Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
December 07, 2009

Trends and Factors Associated With Infant Sleeping PositionThe National Infant Sleep Position Study, 1993-2007

Author Affiliations

Author Affiliations: Department of Pediatrics, Yale University School of Medicine, New Haven, Connecticut (Dr Colson); Data Coordinating Center (Mr Rybin and Dr Colton), and Department of Epidemiology, Boston University School of Public Health, Boston, Massachusetts (Dr Colton); Massachusetts Department of Health, Boston (Dr Smith); Department of Pediatrics, UT Southwestern Medical Center, Dallas, Texas (Dr Lister); Departments of Pediatrics and Epidemiology, Boston University Schools of Medicine and Public Health, Boston (Dr Corwin).


Copyright 2009 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2009

Arch Pediatr Adolesc Med. 2009;163(12):1122-1128. doi:10.1001/archpediatrics.2009.234

Objective  To determine trends and factors associated with choice of infant sleeping position.

Design  Annual nationally representative telephone surveys from 1993 through 2007.

Setting  Forty-eight contiguous states of the United States.

Participants  Nighttime caregivers of infants born within the last 7 months; approximately 1000 interviews were given each year.

Main Outcome Measure  Whether infant is usually placed supine to sleep.

Results  For the 15-year period, supine sleep increased (P < .001) and prone sleep decreased (P < .001) for all infants, with no significant difference in trend by race. Since 2001, a plateau has been reached for all races. Factors associated with increased supine sleep between 1993 and 2007 included time, maternal race other than African American, higher maternal educational level, not living in Southern states, first-born infant, and full-term infant. The effect of these variables was reduced when variables related to maternal concerns about infant comfort, choking, and advice from physicians were taken into account. Between 2003 and 2007, there was no significant yearly increase in supine sleep. Choice of sleep position could be explained almost entirely by caregiver concern about comfort, choking, and advice. Race no longer was a significant predictor.

Conclusions  Since 2001, supine sleep has reached a plateau, and there continue to be racial disparities. There have been changes in factors associated with sleep position, and maternal attitudes about issues such as comfort and choking may account for much of the racial disparity in practice. To decrease sudden infant death syndrome rates, we must ensure that public health measures reach the populations at risk and include messages that address concerns about infant comfort and choking.

Sudden infant death syndrome (SIDS) remains the leading cause of postneonatal death in the United States.1 Placing the infant to sleep in the supine position has been associated with the dramatic decrease in the SIDS rate since the Back to Sleep Campaign began in 1994.2 Despite this decrease, African American infants continue to have more than twice the incidence of SIDS as white infants.1 African American infants are also less likely than white infants to be placed in the supine position for sleep.35

The National Infant Sleep Position Study (NISP), an annual national telephone survey, provides much of what we know about national trends in infant care practices related to SIDS and the efficacy of the Back to Sleep Campaign.4,6 The objectives of this article are to examine trends in infant sleeping position, understand factors associated with choice of infant sleeping position, and identify barriers to further change in practice using data collected since the inception of the NISP surveys in 1993 through 2007.


The data used in the analysis for this study are part of the NISP. The sample was chosen to represent the 48 contiguous states (not including Alaska and Hawaii). DataStat, Inc (Ann Arbor, Michigan) conducted the telephone interviews to home and cellular telephones by randomly sampling targeted households with infants aged 7 months and younger. A representative list from the 48 contiguous states was purchased from Metromail (Lincoln, Nebraska). The list was based on public information from sources such as birth records, infant photography companies, and formula companies. Interviews were completed if the respondent answered “yes” to the question, “Is there an infant in the house born in the last 7 months, that is, on or after (date)?” More than 80% of the respondents were the infants' mothers. The goal was to complete approximately 1000 calls each year. The number of calls completed ranged from 1012 to 1188 between 1993 and 2007. Response rates were calculated using the American Association for Public Opinion Research standard definitions and formulas. An exact response rate cannot be calculated because eligibility cannot be determined for those who refused to be interviewed. An estimate of the response rate was made, based on the assumption that the eligible proportion of households who refused is the same as the eligible proportion of those for whom we could determine eligibility. The average response rate from 1993 through 2007 was 73%.7

With telephone surveys, participation is limited to those with access to telephones and may not include those who are economically disadvantaged. In 2005, the NISP sample differs from the National Center for Health Statistics (NCHS) in the following maternal characteristics: maternal race, African American (NCHS 15% vs NISP 6%); Hispanic ethnicity (NCHS 24% vs NISP 6%); younger than 20 years (NCHS 10% vs NISP 3%); and education less than 12 years (NCHS 24% vs NISP 4%).8,9


The survey was designed for the NISP study. The results presented here focus on infant sleeping position, with the dependent variable based on the response to the question, “Do you have a position you usually place your baby in?” Factors examined as independent variables that might influence sleep position were chosen based on previous studies identifying factors associated with sleep position including year, maternal demographic variables (maternal age, race, and education, household income, region of country), child variables (child age, sex, prematurity status, and sleep location), maternal concerns about infant comfort and choking, and physician's advice about sleep position (categorized as recommending exclusively supine, not exclusively supine, or no advice).26

Participants were asked about their race and/or ethnicity because of health disparities seen in SIDS and in infant care practices. All participants were asked, “Which of the following best describes [the mother’s] racial or ethnic background?” The interviewer then read, “White, African American, Hispanic, Asian, Native American, and some other race.” If they chose some other race, the participant was asked to specify. It was also documented if the participant refused to answer or said that they did not know.


The 15-year time period was examined. We hypothesized that factors associated with the choice to place the infant in the supine sleep position would change over time. The data were analyzed in 3 equal parts to have sufficient numbers for analysis and to detect changes over time.

Descriptive statistics were calculated including frequencies and percentages. The χ2 test was used to test for differences in proportions across demographic subgroups. Trends over time in the supine, prone, and lateral sleep positions were examined through logistic regression fitting linear change in log-odds over time, controlling for race. Racial differences in trends over time were examined through interaction models. Plots of sleep position over time suggest a change in trend lines at 2001, so separate models examined trends from 1993 to 2000 and from 2001 to 2007. Multiple logistic regression was also used to examine associations between supine sleep position and maternal and infant factors, controlling for categorized year from 1993 to 2007. This study was approved by the Institutional Review Boards at Boston University School of Medicine and at Yale University School of Medicine.


When asked about infant sleep position, “Do you have a position you usually place the baby to sleep,” 90% replied affirmatively. Figure 1 shows the changes in usual sleeping position by race/ethnicity for the period 1993 through 2007. Between 1993 and 2000, there was a clear increase in supine sleeping and a decrease in prone sleeping in each of the racial/ethnic groups. Logistic regression models showed significant increases in supine sleep (P < .001) and significant decreases in prone sleep (P < .001), with no significant differences in the trends by race (tests for race × time interaction gave P = .09 for supine and P = .71 for prone position). However, the African American population has consistently had the lowest use of the supine position for sleep and the highest use of the prone position (P < .001 for comparison with white individuals based on logistic regression analyses). Hispanic individuals did not significantly differ from white individuals regarding prone sleep (P = .73). Use of the lateral sleep position has been relatively stable during the entire 15-year period, with only minor differences between racial/ethnic groups. Since 2001, there has been little change in sleep position practices, with no significant change in sleep position over time (P = .16 for supine sleep; P = .37 for prone sleep) and no significant differences in trends over time by race (P = .44 for interaction between race and time for supine sleep; P = .21 for prone sleep). Supine sleep has reached a plateau of approximately 75% and 58%, and prone sleeping position has reached a plateau of approximately 10% and 20% in the white and African American populations, respectively.

Figure 1.
Image not available

Percentage of study infants usually placed in the supine (A), prone (B), or lateral (C) positions from 1993 through 2007.


To assess the factors associated with usual supine sleep position, we initially performed a multiple logistic regression analysis in which the explanatory variables included survey year, geographic region, and fixed characteristics of mothers and infants (maternal age, education, race, income and parity categories, and infant age and prematurity status categories). The results of these analyses are provided in Table 1, with the adjusted odds ratio (95% confidence interval) for each variable shown. Survey year is the strongest predictor of supine sleep position, with odds ratios, compared with the reference year of 1993, steadily increasing from 1.8 in 1994 to 13 in 2001; however, there was little change between 2001 and 2007. Other characteristics associated with greater likelihood of reporting usual supine sleep position included older maternal age, race other than African American, higher maternal educational level, higher maternal income level, mother not having other children, geographic region other than the Southern United States, older infant age, and infant being born after more than 37 weeks' gestation.

Table 1. 
Image not available
Adjusted Odds Ratios of Usual Supine Sleep Position for 1993-2007a

To assess the extent to which certain maternal attitudes or practices may also affect usual supine sleep position, we performed a second multiple logistic regression analysis in which the following potential explanatory variables were added to those described above: usual sleep location (bassinet, crib, adult bed, other), reported maternal concern about infant choking, reported maternal concern about infant comfort, and reported physician advice received (exclusively supine, not exclusively supine, or no advice). The adjusted odds ratios (95% confidence interval) for this second analysis are shown in Table 1. Three of these 4 added variables (ie, all but usual sleep location) were strongly associated with usual supine sleep position. Although only 10% of mothers reported concerns about choking, mothers who did not report this concern had 5 times the odds of reporting usual supine sleep position. Almost 38% of mothers reported a concern about infant comfort, with those not reporting this concern having 4 times the odds of choosing usual supine sleep position. Only one-third of mothers reported receiving positive advice from their physician to use the supine sleep position, with one-third reporting negative advice regarding supine position and one-third reporting receiving no advice. Mothers who received positive advice had 3 times the odds of reporting usual supine position compared with either negative advice or receiving no advice.

When taking the 4 additional variables into account, the magnitude of the odds ratios of many of the other variables was reduced such that maternal age and income no longer reached statistical significance, and the magnitude of the odds ratios for survey years 1999-2007 were reduced by approximately 50%.


To assess the extent to which the factors associated with usual supine sleep position have changed over time, we repeated the above multiple logistic regression analyses with analyses restricted to each of 3 successive 5-year time windows: 1993-1997, 1998-2002, and 2003-2007. Table 2 shows the results of these analyses. Important findings from these analyses include the following:

Table 2. 
Image not available
Adjusted Odds Ratios for Usual Supine Sleep Position in Consecutive 5-Year Time Windows

  • In the earliest time window, survey year is a strong predictor of supine sleep position; however, the association is less strong in the middle time window, and within the most recent time window there is no longer an increase over time. There is even a suggestion of decreasing use of supine position with time. Comparing 2003 and 2007, in the most recent time period, there are statistically fewer infants placed in the supine position in 2007.

  • In the most recent time period (2003-2007), most of the demographic variables that were associated with supine sleep position in earlier time periods are no longer significantly associated with supine sleep position. The only variables that remain consistently significant from 2003 through 2007 are living in a Midwestern state and Hispanic ethnicity. The differences in use of supine sleep position can be explained almost entirely by the following variables: maternal concern about comfort, maternal concern about choking, and advice from a physician to place the infant in the supine position for sleep.

  • The prevalence of maternal concerns about infant comfort and choking with regard to sleep position have decreased over time. Concerns about infant choking decreased from 16.8% to 7.1% to 6.3% of mothers from the earliest to the most recent time period (P < .001, χ2 test). Similarly, concerns about infant comfort decreased from 49.4% to 34.1% to 30.5% during these same time periods (P < .001). However, the importance of these factors in predicting supine sleep position increased over time such that, in the most recent time period, mothers who were not concerned about choking had 8 times the odds of reporting usual use of supine position, and mothers not concerned about comfort had 12 times the odds or reporting usual use of the supine position.

  • The prevalence of mothers reporting positive advice from a physician regarding supine position increased over time from 5.8% in the earliest time period to 36.9% to 53.6% in the most recent time period (P < .001). Receiving positive advice remains important in the most recent time period, with mothers receiving positive advice having 2.62 times the odds of reporting usual supine position compared with no advice, and those receiving negative advice having 30% lower odds of reporting usual supine position compared with those receiving no advice.


Through this study we learned a number of things about choice of infant sleeping position. First, in examining the graphs of the time trends, it is clear that there is a plateau in supine sleep for all racial/ethnic groups. It is particularly concerning that the trend could be heading in the opposite direction.

Second, fewer African American infants are placed in the supine position for sleep than white infants. Assuming a strong connection between placing an infant in the supine position for sleep and SIDS, African American infants die each year because of lack of adherence to the Back to Sleep message.

It is useful to illustrate this point by making a simple calculation using data from the NISP Web site and SIDS rates from the National Center for Vital Statistics Web site.8,10Figure 2 shows the actual number of infant deaths among African American infants in the US from 1997 through 2001 in the solid line. The dashed line shows the theoretical number of deaths of African American infants if the SIDS rate achieved for the pooled years 2002 to 2004 for African American infants had been achieved as early as 1997. The key point is that if improvement in the rate of supine sleeping had been achieved in African American infants by 1997, as it had in white infants, and if it had been accompanied by the SIDS rate that was actually observed in African American infants from 2002 through 2004, then 719 fewer African American infants would have died during this 5-year period (a decrease of 18%).

Figure 2.
Image not available

Cumulative number of sudden infant death syndrome (SIDS) deaths of African American infants. Solid triangles indicate the actual number of SIDS deaths; open triangles, SIDS deaths calculated using the pooled 2002 to 2004 SIDS rate.

Third, from 2003 and 2007, the difference in supine sleep between African American and white infants can be explained, at least in part, by caregiver concern about choking and comfort. While the prevalence of concern about comfort and choking decreased markedly over time, the relative importance of these attitudes as predictors of sleep position has increased.

Finally, receiving advice from a physician for supine sleep position has remained a strong predictor of supine sleeping over time and has markedly increased in prevalence. However, more than 45% of mothers reported either receiving no advice from their physician or receiving advice in favor of the nonsupine position.

Potential limitations of this study include using telephone surveys, which may result in underrepresentation of minority and low-income care providers. However, we believe that our findings would have remained significant given the strength of concerns about comfort, choking, and advice from a physician. Additionally, all data are based on caregiver report, which may not reflect actual behavior. Finally, fewer parents of young infants were included in the study, likely owing to the delay that occurs in obtaining information about the birth of a child. Despite these limitations, we believe that this large sample is the best available to track infant care practices and likely reflects the practice of many infant caregivers across the United States.

To reduce death rates, we must ensure that public health measures reach the populations at highest risk and include messages that address concerns about infant comfort or choking. We must remain vigilant about tracking trends in and parental attitudes about infant care practices, as we are seeing evidence of slippage in adherence to the recommendations.

Back to top
Article Information

Correspondence: Eve R. Colson, MD, Department of Pediatrics, Yale University School of Medicine, PO Box 208064, New Haven, CT 06520 (

Accepted for Publication: July 15, 2009.

Author Contributions: Drs Colson and Corwin had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Colson, Smith, Lister, and Corwin. Acquisition of data: Colson, Smith, and Corwin. Analysis and interpretation of data: Colson, Rybin, Smith, Colton, Lister, and Corwin. Drafting of the manuscript: Colson, Colton, and Corwin. Critical revision of the manuscript for important intellectual content: Colson, Rybin, Smith, and Lister. Statistical analysis: Rybin and Colton. Obtained funding: Colson and Corwin. Administrative, technical, and material support: Smith and Lister. Study supervision: Smith and Corwin.

Financial Disclosure: None reported.

Funding/Support: This study was supported in part by National Institute of Child Health and Human Development grant U10 HD029067-09A1C.

Additional Contributions: We thank Timothy Hereen, PhD, Marian Willinger, PhD, and Isabelle VonKohorn, MD, for their thoughtful reviews of the manuscript.

Matthews  TJMacDorman  MF Infant mortality from 2004 period linked birth/death data set. Natl Vital Stat Rep 2007;55 (14) 1- 32
American Academy of Pediatrics Task Force on Sudden Infant Death Syndrome, The changing concept of sudden infant death syndrome: diagnostic coding shifts, controversies regarding the sleeping environment, and new variables to consider in reducing risk. Pediatrics 2005;116 (5) 1245- 1255
Colson  ERLevenson  SRybin  D  et al.  Barriers to following the supine sleep recommendation among mothers at four centers for the Women, Infants, and Children Program. Pediatrics 2006;118 (2) e243- e250
Willinger  MKo  CWHoffman  HJKessler  RCCorwin  MJ Factors associated with caregivers' choice of infant sleep position, 1994–1998: the National Infant Sleep Position Study. JAMA 2000;283 (16) 2135- 2142
Brenner  RASimons-Morton  BGBhaskar  B  et al.  Prevalence and predictors of the prone sleep position among inner-city infants. JAMA 1998;280 (4) 341- 346
Willinger  MHoffman  HJWu  KT  et al.  Factors associated with the transition to nonprone sleep positions of infants in the United States: The National Infant Sleep Position Study. JAMA 1998;280 (4) 329- 335
The American Association for Public Opinion Research, Standard Definitions: Final Dispositions of Case Codes and Outcome Rates for Surveys. 3rd ed. Lenexa, Kansas American Association for Public Opinion Research2004;
 National Infant Sleep Position Public Access Web Site. Accessed March 3, 2009 
Martin  JAHamilton  BESutton  PD  et al. Centers for Disease Control and Prevention National Center for Health Statistics National Vital Statistics System, Births: final data for 2005. Natl Vital Stat Rep 2007;56 (6) 1- 103
 Linked Birth and Infant Death Data. Centers for Disease Control and Prevention, National Vital Statistics System Web site. Accessed July 6, 2009