Context: The success and simplicity of the 1994 national "Back to Sleep" campaign to reduce sudden infant death syndrome provides an opportunity to study which elements determine whether a behavior will change in the desired direction in response to a public health intervention.
Objective: To examine sociodemographic characteristics, motivation, and message exposure to ascertain which factors influenced a caregiver's choice of infant sleep position after implementation of the campaign.
Design: Annual nationally representative telephone surveys conducted between 1994 and 1998.
Setting: The 48 contiguous United States.
Participants: Nighttime caregivers of infants born within the 7 months prior to interview between 1994 and 1998. Approximately 1000 interviews were conducted each year.
Main outcome measures: The position the infant was usually placed in for sleep, sleep position recommendations received from specific sources, and reasons reported for position choice.
Results: Between 1994 and 1998, prone placement declined from 44% to 17% among white infants and from 53% to 32% among black infants. Supine placement increased from 27% to 58% among white infants and from 17% to 31% among black infants. During this period, reports of supine recommendations from at least 1 source doubled from 38% to 79%. From 1995 to 1998, 86% of caregivers who placed the infant prone reported receiving only nonprone recommendations. Infant comfort was given as a reason for prone placement by 82% of these caregivers. In multivariate analysis, physician recommendation of "supine not prone" had the strongest influence and was associated with decreased prone placement (odds ratio [OR], 0.25 [95% confidence interval [CI], 0.16-0.39]) and increased supine placement (OR, 3.37 [95% CI, 2.38-4.76]). Recommendations from all 4 sources (the physician, neonatal nurse, reading materials, and radio/television) further increased the probability of supine placement (OR, 6.01 [95% CI, 4.57-7.90]). Other factors independently associated with increased prone and decreased supine placement included maternal black race, parity of more than 1, and living in a southern or mid-Atlantic state.
Conclusions: According to our study, as of 1998, approximately one fifth of infants were still placed prone, and only half were placed supine. Recommendations of supine placement during infancy by physicians at well-baby checks and by neonatal nursery staff and print and broadcast media have increased the proportion of infants placed supine. Caregiver beliefs regarding perceived advantages of prone sleeping should be addressed to attain further reduction in prone placement.