Comparison of Infant Sleep practices in African-American and U.S Hispanic Families: Implications for Sleep-Related Infant Death

Abstract

BACKGROUND

African-American and Hispanic families share similar socioeconomic profiles. Hispanic rates of sleep-related infant death are four times lower than African-American rates.

METHODS

Cross-sectional, multi-modal (surveys, qualitative interviews) study to compare infant care practices that impact risk for sleep-related infant death in African-American and Hispanic families.

RESULTS

We surveyed 422 African-American and 90 Hispanic mothers. Eighty-three African-American and 6 Hispanic mothers participated in qualitative interviews. African-American infants were more likely to be placed prone (p<0.001), share the bed with the parent (p<0.001), and to be exposed to smoke (p<0.001). Hispanic women were more likely to breastfeed (p<.001), while African-American women were more knowledgeable about SIDS. Qualitative interviews indicate that, although African-American and Hispanic parents had similar concerns, behaviors differed.

DISCUSSION

Although the rationale for infant care decisions was similar for African-American and Hispanic families, practices differed. This may help to explain the racial/ethnic disparity seen in sleep-related infant deaths.

BACKGROUND

The incidence of sudden infant death syndrome (SIDS) in the U.S has declined by 50% since 1992, when the American Academy of Pediatrics (AAP) first recommended that infants be placed in a non-prone (i.e., side or back) position for sleep.(1) However, African-American infants still remain twice as likely to die from SIDS;(2) in addition, African-American infants are also twice as likely to die from other sleep-related infant deaths, such as accidental suffocation and strangulation in bed (ASSB) and ill-defined causes of death.(2, 3) All of these causes of death are frequently associated with bedsharing.(3–5)

African-American and U.S Hispanic (hereafter referred to as Hispanic) families share similar socioeconomic profiles and high rates of bedsharing.(6, 7) However, Hispanic rates of SIDS, ASSB and other ill-defined causes of death are much lower than national rates and approximately four times lower than African-American rates.(2, 8)

While case-control and cohort studies have demonstrated that African-Americans are more likely to place their infants prone (on the stomach) for sleep(2, 8) and to bedshare,(6, 9, 10) little is known about infant sleep practices among Hispanic families. The goal of this study was to compare infant care practices pertinent to sleep-related infant deaths, including sleep position, bedsharing, roomsharing, parental smoking and breastfeeding, among African-American and Hispanic families. In addition, knowledge about SIDS was compared. Because rates of sleep-related infant death are much lower in Hispanic infants, we hypothesized that Hispanic mothers are more likely to place their infants supine for sleep, breastfeed exclusively and are more likely to be informed about SIDS.

METHODS

We conducted a multimodal study, utilizing both quantitative and qualitative techniques to ascertain factors, attitudes and beliefs that inform infant care practices among African-American and Hispanic mothers. The techniques included surveys, individual in-depth interviews, and focus group interviews. We chose this multimodal approach because we believe that the strengths of the various techniques complement one another. Surveys permitted us to quantify the distribution of the knowledge, attitudes and practices. Qualitative data are frequently used to better understand motivations and perceptions underlying health decisions(11, 12) and rely on obtaining the widest possible range of perspectives through systematic purposeful sampling.(13, 14) We therefore selected two different qualitative interview formats. Focus groups provide participants in a group of people with similar backgrounds with a comfortable forum to express opinions;(13) however, socially sensitive topics might be more likely to be raised in individual interviews.(14) The institutional review boards at Children's National Medical Center, MedStar Research Institute (Washington, DC), and Holy Cross Hospital (Silver Spring, MD) approved this study.

Participants

We enrolled a cross-sectional sample of African-American and Hispanic mothers of infants <6 months of age. The mothers were 18 years old and lived in Washington DC and Maryland. Mothers were recruited from newborn nurseries, urban pediatric primary care centers, WIC sites, private pediatric practices, advertisements in newsletters and on-hold messages played during calls to Children's National Medical Center. Recruitment has been described previously.(15–17) Briefly, the mothers were eligible if they self-identified as African-American or Hispanic. A mother was excluded if, by parental report, she was not a custodial parent of the child, the infant had a chronic illness that precluded use of the supine sleep position (eg., recent spinal surgery), or the infant was born prematurely (gestational age<36 weeks).

We recruited parents of lower and higher socioeconomic status (SES) to assure a broad range of experience, influences and attitudes. SES was determined by parental educational level and eligibility for Medicaid and Women, Infants, and Children (WIC) nutritional program. Because the latter 2 are easily verifiable and do not rely upon self-report, they were used as proxies for family income. A parent was defined as being of lower SES if the family had Medicaid or other public health insurance, or the family received WIC benefits. The family was defined as being of higher SES if the family had commercial health insurance and did not receive WIC benefits.

Data Collection

After written informed consent was obtained, qualified and interested parents participated in a 15-minute staff-administered survey that asked about knowledge, attitudes and practices regarding infant care, sleep environment and family demographics. The survey instrument was validated, and the questions have been used in previous published studies.(18, 19) Based on these responses, a purposeful sample of parents who were predicted to have a wide range of opinions was asked to participate in either a focus group or individual interview, depending on timing and individual schedules.

All interviews were conducted by trained facilitators (R.P.O., B.L.J., I.A.), who used the same interview guide for focus groups and individual interviews. Interviews with African-Americans were all conducted in English; interviews with Hispanic mothers were conducted in English or Spanish, as determined by participant preference. Questions were asked about infant care and infant sleep environment, including sleep position, bedsharing and the use of soft bedding. In both formats, broad, open-ended questions were followed by more-specific, probing questions to elucidate responses. We asked parents about their infant care practices. For sleep practices, we asked about usual practice (during the previous week) and practice during the night before the survey. This strategy is frequently utilized in SIDS research to encourage honest answers, particularly when practices might be perceived to be not socially acceptable.(14)

For the survey, we estimated that the proportion of families that would be bedsharing would be 60% in Hispanic families and 80% in African-American families. Based on those proportions, we calculated that a sample size of 81 in each group would be sufficient to provide statistical power of 80% with an alpha level of 0.025. We also anticipated that a minimum of 10 focus groups and 10 individual interviews would be conducted, as we assumed that 3 to 4 semi structured interviews and/or 3 to 4 focus groups with any 1 type of participant would be necessary(20) to allow for thematic saturation (the point at which no new themes are emerging) and for analysis across groups for themes and patterns.

Measures and Analysis

The primary outcome measures, based on parental report, included infant sleep position (categorized as supine [back], side, prone), roomsharing (with one or both parents), bedsharing (with parent[s] or other people), any breastfeeding (which included partial and exclusive breastfeeding), smoke exposure (from mother and others) and the knowledge of SIDS. Knowledge of SIDS was assessed by the parent's response to the question, “Can you tell me what SIDS is?” Racial/ethnic group was the primary predictor variable.

Descriptive statistics (mean, range, SD units) of demographics were calculated. Demographic variables were analyzed for correlation with primary outcome measures, using the Pearson correlation coefficient. The relationship between outcome measures and predictor variables also were analyzed, using a multiple logistic regression model. All variables significant in the univariate analysis were included in the multivariate analysis. All analyses were conducted with STATA software.(21)

All qualitative interviews were video- and audio-recorded and transcribed by the authors. Video recordings allowed identification of the speakers, so that all responses could be appropriately attributed to the correct speaker; in addition, facial expressions could be noted. After initial transcription, the transcript was checked by 2 additional authors for accuracy. If there was disagreement about the transcription, all authors listened to the recordings to reach consensus. This multistep process was used to maximize accuracy and eliminate bias from the transcription process.

Qualitative analysis software (NVivo 8)(22) was used to organize, sort, and code the data (quotes). Using grounded theory methodology, themes were developed and revised in an iterative manner as patterns became apparent.(23) In weekly meetings, authors discussed emerging themes and patterns in the data and reached consensus on the major themes. Individual interviews and focus groups were analyzed separately, but by the same investigators, after which emerging themes were compared. Matrix analysis was conducted to determine the distribution of themes, in particular whether particular themes recurred more often with participants with specific characteristics. For instance, matrix analysis would reveal whether concerns about infant comfort recurred more frequently in breastfeeding mothers than in non-breastfeeding mothers. Concurrent triangulation, or use of multiple sources for verification of findings,(20) of the focus groups and individual interviews was used to confirm findings. Our findings were additionally corroborated through peer review and feedback during presentations to pediatric and SIDS researchers, and maternal and child health professionals.

RESULTS

Participant Demographics (Table 1)

TABLE 1.

Demographics for African-American and Hispanic Women and Infants

	African-American (N=422)	Hispanic (N=90)	p-value
Mean baby age (months)	2.9	1.0	P<0.001
Mean Parent age (years)	26.7	28.9	P=0.002
Baby gender
Male	207 (49.1%)	41 (45.5%)
female	214 (50.8%)	49 (54.4%)	P=0.53
Marital status
Married	90 (21.4%)	41 (45.6%)	P<0.001
Never Married	316 (75.2%)	48 (53.3%)
Other	14 (3.3%)	1 (1.1%)
Education
8th Grade or less	1 (0.2%)	21(23.3%)
Grades 9–11	41 (9.8%)	16 (17.8%)	P<0.001
High School Graduate/GED	170 (40.5%)	34 (37.8%)
Some technical /vocational school	101 (24.1%)	7 (7.8%)
Some College	26 (6.2%)	3 (3.3%)
4 year College graduate	47(11.2%)	9 (10.0%)
Postgraduate training	34(8.1%)	0 (0.0%)

Between July 2006 and July 2012, we collected survey data from 422 African-American and 90 Hispanic mothers. We conducted 14 focus groups and 11 individual interviews with 83 African-American and 6 Hispanic mothers and reached thematic saturation. Focus group attendance averaged 4.9 (range: 3–7) participants. Participant demographic characteristics are described in Table 1. There was a statistically significant difference in demographics between the African-American and Hispanic women and infants in several variables. Mean maternal age was 26.7 years (range: 18–42 years, SD 6.0) for African-American mothers and 28.9 years (range 18–24 years, SD 5.6) for Hispanic mothers. Three-quarters (75.2%) of African-American mothers and 53.3% of Hispanic mothers were never married. Forty percent of African-American mothers and 38% of Hispanic mothers had a high school diploma; an additional 11.0% of African-American mothers and 10.0% of Hispanic mothers had a 4 year college degree. At the time of the focus group or individual interview, mean infant age was 2.5 months (range: 1 to 6 months, SD 1.6). Hispanic mothers in this study self-identified as being from Bolivia, Columbia, Costa Rica, El Salvador, Guatemala, Honduras, Nicaragua, Puerto Rico, Peru, and Venezuela. Because responses did not correlate with country of origin, we did not further stratify the Hispanic cohort.

Survey results (Table 2)

TABLE 2.

Infant Care Practices in African-American and Hispanic Families: Univariate and Multiple Logistic Regression Analysis

Risk factor	African-American # (%)	Hispanic (Reference) # (%)	Univariate p-value	Multivariate p-value*	Adjusted odds ratio (95% Confidence interval)
Usual sleep position
Back	267(63.2%)	75 (83.3%)			reference
Side	90 (21.3%)	13 (14.4%)	P<0.001	P=0.01	2.8 (1.3–6.0)
Stomach	65 (15.4%)	2 (2.2%)		P=0.01	9.2 (1.0–43.0)
Sleep position last night
Back	279 (66.2%)	75 (83.3%)			reference
Side	79 (18.7%)	15 (16.6%)		P=0.4 (NS)	1.4 (0.7–2.9)
Stomach	63 (14.9%)	0 (0.0%)	P<0.001	P=1.0 (NS)	unable to calculate
Unknown	1 (0.2%)
Usual supine placement
No	155 (36.7%)	15 (16.7%)	P=0.002	P<0.001	3.6 (1.8–7.3)
Yes	267 (63.3%)	75 (83.3%)			reference
Supine placement last night
No	142 (33.7%)	15 (16.6%)	P<0.001	P=0.001	2.4 (1.2–4.8)
Yes	279 (66.3%)	75 (83.3%)			reference
Unknown	1 (0.2%)
Usually shares room with
Nobody	17 (4.0%)	0 (0%)
Mother only	146 (34.5%)	28 (31.1%)
Father only	1 (0.2%)	1 (1.1%)	P=0.2 (NS)
Both parents	173(40.9%)	42 (46.6%)
Other	85 (20.1%)	19 (21.1%)
Shared room last night with
Nobody	19 (4.5%)	0 (0%)
Mother only	155(36.7%)	27 (30.0%)	P=0.8 (NS)
Father only	7(1.7%)	0 (0%)
Both parents	162(38.4%)	41 (45.6%)
Other	78(18.5%)	22 (24.4%)
Unknown	1 (0.2%)	0 (0%)
Usual bedsharing
No	273 (64.7%)	73 (81.1%)			reference
Yes	149 (35.3%)	17 (18.8%)	P=0.003	P=0.02	2.3(1.2–4.3)
Bedsharing last night
No	276 (65.4%)	79 (87.8%)			reference
Yes	146 (34.6%)	11 (12.2%)	P<0.001	P<0.001	3.5 (1.7–7.4)
Any Breastfeeding
None	262 (62.1%)	8 (8.9%)	P<0.001	P<0.001	11.0 (4.1–29.8)
Partial	86 (20.4%)	60 (66.7%)		P<0.001	0.3 (0.2–0.7)
Exclusive	74 (17.5%)	22 (24.4%)		P=0.13 (NS)	reference
Mom smokes after pregnancy
No	374 (88.6%)	87 (96.6%)	P=0.02	P=0.01	reference
Yes	48 (11.3%)	3 (3.3%)			0.2 (0.04–0.7)
Other smoke exposure
No	305 (72.2%)	85 (94.4%)	P<0.001	P<0.001	reference
Yes	117(27.7%)	5 (5.5%)			0.2 (0.05–0.4)
Knowledge of SIDS
No	52 (12.3%)	35(38.8%)			reference
Yes	370(87.6%)	55(61.1%)	P<0.001	P=0.004	2.7 (1.4–5.3)

^*Controlled for infant age, maternal age, maternal educational level, and marital status.

More than 80% of Hispanic mothers placed their infants supine (on the back) for sleep, compared with approximately 60% of African-American mothers (p<0.001). 15% of African-American infants had been placed prone (on the stomach) for sleep during the night before the survey, compared with no Hispanic infants (p<0.001). With regard to infant sleep location, there were no statistically significant differences with regards to the room where the infant slept; the vast majority in both groups slept with one or more adults (usually the parents). However, African-American infants were significantly more likely to bedshare (i.e., share the same sleep surface) with a parent (p=0.003). Hispanic infants were more likely to be at least partially breastfed (p<0.001), whereas African-American infants were significantly more likely to be exclusively formula-fed (p<0.001). African-American infants were more likely to be exposed to secondhand smoke from the mother (p=0.02) as well as from other individuals in his or her environment (p<0.001). African-American mothers were significantly more knowledgeable about SIDS than Hispanic mothers (p<0.001).

In the multiple logistic regression model, controlling for infant age, maternal age, marital status, and maternal educational level, we found that Hispanic families were more likely to place the infant supine for sleep, both usually (p<0.001) and last night (p=0.001); less likely to bedshare, both usually (p=0.02) and last night (p<0.001); more likely to at least partially breastfeed (p<0.001); and less likely to expose the infant to secondhand smoke from the mother (p=0.01) or other household members (p<0.001). African-American mothers continued to be significantly more knowledgeable about SIDS than Hispanic mothers (p=0.004).

Qualitative (Focus Group and Individual Interview) Findings

Results from the qualitative interviews indicate that although African-American and Hispanic parents often had the same concerns, the actual behavior was sometimes different. Illustrative quotes (Q) are included in Table 3. For instance, infant safety and infant comfort were recurrent themes for placing infants prone by African-American parents (Q1,Q2). However, these were cited by Hispanic parents as reasons for placing infants supine (Q3,Q4). Similarly, convenience for the purpose of being vigilant was a major theme for African-American parents when they discussed reasons for bedsharing (Q5). However, Hispanic parents cited convenience for feeding as a reason for roomsharing without bedsharing (Q6). Reasons for breastfeeding, roomsharing, and not smoking were similar in the two groups. A major theme in the Hispanic interviews was the belief that breastfeeding was beneficial (Q8) for the babies and economical (Q10); this was also true for African-American mothers who were breastfeeding (Q7, Q9). Both Hispanic and African-American mothers cited similar reasons for sleeping in the same room with the baby, including convenience for monitoring and feeding the infant, and lack of space for the infant to sleep in a separate room (Q11–Q14). The safety and health of the baby was a recurrent theme when both African-American and Hispanic mothers discussed reasons to abstain from smoking (Q15,Q16). Matrix analysis of maternal birth parity (i.e., prior experience with infants) or country of origin for Hispanic participants did not reveal any difference in recurrence of themes.

TABLE 3.

Comparison of Qualitative Themes and Quotations between African-American and Hispanic mothers

African-American Women	Hispanic Women
Sleep Position
Safety (Prone position) Q1. “That's what I am afraid of, too. Like laying on his back. I did that a few times with the newest baby….(until one time, when) he was (making choking sounds) like he was choking and that scared me, you know. It was like he was trying to throw something up, but it was going all back down his throat… When I heard that, that really scared me and I just….cleaned him up, and rolled him on his stomach.” (36 year old single higher SES mother of 5 month old)	Safety (Supine position) Q3. “On his back. I think it's the best position based on what I have been told at the doctor's office. Sleeping on his side and stomach is dangerous because he can turn and suffocate.” (32 year old married, El Salvadoran mother of 1 month old)
Comfort (Prone position) Q2. “We tried. We slept him on his back for maybe almost a month, but it was just like he was uncomfortable.. .The next thing I know, he just likes to be on his stomach. And we noticed that he slept longer.. .And that's the way it's been, maybe about 2 months, since he was 2 months old, that he always slept on his stomach.” (36 year old married higher SES mother of 6 month old)	Comfort (Supine position) Q4. “I base it on whatever position gets him to sleep the longest through the night. At this time it's the position that I described….straight, On his back, propped on his back, on an incline.” (30 year old married, Peruvian mother of 2 month old)
Bedsharing
Convenience (Bedsharing) Q5. “He won't lay on his back, he won't lay on his side, but he'll lay on his stomach…And the things is, yea, and it's like if I lay him on his stomach, it's like ok, I gotta go in here and run in here, jump to make sure he is ok…so he sleeps with me.” (28 year old single lower SES mother of 2 month old)	Convenience (Not Bedsharing) Q6.“In my room, right next to us (in a crib) because right now he's sleeping 5 hours straight through the night and I always want to be next to him when I can be ready to breastfeed him.” (31 year old married, Peruvian mother of 2–1/2 month old)
Breastfeeding
Health Benefits for Infant (Breastfeeding) Q7. “I think it helps when, as far as the cold weather now. When he was a baby he didn't get sick as much as my friend's baby did. We had our babies 1 day apart. Her baby stayed sick and mine didn't. That's how my baby is now. Also, my cousin's baby and my baby were 5 days apart. She keeps getting sick. So, I think it helps their immune system when they are young but as far as the development of the brain, I think goes that plays, part of the mother teaching them with as far as what you teach your child.” (21 year old single, lower SES mother of 6 month old) Less Expensive (Breastfeeding) Q9. “I agree with breastfeeding because your baby stays healthy, he don't get sick and its easier, it's cheaper. You don't have to pay for it and it's always there.” (24 year old single, lower SES mother of 4 month old)	Health Benefits for Infant (Breastfeeding) Q8. “Yes, because science has tried to recreate what breastfeeding provides for the baby but they cannot recreate the immunological defenses it provides.” (29 year old single, El Salvadoran mother of 2 month old) Less Expensive (Breastfeeding) Ql0.“It saves me money and it's the best nutrients for my baby.” (23 year old married, El Salvadoran mother of 2 month old)
Roomsharing
Space (Roomsharing) Q11. “Because I don't have enough room where I'm at. I mean, once I get my own, then that'll be something different.” (24 year old single, lower SES mother of 4 month old)	Space (Roomsharing) Q12. “Because I'm breastfeed him and I did not have space to put the bassinet in my room since I have this huge dresser I need to get rid of to make room.” (37 year old married, Puerto Rican mother of 1 month old)
Convenience (Roomsharing) Q13. “My son has his own room but he's in the room with me now…that was like to help prevent SIDS because I could hear and check and make sure he's breathing, `cause he's right there versus, you know… like if he's down the hall, you know, even though you have the monitor, you know, it's not the exact same but if he's in the bassinet right beside me, I can just check him to make sure, you know, he's breathing…notice anything that was going wrong.” (37 year old single, higher SES mother of 2 month old)	Convenience (Roomsharing) Q14. “She does have a crib in my other child's room but I do not put her to sleep there because the room is too far away from my room and I want to be near her when she sleeps.” (35 year old single, Puerto Rican mother of 2 month old)
Reasons for Not Smoking
Safety (Not Smoking) Q15.“So basically I went my whole nine months without smoking or drinking…So once I found out I was pregnant and stuff, I left it alone. Because like I said, I didn't want no health problems and you know what I'm saying? And I wasn't even trying to make the same mistake my cousin did when she kept smoking and drinking and stuff. Now the baby got bad asthma and stuff. So you know, I just left it alone. Because like I said, it's going to always be on this earth; there's nothing that I'm missing. I want a healthy baby, well, babies.” (22 year old single, lower SES mother of a 4 month old)	Safety (Not Smoking) Q16. “No one smokes in my house, but if someone did bring someone else who does , they cannot smoke in my house and I would not allow them to carry my baby, even if they washed their hands.” (23 year old single, El Salvadoran mother of 2 month old)

DISCUSSION

The racial and ethnic disparities in SIDS and other sleep-related infant deaths persist, despite increases in supine positioning. Few epidemiologic analyses of SIDS stratify risk factors by race and ethnicity. In fact, to our knowledge, no other study has compared infant care practices of African-American and Hispanic families. Our study found significant differences in the rates of supine sleep positioning, bedsharing, breastfeeding, and smoke exposure. In all of these practices, Hispanic families were more likely to adhere with AAP recommendations. However, despite lower likelihood of adherence with AAP recommendations, African-American families were more likely to be knowledgeable about SIDS. In addition, preliminary results from our qualitative interviews suggest that African-American and Hispanic families often had similar concerns (e.g., infant safety or infant comfort); however, the 2 groups responded to these concerns with different behaviors.

Non-supine (side or prone) sleep positioning is one of the most significant risk factors for SIDS. Our study confirms others(24–27) that have found that African-American parents are more likely than other racial/ethnic groups to place infants prone. Hispanic families, on the other hand are more likely to choose the supine position for sleep, and this has been demonstrated in prior investigations.(8, 28) It was interesting that both groups cited infant comfort and infant safety as reasons for their sleep positioning decisions, although the decisions were often different (i.e., prone for African-American families and supine for Hispanic families). This may reflect cultural norms in the respective groups. In many African-American communities, prone sleep position is considered the norm; most parents cite that their primary trusted source (usually a senior caregiver in the family, such as a grandmother)(17, 29, 30) used prone positioning many years ago, and that this was considered the best way. In contrast, in many Hispanic cultures, supine positioning is considered the norm;(26) furthermore, Hispanics are more likely to trust the medical profession most for advice about infant care.(26) African-American mothers who placed their infants prone often did so because of specific concerns about the possibility of suffocation or aspiration, whereas Hispanic mothers who were concerned about suffocation often placed their infants supine. Similarly, African-American mothers who cited infant comfort as a primary reason for their choice of sleep position often placed their infants prone because they felt that the baby slept better and for a longer duration. However, Hispanic mothers believed that the supine position was the more ideal position for sound sleep.

We did not find a significant difference between the 2 groups in the percentage of mothers that shared the room with their infants. Roomsharing without bedsharing reduces the risk of SIDS(31–34) and is safer than bedsharing or sleeping in a separate room from the parents.(31–34) Both Hispanic and African-American families described similar issues with lack of space, because of large families or financial constraints. Mothers in both groups also chose to roomshare initially because they felt that it was more convenient to feed the baby and allowed the mother to more easily monitor the baby.

We found that African-American women were significantly more likely to bedshare with their infants than Hispanic women. Other studies have demonstrated that African-American mothers have higher rates of bedsharing than Caucasian families,(6, 9, 10) and the SIDS risk during bedsharing is greater when combined with other risk factors such as smoking, consumption of alcohol, medications, or drugs that alter arousal, young (<3 months) infant age, use of bedding and soft surfaces, bedsharing with persons other than the parents, and extended bedsharing (i.e., all night long).(2, 35–38) Bedsharing among African-American women is viewed as a very efficient way to provide physiological comfort to the mother and child, while being able to monitor the infant's breathing and general health.(15) In addition, some African-American women bedshared with their infant because there was no other space for the infant. Hispanic women on the other hand had much lower rates of bedsharing, despite having similar financial and space constraints. For them, roomsharing without bedsharing was more common. Even Hispanic breastfeeding mothers described in the focus groups that their infants slept next to them in cribs, because this facilitated breastfeeding.

Breastfeeding rates among Hispanic women have been reported in other studies to be consistently higher than those among African-American mothers,(39–42) so we were not surprised to find high rates of any (partial or exclusive) breastfeeding in Hispanic mothers. While the rates of exclusive breastfeeding were similar in the two groups, Hispanic women were more likely to partially breastfeed their infants. Breastfeeding is perhaps more of a cultural norm for Hispanic families than for African-American families.(41, 43) Focus group discussions about breastfeeding centered on the benefits of breastfeeding for the infants. Both Hispanic and African-American mothers who were breastfeeding cited benefits such as improved immunity and higher intelligence. Additionally, breastfeeding mothers in both groups felt that breastfeeding could be practiced easily in any setting and was more economical than formula feeding.

African-American infants were significantly more likely to be exposed to secondhand smoke, both from their mother and other individuals in the infant's environment. This finding is not surprising since smoking rates are higher among African-American women than Hispanic women.(44, 45) Secondhand smoke exposure increases the infant's risk of sleep-related death, particularly in bedsharing situations.(31, 46–49)

We found that African-American mothers generally had a greater awareness of SIDS and were more familiar with the Back to sleep campaign than Hispanic mothers. Because many of the Hispanic women in our study were first generation immigrants, it is possible that they had not been exposed to the information provided in the Back to Sleep campaign. Therefore, it was surprising that Hispanic women were more likely to adhere to AAP-recommended sleep practices, although they were not aware that these practices were recommended. It appears that some of the protective practices, such as partial breastfeeding, are cultural norms in the native communities for the Hispanic families.(41, 50) In addition, there continues to be a gap in African-Americans between knowledge of SIDS and the practice of protective behaviors; in a previous study, we found that many African-American women did not find the relation between sleep practices and SIDS to be plausible, because it did not make sense to them and because they considered a SIDS death to be “God's will,” and that changing their behavior would not prevent the death from occurring.(16)

To our knowledge, this is the first study to compare sleep practices and reasons for these sleep practices in African-American and Hispanic mothers and infants. We found that the results from the quantitative and qualitative data analysis were consistent. However, our study population was limited to African-American and Hispanic mothers in the Washington DC area. In addition, we acknowledge that our sample of Hispanic families in the qualitative analysis, while providing thematic saturation, would be strengthened by additional participants. Our qualitative results cannot be used to define prevalence of any specific attitude or behavior, nor can these findings be construed as generalizable. Furthermore, we did not ask about parental use of alcohol, medications, or illicit drugs, which can increase the risk of sleep-related death during bedsharing.(2)

NEW CONTRIBUTION TO THE LITERATURE

Although the rationale for decisions about sleep positioning, sleep location and breastfeeding is similar in African-American and Hispanic families, the behavior is often different in the two groups. These differences may explain the racial/ethnic disparity seen in sleep-related infant deaths. Knowledge of SIDS does not necessarily mean that safe sleep recommendations are being followed. Some cultural norms in the Hispanic community may be protective.