The objective of this study was to compare rates of infant sleeping position and other risk factors for sudden infant death syndrome from 1991 before the "Back to Sleep" campaign to rates in 1998 after the campaign. We used a cross-sectional risk factor prevalence study of risk factors for the years 1991 and 1998. In North Dakota the prevalence rates of prone sleeping declined 72% for American Indian infants and 62% for Caucasian infants. We were unable to identify a corresponding decline in SIDS in North Dakota for this time period. The relationship between sleeping position and SIDS may be more complex in rural and frontier settings and in American Indian populations than in urban and majority populations. The generalizability of this study is limited by the rural setting and small sample size. Longer term surveillance and additional reports from sites with pre "Back to Sleep" data as a baseline for both SIDS rates and sleeping position will be important to clarify the rate of prone sleeping position and SIDS.