To determine the relationship between the advice mothers receive about infant sleep position and the actual position they place their infants in to sleep and to understand modifiers of that relationship, especially beliefs about infant comfort and safety.
Cross-sectional, face-to-face interviews.
Women, Infants, and Children centers in the United States from 2006 to 2008.
A total of 2299 predominantly African American mothers of infants younger than 8 months.
Advice received and beliefs about infant sleep position.
Usually supine infant sleep position.
Advice for exclusively supine infant sleep position from family (OR, 1.6; 95% CI, 1.17-2.17), doctors (OR, 2.28; 95% CI, 1.77-2.93), nurses (OR, 1.46; 95% CI, 1.15-1.84), or the media (OR, 1.54; 95% CI, 1.22-1.95) was associated with usually placing an infant supine to sleep. Additional sources of advice for exclusively supine position significantly increase the odds that an infant will be placed supine. Mothers who believe an infant is comfortable supine are more likely to place their infants on their backs to sleep (OR, 4.05; 95% CI, 2.51-6.53). Mothers who believe an infant will choke on its back are less likely to place their infants supine (OR, 0.36; 95% CI, 0.24-0.54).
Among predominantly African American mothers, increasing advice for exclusively supine sleep and addressing concerns about infant comfort and choking remain critical to getting more infants on their back to sleep.
African American infants are more likely than other infants in the United States to die in infancy.1,2 Sudden Infant Death Syndrome (SIDS) is the most common cause of infant mortality in the postneonatal period and accounts for a significant proportion of the disparity in deaths between African American and white infants.3,4 Infants who sleep in the supine position are less likely to die of SIDS than infants who sleep in the prone or side positions.5- 12 In 1992, the American Academy of Pediatrics first issued its recommendation that all infants be placed in the supine or lateral position for sleep (now, the recommendation is for exclusively supine sleep position). In 1994, the United States Public Health Service and collaborators initiated the nationwide Back to Sleep campaign to promote supine positioning of all infants when they are put down to sleep. The Back to Sleep campaign has been very successful, raising the rate of supine positioning from 13% in 1992 to 76% in 2006.13- 15 During the same interval, the incidence of SIDS was cut nearly in half.2,16
Despite the success of the Back to Sleep campaign, 2 important problems remain regarding infant sleep position. First, while supine infant sleep has increased in all racial/ethnic groups, African American infants are much less likely than white infants to be placed supine to sleep.15 Likewise, the incidence of SIDS among African Americans has not decreased to the lower rate observed among groups who more commonly place infants supine.4 Second, the national rate of supine sleep has reached a plateau, with nearly one-fourth of all infants (and about half of African American infants) in the United States still sleeping in a nonsupine position.15 The reasons for the racial gap and plateau in supine infant sleep are not well understood. Earlier study of the general population of the United States has shown that advice mothers receive about infant sleep position and their beliefs are important predictors of sleep position.17- 20
The goal of this study was to determine the relationship between advice and actual infant sleep position and to understand the effect of potential modifiers of that relationship, such as beliefs about infant comfort and choking, among a group of low-income, mostly African American mothers who are at high risk of placing their infants nonsupine to sleep. We hypothesized that, in this subset of the population, more advice for exclusively supine sleep would be associated with a higher rate of usually supine infant sleep positioning. We further hypothesized that mothers' beliefs about the position in which an infant is most comfortable or is most likely to choke are important mediators of the relationship between advice and infant sleep position.
In 2006, 2007, and 2008 we conducted face-to-face interviews with mothers of infants younger than 8 months at Women, Infants, and Children (WIC) Supplemental Nutrition Program centers in 6 cities: Birmingham, Alabama; New Haven, Connecticut; Detroit, Michigan; Clarksdale, Mississippi; Jackson, Mississippi; and Dallas, Texas. The WIC centers are a federal program administered by states that provide “supplemental foods, health care referrals, and nutrition education for low-income pregnant, breastfeeding, and non-breastfeeding postpartum women, and to infants and children up to age five who are found to be at nutritional risk” (http://www.fns.usda.gov/wic/). The WIC centers in our study were chosen based on geographic location, availability, and willingness to participate. Most importantly, the WIC centers had to serve at least 50% African American clients. Mothers were eligible to participate in the study if they received benefits from WIC, had an infant younger than 8 months, and spoke English. We confirmed with the WIC center directors that no formal SIDS educational activities occurred during or before our interviews.21
Interview development and interviewer training is described in detail elsewhere.21 Briefly, the interviews were conducted by 2 research assistants local to each WIC center who were trained extensively by our study investigators. Key interview questions are listed in Table 1. After the interview, all participants were given information about safe infant sleep practices using current Back to Sleep guidelines. Institutional review board approval was obtained for all locations. Each participant gave informed consent and received a $10 gift certificate at the end of the interview.
The survey was designed for this study. Our outcome of interest was usually supine infant sleep position for all sleep periods (daytime and nighttime) during the past 2 weeks. Our predictor variables were the advice mothers reported receiving about infant sleep position and maternal perceptions about infant comfort and choking. These variables were chosen based on prior study identifying key concerns of caregivers of young infants regarding sleep position.22,23 We asked mothers from whom they received advice (friends, family, the media, nurses, or doctors) and what advice they received from each source (advice for exclusively supine sleep position, advice for not exclusively supine sleep position, or no advice about sleep position).
To assess the cumulative effect of advice we created a variable, advice score, to determine whether each mother got more advice supporting exclusively supine or not exclusively supine sleep from all the sources combined. We counted exclusively supine sleep advice as +1, not exclusively supine sleep advice as −1, and no advice as 0. The range of possible advice scores was −5 to +5 because there were 5 sources of advice (friends, family, the media, nurses, or doctors).
Covariates that might influence choice of infant sleep position were chosen based on previously identified factors associated with sleep position: maternal variables (race/ethnicity, age, education, and parity), infant variables (sex, age, and prematurity), and interview variables (site and year). Maternal parity and infant prematurity were not associated with usually supine sleep position in bivariate analysis and were not included in multivariate models. Although not significantly associated with our outcome in bivariate analysis, we retained basic demographic variables (infant sex and maternal age) in our multivariate model to allow comparison with other studies.
Descriptive statistics were calculated including frequencies and χ2 tests for differences in proportions. Multiple logistic regression was used to examine associations between variables of interest and usually supine infant sleep position. To evaluate the relative effect of individual sources of advice (friends, family, the media, nurses, or doctors) vs advice score on usually supine sleep, these 2 types of advice variable were entered into separate models. The Cochran-Armitage test of trend was used to evaluate the relationship between amount of positive advice (exclusively supine placement) and reported usual supine placement. All analyses were conducted with SAS 9.1 software (SAS Institute, Cary, North Carolina); α for all tests was set at .05.
We screened 5388 mothers to determine eligibility for this study. A total of 2521 mothers were eligible; of these, 2353 (93%) were interviewed and 2299 (91% of eligible subjects) had complete data for all variables of interest and were included in our analysis. Mothers excluded from the analysis were similar in all demographics and covariates to those included. Of the mothers who participated, 1689 (74%) were African American (Table 2).
A total of 1408 (61%) mothers reported usually placing their infants supine to sleep; 489 (21%), lateral; 390 (17%), prone, and 12 (0.5%), another position. Bivariate associations between usually supine sleep position and demographic variables are shown as unadjusted odds ratios in Table 2.
Figure 1 shows the sources (who) and nature (what) of advice that mothers received. Doctors were reported most commonly to offer exclusively supine advice (56% of the time). Most mothers received no advice from family, friends, or the media about infant sleep position. Most of those who reported advice from family and friends said that the advice was for not exclusively supine sleep. Most of those who reported receiving advice from the media said that the advice was for the exclusively supine position.
Percentage of mothers who received exclusively supine, not exclusively supine, or no advice for infant sleep position from 5 sources.
Figure 2 shows that the percentage of mothers who place their infants supine increased with higher advice score (number of sources of advice for exclusively supine sleep minus the number of sources of advice for not exclusively supine sleep). For example, of the 559 mothers who had a negative advice score, 36% (202 mothers) placed their infants supine, whereas of the 439 mothers with an extremely positive advice score, 85% (373 mothers) usually placed their infant supine to sleep (P < .001).
Percentage of mothers who place infants usually supine for sleep increases with higher advice score (number of sources of advice for supine sleep minus the number of sources of advice for nonsupine sleep).
Most mothers had well-defined beliefs about the position in which their infants were most comfortable or likely to choke. Most mothers (1443; 63%) believed that their infants were most comfortable in a nonsupine position (usually prone). Most mothers 1280 (56%) thought that their infants were most likely to choke when supine. Bivariate associations between beliefs and usually supine sleep are shown as unadjusted odds ratios in Table 2.
To assess the factors associated with usually supine sleep position, we first performed a multiple logistic regression analysis in which the explanatory variables included demographics of mothers and infants as well as survey variables (site and year). Hispanic mothers, compared with not Hispanic African American mothers, were significantly more likely to place their infants supine to sleep. Infants older than 1 month were less likely to be placed supine, as were infants from Dallas, Jackson, and Clarksdale, compared with New Haven (regression not shown).
To evaluate how maternal beliefs about infant comfort and choking influence frequency of supine sleep, we performed a second multiple logistic regression analysis in which we added 2 variables: maternal beliefs about (1) the position in which an infant is most comfortable and (2) the position in which an infant is most likely to choke. The results of this second analysis are in Table 2 (beliefs). Mothers who believed that infants are most comfortable supine were significantly more likely to usually place their infants supine. Those who believed that infants are most likely to choke supine were significantly less likely to usually place their infants supine. In this model, Hispanic mothers were no longer more likely than other mothers to place their infants supine.
To measure the relationship between advice that mothers reported receiving and usually supine infant sleep, we performed 2 more multiple logistic regression analyses. First, we assessed the effect of each of the 5 sources of advice (family, friends, the media, doctors, and nurses) according to the advice they gave (not exclusively supine, no advice, or exclusively supine). Second, we removed these 5 variables and added the advice score variable. A negative advice score indicates that a mother received more advice not supporting supine rather than supporting exclusively supine sleep from the 5 sources; a positive score indicates more advice for exclusively supine sleep. The results of these 2 analyses are found in Table 2 (advice sources and advice score).
Advice for exclusively supine sleep position from most sources (family, the media, doctors, and nurses) was associated with higher odds of usually supine sleep position after adjustment for covariates. Advice for not exclusively supine position from a nurse was associated with lower odds of usually supine sleep position.
The advice score was significantly associated with supine infant sleep. Compared with the reference group (mothers who received an advice score of zero, indicating that they received exactly the same amount of advice for exclusively supine and not exclusively supine sleep), mothers who received more advice for exclusively supine sleep (positive advice score) were more likely to place their infants supine to sleep. On the other hand, mothers who received more advice overall for not exclusively supine sleep (negative advice score) were significantly more likely to place their infants nonsupine.
Finally, we performed secondary interaction analyses to determine if advice and beliefs varied depending on the age of the infant. There was no significant interaction between age and comfort, choking risk, or advice from friends or the media. There were significant interactions between infant age and family (P = .05), doctor (P = .04), and nurse advice (P = .03), indicating that the associations between these variables and sleep position may vary with age. All of the significant interactions suggested that the associations between advice and sleep position are strongest for younger children.
Throughout our analyses, mothers' beliefs about infant comfort and choking remained significant. Race/ethnicity, a significant predictor of usually supine infant sleep in bivariate analysis, became nonsignificant once we included mothers' beliefs and remained nonsignificant in all subsequent models. The Clarksdale interview site was no longer statistically significant in either advice model.
Since the Back to Sleep campaign began in 1994, many more infants are sleeping on their backs. In the past few years, though, the percentage of infants placed supine to sleep has reached a plateau, with African American infants at highest risk of nonsupine sleep and SIDS.3,15 We tried to understand what factors most strongly influence the behavior of the residual group of mothers who still place their infants nonsupine for sleep. We have shown in this population of 2299 low-income, mostly African American mothers that the amount and source of advice supporting exclusively supine sleep position strongly predict actual choice of infant sleep position. Furthermore, mothers' beliefs about infant comfort and choking remain important predictors of usual sleep position, even when accounting for advice.
When accounting only for demographics, it appears that Hispanic mothers are more likely to place their infants supine to sleep. This is consistent with prior studies.17,21 When accounting for mothers' beliefs about infant comfort and choking, the ethnic difference is no longer statistically significant. In other words, the ethnic difference in supine positioning may be owing to the fact that Hispanic mothers are more likely than others to think that infants are comfortable, or not at risk of choking, on their backs.
Importantly, we did not find a difference in the rate of supine sleep between white and African American infants, as has been seen in other studies. We can think of 3 reasons that we did not find a difference. First, the number of white mothers in our sample may have been too small to detect the difference. Second, practices reported in face-to-face interviews about infant sleep position may differ from practices reported in telephone interviews. Finally, though previously unreported, it is possible that low-income white mothers are less likely than those in the general population to place their infants supine for sleep.
In contrast to earlier studies, we found that mothers of older infants were less likely to place their infants supine.17,21 Our secondary analyses revealed that there is an association between infant age and advice from family, doctors, and nurses. The associations are generally stronger for younger infants. These results suggest that mothers' choice of infant sleep position is most strongly influenced by family, doctor, and nurse advice during the first month of the infant's life. Alternately, there may be an attenuation of the effect of advice on mothers' choice of sleep position for older infants. These results could have implications for planning interventions to improve supine positioning among older infants.
As noted by others,15,17 we found regional differences; those living in southern states were more likely to place their infants in the nonsupine position for sleep. Any generalization of our data to a nationwide population should account for these variations.
Mothers' beliefs about infant comfort and choking are strongly associated with sleep position, regardless of advice. These data suggest that health care providers may be more effective in promoting supine infant sleep position if they understand beliefs about infant comfort and choking when giving advice about sleep position. On the public health level, campaigns that address beliefs about infant comfort and choking may be most effective. There are some data from the literature about comfort measures that also improve safety. For example, using a pacifier has been shown to decrease SIDS.24- 26 More research is needed, though, about explanations and safe alternatives to offer parents and caregivers who are concerned about infant comfort and choking in the supine position.
Advice for exclusively supine infant sleep position from family, the media, doctors, or nurses significantly increases the likelihood that an infant will be placed on their back to sleep. Disappointingly, though, only 56% of doctors and 44% of nurses were reported to give advice supporting exclusively supine infant sleep position. Although many doctors and nurses were reported not to give any advice about sleep position, an alarming proportion (about 1 in 7) were reported to give incorrect advice. These data are consistent with a recent study by Moon et al27 in which only 69% of doctors who care for infants self reported advising exclusively supine sleep. This incorrect advice may have an important effect on infant sleep position; we found that advice not exclusively supporting supine sleep position from a nurse is associated with lower odds that an infant will be placed on its back to sleep. This is consistent with prior data indicating that changing infant positioning practices by nurses in well-baby nurseries influences how mothers place their infants to sleep.19
We were particularly interested to find that although the media were a source of advice for only 43% of the mothers in our study, a very high percentage of mothers who reported receiving advice from the media said the advice was for exclusively supine sleep position (77%). This is comparable with the percentage who report exclusively supine advice from doctors and nurses. Thus, the media appear to be a relatively untapped, likely beneficial resource for distributing the Back to Sleep message. The most effective content, distribution, and location of media messages should be further explored, with a focus on addressing perceptions of infant comfort and choking.
Importantly, we found a dose-dependent increase in supine sleep for additional sources of advice exclusively supporting the supine position. Mothers who receive more advice for exclusively supine than not exclusively supine sleep position are significantly more likely to place their infants on their backs to sleep.
Although prior study has shown that advice and beliefs are crucial predictors of infant sleep position,21 this study is important because of the more recent data, the larger sample size, and the in-depth exploration of advice and beliefs. Our findings confirm and reveal factors that may be important in improving supine infant sleep positioning among the residual group of mothers and, more specifically, those who are African American who still place their infants nonsupine to sleep. These factors include increasing exposure to many different sources of advice for supine sleep, combating incorrect advice, addressing beliefs about comfort and choking, and being sensitive to uptake of Back to Sleep advice as infants age.
Our findings may not be applicable to all mothers and infants because our sample of mostly African American mothers at WIC centers was designed to describe infant sleep practices, advice, and beliefs in a group of mothers at high risk of placing their infants in a nonsupine position for sleep. Our data may be biased if mothers reported what they thought they ought to say rather than actual behaviors, beliefs, and advice. We are reassured by the percentage of mothers in our sample who reported usually supine sleep (61%), which is consistent with the percentage from previous studies of African American mothers.15,21
In this study of predominantly African American mothers, we have shown that advice supporting exclusively supine position from many different sources increases the chance that a mother will place her infant supine for sleep and that there is a dose-dependent effect of advice for exclusively supine sleep. We have also demonstrated that mothers' beliefs about the position in which an infant is most comfortable and likely to choke influence infant sleep position, no matter how much advice she receives or from whom. When accounting for mothers' beliefs, we do not see a racial or ethnic difference in propensity for supine infant sleep. Advice about infant sleep position may be less potent for mothers of older infants. Increasing advice for exclusively supine sleep, especially through the media, and addressing mothers' concerns about infant comfort and choking are critical to getting more infants on their back to sleep.
Correspondence: Isabelle Von Kohorn, MD, c/o Eve R. Colson, MD, 20 York St, WP 1114, New Haven, CT 06510 (firstname.lastname@example.org).
Accepted for Publication: December 2, 2009.
Author Contributions:Study concept and design: Von Kohorn, Corwin, Heeren, Lister, and Colson. Acquisition of data: Colson. Analysis and interpretation of data: Von Kohorn, Corwin, Rybin, Heeren, Lister, and Colson. Drafting of the manuscript: Von Kohorn, Lister, and Colson. Critical revision of the manuscript for important intellectual content: Von Kohorn, Corwin, Rybin, Heeren, Lister, and Colson. Statistical analysis: Von Kohorn and Rybin. Obtained funding: Corwin, Heeren, and Lister. Administrative, technical, and material support: Von Kohorn and Corwin. Study supervision: Corwin, Lister, and Colson.
Financial Disclosure: None reported.
Funding/Support: This study was supported by National Institute of Child Health and Human Development grant U10 HD029067-09A1 and training grant T32HD07094.
Additional Contributions: The authors would like to thank Marian Willinger, PhD, of National Institute of Child Health and Human Development for her thoughtful comments during the preparation of the manuscript.
Von Kohorn I, Corwin MJ, Rybin DV, Heeren TC, Lister G, Colson ER. Influence of Prior Advice and Beliefs of Mothers on Infant Sleep Position. Arch Pediatr Adolesc Med. 2010;164(4):363-369. doi:10.1001/archpediatrics.2010.26