journal homepage:

Original Article

Association between rooming-in policy and

neonatal hyperbilirubinemia
Ken-Hsyuan Shan a, Teh-Ming Wang a, Ming-Chih Lin a,b,c,*

a Department of Pediatrics, Taichung Veterans General Hospital, Taiwan b Department of

Pediatrics, School of Medicine, National Yang-Ming University, Taiwan c Department of Food and
Nutrition, Providence University, Taichung, Taiwan

Received Apr 28, 2017; received in revised form Dec 8, 2017; accepted Jun 8, 2018
Available online 12 June hyperbilirubinemia has seldom been reported. The aim of this study was to evaluate
2018 the association between rooming-in and neonatal hyperbilirubinemia.
Methods: This was a retrospective cohort study. Term neonates were consecutively
enrolled from the nursery of a medical center from January 2011 to December
2013. During the study period, rooming-in care was strongly encouraged according
Key Words to the World Health Organization guidelines, if the parents agreed. The endpoint
breastfeeding; was defined as admission for phototherapy. Risk of neonatal hyperbilirubinemia in
neonatal rooming-in neonates was calculated. Potential confounding factors, including
hyperbiliru- exclusive breastfeeding, potential ABO incompatibility, Glucose6-Phosphate
binemia; Dehydrogenase (G6PD) deficiency, and body weight loss (BWL), were adjusted by
rooming-in multiple logistic regression models.
Results: Totally, 3341 infants were enrolled in this study after excluding 40 infants
admitted for other reasons. The rooming-in rate increased yearly during the study
Background: The practices period. However, the rate of neonatal hyperbilirubinemia also increased
promoted by the Baby- simultaneously. The odds ratio ( OR ) of neonatal hyperbilirubinemia in the
friendly Hospital Initiative rooming-in group was 7.04 (95% CI, 4.41w11.24). The rooming-in group
have become a part of demonstrated a higher percentage of exclusive breastfeeding and BWL >10% at 3
current mainstream days of age. After adjusting for potential confounding factors, rooming-in was still a
postpartum infant care. significant risk factor for neonatal hyperbilirubinemia (OR: 8.48; 95% CI:
Rooming-in to facilitate 5.04w14.25).
skin-to-skin contact and Conclusions: The practice of rooming-in is now part of the mainstream postpartum
breastfeeding is a major newborn care. However, the increased incidence of neonatal hyperbilirubinemia is a
component of this initiative. potential side effect of which healthcare providers should be aware. Further
However, whether research is needed to confirm the role of rooming-in in neonatal
rooming-in is associated hyperbilirubinemia.
with admission for neonatal
* Corresponding author. 1650 Taiwan Boulevard Sec. 4, Taichung, 40705, Taiwan. Fax: þ886 4
23595046. E-mail address: mingclin@gmail.com (M.-C. Lin).

https://doi.org/10.1016/j.pedneo.2018.06.002
1875-9572/Copyright ª 2018, Taiwan Pediatric Association. Published by Elsevier Taiwan LLC. This is an open access article
under the CC BYNC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Copyright ª 2018, Taiwan Pediatric Association. Published by Elsevier Taiwan
LLC. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/ by-nc-nd/4.0/).
1. Introduction in 1991,1 the practice of BFHI has become a
mainstream concept in postpartum infant care. In
2010, it was estimated there were 21,328 hospital or
Since the World Health Organization (WHO)
launched the Baby-friendly Hospital Initiative (BFHI)
Rooming-in and neonatal hyperbilirubinemia 187

birthing centers designated as a Babyfriendly American Academy of Pediatrics clinical practice

Hospital (BFH) in 160 countries around the world. 2,3 guideline on the management of hyperbilirubinemia
Among the “Ten Steps” of the BFHI, the policy of in the newborn infant.13 The study endpoint was
rooming-in plays a very important role as it is defined as admission for phototherapy. The study
intended to promote breastfeeding, which is the protocol was approved by Taichung Veterans
main purpose of the BFHI.46 The rooming-in policy General Hospital’s Institutional Review Board, which
has various advantages which include promoting and waived the requirement for obtaining informed
supporting breastfeeding, facilitating skin-to-skin consent.
contact, and limiting the number of infants infected The demographic data were
in nurseries due to contact with other infants, compared by
doctors, and nurses.7,8 ManneWhitney U tests and chi-square tests. Each
However, while the BFHI was promoted to odds ratio and 95% confidence intervals for
enhance breastfeeding and rooming-in in Taiwan, phototherapy of rooming-in and other factors
where it became the mainstream approach in were calculated first. Then possible confounding
postnatal care, the incidence of neonatal factors including exclusive breastfeeding,
hyperbilirubinemia also increased simultaneously. 9 potential ABO incompatibility, G6PD deficiency,
This phenomenon might be attributed to the and BWL were adjusted in the multiple logistic
increased breastfeeding rate, but whether rooming- regression models. Then stratified analysis was
in is an independent risk factor for neonatal further performed to clarify the interaction
hyperbilirubinemia has seldom been studied. 1012 As a between rooming-in, exclusive breast feeding,
result, we conducted this retrospective hospital and body weight loss. The statistical analyses
based cohort study to elucidate the association were performed using commercially available
between the rooming-in policy and neonatal computer software programs SAS 9.4 for
hyperbilirubinemia. Windows ( SAS
Institute, Inc., Cary, NC, USA).

2. Material and methods 3. Results

This was a hospital-based retrospective cohort study.

During the study period, after excluding 40 infants
Medical records of full-term newborn infants born
who were transferred to the sick baby room or
between January 1, 2011 and December 31, 2013,
the neonatal intensive care unit due to diseases
who were admitted to the well-baby nursery of
other than hyperbilirubinemia, the medical
Taichung Veterans General Hospital, were reviewed
records of 3341 term neonates from our nursery
retrospectively.
were reviewed. The neonates were divided into
During the study period, rooming-in care was rooming-in and non-rooming-in groups.
encouraged according to the WHO guidelines: each Demographic data are listed in Table 1. Among
newborn stayed in the mother’s room for the entire them, 786 newborn infants (23.5%) were roomed-
day after birth if the parents agreed.1 Exclusive in within the study period. The infants of room-in
breastfeeding was defined as infants who were group had higher portion of natural spontaneous
breastfed only, without formula supplementation. delivery, potential ABO incompatibility, exclusive
The feeding policy in our nursery was mainly under breast feeding, and body weight loss >7%.
the principle of the Baby-friendly Hospital Initiative. Although the birth weight and gestational age
Unless under special circumstances or the parents’ reached statistical significance, the difference
request, infants would not be fed with anything might not be clinically significant (71 gm and 2.1
other than breast milk. Body weight loss (BWL) at 3 days).
days of age was defined as [(birth body weight 3-
The rooming-in ratio progressively increased
day-old body weight)/(birth body weight x 100%)].
following the promotion of the BFHI, which was
Factors which might be associated with neonatal
paralleled by a rise in the rate of neonatal
hyperbilirubinemia were also collected, such as
hyperbilirubinemia (Fig. 1A). Babies of rooming-in
Glucose-6-Phosphate Dehydrogenase (G6PD)
care contributed more to the raised admission
deficiency, and ABO blood types were also collected.
rate than those not rooming-in (Fig. 1B). The rates
Because the Coombs test was not available for most
of neonatal hyperbilirubinemia group were higher
of the infants, potential ABO incompatibility was
among babies of rooming-in care, exclusive
defined as follows: a mother’s blood type was O and
breastfeeding, G6PD deficiency, and potential
the infant’s blood type was A or B. We checked
ABO incompatibility in the univariate analysis.
every baby’s transcutaneous bilirubin level (TcB)
Body weight loss of 7% or 10% did not significantly
every morning. If TcB >11 mg/dl, we checked total
increase the risk of hyperbilirubinemia (Table 2).
serum bilirubin from heel stick to evaluate if the
baby needed phototherapy according to 2004
 188

Table 1 The demographic data of the study population.

Rooming-in (n Z 786) Non-rooming-in (n Z 2555)
Gestational age (weeks)* Birth 38.9 1.2 38.6 1.4
weight (g)* 3161.7 377.8 3090.7 414.5
Gender
Male 406 (51.7%) 1368 (53.5%)
Female 380 (48.4%) 1187 (46.5%)
NSD* 604 (76.8%) 1544 (60.4%)
G6PD deficiency 12 (1.5%) 49 (1.9%)
Potential ABO incompatibilitya,* 369 (47.1%) 1082 (42.4%)
Exclusive breast feeding* 576 (73.3%) 741 (29.0%)
Body weight loss >7%* 537 (68.3%) 1200 (47.0%)
Body weight loss >10%* 42 (5.3%) 79 (3.1%)
Values Z Mean SD.
NSD: natural spontaneous delivery.
*p < 0.05.
a
Z 4).
Missing data (N

We further analyzed the independent effect of body weight loss of more than 7% or not, the odds
roomingin by multiple logistic regression. The ratio for neonatal hyperbilirubinemia was 6.69 (95%
crude odds ratio of neonatal hyperbilirubinemia in confidence interval: 3.44w13.03) in body weight loss
the rooming-in group was >7% group and 8.23 (95 % confidence interval:
7.04 (95% confidence interval, 4.41w11.24) in model 4.18w16.20) in the non-rooming-in group. The
0. After adjusting for exclusive breastfeeding and common odds ratio was 7.28 (95% confidence
BWL percentages exceeding 10% at 3 days of age in interval: 4.48w11.81). The homogeneity test was
model 1, the odds ratio of neonatal not significant. The effects were not different
hyperbilirubinemia in rooming-in neonates became between strata.
8.55 (95% confidence interval, 5.10w14.32). We Table 3 Multiple logistic regression models for risk
further adjusted for all possible confounding factors of neonatal hyperbilirubinemia between rooming-in
including exclusive breastfeeding, BWL of >10% at 3 and non-rooming-in babies.
days of age, G6PD deficiency, potential ABO
Odds ratio 95% CI
incompatibility, and NSD in the saturated model 2,
and the odds ratio of neonatal hyperbilirubinemia Model 0 7.04 (4.41w11.24)
among roomingin neonates was still robustly 8.48 Model 1 8.55 (5.10w14.32)
(95% confidence interval, 5.04w14.25) (Table 3). Model 2 8.48 (5.04w14.25)
Because exclusive breastfeeding is a potential
confounder and effect modifier, we further Model 0: Univariate logistic regression.
conducted a stratified analysis for exclusive Model 1: Adjusted for exclusive breastfeeding and body
weight loss >10%.
breastfeeding and roomingin. For infants who were
exclusively breastfed, the odds ratio of rooming-in Model 2: Adjusted for exclusive breastfeeding, body
weight loss >10%, G6PD deficiency, potential ABO
for neonatal hyperbilirubinemia was 9.81 (95%
incompatibility and NSD. CI: confidence internal.
confidence interval: 3.83w25.17), and in
nonbreastfeeding infants the odds ratio of neonatal
hyperbilirubinemia in rooming-in neonates was 8.10
(95% confidence interval: 4.31w15.24). The common The risk of rooming-in for hyperbilirubinemia was
odds ratio was independent of body weight loss (Table 6).
8.95 (95% confidence interval: 5.07w15.82). Because
the homogeneity test was non-significant, rooming-
in might be a risk factor for admission due to
hyperbilirubinemia that is independent of
breastfeeding (Table 4). When babies were stratified
according to rooming-in care or not, the odds ratio
for neonatal hyperbilirubinemia was 0.67 (95%
confidence interval: 0.38w1.20) in the rooming-in
group and 0.55 (95% confidence interval:
0.21w1.47) in the nonrooming-in group. The
common odds ratio was 0.63 (95 % confidence
interval: 0.38w1.04). The homogeneity test was not
significant. The effects were not different between
strata. The risk of exclusive breast feeding for
hyperbilirubinemia was independent of rooming-in
(Table 5). When babies were stratified according to
Rooming-in and neonatal hyperbilirubinemia 189
Odds ratio (95% CI)a Common odds ratio (95%
CI)b

EBF(þ) Rooming-in (þ) 36/576 (6.3%) 9.81 (3.83e25.17) 8.95 (5.07e15.82)

Rooming-in () 5/741 (0.7%)

EBF() Rooming-in (þ) 19/210 (9.1%) 8.10 (4.31e15.24)
Rooming-in ( ) 22/1814 (1.2%)
Gynecologists have also endorsed rooming-in
4. Discussion policy for
neonates after birth.1820
Since the implementation and promotion of
Table 4 Risk of hyperbilirubinemia for rooming-in stratified by exclusive breastfeeding.

Table 2 Proportion and odds ratio for hyperbilirubinemia of rooming-in, exclusive breastfeeding, body weight loss, G6PD deficiency,
potential ABO incompatibility, and delivery type.

Hyperbilirubinemia (þ ) Hyperbilirubinemia ( ) OR (95% CI)

CI: confidence interval; EBF: exclusive breastfeeding. a Z 0.72.
Breslow-Day odds ratio homogeneity test, p n Z 82 (2.5%) n Z 3259 (97.6%)
Rooming-in*
b 55/82 (67.1%) 731/3259 (22.4%) 7.1
Common odds ratio was estimated by the Mantel-Haenszel common odds ratio estimation under the common odds ratio of 1.000
(4.41e11.25)
assumption.
Exclusive breast feeding* 41/82 (50.0%) 1276/3259 (39.2%) 1.6 (1.00e2.41)
Body weight loss >7% 46/82 (56.1%) 1691/3259 (51.9%) 1.2 (0.76e1.84)
Body weight loss >10% 5/82 (6.1%) 116/3259 (3.6%) 1.8 (0.70e4.43)
G6PD deficiency* 5/82 (6.1%) 56/3259 (1.7%) 3.7 (1.45e9.53)
Potential ABO incompatibility* 49/82 (59.8%) 1402/3255 (43.1%)a 2.0 (1.26e3.07)
NSD 56/82 (68.3%) 2092/3259 (64.2%) 1.2 (0.75e1.92)
*p < 0.05.
CI: confidence interval, G6PD: Glucose-6-Phosphate Dehydrogenase, NSD: natural spontaneous delivery, OR: odds ratio.
a
Z 4). missing data ( n

rooming-in in the period from 2011 to 2013, the

incidence of admission for phototherapy because of
neonatal hyperbilirubinemia increased. Although
exclusive breastfeeding and greater body weight
loss partially explained the increased risk, rooming-
in was still a significant independent risk factor for
neonatal hyperbilirubinemia.
Rooming-in is a method of caring for newborns
which involves the infant staying with the mother
in the same room, with the mother taking care of
her baby herself. Rooming-in provides numerous
advantages that include facilitating skin-to-skin
contact, reducing the number of infants infected
by contact with other infants, doctors, and nurses,
and, most importantly, facilitating breastfeeding.
Since the 1940s, in contrast to the conventional
separation of the mother and infant with isolated
baby nurseries in hospitals, this “new” concept
was proposed to facilitate breastfeeding and
maternal-infant bonding. 14e16 In 1991 , the WHO
and United Nations Children’s Fund launched the
Baby-friendly Hospital Initiative. Rooming-in is
one of the major components of this initiative. 17 A
number of major professional organizations such
as the American Academy of Pediatrics and the
American College of Obstetricians and
 a a
Odds
Oddsratio
ratio
(95%
(95%
CI)CI) Common
Commonodds
odds ratio
ratio (95%
(95%
b b
CI)CI)
BWL>7%
Rooming-in (þ) Rooming-in (þ)
EBF(þ) 36/576 34/537 (6.3%)
(6.3%) EBF() 6.69
0.67 (3.44e13.03)
(0.38e1.20) 7.28
0.63(4.48e11.81)
(0.38e1.04)
Rooming-in ()19/210 (9.1%) 12/1200 (1.0%)
Rooming-in ()
BWL7% Rooming-in (þ)
EBF(þ) 5/741 (0.7%)
21/249 (8.4%) 0.55
8.23
(0.21e1.47)
(4.18e16.20)
Rooming-in ( EBF(
) ) 22/181415/1355
(1.2%) (1.1%)
Breastfed babies have a higher risk of neonatal it is currently a widely accepted and
hyperbilirubinemia compared with formula-fed recommended approach to neonatal care. The
babies because there is a greater likelihood of infants in the rooming-in group obviously had a
insufficient feeding, greater body weight loss, or higher breastfeeding rate compared with the non-
decreased calorie intake.2124 The rooming-in policy rooming-in group. Also, a greater proportion of
facilitates breastfeeding. Thus, the rate of the rooming-in infants exhibited body weight loss
neonatal hyperbilirubinemia may increase in compared with nonrooming-in infants.
parallel with the increased adoption of rooming- Nevertheless, in the multiple logistic regression
in. Previous studies concluded that rooming-in model and stratified analysis, breastfeeding and
was not an independent risk factor for neonatal body weight loss could not fully explain the
hyperbilirubinemia among healthy, nonpremature increased hyperbilirubinemia rate. Body weight
newborns.1012 However, those studies were loss may be a late index of inadequate feeding.
conducted over 20 years ago, so the newborn Most babies receiving rooming-in are nipple
feeding or care policies and criteria for feeding, so it is difficult to accurately estimate the
phototherapy might be different. In this study, amount of breast milk intake. Neonatal
however, after adjustment for potential hyperbilirubinemia may happen before body
confounding factors in the multiple logistic weight loss due to increased enterohepatic
regression models, rooming-in was still an circulation. Nevertheless, we did not determine
independent risk factor for neonatal which component of the practice of rooming-in
hyperbilirubinemia. This result is therefore increased the prevalence of neonatal
inconsistent with the findings of previous reports. hyperbilirubinemia in this retrospective study
This inconsistency is possibly due to the larger because of the limited information available from
sample size and longer study period in our study charts. Different duration of hospital stay due to
compared with those of the three delivery type could influence the diagnosis of
aforementioned studies (3341 healthy term neonatal hyperbilirubinemia. According to the
infants versus 903 vs. 204 vs. 414; 3 years versus 6 reimbursement rule of Taiwan National Health
months vs. 6 months vs. 6 months). By stratified Insurance, the length of stay is 3 days for NSD and
analysis in Table 5, exclusive breast feeding was 5 days for CS. The variance among length of stay
not a significant risk factor for neonatal was almost completely explained by the delivery
hyperbilirubinemia in either rooming-in or type. Therefore, we did not collect the data of
nonrooming-in strata. The homogeneity test was admission duration at data collection step. The
not significant, which might mean that the effect median age of diagnosis was not collected either,
of breast feeding was independent to rooming-in. so we cannot provide further analysis on this
This was a single institutional study, so the results which is one limitation of this study. The real
might not be applicable in other hospitals and cause of increased hyperbilirubinemia rate should
other countries. However, our study be further studied in larger-scale studies with
demonstrated that the rooming-in policy could prospective design. Parity could be a factor
still lead to an adverse health effect even though related to maternal experiences of breast feeding
190

Table 5 Risk of hyperbilirubinemia for exclusive breastfeeding stratified by rooming-in.

CI: confidence interval; EBF: exclusive breastfeeding. a Z 0.74.

Breslow-Day odds ratio homogeneity test, p
b
Common odds ratio was estimated by the Mantel-Haenszel common odds ratio estimation under the common odds ratio of 1.000
assumption.

Table 6 Risk of hyperbilirubinemia for rooming-in stratified by body weight loss.

CI: confidence interval; BWL: body weight loss. a Z 0.67.

Breslow-Day odds ratio homogeneity test, p
b
Common odds ratio was estimated by the Mantel-Haenszel common odds ratio estimation under the common odds ratio of 1.000
assumption.
Rooming-in and neonatal hyperbilirubinemia 191

and room-in care. However, the data were not term newborns in partial and full rooming-in. J Matern
analyzed in this study, which is another limitation Fetal Neonatal Med 2009;22:801e5.
of this study. 13. American Academy of Pediatrics Subcommittee on
Hyperbilirubinemia. Management of
The practice of rooming-in is now part of the
hyperbilirubinemia in the newborn infant 35 or more
mainstream postpartum newborn care. However, weeks of gestation. Pediatrics 2004; 114:297e316.
the increased incidence of neonatal 14. Ungerer RL, Miranda AT. Rooming-in history. J Pediatr
hyperbilirubinemia is a potential side-effect of which (Rio J) 1999;75:5e10 [Article in Portuguese].
healthcare providers should be aware. Further 15. Jackson EB. General reactions of mothers and nurses
research is needed to confirm the role of roomingin to rooming-in. Am J Public Health Nations Health
in neonatal hyperbilirubinemia. 1948;38: 689e95.
16. Jelliffe EF. Programmes to promote breastfeeding.
Conflict of interest Med J Malaysia 1986;41:64e71.
17. WHO baby-friendly hospital initiative. A worldwide
commitment. Midwives Chron 1992;105:352.
The authors have no conflicts of interest relevant to 18. Szucs KA. American Academy of Pediatrics section on
this article. breastfeeding. J Hum Lact 2011;27:378e9.
19. Committee on Health Care for Underserved Women,
American College of Obstetricians and Gynecologists.
References
ACOG committee opinion No. 361: breastfeeding:
maternal and infant aspects. Obstet Gynecol
1. Baby-friendly hospital initiative: revised, updated and 2007;109:479e80.
expanded for integrated care. 2009. Geneva. 20. Gartner LM, Morton J, Lawrence RA, Naylor AJ, O’Hare
2. Pe´rez-Escamilla R, Martinez JL, Segura-Pe´rez S. D, Schanler RJ, et al. Breastfeeding and the use of
Impact of the Baby-friendly Hospital Initiative on human milk. Pediatrics 2005;115:496e506.
breastfeeding and child health outcomes: a systematic 21. Maisels MJ, Gifford K. Normal serum bilirubin levels in
review. Matern Child Nutr 2016; 12:402e17. the newborn and the effect of breast-feeding.
3. Labbok MH. Global baby-friendly hospital initiative Pediatrics 1986;78: 837e43.
monitoring data: update and discussion. Breastfeed 22. Maisels MJ, Gifford K. Breast-feeding, weight loss, and
Med 2012;7:210e22. jaundice. J Pediatr 1983;102:117e8.
4. Kramer MS, Fombonne E, Igumnov S, Vanilovich I, 23. Maisels MJ, Gifford K. Neonatal jaundice in full-term
Matush L, Mironova E, et al. Effects of prolonged and infants. Role of breast-feeding and other causes. Am J
exclusive breastfeeding on child behavior and Dis Child 1983; 137:561e2.
maternal adjustment: evidence from a large, 24. Rubaltelli FF. Unconjugated and conjugated bilirubin
randomized trial. Pediatrics 2008;121:e435e40. pigments during perinatal development. IV. The
5. Kramer MS, Aboud F, Mironova E, Vanilovich I, Platt influence of breastfeeding on neonatal
RW, Matush L, et al. Breastfeeding and child cognitive hyperbilirubinemia. Biol Neonate 1993; 64:104e9.
development: new evidence from a large randomized
trial. Arch Gen Psychiatry 2008;65:578e84.
6. Kramer MS, Chalmers B, Hodnett ED, Sevkovskaya Z,
Dzikovich I, Shapiro S, et al. Promotion of
breastfeeding intervention trial (PROBIT): a
randomized trial in the Republic of Belarus. JAMA
2001;285:413e20.
7. Cantey JB, Bascik SL, Heyne NG, Gonzalez JR, Jackson
GL, Rogers VL, et al. Prevention of mother-to-infant
transmission of influenza during the postpartum
period. Am J Perinatol 2013;30:233e40.
8. Lee YM, Song KH, Kim YM, Kang JS, Chang JY, Seol HJ,
et al. Complete rooming-in care of newborn infants.
Korean J Pediatr 2010;53:634e8.
9. Huang HC, Yang HI, Chang YH, Chang RJ, Chen MH,
Chen CY, et al. Model to predict hyperbilirubinemia in
healthy term and near-term newborns with exclusive
breast feeding. Pediatr Neonatol 2012;53:354e8.
10. Yamauchi Y, Yamanouchi I. The relationship between
roomingin/not rooming-in and breast-feeding
variables. Acta Paediatr Scand 1990;79:1017e22.
11. Syafruddin M, Djauhariah AM, Dasril D. A study
comparing rooming-in with separate nursing. Paediatr
Indones 1988;28: 116e23.
12. Zuppa AA, Sindico P, Antichi E, Carducci C, Alighieri G,
Cardiello V, et al. Weight loss and jaundice in healthy