International Child Care Practices Study: infant sleeping environment

Abstract

The International Child Care Practices Study (ICCPS) has collected descriptive data from 21 centres in 17 countries. In this report, data are presented on the infant sleeping environment with the main focus being sudden infant death syndrome (SIDS) risk factors (bedsharing and infant using a pillow) and protective factors (infant sharing a room with adult) that are not yet well established in the literature.
Using a standardised protocol, parents of infants were surveyed at birth by interview and at 3 months of age mainly by postal questionnaire. Centres were grouped according to geographic location. Also indicated was the level of SIDS awareness in the community, i.e. whether any campaigns or messages to "reduce the risks of SIDS" were available at the time of the survey.
Birth interview data were available for 5488 individual families and 4656 (85%) returned questionnaires at 3 months. Rates of bedsharing varied considerably (2-88%) and it appeared to be more common in the samples with a lower awareness of SIDS, but not necessarily a high SIDS rate. Countries with higher rates of bedsharing appeared to have a greater proportion of infants bedsharing for a longer duration (>5 h). Rates of room sharing varied (58-100%) with some of the lowest rates noted in centres with a higher awareness of SIDS. Rates of pillow use ranged from 4% to 95%.
It is likely that methods of bedsharing differ cross-culturally, and although further details were sought on different bedsharing practices, it was not possible to build up a composite picture of "typical" bedsharing practices in these different communities. These data highlight interesting patterns in child care in these diverse populations. Although these results should not be used to imply that any particular child care practice either increases or decreases the risk of SIDS, these findings should help to inject caution into the process of developing SIDS prevention campaigns for non-Western cultures.

Full text

Ž.
Early Human Development 62 2001 43–55
www.elsevier.comrlocaterearlhumdev
International Child Care Practices Study: infant
sleeping environment
E.A.S. Nelsona,), B.J. Taylor b
aDepartment of Paediatrics, The Chinese UniÕersity of Hong Kong, 6rF Clinical Science Building, Prince of
Wales Hospital, Shatin, Hong Kong, People’s Republic of China
bDepartment of Paediatrics, Dunedin School of Medicine, UniÕersity of Otago, Dunedin, New Zealand
ICCPS Study Group, Alejandre Jenik c, John Vance d,
Karen Walmsley d, Katie Pollard d, Michelle Freemantle d,
Dot Ewing d, Christa Einspieler e, Heidemarie Engele e,
Petra Ritter f, G. Elske Hildes-Ripstein g, Monica Arancibiah,
Xiaocheng Ji i, Haiqi Li j, E.A.S. Nelson k, Crystal Bedard k,
Karin Helweg-Larsenl, Katrine Sidenius l, Susan Karlqvist l,
Christian Poets m, Eva Barko n, Bernadette Kiberd o,
Mary McDonnell o, Gianpaolo Donzelli p, Raffaele Piumelli p,
Luca Landini p, Arturo Giustardi q, Hiroshi Nishida r,
Stephanie Fukui r, Toshiko Sawaguchi r, Masataka Ino r,
Takeshi Horiuchi s, Koki Oguchi s, Barry J. Taylor t,
Sheila Williamst, Yildiz Perk u, David Tappin v, Joseph Milerad w,
Maria Wennborg w, N. Aryayev x, V. Nepomyashchaya x
cBuenos Aires, Argentina
dBrisbane, Australia
eGraz, Austria
fInnsbruck, Austria
gManitoba, Canada
hSantiago, Chile
iBeijing, People’s Republic of China
jChongqing, People’s Republic of China
kHong Kong, People’s Republic of China
lCopenhagen, Denmark
)Corresponding author. Tel.: q852-263-228-59; fax: q852-263-600-20.
Ž.
E-mail address: tony-nelson@cuhk.edu.hk E.A.S. Nelson .
0378-3782r01r$ - see front matter q2001 Elsevier Science Ireland Ltd. All rights reserved.
Ž.
PII: S0378-3782 01 00116-5

()
E.A.S. Nelson et al.rEarly Human DeÕelopment 62 2001 43–5544
mHannoÕer, Germany
nBudapest, Hungary
oDublin, Ireland
pFlorence, Italy
qNaples, Italy
rTokyo, Japan
sYokohama, Japan
tDunedin, New Zealand
uIstanbul, Turkey
vGlasgow, Scotland, UK
wKarolinska Institute, Sweden
xOdessa, Ukraine
Received 12 August 2000; received in revised form 19 December 2000; accepted 20 December 2000
Abstract
Ž.
Background: The International Child Care Practices Study ICCPS has collected descriptive
data from 21 centres in 17 countries. In this report, data are presented on the infant sleeping
Ž.
environment with the main focus being sudden infant death syndrome SIDS risk factors
Ž.Ž.
bedsharing and infant using a pillow and protective factors infant sharing a room with adult
that are not yet well established in the literature.
Methods: Using a standardised protocol, parents of infants were surveyed at birth by interview
and at 3 months of age mainly by postal questionnaire. Centres were grouped according to
geographic location. Also indicated was the level of SIDS awareness in the community, i.e.
whether any campaigns or messages to Areduce the risks of SIDSBwere available at the time of
the survey.
Ž.
Results: Birth interview data were available for 5488 individual families and 4656 85%
Ž.
returned questionnaires at 3 months. Rates of bedsharing varied considerably 2–88% and it
appeared to be more common in the samples with a lower awareness of SIDS, but not necessarily
a high SIDS rate. Countries with higher rates of bedsharing appeared to have a greater proportion
Ž. Ž .
of infants bedsharing for a longer duration )5 h . Rates of room sharing varied 58–100% with
some of the lowest rates noted in centres with a higher awareness of SIDS. Rates of pillow use
ranged from 4% to 95%.
Conclusions: It is likely that methods of bedsharing differ cross-culturally, and although
further details were sought on different bedsharing practices, it was not possible to build up a
composite picture of AtypicalBbedsharing practices in these different communities. These data
highlight interesting patterns in child care in these diverse populations. Although these results
should not be used to imply that any particular child care practice either increases or decreases the
risk of SIDS, these findings should help to inject caution into the process of developing SIDS
prevention campaigns for non-Western cultures. q2001 Elsevier Science Ireland Ltd. All rights
reserved.
Keywords: Child care; Infant sleeping environment; SIDS
1. Introduction
Ž.
The International Child Care Practices Study ICCPS collected comparative informa-
tion on child care practices from 21 centres in 17 countries. These include countries with

()
E.A.S. Nelson et al.rEarly Human DeÕelopment 62 2001 43–55 45
Ž.
recognised low rates of sudden infant death syndrome SIDS such as Hong Kong and
wx
Japan 1 , as well as information from other countries that previously had higher rates of
wx
SIDS such as New Zealand 2 . This report provides details on some of the less well
Ž.
established SIDS risk factors bedsharing, use of a pillow and protective factors
Ž.
sleeping in the same room as an adult . These practices have been associated with an
increased or decreased risk of SIDS in one or more studies, but there is still debate about
what advice should be given to parents. Information is also presented on the main
daytime and night-time caregivers.
2. Methods
Ž.
The International Child Care Practices Study ICCPS methods have been previously
wx
described 3 . Subsequent to this publication, data from Odessa in the Ukraine has been
included. The study was designed to recruit families so that the infants would be 3
months old during the coldest 2 months of the year. Invitation to participate was made
during the week after birth. A Abirth questionnaireBwas completed at the time of
recruitment by interview and collected mainly socio-demographic data. A Ahome
questionnaireBwas posted to participating families when the baby was 12 weeks old.
Telephone reminders were given if questionnaires were not returned and a second
questionnaire posted if required. Some centres administered the home questionnaire by
face-to-face interview. The home questionnaire was designed so that it should be
completed on the day that it was received, with many questions referring to Alast nightB.
Centres were advised to recruit 250 families. Not all centres achieved this target,
whereas other centres recruited more families. Not all centres recorded rates of refusal to
participate and some centres only submitted data where both birth and home question-
naires had been completed. A number of variations in the methods between different
centres occurred, e.g. some centres conducted face-to-face interviews and others used
wx
retrospectively collected birth information data 3 . Centres coded and entered their data
Ž
using Epi Info data entry and analysis programmes provided Epi Info statistical
.
software version 6, Centers for Disease Control and Prevention, Atlanta . Analysis of the
data was undertaken with Epi Info and has been primarily descriptive and centres were
grouped according to geographic location. Also indicated is whether SIDS was thought
to be a problem in the community and whether any specific Areducing the risks of
SIDSBmessages were being propagated at the time of the survey. Group A were
Ž
considered to be higher awareness centres or countries some form of SIDS awareness
.Ž
messages available and Group B to be the lower awareness centres or countries no
.
specific messages or campaigns available at the time of the survey .
3. Results
Details of room sharing are shown in Table 1. Two questions related to room sharing
were asked. The first was which room the infant slept in: the parents’ room, the infant’s

()
E.A.S. Nelson et al.rEarly Human DeÕelopment 62 2001 43–5546
Table 1
International Child Care Practices Study: infant sleeping environment at 3 months of age
nSIDS rate Room sharing Infant checked
1995 after parents
Parents room, Yes One adult had gone
Ž. Ž. Ž.
n% % only % to bed
% Median
America aŽ.
Manitoba, Canada 230 0.8 112 49 66 21 80 2
bŽ.
Buenos Aires, Argentina 81 1.4 55 68 69 4 86 3
Ž.
Santiago, Chile 226 – 216 96 95 17 98 –
Northern Europe ac
Ž.
Copenhagen, Denmark 361 0.3 318 88 91 10 79 2
aŽ.
Graz, Austria 199 0.3 134 67 70 14 96 2
aŽ.
InnsbuckrVienna, Austria 200 0.7 147 74 76 14 89 2
aŽ.
Hannover, Germany 122 0.9 61 50 59 14 84 2
aŽ.
Dublin, Ireland 321 0.6 282 88 86 12 83 2
a
Ž. Ž.
Scotland three cities 221 0.6 173 78 83 15 85 2
aŽ.
Stockholm, Sweden 241 0.3 228 95 97 16 78 2
Southern and eastern Europe
b
Ž. Ž.
Hungary Budapestrother 32 0.15 22 69 69 0 94 2
bŽ.
Istanbul, Turkey 94 – 72 77 92 10 91 3
Ž.
FlorencerNaples, Italy 200 – 174 87 88 4 98 3
Ž.
Odessa, Ukraine 488 0.6 418 86 96 26 82 3
Asia
Ž.
Beijing, China 306 – 299 98 100 13 95 2
Ž.
Chongqing, China 250 – 208 83 99 23 96 3
Ž.
Hong Kong SAR, China 198 0.1 141 71 85 26 91 3
aŽ.
Hong Kong Caucasian 117 – 64 55 61 15 71 2
aŽ.
TokyorYokohama, Japan 288 0.4 222 77 93 25 91 2
Oceania ac
Ž.
Brisbane, Australia 225 0.9 141 63 61 11 79 2
aŽ.
Dunedin, New Zealand 249 1.1 115 46 58 14 78 2
aCentres with some form of SIDS awareness messages available considered as Group A, and the remainder
considered as Group B.
bCentres with fewer than 100 respondents for the 3-month home questionnaire.
c1994.
own room, or another room. The second question asked more directly the number of
Ž. Ž .
adults over 12 years and children under 12 years who shared a room with the infant.
Ž.
High rates of room sharing one or more adults in the room with the infant were noted
Ž.
in both Group A 97% Stockholm, 93% TokyorYokohama, 91% Copenhagen and
Ž
Group B 88% FlorencerNaples, 96% Odessa, 100% Beijing, 99% Chongqing, 85%
.
Hong Kong . Relatively lower rates of room sharing were noted in some of the Group A
Ž.
centres 58% Dunedin, 59% Hannover, 61% Brisbane, 61% Hong Kong Caucasian . A

()
E.A.S. Nelson et al.rEarly Human DeÕelopment 62 2001 43–55 47
further question asked whether the infant had been checked after the parents had gone to
bed, and if checked, the number of times the infant had been checked was recorded. The
data in Table 1 shows that in the majority of centres, infants had been checked and the
median number of times was 2–3. SIDS rates, where known, of the centres for 1995 or
1994 are also presented in Table 1.
Information on bedsharing is detailed in Table 2. Chongqing infants had the highest
Ž.
rate of sharing a bed 88% and in the majority of cases it was with both parents. The
Ž. Ž .
lowest rates of bedsharing were noted in Odessa 9% , Dunedin 19% and the Hong
Ž.
Kong Caucasian 19% samples. In some of the samples, bedsharing was predominantly
Ž.
with mother and infant only Copenhagen, Dublin, TokyorYokohama, Odessa , whereas
Ž
in others, it was mainly with both parents Brisbane, Stockholm, FlorencerNaples,
.
Chongqing . Data on total bedsharing frequency compared with the percentage of the
bedsharers in the mother–infant only group is shown in Fig. 1. Further questions were
asked to obtain more details of method of bedsharing. These questions included
drawings representing an infant placed beside one person, beside two people, or between
two people. These questions were further subdivided into whether the infant was placed
directly on the bed, raised on a separate surface or placed in a container within the bed.
This was a complicated question and it was difficult to determine any patterns regarding
methods of bedsharing. In some centres, infants were more commonly placed beside one
or more persons rather than between two persons, e.g. TokyorYokohama, whereas in
other centres, bedsharing infants were more commonly between two persons, e.g.
Stockholm. Information on whether bedsharing infants were kept Aat arms lengthB, were
Aclose but not touchingBor were Ain direct contactBis also shown in Table 2. About
Fig. 1. International Child Care Practices Study: percentage of infants bedsharing at 3 months of age
Ž)
contrasted with the percentages of bedsharing situations with mother–infant only. Centres with fewer than
.
100 respondents for the 3-month home questionnaire.

()
E.A.S. Nelson et al.rEarly Human DeÕelopment 62 2001 43–5548
Table 2
International Child Care Practices Study: bedsharing at 3 months of age
bc
nShare Bedshared Bedsharing method Closeness Duration of bedsharing
a
Ž.
% with whom Ž. Ž. Ž.
PQ R% SACD% -2 h 2–5 h )5h %
Ž.
M % B Other
America dŽ. Ž. Ž. Ž. Ž.
Manitoba, Canada 230 53 23 23 43 21 9 21 12 13 25 7 5 27 19 36 10 8 34 65
Ž.
Buenos Aires, 81 12 15 4 7 1 4 5 2 1 0 5 7 8 2 2
Argentina
Ž. Ž. Ž. Ž. Ž.
Santiago, Chile 226 144 64 58 41 66 18 65 53 20 14 6 8 61 74 52 3 6 134 94
Northern Europe
Ž. Ž. Ž. Ž. Ž.
Copenhagen, 360 140 39 125 89 0 15 40 52 41 29 7 8 72 59 42 47 30 63 45
d
Denmark
dŽ. Ž. Ž. Ž. Ž.
Graz, Austria 199 50 25 27 54 21 2 15 12 21 42 2 1 22 27 54 19 16 15 30
Ž. Ž. Ž. Ž. Ž.
InnsbruckrVienna, 200 69 35 24 35 42 3 15 14 39 57 1 5 31 33 48 18 25 26 38
d
Austria
dŽ. Ž. Ž. Ž. Ž.
Hannover, Germany 121 28 23 14 50 10 4 6 5 14 50 3 4 12 12 43 10 10 8 29
dŽ. Ž. Ž. Ž. Ž.
Dublin, Ireland 322 66 21 57 89 0 7 12 28 21 34 1 11 22 27 43 22 24 16 26
Ž. Ž. Ž. Ž. Ž.
Scotland 220 54 25 19 35 28 7 17 17 19 35 1 1 33 19 36 19 21 14 26
d
Ž.
threecities
dŽ. Ž. Ž. Ž. Ž.
Stockholm, Sweden 241 157 65 39 25 93 25 32 38 85 54 2 8 76 73 47 7 34 116 74

()
E.A.S. Nelson et al.rEarly Human DeÕelopment 62 2001 43–55 49
Southern and eastern Europe
Ž.
Hungary 32 5 16 2 3 0 1 0 1000 20 0 5
Ž.
Budapestrother
Ž.
Istanbul, Turkey 87 2 2 1 0 2 0 1 0002 01 1 0
Ž. Ž. Ž. Ž. Ž.
Florencer200 47 24 8 17 36 3 9 23 15 32 0 18 12 17 36 4 3 40 85
Naples, Italy
Ž. Ž. Ž. Ž. Ž.
Odessa, Ukraine 489 45 9 35 78 1 9 29 6 4 9 6 3 33 8 18 6 11 28 62
Asia
Ž. Ž. Ž. Ž. Ž.
Beijing, China 306 164 54 51 31 110 3 50 69 45 27 0 24 73 67 41 4 4 156 95
Ž. Ž. Ž. Ž .
Chongqing, China 250 221 88 51 23 126 44 62 77 49 22 31 – – – 0 0 221 100
Ž. Ž. Ž. Ž.
Hong Kong 197 73 37 43 59 19 11 25 22 13 20 5 – – – 5 7 58 83
SAR, China
dŽ. Ž. Ž. Ž.
Hong Kong Caucasian 117 22 19 7 32 12 3 5 2 13 59 2 2 15 5 6 6 10 46
Ž. Ž. Ž. Ž. Ž.
TokyorYokohama, 287 107 37 84 79 16 7 48 29 21 21 0 11 62 32 31 6 18 83 78
d
Japan
Oceania dŽ. Ž. Ž. Ž. Ž.
Brisbane, Australia 225 68 30 14 21 44 10 15 19 23 39 2 3 32 31 47 14 26 27 40
Ž. Ž. Ž. Ž. Ž.
Dunedin, 249 47 19 18 38 23 6 14 17 6 13 10 3 27 16 35 18 13 16 34
d
New Zealand
aBedsharer: Msmother only, Bsboth parents.
bMethods of bedsharing: PsBabyqone person ONLY, QsBaby at SIDE of two people, R sBaby BETWEEN two people, SsOther combination.
cCloseness of bedsharing: AsAt arms length, CsClose but not touching, DsIn direct contact.
dCentres with some form of SIDS awareness messages available considered as Group A, and the remainder considered as Group B.

()
E.A.S. Nelson et al.rEarly Human DeÕelopment 62 2001 43–5550
Fig. 2. International Child Care Practices Study: percentage of infants bedsharing at 3 months of age
Ž)
contrasted with percentages of bedsharing infants who bedshared for more than 5 h. Centres with fewer than
.
100 respondents for the 3-month home questionnaire.
one-third to one-half of infants were placed in direct contact with the bedsharer.
Duration of bedsharing is compared with total bedsharing frequency in Fig. 2. Informa-
tion on pillow use in these 3-month-old infants is shown in Table 3. Pillows were least
Ž. Ž. Ž.
popular in Scotland 4% , Manitoba 8% and Dunedin 9% and most popular in the
Ž.
Chinese samples Beijing 95%, Chongqing 95% and Hong Kong 80% . Details of the
Ž
materials used to make the pillows were sought but suggested options foam chip, foam
.
block, cotton and feather were probably inadequate as many centres recorded the
material of the pillow as AotherB. The Japanese samples used additional codes, which
showed that folded towels, bean chips, and AdoughnutB-shaped pillows were commonly
wx
used. The doughnut-shaped pillows have also been shown to be used in Hong Kong 4 ,
although the specific data on pillow type was not recorded for the Hong Kong sample in
this study. Details of the pillow size were not sought so it is not known what proportion
of these pillows would be large adult pillows or small pillows specifically for the infant.
Parents were the main caregiver of the infant during the day and at night. In the
Ž.
higher SIDS aware Group A centres, with the exception of the sub-sample of mainly
Hong Kong Caucasians, 93% or more of infants were cared for by parents during the
day and nearly 100% at night. Caregivers, other than the parents, were more common in
Ž.
the lower SIDS aware Group B centres with the highest percentages in the main Hong
Ž.Ž
Kong sample 42% during the day and 22% at night and the Chongqing sample 31%
.
during the day and 18% at night . 22% of infants in the sub-sample of mainly Hong
Kong Caucasian families were looked after by child minders during the day, as were 8%
of the infants at night.
4. Discussion
This study collected descriptive data on child care from a range of different countries
and cultures. Although every attempt was made to standardise the methods, significant

()
E.A.S. Nelson et al.rEarly Human DeÕelopment 62 2001 43–55 51
Table 3
International Child Care Practices Study: use of a pillow at 3 months of age
a
nPillow Whole body Type of pillow
Ž. bbbcc
used % on pillow FC FB CF FE BC DO TO BO FJ OT
America dŽ.
Manitoba, Canada 227 19 8 4 0 1 8 2 0 0 0 0 0 2
Ž.
Buenos Aires, Argentina 81 18 22 0 0 8 3 2 0 0 0 0 0 4
Ž.
Santiago, Chile 226 135 60 4 57 3 14 11 0 0 0 0 0 43
Northern Europe dŽ.
Copenhagen, Denmark 361 51 14 43 0 0 0 0 0 0 0 17 21 4
dŽ.
Graz, Australia 199 20 10 5 4 0 11 5 0 0 0 0 0 0
dŽ.
Hannover, Germany 120 16 13 5 2 0 4 6 0 0 0 0 0 3
Ž.
InnsbruckrVienna, 200 29 15 2 0 2 18 8 0 0 0 0 0 1
d
Austria
dŽ.
Dublin, Ireland 322 46 14 4 19 10 14 1 0 0 0 0 0 0
d
Ž. Ž.
Scotland three cities 220 9 4 0 1 0 3 3 0 0 0 0 0 2
dŽ.
Stockholm, Sweden 241 50 21 6 16 0 31 1 0 0 0 0 0 0
Southern and eastern Europe
Ž.
Hungary 32 0 0 0
Ž.
Budapestrother
Ž.
Istanbul, Turkey 88 46 52 6 1 1 11 1 0 0 0 0 0 14
Ž.
FlorencerNaples, Italy 200 101 51 3 23 4 57 17 0 0 0 0 0 0
Ž.
Odessa, Ukraine 489 183 37 61 1 2 48 84 0 0 0 0 0 48
Asia
Ž.
Beijing, China 305 289 95 0 0 6 53 0 0 0 0 0 0 231
Ž.
Chongqing, China 250 237 95 7 1 0 156 3 0 0 0 0 0 74
Ž.
Hong Kong SAR, China 196 156 80 3 2 19 100 3 0 0 0 0 0 12
dŽ.
Hong Kong Caucasian 116 11 10 0 1 1 7 1 0 0 0 0 0 0
Ž.
TokyorYokohama, 288 162 56 2 4 0 48 0 16 28 52 0 0 5
d
Japan
Oceania dŽ.
Brisbane, Australia 225 53 24 3 5 1 24 1 0 0 0 0 0 12
dŽ.
Dunedin, New Zealand 246 22 9 4 2 2 14 0 0 0 0 0 0 4
aFCsFoam chip, FBsfoam block, CF sCotton, FEsfeather, BCsBean chips, DOsdoughnut-shaped
pillow, TOsfolded towel, BOsflat layer of cotton sheet, FJsflat pillow of feathers, OTsother.
bCodes only used for the sample from Japan.
cCodes only used for the Copenhagen sample.
dCentres with some form of SIDS awareness messages available considered as Group A, and the remainder
considered as Group B.
variations in data collection occurred. Thus, although we consider that the international
comparison of child care practice is possible using these standardised methods, we also
note that, in view of the heterogeneity of the samples, it is important to avoid
over-interpretation of any differences identified, and to view such differences within the

()
E.A.S. Nelson et al.rEarly Human DeÕelopment 62 2001 43–5552
wx
qualitative context of each individual sample 3 . Provided the limitations of such
ecological studies are acknowledged, we believe that the information generated can be
useful for hypothesis generation, as well as providing useful baseline descriptive data.
This report looks at data on some of less well-established SIDS risk factors
Ž. Ž
bedsharing and use of a pillow and a possible SIDS protective factor sharing a room
.
with an adult . Information is also presented on the main daytime and night-time
caregivers, which showed considerable variation between the different samples. In
particular, a relatively high percentage of Hong Kong infants were cared for by people
other than the mother or father during both the day and night. This may reflect the fact
that a significant number of mothers work outside the home during the day. However, it
may also be the result of the phenomenon of separated families in Hong Kong, i.e.
father, grandparents and child live in Chinese Hong Kong whereas, the mother lives in
the Chinese mainland due to immigration restrictions. However, even in the Chongqing
sample in the Chinese mainland, a significant percentage of infants were cared for by
Ž. Ž. Ž
relatives 21% or other carers 10% during the day and at night relatives and other
.
carers 18% . This suggests a Chinese cultural practice involving the extended family in
the infant’s care. In contrast, 22% and 8% of infants in the sub-sample of mainly Hong
Ž.
Kong Caucasians were cared for by paid child minders amahs during the day and at
night, respectively. Most amahs in Hong Kong are migrant workers from the Philip-
pines. It is interesting to speculate how this additional cultural influence might modify
child care practices in this particular population.
Ž
Studies have shown that infants who sleep in the same room as an adult but not the
.wx
same bed have a lower risk of SIDS 5 . Our data noted wide variations in the
frequency with which infants shared a room with one or more adults. Almost no infants
in Chongqing and Beijing and very few infants in the samples in TokyorYokohama,
Ž.
Stockholm, Odessa, and Santiago slept alone Table 1 . However, in some of the
Ž
Caucasian cultures, solitary sleeping for infants was relatively common Dunedin 42%,
.
Hannover 41%, Brisbane 39%, Hong Kong Caucasian 39% . It has been suggested that
in western culture, parents appear to place value on infants becoming independent
whereas, in Asian cultures infants are seen as vulnerable and need of constant attention
wx
and care 6 . High rates of room sharing have been noted previously in Asian
wx
populations in the United Kingdom 7 . It is not clear whether the high rate of room
sharing in the Stockholm sample is a normal cultural practice or the result of increased
awareness of SIDS risk factors in the population. A factor that was not addressed in this
study was the size of the home and the number of rooms. Limited space may make room
sharing a necessity rather than a desired state.
The practice of bedsharing in relation to SIDS risk has generated considerable debate
wx
4,8–13 . Some of the highest rates of bedsharing appeared to be in countries with low
SIDS rates. There also appeared to be a pattern for those countries with higher rates of
Ž
bedsharing to have a greater percentage of infants bedsharing for more than 5 h Beijing,
.
Chongqing, Santiago and Stockholm . In contrast, the infants who bedshared in those
Ž.
countries with lower rates of bedsharing F30% appeared to bedshare for a shorter
Ž.Ž.
duration Brisbane, Dublin, Dunedin, Graz, Hannover, three cities in Scotland Fig. 2 .
This suggests that bedsharing may be a routine cultural practice in those centres with
higher rates of bedsharing of long duration, whereas in the other centres it may be purely

()
E.A.S. Nelson et al.rEarly Human DeÕelopment 62 2001 43–55 53
a process whereby the infant is brought into the bed for a feed or comforting for a part
of the night. A previous Australian study noted relatively low rates of bedsharing in
Ž.
Australian-born mothers of Anglo–Celtic origin 2.5% and in mothers born in southern
Ž. Ž .
Europe 5.4% , moderate rates in mothers born in Asia 25.4% and higher rates in
mothers whose infants were born at home and considered to be a high risk group for
Ž.wx
SIDS 55.9% 14 . In New Zealand, bedsharing was identified as a risk factor for SIDS
wx
when the mother was also a smoker 15 . The mechanism by which bedsharing and
smoking interact to increase the risk of SIDS is not known. However, if it is assumed
that this interaction is in some way casual, it might be argued that high rates of
Ž.
bedsharing in Chongqing 88% do not increase the risk of SIDS in this population
because none of the mothers smoked. However, it should be noted that no SIDS
mortality data were available for Chongqing. This argument was used to explain the
wx
paradox of a low SIDS rate and a moderately high bedsharing rate in Hong Kong 4,16 .
However, It is also likely that the exact method of bedsharing may differ significantly
between different cultures and it may be speculated that high rates of bedsharing in
populations with low rates of SIDS might indicate AsafeBmethods of bedsharing. Since
Ž.
bedsharing has potential beneficial effects e.g. promoting breastfeeding it would seem
to be undesirable to recommend against this practice in populations where it is common
practice and SIDS incidence is low. Table 2 indicates the considerable variability in
Ž
bedsharing in terms of who shared the bed mother only, both parents, or other
.
combination . Additional questions were asked on the method of bedsharing, i.e.
whether infant was placed beside one person, infant was placed beside two people,
infant was placed between two people, or other combination and whether infant was
placed directly on the bed, or on a raised surface or in a container. However, inspection
of the data suggests that there may have been some difficulties with this question, e.g.
Ž.
140r360 infants in Copenhagen shared a bed and in 126 89% , it was mother only, but
the method of bedsharing was described as infant beside one person in 41 cases and
Ž. Ž.
either beside two people 53 cases or between two people 41 cases . Bedsharing is a
very complex and variable process and this appears to be reflected in the difficulties that
some participants appeared to have in answering these questions. Bedsharing will also
wx
influence the thermal environment of the infant 14 . This issue of complexity should
also be considered when interpreting the results of studies undertaken in sleep laborato-
ries on bedsharing mother–infant pairs, which may differ to the types of bedsharing
wx
practices, which occur in the home 13,17,18 .
It has been noted in Hong Kong that small circular doughnut-shaped pillows are
wx
commonly used 4 . Similar pillows are used in our Japanese samples, as are folded
towels or cloths. The data presented in Table 3 indicates a very wide variation in the use
of pillows by 3-month-old infants. In some Western studies, SIDS has been associated
wx
with pillow use 19,20 . Our data show low rates of pillow use in some of the samples
ŽŽ . .
Scotland three cities 4%, Manitoba 8% and Dunedin 9% and high rates in others
Ž.
Beijing 95%, Chongqing 95%, Hong Kong 80% . As many as 56% of the
TokyorYokohama sample used a pillow and additional details of the types of pillows
Ž
used were collected folded towels 32%, cotton 30%, doughnut-shaped pillows 17%,
.
bean chips 10% . Pillow use appeared to be more common in the Asian samples where
SIDS rates have been documented to be low or are thought to be low. We might

()
E.A.S. Nelson et al.rEarly Human DeÕelopment 62 2001 43–5554
speculate that the type of pillows used in the Chinese and Japanese populations are not
dangerous from a SIDS perspective. In one study where pillow use has been implicated
wx
with SIDS, they were large adult V-shaped pillows 20 . It might be expected that
placing an infant on top of a large soft adult pillow will result in a very different level of
risk to that of placing the back of an infant’s head on a small round doughnut-shaped
pillow. As noted, details of the pillow size were not sought, so the proportion of large
adult pillows to small infant pillows is not known, and thus no conclusions can be drawn
from these data. However, on the basis of present data, it is suggested that no attempt be
made to advise against the use of pillows in those populations where this is a common
practice and the risk of SIDS is low. It can also be speculated that the types of pillows
used in these populations could have some beneficial effects. A supine-sleeping infant
whose head is placed on a small doughnut-shaped pillow might have more difficulty
acquiring the momentum to roll over into a prone sleeping position. There is also the
Ž.
issue of flattened head shape occipital plagiocephaly from supine sleep position
wx
21,22 . It has been suggested that for habitual supine sleeping infants, the use of small
doughnut-shape pillows may help prevent flattening of the occipit. Unfortunately, our
data do not provide information on the reasons why parents choose to use pillows, i.e.
whether to promote sleep or to prevent head flattening or whether it is purely an
unquestioned cultural practice.
In conclusion, these results provide interesting descriptive data on child care practices
that have both been associated with SIDS and have generated some debate. Considerable
variation was noted in all the practices described. Bedsharing was common in some of
the samples that were considered to have a low SIDS incidence or to be low SIDS aware
countries. Interaction with smoking may help explain this paradox but further study is
needed to understand the exact methods and complexity of bedsharing both between and
within different cultures. Room sharing with one or more adults was almost universal in
some samples but less common in others, particularly the more western cultures. Similar
variation was noted with pillow use, which was particularly common in the Chinese
samples. It is possible that the types of pillow used in these populations differ
significantly from those implicated with SIDS. A significant proportion of the caregivers
in some regionsrcultures, both during daytime and at night, were relatives and paid
child minders. The results from this study are not intended to be used to imply that any
particular child care practices either increase or decrease the risk of SIDS, but instead to
better understand the complexity of child care in these different cultures. These data
provide useful baseline information and should help to inject some additional caution
into the process of developing AReduce the RisksBmessages for SIDS prevention
campaigns.
Acknowledgements
This project was an initiative of the SIDS Global Strategy Task Force whose support
is greatly appreciated. The Research Grants Council and the Society for the Relief of
Disabled Children in Hong Kong provided some funding.

()
E.A.S. Nelson et al.rEarly Human DeÕelopment 62 2001 43–55 55
References
wx
1 N.N. Lee, Y.F. Chan, D.P. Davies, E. Lau, D.C. Yip, Sudden infant death syndrome in Hong Kong:
Ž.
confirmation of low incidence, BMJ 298 1989 721.
wx
2 E.A. Nelson, S.M. Williams, B.J. Taylor, B. Morris, R.P. Ford, V.M. Binney et al., Prediction of possibly
preventable death: a case-control study of postneonatal mortality in southern New Zealand, Paediatric and
Ž.
Perinatal Epidemiology 4 1990 39–52.
wx
3 E.A. Nelson, B.J. Taylor, International child care practices study: methods and study population, Early
Ž.
Human Development 55 1999 149–168.
wx
4 E.A. Nelson, P.H. Chan, Child care practices and cot death in Hong Kong, New Zealand Medical Journal
Ž.
109 1996 144–146.
wx
5 E.A. Mitchell, J.M. Thompson, Co-sleeping increases the risk of SIDS, but sleeping in the parents’
Ž.
bedroom lowers it, in: T.O. Rognum Ed. , Sudden Infant Death Syndrome. New Trends in the Nineties,
Scandinavian Univ. Press, Oslo, 1995, pp. 266–269.
wx
6 M. Gantley, D.P. Davies, A. Murcott, Sudden infant death syndrome: links with infant care practices,
Ž.
BMJ 306 1993 16–20.
wx
7 S. Farooqi, Ethnic differences in infant care practices and in the incidence of sudden infant death
Ž.
syndrome in Birmingham, Early Human Development 38 1994 209–213.
wx
8 R.K. Scragg, E.A. Mitchell, Side sleeping position and bed sharing in the sudden infant death syndrome,
Ž.
Annals of Medicine 30 1998 345–349.
wx
9 E.A. Mitchell, J.M. Thompson, R.P. Ford, B.J. Taylor, Sheepskin bedding and the sudden infant death
Ž.
syndrome. New Zealand Cot Death Study Group, Journal of Pediatrics 133 1998 701–704.
wx
10 F.R. Hauck, J.S. Kemp, Bedsharing promotes breastfeeding and AAP Task Force on Infant Positioning
Ž.
and SIDS, Pediatrics 102 1998 662–664.
wx
11 D.J. Henderson-Smart, A.L. Ponsonby, E. Murphy, Reducing the risk of sudden infant death syndrome: a
Ž.
review of the scientific literature, Journal of Paediatrics and Child Health 34 1998 213–219.
wx
12 J. Kattwinkel, J. Brooks, M.E. Keenan, M. Malloy, M. Willinger, A. Mitchell et al., Does bed sharing
Ž.
affect the risk of SIDS? Pediatrics 100 1997 272.
wx Ž.
13 J.J. McKenna, S.S. Mosko, C.A. Richard, Bedsharing promotes breastfeeding, Pediatrics 100 1997
214–219.
wx
14 L. Watson, A. Potter, R. Gallucci, J. Lumley, Is baby too warm? The use of infant clothing, bedding and
Ž.
home heating in Victoria, Australia, Early Human Development 51 1998 93–107.
wx
15 E.A. Mitchell, P.G. Tuohy, J.M. Brunt, J.M. Thompson, M.S. Clements, A.W. Stewart et al., Risk factors
for sudden infant death syndrome following the prevention campaign in New Zealand: a prospective
Ž.
study, Pediatrics 100 1997 835–840.
wx
16 R.K. Scragg, E.A. Mitchell, A.W. Stewart, R.P. Ford, B.J. Taylor, I.B. Hassall et al., Infant room-sharing
and prone sleep position in sudden infant death syndrome. New Zealand Cot Death Study Group, Lancet
Ž.
347 1996 7–12.
wx
17 S. Mosko, C. Richard, J. McKenna, S. Drummond, D. Mukai, Maternal proximity and infant CO2
environment during bedsharing and possible implications for SIDS research, American Journal of
Ž.
Physical Anthropology 103 1997 315–328.
wx
18 J. McKenna, S. Mosko, C. Richard, S. Drummond, L. Hunt, M.B. Cetel et al., Experimental studies of
infant–parent co-sleeping: mutual physiological and behavioral influences and their relevance to SIDS
Ž. Ž.
sudden infant death syndrome , Early Human Development 38 1994 187–201.
wx
19 N.J. Scheers, C.M. Dayton, J.S. Kemp, Sudden infant death with external airways covered: case-compari-
Ž.
son study of 206 deaths in the United States, Archives of Pediatrics and Adolescent Medicine 152 1998
540–547.
wx
20 R.W. Byard, S.M. Beal, V-shaped pillows and unsafe infant sleeping, Journal of Paediatrics and Child
Ž.
Health 33 1997 171–173.
wx
21 L.C. Argenta, L.R. David, J.A. Wilson, W.O. Bell, An increase in infant cranial deformity with supine
Ž.
sleeping position, Journal of Craniofacial Surgery 7 1996 5–11.
wx
22 A.A. Kane, L.E. Mitchell, K.P. Craven, J.L. Marsh, Observations on a recent increase in plagiocephaly
Ž.
without synostosis, Pediatrics 97 1996 877–885.