Vital Signs

Trends and Disparities in Infant Safe Sleep Practices — United States, 2009–2015

Jennifer M. Bombard, MSPH; Katherine Kortsmit, PhD; Lee Warner, PhD; Carrie K. Shapiro-Mendoza, PhD; Shanna Cox, MSPH; Charlan D. Kroelinger, PhD; Sharyn E. Parks, PhD; Deborah L. Dee, PhD; Denise V. D'Angelo, MPH; Ruben A. Smith, PhD; Kim Burley; Brian Morrow, MA; Christine K. Olson, MD; Holly B. Shulman, MA; Leslie Harrison, MPH; Carri Cottengim, MA; Wanda D. Barfield, MD


Morbidity and Mortality Weekly Report. 2018;67(1):39-46. 

In This Article


Data Source

PRAMS[10] collects state-specific, population-based data on self-reported maternal behaviors and experiences before, during, and shortly after pregnancy. In each participating state, a stratified random sample of women with a recent live birth is selected from birth certificate files, and women are surveyed 2–6 months postpartum using a standardized protocol and questionnaire. PRAMS data for each site are weighted for sampling design, nonresponse, and noncoverage to produce a data set representative of the state's birth population. PRAMS sites were included in this report if their weighted response rate was ≥65% for years 2009–2011, ≥60% for 2012–2014, and ≥55% for 2015.

PRAMS sites included the question, "In which position do you most often lay your baby down to sleep now?" (check one answer): "on side; on back; on stomach." Respondents who selected "on side" or "on stomach" were classified as placing their infant in a nonsupine sleep position.* Analyses on nonsupine sleep positioning were conducted using 2015 data from 32 PRAMS states and New York City. To explore trends in nonsupine sleep position, CDC analyzed PRAMS data from 2009–2015 in 15 states.§ Analyses of bed sharing used 2015 data from 14 states that included the optional question on their state-specific PRAMS survey: "How often does your new baby sleep in the same bed with you or anyone else?" Respondents who indicated "always," "often," "sometimes," or "rarely" were classified as having bed shared and were compared with respondents who indicated "never." Bed sharing was also categorized as: "rarely or sometimes," and "often or always." Analyses of soft bedding used 2015 data from 13 states** and New York City that included the following optional question on their state-specific survey: "Listed below are some things that describe how your new baby usually sleeps." Respondents were asked to select "yes" or "no" for the following soft bedding items: "pillows," "thick or plush blankets," "bumper pads," "stuffed toys" and "infant positioner." Respondents who selected "yes" to one or more items were defined as using any soft bedding.

Statistical Analysis

The weighted prevalence and 95% confidence intervals of unsafe sleep practices were calculated overall and by state for 2015. Chi-square tests and 95% confidence intervals†† were used to determine differences in unsafe sleep practices by maternal characteristic (i.e., race/ethnicity, age, education level, and participation in the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) program during pregnancy), gestational age at birth (i.e., preterm, <37 weeks' gestation, compared with term, ≥37 weeks' gestation) and any breastfeeding at 8 weeks postpartum. CDC tested for linear trends in nonsupine sleep position overall and by maternal characteristics and state, from 2009 to 2015, using logistic regression. Analyses accounted for the complex survey sampling design of PRAMS.

*A small percentage of respondents (<4%) selected more than one sleep position. Respondents selecting multiple positions were classified as placing their infant in a nonsupine sleep position. Denominator includes supine, on stomach, on side only, and combinations of any of the three positions.
The 32 states include Alabama, Alaska, Arkansas, Colorado, Connecticut, Delaware, Hawaii, Illinois, Iowa, Louisiana, Maryland, Massachusetts, Michigan, Missouri, Nebraska, New Hampshire, New Jersey, New Mexico, New York (excluding New York City), Ohio, Oklahoma, Oregon, Pennsylvania, Tennessee, Texas, Utah, Vermont, Virginia, Washington, West Virginia, Wisconsin, and Wyoming.
§The 15 states include Delaware, Hawaii, Illinois, Maryland, Massachusetts, Missouri, Nebraska, New Jersey, Oklahoma, Pennsylvania, Utah, Vermont, Washington, West Virginia, and Wyoming.
The 14 states include Alaska, Connecticut, Delaware, Louisiana, Nebraska, New Jersey, Pennsylvania, Tennessee, Texas, Vermont, Virginia, Washington, West Virginia, and Wisconsin.
**The 13 states include Alaska, Illinois, Iowa, Louisiana, Maryland, Michigan, Missouri, New Jersey, New York (excluding New York City), Pennsylvania, Tennessee, West Virginia, and Wyoming.
††To provide general guidance on the statistical differences, 95% CIs for the prevalence were compared across groups, with an emphasis on identifying differences (i.e., nonoverlap of CIs) between categories within the selected variables. This typically conservative approach might fail to note differences between estimates more often than formal statistical testing. Overlap between confidence intervals does not necessarily mean there is no statistical difference between estimates.