ARTICLES
Effects of Traditional Swaddling on Development:
A Randomized Controlled Trial
AUTHORS: Semira Manaseki-Holland, BMedSci, MBBS,
MPH, MRCP, MFPHM, PhD,a Elizabeth Spier, BA, MA,
PhD,b Bayasgalantai Bavuusuren, MD, MSc, PhD,c
Tsogzolma Bayandorj, MD, MSc,d Susan Sprachman, BA,
MA,e and Tom Marshall, BA, MSc,a
a
Department of Nutrition and Public Health INTERVENTION
Research, London School of Hygiene and Tropical Medicine,
London, United Kingdom; bAmerican Institutes for Research,
Washington, District of Columbia; cChild DEVELOPMENT Unit,
Maternal and Child Medical Research Centre, Ulaanbaatar,
Mongolia; dPublic Health Institute, Ulaanbaatar, Mongolia; and
e
Mathematica Research Policy, Inc, Princeton, New Jersey
KEY WORDS
children, infants, BSID-II, Bayley, mental score,
psychomotor score, Mongolia
ABBREVIATIONS
RCT—randomized controlled trial
BSID-II—Bayley Scales of Infant Development, Second
Edition CI—confidence interval
This trial has been registered with the ISRCTN Register (http://
isrctn.org) (ISRCTN41832812).
www.pediatrics.org/cgi/doi/10.1542/peds.2009-1531
doi:10.1542/peds.2009-1531
Accepted for publication Aug 24, 2010
Address correspondence to Semira Manaseki-Holland,
BMedSci, MBBS, MPH, MRCP, MFPHM, PhD, School of Health
and Population Sciences, College of Medical and Dental
Sciences, University of Birmingham, 90 Vincent Dr, Edgbaston,
Birmingham B15 2TT, United Kingdom. E-mail:
s.manasekiholland@bham.ac.uk
PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-
4275).
Copyright © 2010 by the American Academy of Pediatrics
FINANCIAL DISCLOSURE: The authors HAVE indicated they HAVE
no financial relationships RELEVANT to this article to disclose.
Tight, prolonged swaddling currently is a de- veloping countries, in which 3 to 4
common practice in many temperate million infants are swaddled each year.1
PEDIATRICS Volume 126, Number 6, December
2010 Downloaded from by guest on October 13, 2016
WHAT’S KNOWN ON THIS SUBJECT: Swaddling is a child care
method that has been practiced for thousands of years in
numerous cultures and is still used to care for millions of infants
annually. Its effects on motor and mental development remain
uncertain and have only been investigated in small ethnographic
studies.
WHAT THIS STUDY ADDS: Traditional, tight, and prolonged
swaddling has no deleterious effects on child psychomotor and
mental development scores at 13 months of age, as assessed by
the Bayley Scales of Infant Development in a randomized
controlled trial.
abstract +
OBJECTIVE: Evidence of the effects of tight, prolonged binding of in-
fants on development is inconclusive and based on small
ethnographic studies. The null hypothesis was that Mongolian infants
not swaddled or swaddled tightly in a traditional setting (to >7 months
of age) do not have significantly different scores for the Bayley Scales
of Infant Devel- opment, Second Edition (BSID-II).
PATIENTS AND METHODS: In a randomized controlled trial, 1279
healthy newborns in Ulaanbaatar, Mongolia, were allocated at birth to
traditional swaddling or nonswaddling. The families received 7 months
of home visits to collect data and monitor compliance. At 11 to 17
months of age, the BSID-II was administered to 1100 children.
RESULTS: No significant between-group differences were found in
mean scaled mental and psychomotor developmental scores. The un-
adjusted mean difference between the groups was —0.69 (95%
confi- dence interval [CI]: —2.59 to 1.19) for psychomotor and —0.42
(95% CI:
—1.68 to 0.84) for mental scores in favor of the swaddling group. A
subgroup analysis of the compliant sample produced similar results.
BSID-II–scaled psychomotor and mental scores were 99.98 (95% CI:
99.03–100.92) and 105.52 (95% CI: 104.89 –106.14), respectively. Back-
ground characteristics were balanced across the groups.
CONCLUSIONS: In the Mongolian context, prolonged swaddling in
the first year of life did not have any significant impact on children’s
early mental or psychomotor development. Additional studies in other
set- tings need to confirm this finding. The Mongolian infants in this
trial had scaled BSID-II mental and psychomotor scores comparable
to United States norms. Pediatrics 2010;126:e1485–e1492
Despite frequent speculation as to the risks
and benefits of swaddling, there is a paucity
e148

of empirical evidence on the develop-
mental effects of this common practice.
Some evidence points to the calming or
sleep-inducing effect of swaddling.2– 4
Ev- idence on infant development is
mostly from observations conducted
as part of small-scale,
nonexperimental an- thropological, or
psychological studies that did not use
standardized develop- mental
assessments or describe the degree
or duration of swaddling with any
specificity.5–13
The results of these studies have been
conflicting. There is some evidence for
negative effects on development.5,9,13 At
16 months, swaddled Guatemalan in-
fants’ motor development lagged by 3 to
4 months compared with their American
peers, but their development was com-
parable by 5 to 11 years of age.9 In a
study of typically swaddled Navajo
children, Chisholm7 also concluded that
initial de- velopmental delays related to
swaddling were not permanent.
Others,6– 8,10–12
re- porting
predominantly on Native Amer- ican
cultures, have shown no effect of
swaddling on infant gross motor de-
velopment, and the author of a retro-
spective study of Kurdistani infants6
reported no significant difference be-
tween swaddled and nonswaddled in-
fants with regard to the ages at which
they attained sitting, standing, and
walking.
Given the prevalence of swaddling inter-
nationally, any association with develop-
ment could have large public health im-
plications. We investigated the effect of
traditional tight and prolonged swad-
dling on infant health and development
in Mongolia in a large-scale randomized
controlled trial (RCT). The first phase of
this trial revealed no harmful effects of
swaddling on pneumonia rates in
infants
PEDIATRICS Volume 126, Number 6, December
2010
ARTICLES
were recruited14 (Fig 1 and Table 1).
up to 7 months of age.14 Our study
repre- sents the second phase, in which
we in- vestigated the null hypothesis
that tightly swaddled Mongolian infants
would not have significantly different
Bayley Scales of Infant Development,
Second Edition (BSID-II) mental and
motor scale scores at around 13
months of age compared with those not
swaddled. There have not been any
measurements of BSID-II scores at the
population level in Mongo- lia.
Therefore, a secondary study objec- tive
was to provide data on the BSID-II
scores for our trial population,
which is a representative sample of
Mongolian infants in Ulaanbaatar.
PATIENTS AND METHODS
Sample
The details of the recruitment and first
stage of follow-up are summarized else-
where.14 In brief, all healthy infants deliv-
ered at the only 4 maternity hospitals in
Ulaanbaatar, Mongolia, were eligible to
be recruited within 48 hours of birth.
More than 95% of Ulaanbaatar births
took place in these 4 hospitals.15 Exclu-
sions were less than 36 weeks’ gesta-
tion, less than 2500 g birth weight, obvi-
ous congenital abnormalities, newborns
with need for intensive care treatment,
and residents in apartments that were
kept too warm for the infant to be swad-
dled during the daytime (as defined by
the mothers). Written informed consent
was obtained from the mother and next
of kin. Random assignment to
swaddling and nonswaddling groups
was through a statistician who was not
involved with recruitment and used a
previously gen- erated Excel (Microsoft,
Redmond, WA) random-study-number
list (equal proba- bility, without
stratification). For alloca- tion, after
consent recruitment doctors telephoned
a centrally located manager who
allocated subjects moving down this list.
The trial profile (Fig 1) summa- rizes
the follow-up and attrition.
In phase 1 of the trial, 1279 infants
e148
Downloaded from by guest on October 13, 2016

For the current child-
development part of the trial, to
detect a 5-point dif- ference in
the development score, with 90%
power and a statistical signifi-
cance of P = .05, we required a
mini- mum sample requirement
of 303 per group. A 5-point score
difference be- tween the groups
(or 5 SDs) was used because
score differences of more than
15 are thought to be clinically
significant,16 whereas Grantham-
McGregor et al17 and Hamadani
et al18 found a decrease in the
average men- tal score with low
birth weight, after adjustment, of
5.9 points. However, to take
advantage of the large phase 1
sample, to account for the
All hospital births from Sep 9–Dec 25, 2002 eligible for assessment (N = 4360)
Enrolled = 1279
Randomization = 1279
Allocated and received clothing as intervention (n = 640)
Lost to follow-up (n = 103 [16%])
6 died
84 could not be found
13 refused further participation
BSID-II tested (n = 537)
Analyzed (n = 537)
3 excluded from psychomotor
4 excluded from mental
FIGURE 1
Trial profile.
PEDIATRICS Volume 126, Number 6, December
2010
ARTICLES
secondary objective and to those in the nonswaddling group.
possibility of some not For 7 months, families were given free
consenting to BSID-II testing, we and easy access to family doctors
con- tacted all families from who re- ferred to central hospital
phase 1. After excluding those pediatricians in case of sickness.
who were lost to follow-up or did Medication was free, and some
not consent to addi- tional doctor-related transportation costs
participation, 536 from the were compensated. Trained field-
swaddling and 564 from the workers made home visits until the in-
nonswad- dling groups signed fants were 7 months of age, collected
new consent forms and attended illness, compliance, family, and
the testing (Fig 1). environ- mental information, and
encouraged the families to follow
Procedures study protocol. There was no more
During the first phase of the trial, contact with the families until the
infants assigned to the swaddling infants were 12 months of age, when
group were given clothes and a field-workers collected informa- tion
blanket, and clothes, sheepskin on the children and invited them for
(traditional to Mongolia), and BSID-II testing at close to 13 months
cotton sleeping bags were given of age.
complier” variable was defined by using
the 2 diary sources: Over the first 2
3081 (74%) not enrolled:
- 69% infants not meeting study criteria including:
months and then from 3 to 7 months, a
good complier in the swaddling group
o 18% resident
outside Ulaanbaatar had a mean of more than 21 of 24
o 38% woman not hours swaddled and in the nonswaddling
well for assessment by 48 h or recruiting
group had a mean of less than 3 hours
doctor not available swaddled.
- 5% of approached refused to participate
Four trained Mongolian testers con-
ducted all tests at 1 center. Their rigor-
ous training for test administration is de-
scribed elsewhere.19 The protocol was
Allocated and received swaddling as control (n = 639) adapted from the model originally used
Allocation in the Early Head Start Research and
Evaluation Project,20 and training was
conducted by its expert staff. Accurate
administration of the BSID-II was moni-
tored weekly through random videotap-
Lost to follow-up (n = 76 [14%]) ing and supervisor observation. Ameri-
12-mo follow-up 3 died
63 could not be found can trainers reviewed the performance
10 refused further participation of each tester by assessing the
videotap- ing of a random 10%
proportion of the tests. The effect of
tester variation was assessed
statistically, and to eliminate any such
Analysis
BSID-II tested (n = 563) effect, “tester,” as an interac- tion and
Analyzed (n = 563) then as an independent term, was
1 excluded from psychomotor
6 excluded from mental applied in multivariate analysis. All
project staff who administered the
BSID- II, computed BSID-II scores,
supervised, and cleaned the data were
blind to ran- domization groups. An
e148
Downloaded from by guest on October 13, 2016

Access (Microsoft) was
database used for data
For infants assigned to the swaddling
yond this age, swaddling mostly was
group, families were instructed to follow
restricted to times when the infant
a pattern of swaddling based on the tra-
was asleep. By 13 months of age, the
ditional Mongolian method. For the first
majority of children were no longer
2 months, this pattern involved wrapping
swaddled (Table 2). Among those
the infant from head or neck down in 2
to 3 layers of thin cotton cloths, covered who did continue swaddling, it rarely
by layers of warm blankets (Fig 2A–D). oc- curred when infants were awake.
In- fants were swaddled during the To monitor compliance, field-workers di-
day and night, with the exception of a rectly observed the infants during every
brief few minutes to change the home visit and collected retrospective
infant’s soiled clothes 3 to 5 times per 24-hour diaries of swaddling. When the
day. From 2 to 7 months of age, infants were 6 and 12 weeks of age,
swaddling still took place during both mothers completed 2 sets of 4-day pro-
day and night, but the number of spective diaries. Unscheduled visits for
hours in the day was gradually direct observations and other data
reduced. At this point, the infant’s sources were triangulated with these di-
arms were some- times freed in the aries to confirm compliance. A “good
daytime (Fig 2E). Be-
TABLE 1 Characteristics of Children in Swaddling and Nonswaddling Arms
Background Indices Swaddling (N = 536),
n (%) or Mean
Nonswaddling (N = 564),
n (%) or Mean (SD
(SD Range)
Range)
Age of BSID-II testing
11–12 mo 5 (1) 2 (0.5)
13 mo 422 (79) 465 (83)
14 mo 98 (18) 79 (14)
15–17 mo 12 (2) 17 (3)
PEDIATRICS Volume 126, Number 6, December
2010 Downloaded from by guest on October 13, 2016
ARTICLES
entry, and Stata (Stata Corp, College
Station, TX)
was used for intention-to-treat univari-
ate analysis and linear multiple regres-
sion analyses.
For secondary analysis, in which the
effect of exposure was examined, first
compliance (good-compliers variable)
was tested as an interaction term and
then as a confounder for the whole
sample in a multivariate regression
analysis. Subsequently, it was used to
define a per-protocol subgroup (good-
complier group) in which the hypothe-
sis was further tested.
BSID-II Instruments, Translation, and
Adaptation
The BSID-II has been extensively vali-
dated and standardized in the United
scale evaluates the degree of body
control, large muscle coordination, fine
manipulatory skills of the hands and
fingers, dynamic movement, pos- tural
imitation, and stereognosis (abil-
ity to recognize objects by sense of
touch). Test scores are reported as
e148
 ARTICLES

Gestational age at birth

19 (4) scaled scores based on the infant’s
36–376⁄7 wk 17 (3) age, with a normed mean of 100 (SD: 15
S38 wk 519 (97) 543 (96)
Type of delivery points).16 Both scales have high corre-
461 (82)
Vaginal 447 (83) lation coefficients (0.83 and 0.77 for
Cesarean 85 (16) 97 (17)
Birth weight psychomotor and mental tests, re-
2400–2950 g 58 (11) 66 (12) spectively) for test-retest reliability.16
3000–3450 g 233 (43) 233 (41)
3500–3950 g 186 (35) 191 (34)
68 (12) The BSID-II training materials were me-
>4000 g 57 (11) ticulously translated into Mongolian (in-
Gender of infant
Female 265 (49) 286 (51)
274 (48) cluding back-translations), and transla-
Male 264 (49) tions were further refined in the course
Breastfeeding 0–4 mo
Not breastfed 39 (7) 51 (9) of tester training to ensure comparabil-
Nonexclusively breastfed 485 (90) 500 (89) ity with original BSID-II instructions. Al-
Exclusively breastfed 7 (1) 9 (1)
No. of children admitted for severe 71 (13) 78 (14) though concerns about cross-cultural
pneumonias defined by IMCI
variation cannot be eliminated,
Anemia, ever diagnosed a 33 (6) 37 (7)
Rickets, ever diagnosed a 204 (38) 295 (35) minimal test adaptation was required,
Maternal age, y 27 (26–27) 27 (26–27) because items administered to
Parity
0 231 (43) 263 (47) children younger than 18 months
1 or 2 90 (17) 86 (15) embodied few linguistic or conceptual
>3 143 (27) 154 (27)
Maternal marital status
features that could vary by culture.
Single 15 (3) 12 (2) This project was approved by the Min-
Common law 235 (44) 264 (47)
Married 279 (52) 282 (50) istry of Health of Mongolia and the
Maternal education eth- ics committee of the London
Primary school 60 (11) 60 (11)
Secondary school 291 (54) 305 (54) School of Hygiene and Tropical
Tertiary education 178 (33) 193 (34) Medicine.
Mother’s employment
Ever worked 410 (76) 439 (77)
Paternal education RESULTS
Primary school 76 (14) 72 (13)
The BSID-II was successfully adminis-
Secondary school 285 (53) 302 (54) tered to 1100 of the trial children. Table 1

PEDIATRICS Volume 126, Number 6, December e148

2010 Downloaded from by guest on October 13, 2016
 ARTICLES

Tertiary education 135 (25) 156 (28) lists the characteristics of the groups.
Father’s employment
Ever worked 389 (73) 419 (74) Intention-to-treat linear regression
Type of residence at recruitment a analyses revealed that there were no
Ger/wood-cement house 170 (32) 191 (34)
Ger/apartment mixed 165 (36) 158 (28) significant differences between the
Apartment 195 (36) 208 (37) swaddling and nonswaddling groups
No. of people sleeping in the infant’s room 4.2 (4.0–4.5) 4.3 (4.1–4.4) in their mental and psychomotor
All variables remained not statistically different between the 2 arms. IMCI indicates the World Health Organization and
United Nations Children’s Fund Integrated Management of Childhood Illnesses. scores (Table 3). Mean scaled BSID-II
a
Also an indicator of indoor air pollution because gers and wooden houses have furnaces, whereas apartments have mental and psychomotor scores for
central or gas heating.
the swaddling versus nonswaddling
groups were 105.73 (95% confidence
States16 and validated and learning and problem-solving; vocal- interval [CI]: 104.85–106.62) versus
translated for other cultural ization and beginning of verbal com- 105.31 (95% CI: 104.41–106.21) and
settings.21–23 The BSID-II mental munication; basis of abstract thinking; 100.34 (95% CI: 98.99 –101.42) versus
scale evaluates chil- dren’s habituation; mental-mapping; complex 99.64 (95% CI: 98.31–100.92), respec-
sensory/perceptual acuities, language; and mathematical concept tively. With “good-complier” as an inde-
discriminations, and response; formation. The BSID-II psychomotor pendent term and its interaction with a
acqui- sition of object constancy;
memory

PEDIATRICS Volume 126, Number 6, December e149

2010 Downloaded from by guest on October 13, 2016
FIGURE 2
A–D, Series of photographs showing the sequence of swaddling a newborn infant; E, a 6-month old infant swaddled from the waist down.
randomized group in a regression
model, neither was statistically signif-
icant (P = .7 and 0.5, respectively),
giv- ing no evidence for any impact of
dif- ferential compliance on
psychomotor or mental scores. Per-
protocol analy- ses of the separate
good-complier sub- sample (see Table
3) also revealed no difference
between the 2 groups in terms of their
scaled scores. As ex- pected, there
were fewer nonswad- dling infants
who were good-
compliers compared with swaddling
infants (Table 3), but power still was
preserved at more than 90%.
Mean BSID-II mental scale scores
(105.52 [95% CI: 104.89 –106.14] vs
100) for the Mongolian children par-
ticipating in this trial were slightly
higher than those in the US normal
sample, 24 and BSID-II psychomotor
scores (99.98 [95% CI: 99.03–
100.92] vs 100) among this
Mongolian sam-
ple were essentially equivalent to US
norms.
Randomization and Bias
At recruitment, in 3 hospitals 25% and
in the fourth 26% of births between
September and December 2002 were
recruited. No differences were found
between hospitals or mean date of re-
cruitment according to allocation to
groups. Randomization was success-
ful, and univariate analyses, applying
a
TABLE 2 Age of Last Swaddling of the Infant, Even if for <1 Hour on a Rare Occasion

Unadjusted Mean Difference Between Groupsa

Age Last Swaddled (Irrespective of the Swaddling Arm, Nonswaddling Arm,

0.03 (—1.06 to 1.12)

0.07 (—1.46 to 1·61)
Duration and Time of Swaddling) n (% of Swaddling Arm) n (% of Nonswaddling Arm)
Never swaddled 0 (0) 301 (53)

to 106.54)
0 to end of 3rd month of life 17 (3) 239 (42)a

to 101.44)
4th to end of 5th month of life 17 (3) 9 (2)

to 106.83) 105.58 (104.61491

to 102.17) 100.03 (98.6 497
Nonswaddling Arm
6th to end of 7th month of life 322 (60) 5 (1)
8th to end of 11th month of life 155 (29)a 6 (1)

Subsample of Good Compliers Tested

—
—
12th month of life until BSID-II testing was 24 (5) 4 (1)
performed between 13th and 14th
month
No data 1 (0.1) 0 (0)

TABLE 3 Mean Scaled BSID-II Scores Based on US Standardized Norms for the Intention-to-Treat Analysis of the Whole Sample and for the Good-Compliers Subsample
Total 536 (100) 564 (100)

105.88 (104.94464
100.74 (99.31465
a Compliance data demonstrated that the majority of the nonswaddling-group children (198 of 239) who were swaddled in

Swaddling Arm
the early months spent an average of less than 3 hours per 24 hours swaddled and usually did so during the sleeping
period. The majority of the swaddling-group children (465 of 536) were swaddled more than 21 hours per 24 hours until 6
months, and then swaddling was gradually reduced during waking hours.

large range of variables (see Table swaddling, we found no significant impact of swaddling on children’s mental or psy-
1), demonstrated that the only back-
ground or risk-factor characteristic
that was slightly different between
the groups at recruitment was the
fathers’ education.14 In the phase 2
trial popu- lation (Table 1), fathers’
education was balanced between
the groups. Partici- pant attrition
rates (Fig 1) did not differ on the
basis of any sample char-
acteristics. Survival analysis (Cox
re- gression modeling) compared
non– acute respiratory illness and all
acute respiratory illness cases as
well as pneumonia cases in the 2
groups and revealed no difference
of rates be- tween the 2 arms of the
trial.
Test scores did not vary significantly
according to the tester. On the basis
of BSID-II age calculations, a total of
883 children were assessed at 13
months of age, 176 children at 14 to
16 months of age, and the remaining
41 children at 11, 12, or 17 months of
age. Age of testing did not vary
according to group assignment.
Almost all of the children ( n = 1049
[95.5%]) were tested at an
assessment center, and the remaining
51 children were tested at home.

DISCUSSION
In this first large-scale RCT of the ef-
fects of tight and prolonged
 to 106.40)
to 101.38)

—to 0.77) 105.72 (105.05955

—to 1.22) 100.38 (97.37962
c y developing countries in which

Total
h millions of infants are swad- dled
o annually. In most of these coun- tries,
m swaddling resembles the Mongo- lian

—to 1·19) —0.23 (—1.68

—to 0.84) —0.24 (—1.26

o practice in that the infant is bound in a

Adjusted Difference Between Groupsa

t straight position almost continu- ously

Group differences in mean scaled scores, with adjustment for testers (interaction term) and age at testing by linear multiple regression.
o in the first few months of life through
r the use of sheets and blankets or with
a cradleboard and thereafter

to 106.21)—0.42 (—1.68
to 100.92) —0.69 (—2.59
Unadjusted Difference Between Groupsa
d increasingly let free for waking hours
e and eventually stopping swaddling.
v The findings of this trial mirror those

Full Sample Tested, Intention-to-Treat Analysis

e found in smaller ethnographic and
l

to 106.62) 105.31 (104.41557

ob- servational studies6– 8,10–12 by

99.64 (98.31 562

Nonswaddling Arm
o suggest- ing that tight and prolonged
p swaddling (traditional style) neither
m harms nor enhances infant mental
e and psy- chomotor development.

to 101.42)
n

to 106.14) 105.73 (104.85533

Despite having fewer stimulating objects

100.34 (98.99534
t

Swaddling Arm
(ie, toys) to experiment with than their
.
American counterparts and less free-
dom of motion for many of the initial
T
months of life, the Mongolian infants in

to 100.92)
h
our sample performed as well as Amer-

1090
99.98 (99.031096
i

105.52 (104.89
ican infants on the BSID-II psychomotor
s
scale and mental scale. The slightly bet-

Total
ter mean mental scale score was less
fi
than 10 points and, therefore, was not
n
considered to be clinically significant in
d
an individual child.16 However, such a Scaled psychomotor score
i
rel- atively small shift in population
n
mean will correspond to a definite
g
change in the proportion of those whose
excess a
i
s

r
e
l
e
v
a
n
t

t
o

m
a
n
 over the American norm was 10 or
more points.25
The findings that swaddled Guatemalan
infants from a remote region had the
same levels of development as non-
swaddled American children7 were ex-
plained by Lewis and Freedle,26 who de-
scribed Guatemalan infants as being
spoken to or played with 25% of the
time compared with 6% of the time for
middle-class American children. Our ob-
servations in Mongolia indicate that the
infant (swaddled or not) is rarely out of
a caretaker’s arms during waking hours
and is at the center of family interac-
tions. The typical Mongolian household
contains an extended family, and similar
to many non-Western swaddling cul-
tures,27 Mongolian culture is oriented to-
ward providing infants with a high
amount of social attention. These and
other contributing explanatory factors in
Mongolian infants’ care practices have
enabled infant development that is com-
parable to US norms despite intense
mo- tor restriction and a minimal
exposure to toys.
Such findings raise fundamental ques-
tions in developmental child health re-
garding the importance and the nature
of early motor stimulation and experi-
ence with object manipulation required
for normal infant mental and psychomo-
tor development to occur, at least in a
cultural context with a high degree of
so- cial stimulation. Our results suggest
that the promotion of positive infant
mental and psychomotor development
may need to focus on social interactions
rather than emphasizing the availability
of motor and sensory stimulation. Find-
ings also would be of particular impor-
tance to Western families who may not
be able to afford early childhood toys
but who may live in extended family
units. Driven by rigorous publicity and
social pressures linking good parenting
to pro- vision of toys, these families may
feel disempowered in contributing to
their children’s development. Although
the
e1492 MANASEKI-HOLLAND et al
Downloaded from by guest on October 13, 2016
stressors of poverty exist in Mongolia
just as they do in Western countries,
the impact of this poverty on
development in young children seems
to be mitigated by their high exposure
to child-friendly and child-focused
social interactions.
This article primarily presents the re-
sults of an RCT that investigated the ef-
fect of swaddling on the psychosocial
de- velopment of children. Given the
RCT nature of the study, like is
compared with like, because
randomization was successful for both
the recruited sam- ples in phase 1 and
because low attrition before the second
phase still left the arms balanced in all
background and risk-factor
characteristics (Table 1). This design
eliminates the effect of an- ticipated
and unknown factors.
However, for the secondary study ques-
tions, generalizability may be
questioned when presenting the overall
mean scores for BSID-II in this study
and com- pared with the US norms. A
limitation of our trial includes an inability
to identify the characteristics of families
who re- fused to participate during
recruitment. However, given the small
number of re- fusals, it is unlikely that
this introduced biases to the overall
developmental sta- tus of the infants in
the sample relative to the entire
Mongolia population. The ur- ban
Ulaanbaatar setting may make gen-
eralizability to rural populations ques-
tionable. Although differences in child
care cannot be excluded, recent political
changes have meant that a large num-
ber of residents of Ulaanbaatar are re-
cent migrants from the rural areas living
in gers or wooden-cement housing dis-
tricts. The high rate of these participants
in our trial demonstrates the high repre-
sentation of rural families and, thus, ru-
ral practices. It is important to note that
despite the high rate of literacy of moth-
ers, our population was poorer than av-
erage in Ulaanbaatar because those liv-
ing in better apartment housings
(warm) were not included. A high liter-
acy rate is a feature of Mongolia as a
postsocialist country in which more
than 98% of women are literate (in
rural and urban settings) and have a
higher rate of tertiary education than
men.28,29
In terms of the inclusion and exclusion
criteria, perhaps if our sample included
children with gross congenital abnor-
malities, with newborn illness, or born at
less than 36 weeks’ gestation or less
than 2500 g birth weight, BSID-II scores
would have been slightly lower.
However, we do not speculate a large
effect if such a cohort were included
because Mongo- lian infants who have
severe prematu- rity, very low birth
weight, or major neo- natal problems
rarely survive to 13 months given the
relative lack of ad- vanced neonatal and
pediatric medical services in Mongolia.
Furthermore, in- creased rates of infant
morbidities in Mongolia compared with
those in the United States would be
expected to delay development and
scores, thus balancing out the effect of
a higher percent of in- fants being
healthy at birth in our sam- ple.
Summarizing issues of generalizabil- ity,
it can be concluded that the effect of
swaddling on development can be
gener- alized, whereas the inclusion and
exclu- sion criteria in our population
may have slightly improved the overall
mean score of our sample compared
with the real mean in Mongolia. The
effect of our se- lection criteria on the
comparison of our total study population
score with the to- tal Mongolian
population’s BSID-II score cannot be
ascertained through the cur- rent study.
CONCLUSIONS
Our trial was the first RCT of the
health effects of tight and prolonged
tradi- tional swaddling. We
demonstrated no harmful effect of
swaddling on child- hood mental and
psychomotor devel- opment at 11 to
17 months of age among Mongolian
children, which is a significant
finding for millions of par- ents in
developing countries in which
swaddling is a convenient and
cheap
 and develop Harvard of the New York,
Ross; ment. We thank Universi Ef- fect NY:
PEDIATRICS Volume 126, Number 6, December 1976 Am all 270 ty of Psychologi
2010 6. Bloch A. Psychol project Press; Swaddl cal
The . 1948 ing on Corporatio
method of CES ne- ous staff
arousals in Kurdista 1973;28 12. Lipton Lower n; 1993
child (11):94
members EL, Respir 17. Grantham-
1. Nelson supine ni cradle
care. EAS, story: a 7–961 who Steinsc atory McGregor
infants
These Schiefe mod- 10. Kluckh followed- hneider Tract SM, Lira
while
nhoevel up the A, Infectio PIC,
findings swad- dled ern ohn C.
n in
W, analysis Some Richmo Ashworth
merit and families Infants
Haimerl unswaddle of this aspects nd JB. A, Morris
additional and im- Swaddli From SS,
F. Child d during centurie of
investigati care rapid eye s-old Navajo plemented ng, a Mongol Assuncao
the project childcar ia [PhD MAS. The
on in practice movement infant in-
e thesis]. develop-
other s in and quiet swaddli fancy in its first 7 London
unindus ng and practice ment of
cultural sleep. months, to , United
trialized practice. early : low birth
Pediatrics .
settings. societie Clin childho
the 22 who historica Kingdo weight
2002;110( m: Lon-
s. 6). Pediatr od. did the l, term
cultural don infants
ACKNOW Pediatri Available (Phila). Psycho same Univers
cs. 1966;5( anal and and the
LEDGMEN at: during the experim ity; effects of
2000;1 www.pedia 10):641 Soc
TS BSID-II ental 2005 the
05(6). trics. – 645 Sci.
This Avail- 1947;1( testing observa 15. World environme
org/cgi/con 7. Chishol
project stage of tions. Health nt in
able at: tent/full/11 m JS. 1):37–
Pe- Organiz northeast
was www.pe 0/6/e70 NAVAJO 86 the work, diatrics. ation, Brazil. J
funded by diatrics. 4. Ohgi S, Infancy: 11. Leighto and to the 1965;35 Mongoli Pediatr.
org/cgi/ Akiyama T, An n D,
the following (suppl): an 1998;132(
content Arisawa K, Ethologi Kluckho
Clinical Mongolian 519 – Gov- 4):661–
/ Shigemori cal hn C.
567 ernmen 666
Epidemiol full/105 K. Study of Childre collaborato t.
n of the 13. Tronick 18. Hamadani
ogy /6/e75 Randomise Child rs: Prof EZ, Mongol J, Fuchs
Program 2. Franco d controlled DEVELO Jarghansai Thomas ia G,
P, Seret trail of PMENT. Health
of the khan and RB, Osendarp
N, van swaddling New Daltabuit Sector S, Huda
Well- Hees J, York,
Dr
versus M. The REVIEW S,
come Scaillet massage in NY: Soyogerel Quechu (HSR). Grantham
Trust, S, the Aldine, at the a manta Ulaanb -
Canada Groswas manageme Hawthor Ministry of pouch: a aatar, McGregor
ser J, nt of ex- ne; 1983 caretaki Mongoli S. Zinc
Fund, and Health and
Kahn A. cessive 8. Dennis ng prac- a: suppleme
United Influence crying
Dr World
W, tice for nta- tion
Nations of swad- infants with Dennis Narnatuya, bufferin Health during
Children’s dling on cerebral MG. The the Director g the Organi pregnancy
sleep inju- ries. effect of of Public Peruvia zation and
Fund
and Arch Dis cradling n infant and effects on
Mongolia arousal
Health the
Child. practices against men- tal
Country characte 2004;89(3): upon the
Institute, the Gover developm
Office. ristics of 212–216 onset of who sup- multiple nment ent and
These healthy walking ported and stressor of behavior
5. Baker PT,
infants. in Hopi s of high Mongoli of the
funders Little MA. provided a; 1999
Pediatric Man in the children. alti- infants: a
had no s. local tude. 16. randomize
Andes: A J Genet
other 2005;115 Mul- Psychol. supervision Child Bayley d placebo
involveme (5): tidisciplinar 1940;56: for this DEV. N. controlled
nt with the 1307– y Study of 77– 86 project. We 1994;65( Bayley trial. Lan-
1311 High- 4):1005– Scales cet.
details of 9. Kagan J, are grateful 1013 of 2002;360(
3. Gerard Altitude Klein
trial CM,
for the 14. Manase Infant 9329):290
Que- chua RE.
implement Harris NATIVES. Cross- ki- DEVELO –294
PMENT:
ation or KA, Strousburg cultural Holland 19. Spier ET,
Thach People. S. Second Sprachma
analysis. , PA: perspec-
Edition
BT. Dowden, tives on Cambridg INVESTI n S,
REFEREN Spontaet al e, MA: GATION Manual. Rowand
e1492 MANASEKI-HOLLAND Hutchinson early
Downloaded from by guest on October 13, 2016
 C. n T e syc K - l
Imple- lead h r hos a B Stud
suppor
menting ersh e n o- n ir y.
large- ted. I a cial g t Bost
ip
scale Further m t sti ar h on,
and
studies more, p i mul o - MA:
of sup a o atio o W Ab
we are
children port c n n C e Initio
indebt t ar i
us- ing of a on Inter
clinical ed to o l me e g natio
Prof
assess the f C ntal o h nal;
ments. Choi E e dev n t 200
expert
In: - a n elo N I 1–
advice
Children jumt rl t pm E n 200
and the and y r ent U - 2.
s
Mediterr directi H e and R f Avail
anean: Got e be a -
on of O
Health, ov, a f hav B n able
Prof
Culture Dire d o iou E t at:
and Ur- Sally S r r of s ww
ctor H
ban McGre t D mal A O w.br
Settings.
of a i no azel
gor VI V
Genoa, the rt a uris O E ton-
and
Italy: Mat . r he R R insti
Fondazi Mrs P r d t tute
erna AL
one Betty ri h chil O h .co
l
Insti- Hutchi n o dre rg e m/
tuto and c e n: a F abini
nson
Mediterr Chil e a a ni ir tio20
aneo at the t l ran z s 02/a
d
per start of o D - at t rt1.h
Heal
L’Infanzi this n i do io Y tml.
a; 2004 th , s mis n, e Acc
trial,
20. Love JM, Med N e ed In a esse
trainin fa r
Kisker ical J: a con d
EE, g M s trol nt o Nov
Res
Ross C, assist a e trial T f emb
earc t . e L er
et al. ance
Making h h R He m if 11,
from
a Cen - e ad p e 201
Differen Dr e s Sci er : 0
tre
ce in Amy m e Bul a T 23. Wrig
Hos a a l m h
the Dama ht
LIVES of pital ti r (Ba e e RO,
st and
Infants , c c ng- nt N Hu
Mrs a h lad a a
and with H,
Tod- Margie P , es n - Silv
out ol B h). d g
Stone, erm
the ic a 20 D a an
and
facili y n 05; e- s EK,
statisti R g 3 a
ta- V et
cal e l 22. Ogi EL ki al.
tion
advice s a S, O U Apol
of e d Fuk P N i-
of Dr
who a e ada M I popr
Kim r s
m M, E V otei
Boller. c h Tak N E n E
this
h . aha T R gen
proj ; E ski in S otyp
ect dler 2 f T, H IT e
woul s 0 f et e Y pred
d and 0 e al. al H icts
Thei 2 c The th o 24-
not t y s
r 21. I Ef- mon
have n o fect L p th
Fam
bee ilies: t f s of o it Bayl
e1492 MANASEKI-HOLLAND et al p w a
Downloaded from by guest on October 13, 2016
 ey Y: ntrie ri s ildr
Sc P s: a c F en’
al s meth a u s
es y odol n n Fun
Inf c ogic s d d;
an h al re- [i . 200
t ol view n T 4
D o . J h 29. Unit
ev gi Stud a e ed
el c Fam p Nati
op al Plan a S ons
m C n. n t Chil
en or 1986 e a dre
t p ;17( s t n’s
sc or 2):7 e e Fun
or at 8 – ]. d.
e. io 94 S Mo
P n; h o
26. Lewi ngo
e- 1 i f
s M, lia
di 9 n T
Free sta-
at 6 ri h
dle tisti
r 9 g e
MR. cs,
R 25. R Mot a Mo
es o her- k W ngo
. s Infa u o lia
20 s nt K r at a
03 D Dya e l gla
;5 A d. n d nce
4( , Prin k ’ .
6): V ceto y s Ava
81 a n, u ilab
9 u NJ: . C le
– g Edu 1 h at:
82 h catio 9 i ww
5 a nal 9 l w.u
24. Ba n Testi 6 d nic
yle J ng ; r ef.o
y P Serv 6 e rg/i
N. . ice; 7 n nfo
Th H 197 ( . byc
e e 2 4 N oun
B al 27. Mas ): e try/
ay th atak 2 w mo
le in a N. 8 ngo
y te On 5 lia_
Y
Sc rv the – stati
o
al ie funct 2 stic
r
es w ion 9 s.ht
k
of s of 1 ml.
,
Inf ur swa 28. U N Acc
an v ddlin ni Y ess
t e g as t : ed
D y tradi e U Nov
EV s tiona d n em
EL in l N it ber
- d infan a e 11,
op e t- ti d 20
m v care o
en el prac n N
t. o ticed s a
N pi by C ti
e n nativ hi o
w g e ld n
Yo c Sout r s
rk, o h e C
N u Ame n’ h
e1492 MANASEKI-HOLLAND et al
Downloaded from by guest on October 13, 2016