Prevalence and Predictors of the Prone Sleep Position Among Inner-city Infants FREE

RECENT STUDIES in the United States and abroad have shown that the prone sleep position is associated with an increased risk of sudden infant death syndrome (SIDS).^{1 - 4} In countries where the prevalence of the prone sleep position has been reduced to less than 10%, the rates of SIDS have declined by more than 50%.^{5 - 9} Although the mechanism linking prone sleep position to SIDS has not been fully elucidated, research suggests that infants placed in the prone position may rebreathe expired air, leading to increases in blood carbon dioxide levels, hypoxia, and subsequent asphyxia or that the prone sleep position may cause infants to overheat through decreased dissipation of heat.^{10 - 15}

In 1992, the American Academy of Pediatrics (AAP) released a statement recommending that full-term, healthy infants be placed down for sleep in either the back or side position.¹⁶ In June 1994, the US Public Health Service launched the "Back to Sleep" educational campaign in collaboration with the AAP, the SIDS Alliance, and the Association of SIDS Program Professionals.¹⁷ Currently, the AAP recommends the supine sleep position (rather than using a combination of either the supine or side positions) as the preferred infant sleep position.¹⁸ Since release of the initial AAP recommendations, national telephone surveys have documented a decline in the prevalence of the prone sleep position from 70% in 1992 to 24% in 1996.¹⁸

Common reasons that mothers give for choosing a particular sleep position are that the infant sleeps better or is more comfortable in the chosen position or that the mother used the position with previous children.^{19 - 22} In this report, we describe the findings of the first prospective study in the United States to examine risk factors for infant sleep position in a low-income, inner-city population. The primary objectives were (1) to describe infant sleep position practices in a cohort of infants born to predominantly low-income, inner-city mothers, a group known to be at increased risk of SIDS, and (2) to identify predictors of the prone sleep position, particularly risk factors that may be amenable to interventions.

The primary focus of this cohort study was to identify determinants, as measured at birth, of immunization and other preventive health care practices, including infant sleep position. The study was approved by the institutional review boards of all participating study sites and collaborating institutions.

Mothers of singleton births were recruited shortly after delivery from 3 hospitals in the District of Columbia. Two of the selected hospitals had a high prevalence of low-income patients and enrollment at the third site was limited to nonprivate patients (patients cared for by staff physicians) to capture a sample of low-income, inner-city patients. Eligibility criteria included residency in the District of Columbia and the ability to speak and understand either English or Spanish. Infants were excluded if they were not born alive, died during the hospital stay, had birth weights less than 1800 g, had major congenital malformations identified within the first 4 hours of life, or were being placed for adoption. Infants whose mothers were incarcerated at the time of delivery were also excluded. Mothers were enrolled between August 1995, and September 1996. Recruitment days were chosen systematically to ensure a representative sample of weekend and midweek sample days. Sampling was for a 24-hour period on the selected days. The study included 1 interview with the mother shortly after delivery (the baseline interview) and a second interview when the infant was between the ages of 3 and 7 months (the follow-up interview). Mothers approached for enrollment in the study were given a layette set. Additionally, those who completed the follow-up interview were given $25.

Baseline interviews were conducted by trained research assistants and included a face-to-face interview and a short self-administered questionnaire. The face-to-face interview included questions about sociodemographic and psychosocial factors, such as social support and health behaviors. The self-administered questionnaire contained potentially sensitive questions about depression, drug and alcohol use, violence in the home, and whether the pregnancy was wanted. Information about infant birth weight, gestational age, infant sex, type of delivery, parity, and maternal smoking during the pregnancy was abstracted from delivery and newborn hospital records. Of the 452 baseline interviews, 417 (92%) were conducted in the hospital during the postpartum stay. Four hundred five (90%) were completed within 48 hours of delivery and, with the exception of 1 interview that was conducted at 18 days postpartum, all interviews were completed within 2 weeks of delivery.

Mothers were interviewed again between November 1995 and March 1997 when infants were between 3 and 7 months old. Outcomes of interest included infant sleep position, immunizations, and other health care practices. Although most interviews were face-to-face, 28 (7%) were conducted by telephone at the mother's request or due to scheduling difficulties. Interviewers did not have access to the baseline interviews at the time that they conducted the follow-up interviews.

"Intended sleep position" was defined as the sleep position in which mothers planned to place their infants for sleep, as ascertained shortly after delivery. "Usual sleep position" referred to the position in which the mother reported usually placing her infant for sleep when the mother was interviewed at follow-up. "Last night's sleep position" was the position in which the infant was placed on the night prior to the follow-up interview. "Prone" was defined as sleeping on the stomach, and "nonprone" denoted either the side or back position. In bivariate and multivariate analyses, if prone was 1 of the reported sleep positions, the response was coded as prone. We reasoned that infants who were being placed for sleep in the prone position, all or some of the time, were not in compliance with AAP recommendations and were at increased risk for SIDS.

At baseline mothers were asked about the intended infant sleep position and their reasons for choosing that position. The influence of health providers was assessed by asking respondents to recall their observations of hospital nurses putting their infants down for sleep (hereafter referred to as "hospital sleep position") and by asking if a physician or nurse had discussed sleep position with the respondent since the time she had become pregnant.

At follow-up, mothers were asked about the infant's usual sleep position and about last night's sleep position. Respondents were also asked why they chose the reported usual sleep position. To assess the influence of health providers, mothers were asked if, at the time the child was born, anyone from the hospital had discussed sleep position with them. Additionally, for each reported well-child care visit, mothers were asked if sleep position had been discussed. If more than 1 position was cited as an intended or usual sleep position, multiple responses were coded. Mothers could also provide multiple reasons for choosing the intended or usual sleep position.

Bivariate associations between independent variables and outcome variables were assessed using χ² or Fisher exact tests for categorical variables and the Student t test for continuous variables. Factors significantly associated with the outcome at P<.10 in bivariate analyses were introduced into multivariate logistic regression models to determine the independent predictive effects of these factors. A backward elimination algorithm, with an exit P value of <.05, was used to select variables having an independent contribution to the model. All statistical tests were interpreted in a 2-tailed fashion to estimate P values. The sample size required for the study was calculated to address the immunization outcome, and, thus, the assumptions and differences are not relevant to the current study.

Of the 1802 infants born on the sampled days, 518 were eligible for enrollment in the study. The most common reasons for ineligibility were residency outside the District of Columbia (n=704), being a private patient at site C (n=373), and birth weight less than 1800 g (n=91). Four hundred fifty-two eligible mothers (87%) gave informed consent and were successfully enrolled. Based on information abstracted from medical records, eligible mother-infant dyads enrolled in the study did not differ significantly from eligible dyads not enrolled with respect to race, age, marital status, type of insurance, mother's employment status, type of delivery, birth weight, gestational age at delivery, or infant sex. Of the 452 mothers interviewed at baseline, 395 (87%) were located and interviewed at 3 to 7 months postpartum. Mothers lost to follow-up were not significantly different from those who completed the follow-up interview with respect to any of the above baseline characteristics. All but 1 of the 395 mothers who completed the follow-up interview reported that they had lived with their infant most of the time since leaving the hospital. Data from these 394 mother-infant dyads are reported below.

Mothers were predominantly black (334 [85%]), unmarried (285 [72%]), and non-Hispanic (325 [82%]) (Table 1). Two hundred fifty-nine (66%) of the mothers had household incomes below the poverty level²³ (mean household income, $12600) and 171 (43%) had not completed high school. The mean maternal age was 25 years (range, 13-43 years), and 85 (22%) were teenage mothers. The mean age of the infant at the time of the follow-up interview was 129 days (range, 90-228 days).

When interviewed shortly after delivery, 137 mothers (35%) reported that they intended to place the infant down for sleep in a prone position, 243 (62%) said they would place their infants in a nonprone position, and 14 (4%) said they did not know or had not thought about which sleep position they would use (Table 2). At 3 to 7 months postpartum, 143 mothers (36%) usually placed their infants in the prone position, 30 (8%) used the prone position as one of the usual positions, and 220 (56%) usually placed their infants in the side or supine position. On the night prior to the follow-up interview, 157 mothers (40%) had placed their infants in the prone position.

Of the 137 mothers who indicated at baseline that they intended to place the infant in the prone sleep position, 77 (56%) reported that their infants usually were placed for sleep in the prone position, and 72 mothers (53%) reported that the infant slept in a prone position on the night prior to the follow-up interview. Similarly, of the 243 mothers who intended to place the infant in a nonprone position, 154 (63%) reported that the infant usually was placed down for sleep in the nonprone position, and 158 mothers (65%) reported that their infants had slept nonprone on the night prior to the follow-up interview.

At baseline, the most common reason cited for the intended sleep position, both prone and nonprone, was to prevent choking (Table 3). However, at follow-up 100 (58%) of the mothers who chose the prone position cited infant comfort as 1 of the reasons. Common reasons cited for using the nonprone sleep position included the prevention of choking (73 [33%]), infant comfort (52 [24%]), and the prevention of SIDS (37 [17%]). Thirty-nine (18%) of the mothers who placed their infants in nonprone positions said they did so at the advice of medical personnel.

There was high agreement (85%) between usual sleep position and last night's sleep position. We assessed predictors of last night's sleep position because we regarded this as a less subjective outcome. In bivariate analyses, factors measured at baseline that were significantly (P<.05) associated with the use of the prone sleep position included black race, maternal age younger than 20 years, single marital status, poverty, less than 12 years of education, vaginal delivery, the presence of infant's grandmother in the home, prone hospital sleep position, and intent to place the infant prone (Table 4, Part A, and Table 4, Part B). Although 69% of mothers indicated that they had received advice on sleep position while in the hospital and 64% received advice at a well-child care visit prior to 90 days postpartum, neither of these variables was significantly associated with infant sleep position. Other variables of interest not associated with infant sleep position included age of the infant at follow-up, the type of bed in which the infant slept, and whether the infant shared a bed with another person.

Information about whether the mother had observed the infant's hospital sleep position was available for 291 respondents (74%). Presumably, at the time of the baseline interview, the remaining respondents had not yet observed the infant sleep position used by hospital nurses. Information on both hospital sleep position and intended sleep position was available for 284 mother-infant dyads. Of the 43 mothers who had observed their infants in the prone sleep position while in the hospital, 40 (93%) intended to place their infants for sleep in the prone position at home. In contrast, only 58 (24%) of the 241 mothers who observed a nonprone hospital sleep position intended to use the prone position at home (OR, 42.07; 95% CI, 12.55-141.06—data not shown). This finding suggests that observed hospital sleep position may modify maternal intentions and ultimately the choice of a particular sleep position. Information on both hospital sleep position and last night's sleep position was available for 283 mother-infant dyads. Of the 41 mothers in this subset who had observed their infants in the prone sleep position while in the hospital, 23 (56%) indicated at follow-up that they had placed their infants in the prone position last night, as compared with 83 (34%) of mothers who had observed a nonprone sleep position while at the hospital (OR, 2.45; 95% CI, 1.25-4.79) (Table 4, Part A, and Table 4, Part B).

In multivariate analyses (model 1, Table 5), black race (OR, 2.06; 95% CI, 1.05-4.04), poverty (OR, 1.81; 95% CI, 1.10-2.99), presence of the infant's grandmother in the home (OR, 1.83; 95% CI, 1.11-3.00), and intent to use the prone sleep position (OR, 2.28; 95% CI, 1.44-3.60) were independently associated with last night's prone sleep position. Observed hospital sleep position was not included in this model as information was missing for 26% of respondents. We lacked sufficient power in our study to fully examine variations in risk factors for the prone sleep position by race and ethnicity. However, among black mothers there was a suggestion of an independent association between being single and use of the prone sleep position (adjusted OR, 2.61; 95% CI, 1.42-4.78) and delivering vaginally and use of the prone sleep position (adjusted OR, 2.01; 95% CI, 1.13-3.56).

Due to the potential importance of sleep position practices in the hospital as a focus for future interventions and the strong association noted between hospital sleep position and maternal intentions, we reanalyzed predictors of last night's sleep position, including observed hospital sleep position but excluding intended sleep position as an independent variable. In this multivariate analysis (model 2, Table 5), the presence of a grandmother in the home (OR, 1.90; 95% CI, 1.10-3.27) and observed prone hospital sleep position (OR, 2.35; 95% CI, 1.19-4.64) were independently associated with use of the prone sleep position at 3 to 7 months of age. Income and race were not significant in this model.

More than 3 years after the initial AAP recommendations and substantial public education about infant sleep position, a high proportion of infants in the low-income, inner-city population included in this study were placed for sleep in the prone position. A number of factors—including black race, the presence of the infant's grandmother in the house, baseline intentions, and poverty—were found to be associated with an increased risk of using the prone sleep position.

Several study limitations should be noted. Only a limited number of psychosocial variables were examined as possible predictors of sleep position. Although instructional brochures on infant sleep position were reportedly included with the discharge materials at 2 of the 3 hospitals, we did not confirm that mothers received these materials, nor did we obtain details about verbal advice given by providers. Finally, in the current study, although mothers were selected to represent District of Columbia residents of lower socioeconomic status, our sampling scheme was not population based.

Our finding of a 40% prevalence of use of the prone sleep position at 3 to 7 months is higher than national prevalence estimates for 1996 of 24% of infants younger than 8 months being placed in the prone position among all races, but our findings are comparable to the 43% prevalence reported among blacks.²⁴ Consistent with other studies, the most common reason for choosing the prone position was infant comfort.¹⁹ Infants who sleep in a prone position have fewer spontaneous arousals and sleep for longer periods than those who sleep in a supine position.²⁵

The association between low income and use of the prone sleep position identified in the current study is consistent with findings in Tasmania and New Zealand, where low socioeconomic status was found to be associated with the prone sleep position after, but not before, implementation of public health campaigns to avoid use of the prone sleep position.^{1 ,26} In the current study, we may have actually underestimated the magnitude of these associations by examining these associations within a population that was predominantly low income. Such populations may adopt new health practices more slowly than higher-income populations, perhaps because they have fewer information sources from which to learn about innovations.²⁷ In support of this, research in Tasmania has identified an association between markers of low health education uptake and use of the prone sleep position.²⁸

We also identified an independent association between race and sleep position. Blacks were more likely than persons of other races (in this study primarily Hispanics) to use the prone sleep position (adjusted OR 2.06, 95% CI 1.05-4.04). In Washington State, African Americans were more likely to place their infants in a prone position compared with non-African Americas, but this association was not statistically significant (adjusted OR 3.5, CI 0.8-16.3).²⁹ A cross-sectional, practice-based survey in Ohio also found no significant association between either race or family income and infant sleep position.³⁰ Other independent predictors of use of the prone sleep position identified in the current study included the presence of the infant's grandmother in the household, observed hospital sleep position, and baseline intentions, none of which have been reported in previous publications. Studies have, however, documented the important influence that grandmothers can exert on health decisions for their grandchildren, such as the use of an emergency department³¹ or the timeliness of obtaining childhood immunizations.³²

All hospitals in the current study reported that, following the 1992 AAP recommendations, nurses were verbally instructed to place infants in either the supine or side sleep position. One of the hospitals participating in this study had a written policy on infant sleep position in the form of a letter to nursery staff, which was posted in the nursery reminding its staff to use the supine or side sleeping position, and 2 of the 3 hospitals placed stickers on each bassinet reminding nurses to use the nonprone position. Despite these differences in hospital practices, the percentage of mothers who reported that they had observed their infant in the prone sleep position was consistent among the 3 sites, ranging from 14% to 17%. Similarly, in a 1995 national survey of newborn nurseries, 11% of head nurses reported the prone position as 1 of the sleep positions used for healthy newborn infants (Marian Willinger, PhD, National Institute of Child Health and Human Development, written communication, April 1998). In this study, mothers who observed their infants being placed in the prone position by hospital personnel were significantly more likely to indicate the intention to place the infant in the prone position and also more likely to do so compared with mothers who observed their infants in a nonprone position. Because people learn through observation,³³ modeling by health care providers can be a powerful influence on behavior.

Mothers, in this study, who reported that they had discussed sleep position with a health care clinician during the postpartum stay were less likely to place their infants in the prone position than mothers who had not had such discussion (38% vs 46%). Although this association was not statistically significant, at least 1 other study has also suggested that advice given to parents during the postpartum period may be effective in altering infant sleep position.³⁴ Ideally, hospitals should adopt and implement policies on infant sleep position that are consistent with current recommendations and should monitor compliance with policies once adopted. Clinicians should counsel mothers on the use of the correct infant sleep position and demonstrate its use by placing infants in the supine sleep position while in the hospital.

Although sleep position at 3 to 7 months was associated with maternal intentions measured shortly after delivery, approximately a third of mothers who intended to place their infants to sleep in a nonprone position indicated at follow-up that they had placed their infants in the prone position during the previous night. Despite initial intentions, mothers may be influenced to place their infants in the prone position if they perceive that the infant would be more comfortable or would sleep longer in this position. Alternatively, an influential other person, such as a grandmother, might persuade the mother to place the infant in the prone position. Therefore, counseling and reinforcement on the importance of the supine sleep position should continue beyond the initial hospital stay and, when possible, be directed to both the primary caregiver and other extended family members, such as grandmothers. As noted by Ponsonby and colleagues, educational efforts should "focus on the use of appropriate methods to settle an infant without resorting to the prone position."²⁸