• infant care;
  • ethnic group;
  • Born in Bradford;
  • breast feeding;
  • bed sharing;
  • sleeping position;
  • pacifier use


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Ball HL, Moya E, Fairley L, Westman J, Oddie S, Wright J. Infant care practices related to sudden infant death syndrome in South Asian and White British families in the UK. Paediatric and Perinatal Epidemiology 2012; 26: 3–12.

In the UK, infants of South Asian parents have a lower rate of sudden infant death syndrome (SIDS) than White British infants. Infant care and life style behaviours are strongly associated with SIDS risk. This paper describes and explores variability in infant care between White British and South Asian families (of Bangladeshi, Indian or Pakistani origin) in Bradford, UK (the vast majority of which were Pakistani) and identifies areas for targeted SIDS intervention. A cross-sectional telephone interview study was conducted involving 2560 families with 2- to 4-month-old singleton infants enrolled in the Born in Bradford cohort study. Outcome measures were prevalence of self-reported practices in infant sleeping environment, sharing sleep surfaces, breast feeding, use of dummy or pacifier, and life style behaviours. We found that, compared with White British infants, Pakistani infants were more likely to: sleep in an adult bed (OR = 8.48 [95% CI 2.92, 24.63]); be positioned on their side for sleep (OR = 4.42 [2.85, 6.86]); have a pillow in their sleep environment (OR = 9.85 [6.39, 15.19]); sleep under a duvet (OR = 3.24 [2.39, 4.40]); be swaddled for sleep (OR = 1.49 [1.13, 1.97]); ever bed-share (OR = 2.13 [1.59, 2.86]); regularly bed-share (OR = 3.57 [2.23, 5.72]); ever been breast-fed (OR = 2.00 [1.58, 2.53]); and breast-fed for 8+ weeks (OR = 1.65 [1.31, 2.07]). Additionally, Pakistani infants were less likely to: sleep in a room alone (OR = 0.05 [0.03, 0.09]); use feet-to-foot position (OR = 0.36 [0.26, 0.50]); sleep with a soft toy (OR = 0.52 [0.40, 0.68]); use an infant sleeping bag (OR = 0.20 [0.16, 0.26]); ever sofa-share (OR = 0.22 [0.15, 0.34]); be receiving solid foods (OR = 0.22 [0.17, 0.30]); or use a dummy at night (OR = 0.40 [0.33, 0.50]). Pakistani infants were also less likely to be exposed to maternal smoking (OR = 0.07 [0.04, 0.12]) and to alcohol consumption by either parent. No difference was found in the prevalence of prone sleeping (OR = 1.04 [0.53, 2.01]). Night-time infant care therefore differed significantly between South Asian and White British families. South Asian infant care practices were more likely to protect infants from the most important SIDS risks such as smoking, alcohol consumption, sofa-sharing and solitary sleep. These differences may explain the lower rate of SIDS in this population.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Alterations in infant care practices, such as sleep position,1–7 have been accompanied by substantial reductions in sudden infant death syndrome (SIDS) rates.8–10 Ethnic differences in infant care practices have also been associated with variation in SIDS rates.11–18 It is therefore important to understand ethnic variability in infant care in the UK to (i) determine the need for further interventions and (ii) design interventions relevant for multicultural populations.

The general UK SIDS rate is currently 0.3–0.5/1000 livebirths*; however, immigrant mothers from ‘New Commonwealth’ countries (India, Pakistan, Bangladesh and Caribbean) experience a substantially lower SIDS rate than mothers born in the UK.19 In Bradford 1998–2003 the overall SIDS rate was 0.5/1000, being 0.2/1000 among South Asian infants and 0.8/1000 among White British infants (Moya, personal communication). Previous UK studies of South Asian infant care reported different propensities for prone infant sleep position, lone infant sleep (separate room), infant wrapping and covering, and exposure to cigarette smoke,20,21 but these studies were relatively small, and did not control for confounding variables. Furthermore, since they were conducted, SIDS reduction campaigns have addressed various practices designed to change the landscape of infant care in the UK.

We designed the Bradford Infant Care Study (BradICS) to give a detailed picture of South Asian infant care practices, to contrast them with White British families, and thereby to advance the understanding of practices that might contribute to, or protect from, the risk of SIDS.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

With a population of 500 000, Bradford is the fifth largest metropolitan area in the UK, home to the eighth most deprived health community. Twenty-two per cent of Bradford's inhabitants are from minority ethnic groups, predominantly from Pakistan (70 000), India (13 000) and Bangladesh (5000), with 50% of the 6200 births at Bradford Royal Infirmary each year being to parents of South Asian origin. Bradford therefore provides an ideal location for ascertaining cultural variability in infant care practices among well-established ethnic communities in comparison with White British families living in the same or adjacent neighbourhoods and under similar socio-economic conditions.

Bradford Infant Care Study used a sub-population drawn from the Born in Bradford (BiB) longitudinal cohort study.22 In BiB, pregnant women are recruited at 26–28 weeks gestation and asked to complete a baseline questionnaire including ethnicity and family history, social and economic data, behavioural and life style data. Approximately half of the study participants are of Pakistani ethnicity (self-defined). NHS Research Ethics approval for BradICS was obtained as an amendment to the main NHS Research Ethics Committee approval granted for BiB. Recruitment, consent and enrolment procedures for BiB are published elsewhere.22

Inclusion criteria for BradICS were to be enrolled in BiB and have a livebirth at Bradford Royal Infirmary between June 2008 and September 2009. A telephone questionnaire was designed and piloted with White British and South Asian families to determine predicted response rate, average call duration, most effective timing for calls and proportion of respondents requiring a translated questionnaire. Twenty-five questions with yes/no and multiple choice answers were constructed addressing 12 aspects of infant care practices previously identified as positively or negatively related to SIDS risk. Following the pilot study any culturally sensitive or ambiguous questions were reworded. Participants were not asked about their knowledge of SIDS prevention, and reference to SIDS was not made in the telephone questionnaire, to avoid influencing participants to report what they ‘should’ be doing, rather than what they were doing. We anticipated approaching up to 5000 BiB-enrolled families to obtain survey data from approximately 2500.

A team of trained researchers contacted all BiB participants meeting the inclusion criteria by phone when their baby was 2–4 months old. Where phone calls were not answered, up to five repeated attempts were made at various times of day. If a respondent declined participation, a phone number was unavailable, or five unsuccessful attempts were made, the family was classed as a ‘non-completer’. If a respondent was not fluent in English, a bilingual research assistant phoned back. In total, 4131 families were eligible for inclusion; 3082 were contacted, consented and completed the telephone questionnaire; 1049 were designated as ‘non-completers’ of the BradICS questionnaire and 85% of the BradICS completers also completed the BiB baseline questionnaire (reflecting the overall baseline completion rate for BiB of 80%). Analysis of the BradICS participants in this paper is therefore based upon 2560 women who had completed the baseline and telephone interviews and had a singleton birth: 2457 (96.0%) of the respondents were the mothers (Figure 1). The majority of respondents [2243 (87.6%)] completed the telephone interviews in English, 247 (9.7%) in Urdu, 17 (0.7%) in Mirpuri/Punjabi and for 53 (2.1%) this information was missing. The median baby's age at completion of the BradICS telephone survey was 107 days (interquartile range 95–155 days).


Figure 1. Participant flow through the study.

Download figure to PowerPoint

Table 1 compares the demographic characteristics of completers and non-completers of the BradICS survey. BiB data were available for 774 BradICS non-completers who had a singleton birth. Compared with those who completed the BradICS survey, non-completers were generally younger, of White British ethnicity, never married, not living with a partner, and with low income. No differences were observed between completers and non-completers for infant characteristics including parity, birthweight and gestational age, or with regard to alcohol consumption; however, significantly more non-completers smoked cigarettes and consumed drugs before and during their pregnancy.

Table 1.  Comparison of baseline characteristics between completers and non-completers of the Bradford Infant Care Study
CharacteristicCompleted (n = 2560)Not completed (n = 774)P-value
n (%)n (%)
Maternal age <20 years134 (5.2)82 (10.6)<0.0001
White British ethnicity968 (37.8)333 (43.0)<0.0001
Single (never married)655 (25.6)293 (37.9)<0.0001
Not living with partner345 (13.5)171 (22.1)<0.0001
Annual income after tax <£9300420 (16.4)183 (23.6)<0.0001
Parity 3+350 (13.7)90 (11.6)0.16
Male baby1336 (52.2)369 (47.7)0.03
Low birthweight (<2500 g)190 (7.4)52 (6.7)0.51
Preterm (<37 weeks gestation)121 (4.7)43 (5.6)0.35
Mother smoked during pregnancy or 3 months before462 (18.0)194 (25.1)<0.0001
Mother drank alcohol during pregnancy or 3 months before801 (31.3)261 (33.7)0.34
Mother used drugs during pregnancy or 3 months before20 (0.8)18 (2.3)<0.0001

Outcome measures

Outcome measures were the prevalence of self-reported infant care practices positively or negatively related to SIDS risk expressed as binary responses. These items fell into four domains: infant sleeping environment, sharing a sleep surface, breast feeding and dummy use, and maternal life style. Outcomes within the infant sleep environment were: baby sleeps in a room alone, sleeps in an adult bed, positioned feet-to-foot in a cot, positioned prone for night-time sleep, positioned on side for night-time sleep, sleeps with a soft toy, pillow, duvet, in an infant sleeping bag, swaddled for sleep, and whether there was a thermometer in room. Within the sleep-sharing domain the outcomes included: ever bed-shares with mother, regularly bed-shares with mother, ever bed-shares with others, ever sofa-shares with mother. Breast feeding and dummy use outcomes were: baby ever breast-fed, breast-fed for 8 weeks or more, breast-fed at night, dummy used at night. Outcomes within the life style behaviours were: mother currently smokes, father currently smokes, mother drinks alcohol sometimes and more often, father drinks alcohol sometimes and more often.

Overall, the number of missing responses was low with the greatest number of missing values (47, 2.1%) regarding alcohol consumption of the father. All available data for each outcome were used.

Other measures

Ethnicity was self-defined by the mother in the baseline questionnaire. Analysis included four groups – White British (n = 968), Pakistani (n = 1212), Indian (n = 104), Bangladeshi (n = 63); 213 mothers were excluded as they reported another ethnic group or data on ethnicity were missing.

Other key confounders included in analyses were maternal age, education, parity and the language the baseline questionnaire was completed in. Data on confounders were missing for maternal education (47 cases, 2.0%), parity (62 cases, 2.6%) and language (16 cases, 0.7%); cases where any of these variables were missing were excluded from further analysis (n = 123). In total, 2224 mothers were included in the analysis comprising 992 White British, 1143 Pakistani, 100 Indian and 59 Bangladeshi mothers.


Univariable tests of association between the confounders (maternal age, education, parity and language) and ethnic group and each of the outcomes and ethnic group were carried out using chi-squared tests. Fisher's exact test was used to test the association between ethnic group and language at baseline because of empty cells in the cross-tabulation. Logistic regression models were used to examine the relationship between each of the outcomes and ethnic group. Univariable models were fitted only including ethnic group and each of the outcomes of interest. The number of missing values for each outcome varied by outcome and all available data for each outcome were used in the logistic regression models. Multivariable models were then fitted to examine the association between ethnicity and each of the outcomes of interest adjusting for key confounders. The confounders adjusted for were maternal age, education, parity, language and age of baby at completion of study. The logistic regression models were restricted to differences between the White British and Pakistani groups for clarity because of the small numbers in the Indian and Bangladeshi groups.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The sociodemographic characteristics of mothers completing the telephone questionnaires are shown in Table 2. Differences were found across the four groups for all variables (mother's education, age, parity, English competency). Indian and Bangladeshi mothers received the most education. A greater proportion of White British mothers were aged below 20 years. A greater proportion of Pakistani and Bangladeshi mother had higher parities, while a greater proportion of White British and Indian women were first-time mothers.

Table 2.  Maternal characteristics of study population by ethnic group
CharacteristicAllWhite British (n = 922)Pakistani (n = 1143)Indian (n = 100)Bangladeshi (n = 59)P-valuea
n (%)n (%)n (%)n (%)n (%) 
  • a

    Chi-squared P-value.

Mother's education     <0.0001
 None314 (14.1)91 (9.9)213 (18.6)5 (5.0)5 (8.5) 
 School715 (32.2)323 (35.0)358 (31.3)18 (18.0)16 (27.1) 
 Further and higher925 (41.6)354 (38.4)481 (42.1)61 (61.0)29 (49.2) 
 Other270 (12.1)154 (16.7)91 (8.0)16 (16.0)9 (15.3) 
Mother's age (years)     <0.0001
 <20116 (5.2)89 (9.7)26 (2.3)01 (1.7) 
 20–24572 (25.7)241 (26.1)301 (26.3)15 (15.0)15 (25.4) 
 25–29769 (34.6)287 (31.1)425 (37.2)37 (37.0)20 (33.9) 
 30–34486 (21.9)173 (18.8)263 (23.0)33 (33.0)17 (28.8) 
 35+281 (12.6)132 (14.3)128 (11.2)15 (15.0)6 (10.2) 
Parity     <0.0001
 0877 (39.4)451 (48.9)361 (31.6)47 (47.0)18 (30.5) 
 1656 (29.5)289 (31.3)315 (27.6)36 (36.0)16 (27.1) 
 2375 (16.9)107 (11.6)241 (21.1)10 (10.0)17 (28.8) 
 3+316 (14.2)75 (8.1)226 (19.8)7 (7.0)8 (13.6) 
Language at baseline     <0.0001
 English1820 (81.8)922 (100.0)753 (65.9)86 (86.0)59 (100.00) 
 Non-English404 (18.2)0390 (34.1)14 (14.0)0 

Infant sleeping position, location and environment

The prevalence of prone sleeping was very low (<3%) with no significant differences across the groups (Table 3); a greater proportion of South Asians (Pakistani 16%, Indian 11%, Bangladeshi 12%) reported that their infants were placed to sleep on their side compared with White British (3%). Almost a quarter of White British infants (24%) slept in a room alone compared with fewer than 2% of Pakistani infants.

Table 3.  Infant sleep position, location and environment by ethnic group
 AllWhite BritishPakistaniIndianBangladeshiP-valuea
n (%)n (%)n (%)n (%)n (%)
  • a

    Chi-squared P-value.

  • b

    This is based only on those who reported that baby sleeps in a cot.

  • Denominator is based on all completed questions and may vary for each outcome.

Baby placed prone for sleep58 (2.6)21 (2.3)33 (2.9)2 (2.0)2 (3.4)0.79
Baby placed on side for sleep226 (10.2)31 (3.4)177 (15.5)11 (11.0)7 (11.9)<0.0001
Baby positioned feet to footb723 (65.1)402 (79.3)261 (50.8)42 (71.2)18 (60.0)<0.0001
Baby sleeps in a room on own246 (11.1)221 (24.0)21 (1.8)4 (4.0)0 (0)<0.0001
Baby sleeps in an adult bed66 (3.0)4 (0.4)60 (5.3)2 (2.0)0 (0)<0.0001
Baby sleeps with a soft toy430 (19.3)220 (23.9)181 (15.8)25 (25.0)4 (6.8)<0.0001
Baby sleeps with pillow396 (17.8)27 (2.9)325 (28.4)18 (18.0)26 (44.1)<0.0001
Baby sleeps under duvet402 (18.1)78 (8.5)298 (26.0)14 (14.0)12 (20.3)<0.0001
Baby sleeps in an infant sleeping bag653 (29.4)459 (49.8)162 (14.2)22 (22.0)10 (17.0)<0.0001
Baby is swaddled for sleep323 (14.5)127 (13.8)175 (15.3)12 (12.0)9 (15.3)0.685
Thermometer in room1742 (78.8)758 (82.7)856 (75.3)76 (76.8)52 (88.1)<0.0001

Cots, cribs, baskets and prams were commonly reported as infant sleep surfaces; however, a greater proportion of Pakistani than White British infants were reported to sleep in an adult bed (Table 3). Where babies slept in cots, 79% of White British mothers employed ‘feet-to-foot’ in contrast to 51% of Pakistani, 71% of Indian and 60% of Bangladeshi mothers. More Indian (25%) and White British (24%) infants slept with a soft toy than Pakistani (16%) infants; however, the proportion of South Asian infants sleeping with a pillow was greater for all three groups (Table 3). A greater proportion of South Asian (25%) than White British infants (9%) slept under duvets, while 50% of the White British infants utilised sleeping bags. There was no inter-ethnic variation in swaddling for sleep; more White British and Bangladeshi mothers kept a thermometer in their infant's room.

Sharing a sleep surface

Significant differences were observed for ever bed-sharing, regular bed-sharing and sofa-sharing. Although few infants (66/2222; Table 3) were reported to ‘normally’ sleep in an adult bed, many more mothers (346/2221) responded affirmatively to the question ‘Does your baby sleep in your bed when you are asleep?’ (Table 4). Indian families reported the highest proportion (25%) for ever bed-share with 10% of Pakistani and 9% of Indian infants bed-sharing regularly. Very few infants bed-shared with someone other than the mother; however, four times as many White British mothers (16%) reported falling asleep with their infant on a sofa than Pakistani mothers (4%). White British mothers reported a greater proportion of ever sofa-share (16%) than ever bed-share (10%).

Table 4.  Sharing a sleep surface by ethnic group
 AllWhite BritishPakistaniIndianBangladeshiP-valuea
n (%)n (%)n (%)n (%)n (%)
  • a

    Chi-squared P-value.

  • b

    Only when baby has ever bed-shared. Indian and Bangladeshi excluded from this comparison because of small numbers reporting ever bed-sharing.

  • Denominator is based on all completed questions and may vary for each outcome.

Baby ever bed-shared with mother346 (15.6)95 (10.3)220 (19.3)25 (25.0)6 (10.2)<0.0001
Baby regularly bed-shares with mother158 (7.1)27 (2.9)119 (10.4)9 (9.0)3 (5.1)<0.001
Baby ever bed-shared with other26 (1.2)6 (0.7)16 (1.4)3 (3.0)1 (1.7)0.126
Baby ever sofa-shared201 (9.1)145 (15.8)49 (4.3)4 (4.0)3 (5.1)<0.0001
Baby has separate coversb173 (54.9)53 (55.8)120 (54.6)000.84
Baby sleeps in bedb148 (47.0)36 (37.9)112 (50.9)000.034
Baby sleeps on bedb125 (39.7)44 (46.3)81 (36.8)000.114
Baby sleeps in middle of bedb146 (47.3)49 (51.6)100 (45.5)000.32
Baby sleeps on edge of bedb123 (39.1)33 (34.7)90 (40.9)000.30
Baby swaddled in bedb27 (8.6)3 (3.2)24 (10.9)000.024

Breast feeding and dummy use

There was a significant difference by ethnicity in whether infants were ever breast-fed and breast-fed for 8 weeks or more (Table 5). A much greater proportion of White British infants (41%) were receiving solid foods when interviewed (median 107 days of age); more South Asian mothers were still breast feeding their infants during the night (35%). Dummy use at night was more prevalent among White British families.

Table 5.  Breast feeding practice and dummy use by ethnic group
 AllWhite BritishPakistaniIndianBangladeshiP-valuea
n (%)n (%)n (%)n (%)n (%)
  • a

    Chi-squared P-value.

  • Denominator is based on all completed questions and may vary for each outcome.

Breast feeding practice and dummy use      
 Baby ever breast-fed1565 (71.5)574 (63.3)847 (75.3)92 (92.0)52 (89.7)<0.0001
 Baby breast-fed for 8 weeks or more789 (36.0)236 (26.0)456 (40.5)67 (67.0)30 (51.7)<0.0001
 Baby currently receiving solid foods604 (27.3)372 (40.5)207 (18.2)19 (19.0)6 (10.5)<0.0001
 Baby breast-fed during the night599 (27.0)150 (16.3)369 (32.3)57 (57.6)23 (39.0)<0.0001
 Dummy used at all at night996 (44.9)543 (59.0)415 (36.4)21 (21.2)17 (28.8)<0.0001

Parental life style

Smoking was more prevalent among White British than South Asian mothers, but a greater proportion of Pakistani fathers smoked than any other group (Table 6). Maternal and paternal consumption of alcohol was rarely reported among Pakistani and Bangladeshi families. Indian fathers sometimes consumed alcohol, but consumption was rare for mothers. Proportionately more White British mothers and fathers reported drinking alcohol, with 18% of mothers and 26% of fathers doing so regularly.

Table 6.  Parental life style by ethnic group
 AllWhite BritishPakistaniIndianBangladeshiP-valuea
n (%)n (%)n (%)n (%)n (%)
  • a

    Chi-squared P-value.

  • Denominator is based on all completed questions and may vary for each outcome.

Mother currently smokes218 (9.8)194 (21.1)22 (1.9)2 (2.0)0 (0)<0.0001
Father currently smokes409 (18.4)153 (16.6)236 (20.7)12 (12.0)8 (13.6)0.024
Mother drinks alcohol sometimes or more often171 (7.7)164 (17.8)2 (0.2)4 (4.0)1 (1.7)<0.0001
Father drinks alcohol sometimes or more often247 (11.4)228 (25.5)2 (0.2)16 (16.3)1 (1.7)<0.001

Results of logistic regression (Table 7) controlling for maternal age, education, parity, language and baby's age at interview indicate that Pakistani infants were more likely than White British infants to sleep in an adult bed (OR = 8.48 [95% CI 2.92, 24.63]), be positioned on side for sleep (OR = 4.42 [2.85, 6.86]), have a pillow in their sleep environment (OR = 9.85 [6.39, 15.19]), sleep under a duvet (OR = 3.24 [2.39, 4.40]), be swaddled for sleep (OR = 1.49 [1.13, 1.97]), ever bed-share (OR = 2.13 [1.59, 2.86]), regularly bed-share (OR = 3.57 [2.23, 5.72]), be ever breast-fed (OR = 2.00 [1.58, 2.53]), be breast-fed for 8 or more weeks (OR = 1.65 [1.31, 2.07]) and be breast-fed during the night (OR = 2.08 [1.62, 2.67]). Pakistani infants were less likely to sleep in a room alone (OR = 0.05 [0.03, 0.09]), be positioned ‘feet-to-foot’ (OR = 0.36 [0.26, 0.50]), sleep with a soft toy (OR = 0.52 [0.40, 0.68]), sleep in a sleeping bag (OR = 0.20 [0.16, 0.26]), ever sofa-share (OR = 0.22 [0.15, 0.34]), receive solid foods by the time of the survey (OR = 0.22 [0.17, 0.30]), have a dummy at night (OR = 0.40 [0.33, 0.50]), be exposed to maternal smoking (OR = 0.07 [0.04, 0.12]) or alcohol consumption by either parent. No differences in adjusted odds ratios were observed between White British and Pakistani groups for being put down prone for sleep, having a room thermometer, paternal smoking or bed-sharing with an adult who was not the mother.

Table 7.  Logistic regression results for the association between infant care practice and ethnicity, odds ratios (ORs) and 95% confidence intervals [CI]
OR [95% CI]OR [95% CI]
  • a

    Model adjusted for maternal age, maternal education, parity, language questionnaire administered and baby age at completion of interview.

  • b

    This is based only on those who reported that baby sleeps in a cot.

  • c

    Only when baby has ever bed-shared.

  • d

    Models for alcohol were not fitted because of small numbers of drinkers in the Pakistani group.

  • Reference group is the White group.

Sleeping environment  
 Baby sleeps in a room on own0.06 [0.04, 0.09]0.05 [0.03, 0.09]
 Baby sleeps in an adult bed12.71 [4.60, 35.11]8.48 [2.92, 24.63]
 Baby positioned feet to footb0.27 [0.20, 0.35]0.36 [0.26, 0.50]
 Baby positioned prone for sleep1.28 [0.73, 2.22]1.04 [0.53, 2.01]
 Baby positioned on side for sleep5.28 [3.56, 7.81]4.42 [2.85, 6.86]
 Baby sleeps with a soft toy0.60 [0.48, 0.75]0.52 [0.40, 0.68]
 Baby sleeps with pillow13.17 [8.79, 19.72]9.85 [6.39, 15.19]
 Baby sleeps under duvet3.82 [2.92, 4.98]3.24 [2.39, 4.40]
 Baby sleeps in an infant sleeping bag0.17 [0.13, 0.21]0.20 [0.16, 0.26]
 Baby is swaddled for sleep1.13 [0.88, 1.45]1.49 [1.13, 1.97]
 Thermometer in room0.64 [0.51, 0.79]0.91 [0.70, 1.18]
 Baby ever bed-shared with mother2.07 [1.60, 2.68]2.13 [1.59, 2.86]
 Baby regularly bed-shares with mother3.85 [2.51, 5.90]3.57 [2.23, 5.72]
 Baby ever bed-shared with anyone else2.17 [0.84, 5.57]1.89 [0.58, 6.22]
 Baby ever sofa-shared0.24 [0.17, 0.33]0.22 [0.15, 0.34]
 Baby has separate coversc0.95 [0.59, 1.54]1.23 [0.71, 2.13]
 Baby sleeps in bedc1.70 [1.04, 2.78]1.61 [0.92, 2.80]
 Baby sleeps on bedc0.68 [0.41, 1.10]0.66 [0.38, 1.15]
 Baby sleeps in middle of bedc0.78 [0.48, 1.27]0.71 [0.41, 1.22]
 Baby sleeps on edge of bedc1.30 [0.79, 2.15]1.74 [0.98, 3.07]
 Baby swaddled in bedc3.76 [1.10, 12.79]5.65 [1.54, 20.81]
Breast feeding practice and dummy use  
 Baby ever breast-fed1.77 [1.46, 2.14]2.00 [1.58, 2.53]
 Baby breast-fed for 8 weeks or more1.94 [1.60, 2.34]1.65 [1.31, 2.07]
 Baby currently receiving solid foods0.33 [0.27, 0.40]0.22 [0.17, 0.30]
 Baby breast-fed during night2.44 [1.97, 3.03]2.08 [1.62, 2.67]
 Dummy used at all at night0.40 [0.33, 0.48]0.40 [0.33, 0.50]
Life styled  
 Mother currently smokes0.07 [0.05, 0.12]0.07 [0.04, 0.12]
 Father currently smokes1.31 [1.04, 1.64]1.07 [0.82, 1.40]


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This is by far the largest comparative study of ethnic differences in infant care in the UK to date, and the size of the sample is a major strength. The results demonstrate that night-time infant care practices for South Asian and White British infants differ in several key ways. South Asian infant care practices involve some (but not all) conventional SIDS risk reduction measures (keep baby in parental room, breast feed, avoid maternal smoking, alcohol consumption and sofa-sharing). White British families implement other aspects of conventional SIDS risk reduction (place baby feet-to-foot of cot, offer dummy, avoid bed-sharing, avoid pillows and duvets). In 1993 South Asian infants experienced less prone and solitary sleep than White British infants.21 In 2010 few infants were placed prone for sleep in either group, and lone infant sleep among White British infants had declined from 61% to 24%; however, the gap in SIDS rates between South Asian and White British infants has not reduced. The rate of SIDS in Bradford from 1998–2003 was 0.2/1000 for South Asian and 0.8/1000 for White British infants (Moya, personal communication). Data from the Local Register of Deaths on the Child Health Informatics System for Bradford District indicate that during the course of the BradICS project, four sudden unexpected deaths in infancy occurred in Bradford, all to White British infants; all were designated a registered cause of death; none were designated as SIDS and no deaths occurred within the study sample.

With the reduction in prone sleep position, parental smoking is now the most important modifiable risk factor for SIDS. Within the BradICS sample a large discrepancy was found between White British and Pakistani groups for maternal smoking, although the proportions of fathers who were smokers were similar in both groups. We cannot ascertain from the current study whether exposure to cigarette smoke differed between Pakistani and White British infants.

Evidence regarding SIDS risk, in association with duvet and dummy use is equivocal;23,24 the use of duvets and lack of dummies among South Asians are unlikely to substantially contribute to SIDS. The largest adjusted odds ratio for South Asian infant care concerned pillow use, some of which may be explained by the greater prevalence of bed-sharing; however, South Asian families also use ‘baby pillows’ in their infant's cot. These pillows are commonly thin and flat; some are moulded to maintain the infant's head in a face-up position. They cannot be likened to traditional western pillows that have been implicated as SIDS or suffocation risks in young infants25 and are likely to reflect a similar type of benign or beneficial infant pillow use reported for East Asian families.26

The most controversial current issue in SIDS risk reduction is sleep-sharing. Pakistani mothers were more likely to bed-share than White British mothers, and so these data do not support the contention that bed-sharing is a SIDS risk under all circumstances. Bed-sharing occurs in the same types of beds and bedding in both groups (Westman, personal communication). White British mothers were more likely to sofa-share for sleep with their infants. White British mothers were also more likely to consume alcohol, and a significant interaction between sharing a sleep location and alcohol consumption has recently been reported in a study of unsafe infant sleeping environments.27 Obtaining accurate data about bed-sharing practices can be tricky; previous research has shown that parents may tell the researcher where the baby is supposed to sleep, or where the baby starts the night, but not whether the baby spends any portion of their sleep time on the same surface as an adult, or for how long, unless explicitly asked.28

Understanding White British and South Asian infant care practices may contribute to further reduction of SIDS in White British families, and avoid transmission of infant care practices from majority culture to immigrant groups that may cause SIDS rates to increase. Limitations of this study relate to the retrospective self-reporting inherent in a cross-sectional questionnaire and the observational design. A potential limitation is the 75% response rate; however, as BradICS participants were recruited from the larger BiB study, we had access to prenatal baseline data permitting comparison of the majority of completers and non-completers. Non-completers were predominantly White British, young, not partnered and smokers; these mothers are characterised in other studies as having high SIDS risk in the UK.7 Thus, our estimates of the magnitude of differences in infant care practices between White British and South Asian families are likely to be conservative.

This study identifies maternal smoking, non-breast feeding, sofa-sharing and alcohol consumption as clear targets for SIDS risk reduction among White British families. The study also shows that South Asian families prioritise close proximity, breast feeding and maternal behaviours congruent with infant health and low SIDS risk as normal cultural practice. This study therefore calls into question the unproven value of applying SIDS risk reduction advice developed for White British families to those of other ethnicities. In groups where mothers breast feed, do not drink alcohol or smoke, sleep their infants supine and in close proximity to a parent at night, SIDS rates are so low we speculate that alterations in other infant care practices are poor targets for achieving further SIDS reduction.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This research was funded by a project grant from the Foundation for the Study of Infant Deaths. We are grateful to all the families who took part in this study, to the midwives for their help in recruiting them, the paediatricians and health visitors and to the Born in Bradford team, which included interviewers, data managers, laboratory staff, clerical workers, research scientists, volunteers and managers.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    Fleming PJ, Blair PS, Bacon C, Bensley D, Smith I, Taylor E, et al. Environment of infants during sleep and risk of the sudden infant death syndrome: results of 1993–5 case-control study for confidential inquiry into stillbirths and deaths in infancy. British Medical Journal 1996; 313:191194.
  • 2
    Gunn AJ, Gunn TR, Mitchell EA. Is changing the sleep environment enough? Current recommendations for SIDS. Sleep Medicine Reviews 2000; 4:453469.
  • 3
    Hauck FR, Tanabe KO. SIDS. Clinical Evidence 2009; 6:315328.
  • 4
    Tirosh E, Becker T, Mansour Y, Cohen A, Jaffe M. Sleep position, bedding and heating practices in high- and low-risk ethnic groups for unexpected death in infancy (UDI). European Journal of Epidemiology 2000; 16:281286.
  • 5
    American Academy of Pediatrics Task Force on Infant Sleep Position and Sudden Infant Death Syndrome. Changing concepts on sudden infant death syndrome: implications for infant sleeping environment and sleep position. Pediatrics 2000; 105:650656.
  • 6
    Alm B, Lagercrantz H, Wennergren G. Stop SIDS – sleeping solitary supine, sucking soother, stopping smoking substitutes. Acta Paediatrica 2006; 95:260262.
  • 7
    Blair PS, Sidebotham P, Berry PJ, Evans M, Fleming PJ. Major epidemiological changes in sudden infant death syndrome: a 20-year population-based study in the UK. Lancet 2006; 367:314319.
  • 8
    Willinger M, Hoffman HJ, Wu KT, Hou JR, Kessler RC, Ward SL, et al. Factors associated with the transition to non-prone sleep positions of Infants in the United States: the National Infant Sleep Position Study. JAMA 1998; 280:329335.
  • 9
    Beal S. Sudden Infant Death Syndrome in South Australia 1968–97. Part I: changes over time. Journal of Paediatrics and Child Health 2000; 36:540547.
  • 10
    Hill SAR, Hjelmeland B, Johannessen NM. Changes in parental risk behaviour after an information campaign against Sudden Infant Death Syndrome (SIDS) in Norway. Acta Paediatrica 2004; 93:250254.
  • 11
    Tipene-Leach DC. Bedsharing and smoking. New Zealand Medical Journal 1995; 108:300301.
  • 12
    Tuohy PG, Smale P, Clements M. Ethnic differences in parent/infant co-sleeping practices in New Zealand. New Zealand Medical Journal 1998; 111:364366.
  • 13
    Sleuwen BEV, Hoir MPL, Engleberts AC, Schulpen TWJ. Infant care practices related to cot death in Turkish and Moroccan families in the Netherlands. Archives of Disease in Childhood 2005; 88:784788.
  • 14
    Lahr MB, Rosenberg KD, Lapidus JA. Maternal-infant bedsharing: risk factors for bedsharing in a population-based survey of new mothers and implications for SIDS risk reduction. Maternal and Child Health Journal 2007; 11:277286.
  • 15
    Balarajan R, Raleigh VS, Botting B. Sudden Infant Death Syndrome and post-neonatal mortality in immigrants in England and Wales. British Medical Journal 1989; 298:716720.
  • 16
    Li DK, Daling JR. Maternal smoking, low birth weight, and ethnicity in relation to sudden infant death syndrome. American Journal of Epidemiology 1991; 134:958964.
  • 17
    Eades SJ, Read AW. Infant care practices in a metropolitan aboriginal population Bibbulung Gnarneep Team. Journal of Paediatrics and Child Health 1999; 35:541544.
  • 18
    Abel S, Park J, Tipene-Leach D, Finau S, Lennan M. Infant Care Practices in New Zealand: a cross-cultural qualitative study. Social Science & Medicine 2001; 53:11351148.
  • 19
    Office for National Statistics. Trends in cot death. Health Statistics Quarterly 2000; 5:1725.
  • 20
    Gantley M, Davies D, Murcott A. Sudden Infant Death Syndrome: links with infant care practices. British Medical Journal 1993; 306:1620.
  • 21
    Farooqi S, Perry IJ, Beevers DG. Ethnic differences in infant-rearing practices and their possible relationship to the incidence of Sudden Infant Death Syndrome (SIDS). Paediatric and Perinatal Epidemiology 1993; 7:245252.
  • 22
    Born in Bradford Collaborative Group. Born in Bradford, a cohort study of babies born in Bradford, and their parents: protocol for the recruitment phase. BMC Public Health 2008; 8:327.
  • 23
    Mitchell EA, Williams SM, Taylor BJ. Use of duvets and the risk of sudden infant death syndrome. Archives of Disease in Childhood 1999; 81:117119.
  • 24
    Mitchell EA, Blair PS, L'Hoir M. Should pacifiers be recommended to prevent Sudden Infant Death Syndrome? Pediatrics 2006; 117:17551758.
  • 25
    Hauck FR, Herman SM, Donovan M, Iyasu S, CM M, Donoghue E, et al. Sleep environment and the risk of Sudden Infant Death Syndrome in an urban population: the Chicago infant mortality study. Pediatrics 2003; 111:12071214.
  • 26
    Nelson EA, Taylor BJ. International child care practices study: infant sleeping environment. Early Human Development 2001; 62:4355.
  • 27
    Blair PS, Sidebotham P, Evason-Coombe C, Edmonds M, Heckstall-Smith EMA, Fleming P. Hazardous cosleeping environments and risk factors amenable to change: case-control study of SIDS in south west England. British Medical Journal 2009; 339:b3666 (see
  • 28
    Ball HL, Hooker E, Kelly PJ. Where will the baby sleep? Attitudes and practices of new and experienced parents regarding cosleeping with their new-born infants. American Anthropologist 1999; 101:143151.
  • *

    Depending on whether ‘unexplained’ deaths are included as SIDS.