom

e(s)

ofInterest

Summaryof

MainFindings

RelativeRisk

orMeasuresof

Effect

(IfApplicable)

SuggestedMechanism

Tucker

Edmonds

etal32

Periviableinfants(23.0–24.6wk)

inCalifornia,Missouri,and

Pennsylvania,1995–2005.

Resuscitationof

periviable

infants.

Blackrace

apredictorforneonatal

intubation,

even

whencontrolling

forclustering

atthedelivery

hospitallevel.

aORforintubationof

black

infantscomparedwith

white:1.25.

Likelydueto

differences

inpatient

preferences(eg,religious

beliefs,culturalpreferences);

hospital-levelpractices

orresourcescouldalso

beacontributingfactor.

Kangaroo

care

Hendricks-

Muñoz

etal33

42nurses

and143mothers

at2

NewYork

Cityhospitals.

Maternalandprovider

perspectives

onKM

Cand

MCP.

Mothers

ofcolorperceivedthat

they

received

less

educationandaccess

toKM

C;nurses

ofcolorwere

moresupportiveof

MCP

andKM

C.

61%

ofmothers

ofcolor

stronglyidentified

that

access

toKM

Chadbeen

limitedto

them

compared

with

39%

ofwhite

mothers;

77%

ofwhite

mothers

comparedwith

only50%

ofmothers

ofcolorstrongly

perceivedthat

nurses

were

supportivein

helpingthem

provideKM

Cfortheirinfant;

ORof

nurses

ofcolor

encouragingparent

presence

comparedwith

white:2.8.

Culturalandlinguistic

differences

leadingto

communication

difficulties.

Surfactant

and

RDS

Hamvas

etal34

Infantswith

VLBW

born

1987–1989

and1991–1992

inSaintLouisarea;1563infants

intotal,315deaths.

Effectsof

surfactant

approval

onneonatalmortalityam

ong

blackandwhite

infants.

Neonatal

mortalitydecreasedmoreforwhite

than

blackinfantsaftersurfactant

was

approved;after

approval,w

hite

VLBW

mortalitydropped41%

and

VLBW

mortalityforblackinfantsdidnotchange.

Relativerisk

ofdeatham

ong

blacknewbornswith

VLBW

comparedwith

white,

1987–1989:0.7.Increased

to1.3in

1991–1992.

Larger

portionof

neonatal

mortalityin

white

neonates

was

attributed

toRD

S,so

introductionof

surfactant

had

alarger

effect;RDS

moreprevalent

amongwhiteinfantsbecausefetal

pulmonarysurfactant

matures

moreslow

lyinwhiteinfants.

Ranganathan

etal35

44712AfricanAm

erican

infants

and73

942non-Hispanic

white

infants.

Effectsof

surfactant

approval

onneonatalmortalityam

ong

blackandnon-Hispanicwhite

infantswith

VLBW

.

Introductionofsurfactant

therapyin1990

improved

the

outcom

esforwhite

infantswith

VLBW

morethan

black.No

differencebetweenAfricanAm

erican

and

non-Hispanicwhite

inrate

ofdeclineforall

categories

ofmortalitybetween1985

and1988;

DifferencebetweenAfricanAm

erican

andnon-

Hispanicwhite

inrate

ofdeclineforallexcept

nonrespiratory

neonatal

between1988

and1991.

1988–1991:O

ddsof

neonatal

deathfrom

RDSdeclined

34%

innon-Hispanicwhite

infantsandonly16%

inAfricanAm

erican

infants

(P,

.01);oddsof

death

from

allrespiratorycauses

declined

41%

innon-

Hispanicwhite

infantsand

22%

inAfricanAm

erican

infants(P

,.01).

Differentialefficacy

ofsurfactant

onnon-Hispanicwhite

and

AfricanAm

erican

infantswith

VLBW

.Lungs

ofAfricanAm

erican

fetusesmaturemorerapidly

andbegintoproducepulmonary

surfactant

earlier;therefore,

exogenoussurfactant

may

have

fewer

additionalbenefits;

AfricanAm

erican

infantswith

VLBW

may

also

respondless

favorablythan

non-Hispanic

white

infantsof

sameseverity.

PEDIATRICS Volume 144, number 2, August 2019 13 by guest on December 27, 2020www.aappublications.org/newsDownloaded from

that black families were mostdissatisfied with nursing support,lack of compassionate and respectfulcommunication, and less attentivecare of their infants, whereas whiteparents were most dissatisfied withinconsistent nursing care and lack ofinformative exchanges, wantingeducation about their infants’ short-and long-term needs.

Sigurdson et al30 reported onaccounts of disparate care fromproviders and parent advocates viaopen-ended survey. Accountsdescribed disparities related tooverlapping factors (language, cultureor ethnicity, race, immigration statusor nationality, sexual orientation orfamily status, gender, or disability)with race and ethnicity being animportant dimension. The authorsidentified 3 types of disparate care:neglectful care, judgmental care, orsystemic barriers to care, allreflecting suboptimal family-centered care.

Joint Parent-Provider Decision-making

Researchers for 2 articles addressedracial and/or ethnic differences indecision-making.31,32 Van McCraryet al31 reported on 2 cases where caredecisions were influenced by socialand cultural factors that weremisunderstood by NICU medical staffand families. They offered 2 examplesof families who wished for aggressivetreatment when the staff felt the careplan was not appropriate given theburdens of treatment combined withdire prognosis.

Tucker Edmonds et al32 studied racialand/or ethnic differences in use ofintubation for periviable infants,finding black race and/or ethnicity tobe significantly associated withincreased neonatal intubation.

Kangaroo Care

Hendricks-Muñoz et al’s33 survey ofneonatal nurses and NICU mothersrevealed that mothers valuedkangaroo care higher than nurses. Inaddition, more mothers of colorTA

BLE2

Continued

Variable,Author,

andReference

Demographicsof

Study(If

Available)

Outcom

e(s)

ofInterest

Summaryof

MainFindings

RelativeRisk

orMeasuresof

Effect

(IfApplicable)

SuggestedMechanism

Frisbieet

al36

AllUS

infantsborn

1989–1990

(5407166)

and1995–1998

(10809746).

Black-white

disparity

ininfant

mortalitydueto

RDSand

othercauses.

Disparity

inRD

Smortalitybetweenblackandwhite

infantsincreasedafterintroductionof

surfactant.

Absolute

declines

inmortalitygreaterforwhite

infantsthan

blackinfantsam

ongLBWinfantsafter

theintroductionof

surfactant.B

lack

infantshad

arelativesurvival

advantagein

presurfactantera;

thisbecameasurvivaldisadvantage

postsurfactant.

Infant

mortalitydueto

RDSin

blackinfants(versuswhite

infants):1989–1990,O

R5

0.832.1995–1998,O

R5

1.114.

Social

inequalityanddifferential

access

tointerventionbetween

blackandwhite

infants.

Howellet

al37

AllUS

infantsborn

1989–1990

(5407166)

and1995–1998

(10809746).

Black-white

disparity

ininfant

mortalitydueto

RDSand

othercauses.

Disparity

inRD

Smortalitybetweenblackandwhite

infantsincreasedafterintroductionof

surfactant.

Absolute

declines

inmortalitygreaterforwhites

than

blacks

amongLBWinfantsafterthe

introductionof

surfactant.B

lack

infantshad

arelativesurvival

advantagein

presurfactantera;

thisbecameasurvivaldisadvantage

postsurfactant.

Infant

mortalitydueto

RDSin

blackinfants(versuswhite

infants):1989–1990,O

R5

0.832.1995–1998,O

R5

1.114.

Social

inequalityanddifferential

access

tointerventionbetween

blackandwhite

infants.

“African

American”and“black”areoftenused

interchangeablyintheliteraturereview

ed.Inthistable,weusethesamelanguage

asthearticlescited.aHR,adjusted

hazard

ratio;aOR,adjustedodds

ratio;BM,breastm

ilk;BMF,breastmilk

feeding;

CDC,

CentersforDiseaseControlandPrevention;

CI,confidenceinterval;CPQCC,

CaliforniaPerinatalQuality

Care

Collaborative;EI,earlyintervention;

HM,human

milk;KM

C,kangaroo

mothercare;LBW

,lowbirthweight;MCP,maternalcare

partnerships;N

/A,not

available;OR,oddsratio;S

FR,singlefamily

room

.

14 SIGURDSON et al by guest on December 27, 2020www.aappublications.org/newsDownloaded from

reported being discouraged fromkangaroo care than white mothers.Among nurse respondents, nurses ofcolor and foreign-born nurses weremore likely to encourage maternalcare partnerships and kangaroo care.

Surfactant and Respiratory DistressSyndrome

Researchers for 2 studies examinedracial and/or ethnic disparities inoutcomes for infants with VLBW afterthe introduction of surfactant therapy.Hamvas et al34 found that aftersurfactant therapy for respiratorydistress syndrome (RDS) becamegenerally available, neonatalmortality improved more for whitethan black infants, but thesedifferences were explained by thehigher prevalence of RDS amongwhite infants. Ranganathan et al35

reported similar results andhypothesized differential efficacy ofsurfactant on non-Hispanic white andAfrican American infants owing tomore rapid lung maturation andsurfactant production among AfricanAmerican fetuses.

The authors of 1 article lookedspecifically at the disparity inmortality due to RDS between blackand white infants. Frisbie et al36

discovered that disparity increased inthe years after surfactant wasapproved by the US Food and DrugAdministration, indicating that thisnew therapy benefitted white infantswith RDS more than black infantswith RDS. According to this study,black infants had a relative RDSsurvival advantage in thepresurfactant era and a survivaldisadvantage after the introduction ofsurfactant.

In 1 article, researchers explicitlyinvestigated disparities in theadministration of surfactant itself. In2010, in 3 New York hospitals,Howell et al37 found that clinicianswere falling short ofrecommendations for surfactantoverall, but white infants were morelikely to get surfactant treatment than

their black or Hispanic and Latinocounterparts. However, when Hamvaset al34 looked at the same question,they did not find racial and/or ethnicdisparities in surfactant use.

Outcomes

We identified 11 articles thatexamined racial and/or ethnicdisparities in outcomes potentiallydue to differences in quality of caredelivery (Table 3).

Intraventricular Hemorrhage

One article39 examinedintraventricular hemorrhage (IVH)-related mortality, finding that blackinfants had a twofold greater risk ofIVH-related mortality compared withwhite infants, positinga pathophysiological predisposition inblack infants to IVH.

Necrotizing Enterocolitis and IntestinalFailure

Two articles were focused onconditions of the gastrointestinaltract: necrotizing enterocolitis(NEC)40 and survival with intestinalfailure.41 Both revealed disparity thatdisadvantaged infants of color. Guneret al40 noted that Hispanic infantswith NEC were less likely to survivethan non-Hispanic infants, but blackand white infants with NEC hadsimilar survival rates. Analyzing dataon infants who received parenteralnutrition for.60 days, Squires et al41

found that infants of color were lesslikely to receive intestinal transplantby 48 months after the criteria fortransplant was met and were morelikely to die without transplants.

Overall Morbidity or Mortality

Three articles were focused onoverall morbidity or mortality,revealing contradictory trends. In 2articles,42,43 the researchers indicateda disadvantage for infants of color.Jacob et al42 found that AlaskanNative infants had higher rates ofNEC, retinopathy of prematurity(ROP), and chronic lung diseasecompared with non-Native infants.

Pont and Carter43 found a significantdisparity in mortality amongextremely low birth weight infants(,750 g) who died within the first24 hours after birth. When comparing2 different time periods, 1995–1999and 2000–2005, Pont and Carter43

noted that black and Hispanicmortality increased, whereas the rateamong white infants decreased.However, Townsel et al44 noted thatamong very preterm infants (,28weeks’ gestational age [GA]), blackinfants had lower in-hospitalmortality than white infants. Blackand multiracial infants also had lowerrisk of RDS, and multiracial infantshad a lower risk of ROP.

Other Specific Outcomes

Of the 5 articles in this category, 2revealed a disparity that advantagedwhite infants over infants of otherracial and ethnic groups, and 3revealed race and/or ethnicity to notbe predictive of poor outcomes. Collinset al45 found that infant mortality dueto congenital heart disease was higherfor African American than non-Hispanicwhite infants, and Shin et al46 notedsignificant improvement from 1983 to2002 in 1-year survival rates for whiteand Hispanic patients with spina bifidabut not for black infants. Oyetunjiet al,47 however, found that race orethnicity was not an independentpredictor of mortality for infants onextracorporeal membrane oxygenation(ECMO). Similarly, Morris et al48 notedthat race was not a significant predictorof rehospitalization for extremely lowbirth weight. In addition, Collaco et al49

examined whether disparities in careexisted for major NICU therapydecisions, finding that race and/orethnicity did not affect the prescriptionof gastrostomy tubes, supplementaloxygen, or length of initial stay inthe NICU.

DISCUSSION

In this review, we providea comprehensive summary of theliterature in this area and suggest that

PEDIATRICS Volume 144, number 2, August 2019 15 by guest on December 27, 2020www.aappublications.org/newsDownloaded from

TABLE3System

aticReview

Outcom

e

Variable,

Author,and

Reference

Demographicsof

Study(If

Available)

Outcom

e(s)

ofInterest

Summaryof

MainFindings

RRor

Measuresof

Effect

(IfApplicable)

SuggestedMechanism

IVH Qu

reshi

etal39

3249

neonates

and

infantswith

IVH-

related

mortalities.

IVH-relatedmortality.

Incidenceratesof

IVHwerehigher

among

African-Am

erican

infantsthan

white

infants.AfricanAm

erican

infantshad

atwofoldhigher

risk

ofIVHmortality

comparedwith

whiteinfants.Significant

increase

inincidenceover

thestudy

period

forwhite

butnotAfrican

American

infants.

RRof

IVHmortalityforAfricanAm

erican

infantscomparedwith

white:2.0.

Higher

prevalence

ofrespiratorydiseases,

LBWs,andpreterm

deliveriesforAfrican

American

infants.Pathophysiological

characteristicsthat

may

predispose

AfricanAm

erican

infantsto

IVHmay

also

exist.

NECand/or

intestinal

failure

Guneret

al40

3328

infantswith

NEC.

Factorsthat

predictNEC-related

outcom

edisparities.

Overallmortalityforinfantswith

NECwas

12.5%.M

aleandHispanicinfantsless

likelyto

survive.Blackandwhiteinfants

have

similarsurvival

rates.

Odds

ratio

ofmortalityforHispanic

comparedwith

others:1.44(after

adjustmentforBW

).

None

Squires

etal41

272infants(,

12mo)

receiving

parenteral

nutrition

for601

d.

Survival

with

intestinal

failure

andlikelihoodof

receivingan

ITx.

Childrenofcolormorelikelyto

diewithout

ITxandless

likelytoreceiveITxby

48mo

aftercriteriamet.

CIPfordeathwithoutITx:white:0.16,

nonw

hite:0.40.CIPforITx:white:0.31,

nonw

hite:0.07.

Lower

ratesof

breast

feedingandhigher

ratesof

bloodstream

infections

among

blackinfantscouldcontribute,but

the

data

used

here

didnotshow

significant

differences

inhuman

milk

exposure

orsepsis.

Overall

morbidity

ormortality

Jacobet

al42

137Nativeand

469non-Native

infantswith

LBW.

BW-specificdifferences

inneonatal

morbiditiesfor

Alaska

Natives

comparedwith

non-Natives.

MoreNEC,moresevere

ROP,andmoreBPD

amongNativeinfantscomparedwith

non-Native.Forinfants1500–2499

gthat

needed

ventilatory

assistance,there

was

moreIVHoverall,moresevere

IVH,

and

moreacquired

sepsisam

ongNative

infants.

Allstudy

infants,Nativeversus

non-Native-

NEC:13%

vs4.9%

,ROP:12%

vs4.6%

,BPD:

49%

vs34%.Infants

1500–2499

gwith

ventilatory

assistance,N

ativevs

non-Native-IVH:19%

vs7.4%

,severeIVH:

9.5%

vs0.9%

,sepsis:7.1%

vs1.7%

.

Differences

may

bedueto

differences

inaccess

toLevelIIIperinatalcare

and

intrapartum

care.

Pont

and

Carter

43138infants,750g

who

died

within

first24

hof

life.

Characteristicsof

E-ELBW

-NS.

MoreblackandHispanicE-ELBW

-NSin

2000–2005

comparedwith

1995–1999;

increase

inblackinfantswas

not

mirroredby

totalNICU

admissionsor

NICU

admissions,750g.

BlackE-ELBW

-NSfrom

firstto

second

epoch:14.0%to

38.8%;H

ispanic:1.8%

to4.9%

;white:84.2%

to54.3%.

May

bedueto

confoundingvariables(such

asmaternalchorioam

nionitis)

that

this

samplesize

was

toosm

allto

determ

ine;

thereareknow

ndisparities

inoverall

IMR.

16 SIGURDSON et al by guest on December 27, 2020www.aappublications.org/newsDownloaded from

TABLE3

Continued

Variable,

Author,and

Reference

Demographicsof

Study(If

Available)

Outcom

e(s)

ofInterest

Summaryof

MainFindings

RRor

Measuresof

Effect

(IfApplicable)

SuggestedMechanism

Townsel

etal44

4802

LevelIIINICU

admits,G

A,37

wk.

IHM,R

DS,IVH,N

EC,and

ROP,

comparedby

race

and/or

ethnicity.

Nooveralldifferences

inIHM.For

very

preterm

infants(,

28wkGA),black

infantshadlower

IHM

than

white

infants.Comparedwith

white

infants

afterrisk-adjustm

ent,blackand

multiracialinfantshadlower

risk

ofRD

S,Hispanicinfantshadhigher

risk,

andmultiracialinfantshadlower

risk

ofRO

P.

ROP,Hispanicversus

white:aOR

51.70.

ROP,multiracialversus

white:aOR

50.45.R

DS,black

versus

white:aOR

50.57.R

DS,m

ultiracialversus

white:aOR

50.67.IHM

,black

versus

white:

unadjusted

OR5

0.74.

Differences

inretinal

pigm

entationmay

beprotectiveagainstRO

Pformultiracial

infants,especiallyiftheinfantsin

this

samplehave

ahigher

concentrationof

Africanancestry.

Other

Collins

etal45

3684569NH

Wand

782452African

American

term

infants.

IMRdueto

CHD.

IMRdueto

CHDwas

higher

forAfrican

American

than

non-Hispanicwhite

infants.Disparity

was

even

greaterin

thepostneonatal

period.

RRoffirstyear

mortality,AfricanAm

erican

versus

non-Hispanicwhite

mothers

51.36.Postneonatalperiod

RR5

1.53,

neonatal

51.20

(not

significantly

different).

Individual

levelrisk

factors(eg,greater

exposure

tostressorsand/or

toxins

orlack

ofaccess

toquality

health

care

due

tolivingin

animpoverished

area

inAfricanAm

erican

population)

may

explain

PNMRdisparity;h

owever,adjustingfor

maternaldemographicrisk

only

minimallyweakens

theassociation.

Shin

etal46

5165

infantswith

spinabifida.

Survival

with

spinabifida

at1,5,

and20

y.Significant

improvem

entin

1-ysurvival

between1983

and2002

forwhite

and

Hispanicinfantsbutnotblackinfants.

Survival

probabilityhighestforwhite,

then

Hispanic,thenblackinfants.

Survival

forblackinfantsworse

than

white

infantsof

norm

albirthwt,worse

forHispanicthan

whiteinfantswith

LBW.

Change

in1-ysurvival,1983–2002:w

hite

infants:87.9%–95.9%;H

ispanicinfants:

88.3%–92.6%;b

lack

infants:

79.1%–87.5%

(trend

notstatistically

significant);norm

alBW

mortalityfor

blackversus

white,aHR

at1y5

2.1,at

8y5

1.9;LBWmortalityforHispanic

versus

whiteinfants:4.2,aHRat

1mo5

6.0,at

1y5

4.2,at

8y5

3.7.

Possiblybarriers

inaccess

toquality

health

care.

Oyetunji

etal47

827neonates

ECMOoutcom

esEthnicminorities

wereoverrepresentedin

ECMOpopulation,butrace

and/or

ethnicity

was

notamajor

independent

predictorof

mortalityon

ECMO;ratesof

ECMOuseam

ongAfricanAm

erican

and

Hispanicinfantshasincreasedfaster

than

theincrease

inpopulation,

whereas

ECMOam

ongwhite

infantshas

decreased.

RRnotreported.

Moreethnicminority

infantshave

severe

diagnosesthat

requireECMO;may

bedue

tolower

SESandresulting

inadequate

prenatal

care.

PEDIATRICS Volume 144, number 2, August 2019 17 by guest on December 27, 2020www.aappublications.org/newsDownloaded from

disparities in NICU quality of careexist at all levels (structure, process,outcome) and generallydisadvantage infants of color, withsome areas of exception.14,19,34,35,44

Two primary themes driving thesedisparities emerged. First, infants ofcolor, especially black and Hispanicinfants, are more likely to receivecare in quality-challenged hospitals.Second, disparities also exist withinNICUs. Although some overlapexists, strategies to addressdisparities may differ.

The theme of worse care in minority-serving hospitals recurs throughoutthis review.* Although these findingscould be secondary to otherconfounders such as socioeconomicdifferences in influencing disparateoutcomes,39–43,45,46 we also foundracial and/or ethnic disparitiesacross processes9,21,22,24,27–31,33,37

that disadvantage infants of color.This indicates that disparities may beunder the control of providers andamenable to improvement by usingavailable QI tools. Minority-servinghospitals are often underresourced,may lack hospital or system-wide QIinfrastructure, and uncommonlyparticipate in collaborative QI effortsas offered, for example, through theCalifornia Perinatal Quality CareCollaborative50,51 or the VermontOxford Network.52 Theseorganizations have a role to play inorganizing and testing such efforts,potentially in partnership andfunding from governmental agenciesat the federal and state level. Lionand Raphael53 suggest that an idealQI initiative would have an amplifiedeffect on the populationsexperiencing worse quality andwould engage community resourcessuch as frontline providers andcommunity members committed toimproving the care of disadvantagedpopulations. We believe that effortsto specifically target improvementamong the most challenged hospitals

TABLE3

Continued

Variable,

Author,and

Reference

Demographicsof

Study(If

Available)

Outcom

e(s)

ofInterest

Summaryof

MainFindings

RRor

Measuresof

Effect

(IfApplicable)

SuggestedMechanism

Morris

etal48

2446

ELBW

infants

Rehospitalizationbefore

18mo

correctedage(age

from

expected

duedate)andcauses

ofrehospitalization.

Race

was

notasignificant

predictorfor

rehospitalizationoverall.White

infants

weremorelikelytobe

rehospitalized

for

grow

thandnutrition

than

blackinfants.

ORof

rehospitalizationforgrow

thand

nutrition

forwhite

versus

blackinfants:

2.2.

Differencein

parental

expectations

for

grow

th;physician

bias

inreferral

for

rehospitalization.

Collaco

etal49

135patientswith

chroniclung

diseaseof

prem

aturity.

Risk

factorsforuseof

major

therapies(wedidnotinclude

findings

onrespiratory

morbiditiesbecausethey

were

long-term

outcom

es).

Noracialor

ethnicdisparity

foundinmajor

treatm

entdecisions(prescriptionof

homeoxygen,gastrostomytubes,or

initial

length

ofstay).

N/A

Diseaseseveritymay

outweigh

anyother

factors;disparities

may

besubtle;

confoundingfactorsmay

reduce

effectsof

racial

and/or

ethnicfactors.

“African

American”and“black”areoftenused

interchangeablyintheliteraturereview

ed.Inthistable,weusethesamelanguage

asthearticlescited.aHR,adjusted

hazard

ratio;aOR,adjustedodds

ratio;BPD,bronchopulmonarydysplasia;BW

,birthweight;CHD,congenitalheartdisease;CIP,cumulativeincidenceprobability;E-ELBW-NS,earlyextrem

elylowbirthweightnonsurvivors;IHM

,in-hospitalm

ortality;IMR,infant

mortalityrate;ITx,intestinaltransplant;LBW

,low

birthweight;N/A,

notavailable;PNMR,

perinatalmortalityrate;R

R,relativerisk;S

ES,socioeconom

icstatus.

* Refs 9–11,15,17,18,20,21,43,45.

18 SIGURDSON et al by guest on December 27, 2020www.aappublications.org/newsDownloaded from

that serve affected communities holdpromise for reducing disparity.54

Strategies for addressing withinNICU disparities may require thedevelopment of new measures. Manykey aspects of NICU care quality arecurrently not systematicallymeasured but are nonethelessimportant. For example, family-centered care is not currentlyroutinely measured andsystematized across statewide ornationwide NICU data repositories.We are currently developingmeasures for family-centered care inthe NICU that can be standardizedacross NICUs that are meaningful toconsulted minority families and wehope to be sensitive to disparities.30

We hope these efforts will open upnew paths for improvement in familyexperience and quality of care for allfamilies, much in the same way asrecent efforts in assessing andimproving maternity care havedone.55

The term “disparities” is usedthroughout this article; however, wesuggest that this terminology risksnaturalizing racial and/or ethnicdifferences rather than provokingaction. That is, identifying racialdisparities runs the risk of reinforcingthe status quo56,57 and may lead toseeing differences as “natural” andnot amenable to change. For instance,differences can be naturalizedthrough racially biasedpreconceptions about culturalbehaviors51 or biological causes andtheir impact on health disparities.51

This review should be viewed in lightof its design. It may be that somerelevant articles were omitted due tolack of full-text availability,unavailability of publication in theEnglish language, or items notavailable through our search sources.However, articles retrieved in our

study are a representative sample ofliterature in this field that would bereadily accessible to most NICUhealth care practitioners. Additionally,research has been published in thisarea since the time of our search.58–62

Due to study heterogeneity, meta-analysis of data was not possible. Ourdefinition of measurement of qualityof care has relied on input fromexperts in our previous research.8

Although we use this frameworkexplicitly for this review, it isimportant to understand thatdisparities in care and outcomes mayoverlap with biologic andsocioeconomic factors that may bedifficult to untangle.3 We do notaddress these important issues in ourreview. However, there is ampleevidence that racial and ethnicminorities have disadvantages acrossthe spectrum of factors identified asrepresenting quality of NICU care.These disadvantages aremultifactorial but find expression notjust across populations but also inundue variation across hospitals. Weare currently developing a dashboardto track NICU performance over timeby racial and/or ethnic disparities toserve and inform accountable careorganizations in improvingpopulation health and achievinghealth equity.

Gaps remain in the research. Althougha history of institutional racism helpsexplain why racial and ethnicminorities are more likely to be caredfor in worse quality NICUs, what arethe contemporary mechanisms thatcontinue to stratify infants intodifferent care settings? Althoughinterpersonal racism can explainsome within-NICU variation,particularly in process of caremeasures (eg, racially biased bedsidepractices, clinician attitudes), how dodifferences in processes translate into

worse outcomes? We believe that thisresearch requires both quantitativeand qualitative approaches designedto generate hypotheses and testsolutions and should involve affectedcommunities as advisors orpartners.53

CONCLUSIONS

In this review of racial and ethnicdisparities in NICU quality of care, wesuggest that disparities in neonataloutcomes are partly consequences ofdifferential quality of care or access tohigh-quality care. That is, NICUs arenot isolated from racism anddisparities in infant outcomes“increasingly reflect the interactionbetween social forces and technicalinnovation.”63 Researchers are nowlooking at chronic stress caused bysocietal, institutional, orinterpersonal racism as causal factorsfor preterm birth,64,65 and in thisreview, we suggest that these factorsare also causal factors for racial andethnic disparities in NICU quality ofcare. Targeted QI efforts hold promisefor improving racial and ethnic equityin care delivery.

ABBREVIATIONS

BMF: breast milk feedingDHM: donor human milkECMO: extracorporeal membrane

oxygenationGA: gestational ageIVH: intraventricular hemorrhageNEC: necrotizing enterocolitisQI: quality improvementRDS: respiratory distress

syndromeROP: retinopathy of prematurityVLBW: very low birth weightWIC: Supplemental Nutrition

Program for Women, Infants,and Children

PEDIATRICS Volume 144, number 2, August 2019 19 by guest on December 27, 2020www.aappublications.org/newsDownloaded from

Address correspondence to Krista Sigurdson, PhD, Perinatal Epidemiology and Health Outcomes Research Unit, Division of Neonatology, Department of Pediatrics,

Stanford University School of Medicine, MSOB Room x1C19, 1265 Welch Rd, Stanford, CA 94305. E-mail: ksigurd@stanford.edu

PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275).

Copyright © 2019 by the American Academy of Pediatrics

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

FUNDING: Drs Sigurdson, Morton, and Profit and Ms Mitchell are supported by a grant from the Eunice Kennedy Shriver National Institute of Child Health and

Human Development (R01 HD083368-01, principle investigator: Dr Profit). Dr Sigurdson is also supported by a postdoctoral support award from the Stanford

Maternal and Child Health Research Institute. Funded by the National Institutes of Health (NIH).

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

REFERENCES

1. March of Dimes. 2018 Premature BirthReport Card. Available at: https://www.marchofdimes.org/peristats/tools/reportcard.aspx. Accessed June 14, 2019

2. Riddell CA, Harper S, Kaufman JS.Trends in differences in us mortalityrates between black and white infants.JAMA Pediatr. 2017;171(9):911–913

3. Lorch SA. Health equity and quality ofcare assessment: a continuingchallenge. Pediatrics. 2017;140(3):1–2

4. Hassett MJ, Schymura MJ, Chen K,Boscoe FP, Gesten FC, Schrag D.Variation in breast cancer care qualityin New York and California based onrace/ethnicity and Medicaid enrollment.Cancer. 2016;122(3):420–431

5. Dong L, Fakeye OA, Graham G, Gaskin DJ.Racial/ethnic disparities in quality ofcare for cardiovascular disease inambulatory settings: a review. MedCare Res Rev. 2018;75(3):263–291

6. Liberati A, Altman DG, Tetzlaff J, et al.The PRISMA statement for reportingsystematic reviews and meta-analysesof studies that evaluate healthcareinterventions: explanation andelaboration. BMJ. 2009;339:b2700

7. Donabedian A. The quality of care. Howcan it be assessed? JAMA. 1988;260(12):1743–1748

8. Profit J, Kowalkowski MA, Zupancic JA,et al. Baby-MONITOR: a compositeindicator of NICU quality. Pediatrics.2014;134(1):74–82

9. Profit J, Gould JB, Bennett M, et al.Racial/ethnic disparity in NICU qualityof care delivery. Pediatrics. 2017;140(3):e20170918

10. Lake ET, Staiger D, Horbar J, Kenny MJ,Patrick T, Rogowski JA. Disparities in

perinatal quality outcomes for very lowbirth weight infants in neonatalintensive care. Health Serv Res. 2015;50(2):374–397

11. Lake ET, Staiger D, Edwards EM, SmithJG, Rogowski JA. Nursing caredisparities in neonatal intensive careunits. Health Serv Res. 2018;53(suppl 1):3007–3026

12. Bronstein JM, Capilouto E, Carlo WA,Haywood JL, Goldenberg RL. Access toneonatal intensive care for low-birthweight infants: the role ofmaternal characteristics. Am J PublicHealth. 1995;85(3):357–361

13. Gould JB, Sarnoff R, Liu H, Bell DR,Chavez G. Very low birth weight birthsat non-NICU hospitals: the role ofsociodemographic, perinatal, andgeographic factors. J Perinatol. 1999;19(3):197–205

14. Gortmaker S, Sobol A, Clark C, WalkerDK, Geronimus A. The survival of verylow-birth weight infants by level ofhospital of birth: a population study ofperinatal systems in four states. AmJ Obstet Gynecol. 1985;152(5):517–524

15. Hebert PL, Chassin MR, Howell EA. Thecontribution of geography to black/white differences in the use of lowneonatal mortality hospitals in NewYork City. Med Care. 2011;49(2):200–206

16. Featherstone P, Eberth JM, Nitcheva D,Liu J. Geographic accessibility to healthservices and neonatal mortality amongvery-low birthweight infants in SouthCarolina. Matern Child Health J. 2016;20(11):2382–2391

17. Morales LS, Staiger D, Horbar JD, et al.Mortality among very low-birthweightinfants in hospitals serving minority

populations. Am J Public Health. 2005;95(12):2206–2212

18. Howell EA, Hebert P, Chatterjee S,Kleinman LC, Chassin MR. Black/whitedifferences in very low birth weightneonatal mortality rates among NewYork City hospitals. Pediatrics. 2008;121(3). Available at: www.pediatrics.org/cgi/content/full/121/3/e407

19. Kugler JP, Connell FA, Henley CE. Lack ofdifference in neonatal mortality betweenblacks and whites served by the samemedical care system. J Fam Pract. 1990;30(3):281–287; discussion 287–288

20. Howell EA, Janevic T, Hebert PL, EgorovaNN, Balbierz A, Zeitlin J. Differences inmorbidity and mortality rates in black,white, and Hispanic very preterminfants among New York city hospitals.JAMA Pediatr. 2018;172(3):269–277

21. Lee HC, Gould JB, Martin-Anderson S,Dudley RA. Breast-milk feeding of verylow birth weight infants as a function ofhospital demographics. J Perinatol.2011;31:S82–S82

22. Cricco-Lizza R. Black non-Hispanicmothers’ perceptions about thepromotion of infant-feeding methods bynurses and physicians. J Obstet GynecolNeonatal Nurs. 2006;35(2):173–180

23. Fleurant E, Schoeny M, Hoban R, et al.Barriers to human milk feeding atdischarge of very-low-birth-weight infants:maternal goal setting as a key socialfactor. Breastfeed Med. 2017;12:20–27

24. Brownell EA, Smith KC, Cornell EL, et al.Five-year secular trends and predictorsof nonconsent to receive donor milk inthe neonatal intensive care unit.Breastfeed Med. 2016;11(6):281–285

25. Boundy EO, Perrine CG, Nelson JM,Hamner HC. Disparities in hospital-

20 SIGURDSON et al by guest on December 27, 2020www.aappublications.org/newsDownloaded from

reported breast milk use in neonatalintensive care units - United States,2015. MMWR Morb Mortal Wkly Rep.2017;66(48):1313–1317

26. Vohr B, McGowan E, McKinley L, TuckerR, Keszler L, Alksninis B. Differentialeffects of the single-family roomneonatal intensive care unit on 18- to24-month Bayley scores of preterminfants. J Pediatr. 2017;185:42–48.e1

27. Barfield WD, Clements KM, Lee KG,Kotelchuck M, Wilber N, Wise PH. Usinglinked data to assess patterns of earlyintervention (EI) referral among verylow birth weight infants. Matern ChildHealth J. 2008;12(1):24–33

28. Hintz SR, Gould JB, Bennett MV, et al.Referral of very low birth weightinfants to high-risk follow-up atneonatal intensive care unit dischargevaries widely across California.J Pediatr. 2015;166(2):289–295

29. Martin AE, D’Agostino JA, Passarella M,Lorch SA. Racial differences in parentalsatisfaction with neonatal intensivecare unit nursing care. J Perinatol.2016;36(11):1001–1007

30. Sigurdson K, Morton C, Mitchell B, ProfitJ. Disparities in NICU quality of care:a qualitative study of family andclinician accounts [publishedcorrection appears in J Perinatol. 2018;38(8):1123]. J Perinatol. 2018;38(5):600–607

31. Van McCrary S, Green HC, Combs A,Mintzer JP, Quirk JG. A delicate subject:the impact of cultural factors onneonatal and perinatal decisionmaking. J Neonatal Perinatal Med.2014;7(1):1–12

32. Tucker Edmonds B, Fager C, Srinivas S,Lorch S. Racial and ethnic differencesin use of intubation for periviableneonates. Pediatrics. 2011;127(5).Available at: www.pediatrics.org/cgi/content/full/127/5/e1120

33. Hendricks-Muñoz KD, Li Y, Kim YS,Prendergast CC, Mayers R, Louie M.Maternal and neonatal nurse perceivedvalue of kangaroo mother care andmaternal care partnership in theneonatal intensive care unit. AmJ Perinatol. 2013;30(10):875–880

34. Hamvas A, Wise PH, Yang RK, et al. Theinfluence of the wider use of surfactanttherapy on neonatal mortality among

blacks and whites. N Engl J Med. 1996;334(25):1635–1640

35. Ranganathan D, Wall S, Khoshnood B,Singh JK, Lee KS. Racial differencesin respiratory-related neonatalmortality among very low birthweight infants. J Pediatr. 2000;136(4):454–459

36. Frisbie WP, Song SE, Powers DA, StreetJA. The increasing racial disparity ininfant mortality: respiratory distresssyndrome and other causes.Demography. 2004;41(4):773–800

37. Howell EA, Holzman I, Kleinman LC,Wang J, Chassin MR. Surfactant use forpremature infants with respiratorydistress syndrome in three New YorkCity hospitals: discordance of practicefrom a community clinician consensusstandard. J Perinatol. 2010;30(9):590–595

38. Riley B, Schoeny M, Rogers L, et al.Barriers to human milk feeding atdischarge of very low-birthweightinfants: evaluation of neighborhoodstructural factors. Breastfeed Med.2016;11(11):335–342

39. Qureshi AI, Adil MM, Shafizadeh N,Majidi S. A 2-fold higher rate ofintraventricular hemorrhage-relatedmortality in African American neonatesand infants. J Neurosurg Pediatr. 2013;12(1):49–53

40. Guner YS, Friedlich P, Wee CP, Dorey F,Camerini V, Upperman JS. State-basedanalysis of necrotizing enterocolitisoutcomes. J Surg Res. 2009;157(1):21–29

41. Squires RH, Balint J, Horslen S, et al;Pediatric Intestinal Failure Consortium.Race affects outcome among infantswith intestinal failure. J PediatrGastroenterol Nutr. 2014;59(4):537–543

42. Jacob J, Hulman S, Davis RF, PfenningerJ. Racial differences in newbornintensive care morbidity in Alaska.Alaska Med. 2001;43(2):32–37

43. Pont MM, Carter BS. Maternal andneonatal characteristics of extremelylow birth weight infants who die in thefirst day of life. J Perinatol. 2009;29(1):33–38

44. Townsel C, Keller R, Kuo CL, CampbellWA, Hussain N. Racial/ethnic disparitiesin morbidity and mortality for preterm

neonates admitted to a tertiaryneonatal intensive care unit. J RacialEthn Health Disparities. 2018;5(4):867–874

45. Collins JW Jr, Soskolne G, Rankin KM,Ibrahim A, Matoba N. African-American:white disparity in infant mortality dueto congenital heart disease. J Pediatr.2017;181:131–136

46. Shin M, Kucik JE, Siffel C, et al.Improved survival among children withspina bifida in the United States.J Pediatr. 2012;161(6):1132–1137

47. Oyetunji TA, Thomas A, Moon TD, et al.The impact of ethnic populationdynamics on neonatal ECMO outcomes:a single urban institutional study.J Surg Res. 2013;181(2):199–203

48. Morris BH, Gard CC, Kennedy K.Rehospitalization of extremely low birthweight (ELBW) infants: are there racial/ethnic disparities? J Perinatol. 2005;25(10):656–663

49. Collaco JM, Choi SJ, Riekert KA, EakinMN, McGrath-Morrow SA, Okelo SO.Socio-economic factors and outcomesin chronic lung disease of prematurity.Pediatr Pulmonol. 2011;46(7):709–716

50. Gould JB. The role of regionalcollaboratives: the California PerinatalQuality Care Collaborative model. ClinPerinatol. 2010;37(1):71–86

51. Pai VV, Lee HC, Profit J. ImprovingUptake of Key Perinatal InterventionsUsing Statewide Quality Collaboratives.Improving uptake of key perinatalinterventions using statewide qualitycollaboratives. Clin Perinatol. 2018;45(2):165–180

52. Horbar JD, Rogowski J, Plsek PE, et al;NIC/Q Project Investigators of theVermont Oxford Network. Collaborativequality improvement for neonatalintensive care. Pediatrics. 2001;107(1):14–22

53. Lion KC, Raphael JL. Partnering healthdisparities research with qualityimprovement science in pediatrics.Pediatrics. 2015;135(2):354–361

54. Rogowski JA, Staiger DO, Horbar JD.Variations in the quality of care forvery-low-birthweight infants:implications for policy. Health Aff(Millwood). 2004;23(5):88–97

PEDIATRICS Volume 144, number 2, August 2019 21 by guest on December 27, 2020www.aappublications.org/newsDownloaded from

55. Vedam S, Stoll K, Martin K, et al;Changing Childbirth in BC SteeringCouncil. The Mother’s Autonomy inDecision Making (MADM) scale: patient-led development and psychometrictesting of a new instrument to evaluateexperience of maternity care. PLoS One.2017;12(2):e20170918

56. Hetey RC, Eberhardt JL. The numbersdon’t speak for themselves: racialdisparities and the persistence ofinequality in the criminal justicesystem. Curr Dir Psychol Sci. 2018;27(3):183–187

57. Benjamin R. Cultura obscura: race,power, and “culture talk” in the healthsciences. Am J Law Med. 2017;43(2–3):225–238

58. Patel AL, Schoeny ME, Hoban R, et al.Mediators of racial and ethnic disparity

in mother’s own milk feeding in verylow birth weight infants. Pediatr Res.2019;85(5):662–670

59. Parker MG, Gupta M, Melvin P,et alet al. Racial and ethnicdisparities in the use of mother’s milkfeeding for very low birth weightinfants in Massachusetts. J Pediatr.2019;204:134–141.e1

60. Liu J, Sakarovitch C, Sigurdson K, LeeHC, Profit J. Disparities in health care-associated infections in the NICU. AmJ Perinatol. 2019

61. Horbar JD, Edwards EM, Greenberg LT,et alet al. Racial segregation andinequality in the neonatal intensive careunit for very low-birth-weight and verypreterm infants. JAMA Pediatr. 2019;173(5):455

62. Jammeh ML, Adibe OO, Tracy ET,et alet al. Racial/ethnic differences innecrotizing enterocolitis incidence andoutcomes in premature very low birthweight infants. J Perinatol. 2018;38(10):1386–1390

63. Wise PH. The anatomy of a disparity ininfant mortality. Annu Rev Public Health.2003;24:341–362

64. Braveman P, Heck K, Egerter S, et al.Worry about racial discrimination:a missing piece of the puzzle of black-white disparities in preterm birth?PLoS One. 2017;12(10):e0186151

65. McLemore MR, Altman MR, Cooper N,Williams S, Rand L, Franck L. Healthcare experiences of pregnant, birthingand postnatal women of color at riskfor preterm birth. Soc Sci Med. 2018;201(201):127–135

22 SIGURDSON et al by guest on December 27, 2020www.aappublications.org/newsDownloaded from

DOI: 10.1542/peds.2018-3114 originally published online July 29, 2019; 2019;144;Pediatrics 

Henry C. Lee, Nicole Capdarest-Arest and Jochen ProfitKrista Sigurdson, Briana Mitchell, Jessica Liu, Christine Morton, Jeffrey B. Gould,

Racial/Ethnic Disparities in Neonatal Intensive Care: A Systematic Review

ServicesUpdated Information &

http://pediatrics.aappublications.org/content/144/2/e20183114including high resolution figures, can be found at:

Referenceshttp://pediatrics.aappublications.org/content/144/2/e20183114#BIBLThis article cites 63 articles, 8 of which you can access for free at:

Subspecialty Collections

http://www.aappublications.org/cgi/collection/neonatology_subNeonatologysubhttp://www.aappublications.org/cgi/collection/fetus:newborn_infant_Fetus/Newborn Infantsubhttp://www.aappublications.org/cgi/collection/quality_improvement_Quality Improvemente_management_subhttp://www.aappublications.org/cgi/collection/administration:practicAdministration/Practice Managementfollowing collection(s): This article, along with others on similar topics, appears in the

Permissions & Licensing

http://www.aappublications.org/site/misc/Permissions.xhtmlin its entirety can be found online at: Information about reproducing this article in parts (figures, tables) or

Reprintshttp://www.aappublications.org/site/misc/reprints.xhtmlInformation about ordering reprints can be found online:

by guest on December 27, 2020www.aappublications.org/newsDownloaded from

DOI: 10.1542/peds.2018-3114 originally published online July 29, 2019; 2019;144;Pediatrics 

Henry C. Lee, Nicole Capdarest-Arest and Jochen ProfitKrista Sigurdson, Briana Mitchell, Jessica Liu, Christine Morton, Jeffrey B. Gould,

Racial/Ethnic Disparities in Neonatal Intensive Care: A Systematic Review

http://pediatrics.aappublications.org/content/144/2/e20183114located on the World Wide Web at:

The online version of this article, along with updated information and services, is

http://pediatrics.aappublications.org/content/suppl/2019/07/24/peds.2018-3114.DCSupplementalData Supplement at:

by the American Academy of Pediatrics. All rights reserved. Print ISSN: 1073-0397. the American Academy of Pediatrics, 345 Park Avenue, Itasca, Illinois, 60143. Copyright © 2019has been published continuously since 1948. Pediatrics is owned, published, and trademarked by Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it

by guest on December 27, 2020www.aappublications.org/newsDownloaded from