Assessment of Infant Sleeping Position—Selected States, 1996 FREE

MMWR. 1998;47:873-877

1 table omitted

SUDDEN INFANT death syndrome (SIDS) is the leading cause of postneonatal mortality in the United States.¹ In 1992, the American Academy of Pediatrics (AAP) recommended that all healthy babies be put to sleep either on their back or side to reduce the risk for SIDS.² In 1994, a national "Back to Sleep" education campaign was initiated to encourage the public and health-care providers to put babies to sleep on their back or side.³ In November 1996, the AAP modified its policy to preferentially recommend putting infants on their back because of the lower risk for SIDS associated with this position relative to the side position.⁴ To assess adherence to recommendations for infant sleeping position, CDC analyzed population-based data on the usual infant sleeping position for 1996 births by race from 10 states participating in the Pregnancy Risk Assessment Monitoring System (PRAMS). This report summarizes the results of that analysis and indicates that infant sleeping position varied by state and race.

PRAMS is an ongoing, state-based surveillance system of maternal behaviors before, during, and after pregnancy. Each month, PRAMS surveys a random sample of mothers who have given birth during the previous 2-6 months by using stratified, systematic sampling of resident birth certificates. A questionnaire is mailed to each mother, and a second questionnaire is mailed to nonrespondents. Nonrespondents are then contacted by telephone. Most states oversample mothers of low birthweight (<5 lbs, 8 oz [<2500 g]) infants, and four states oversample women of selected racial groups. Details of the survey design, questionnaire, and other operational aspects of the survey have been published.⁵

Mothers were asked, "How do you put your new baby down to sleep most of the time?" Response categories included on the baby's side, back, or stomach. Statistical weights were applied to account for sampling probability, nonresponse, and sampling frame coverage in each state. The state-specific response rate to the entire questionnaire ranged from 71% to 80%. To account for the complex survey design, SUDAAN was used to calculate point estimates and standard errors for each sleeping position by state and maternal race/ethnicity. Women who did not answer the sleeping position question were excluded from the analysis (3.8% of all respondents). Data were analyzed for 15,195 respondents.

The percentage of respondents who reported usually putting their babies to sleep on their stomach varied by state (from 16.0% in Maine to 30.8% in Alabama). In five southern states, the prevalence of the stomach sleeping position was approximately twofold higher than in the states having the lowest percentages (Maine and Washington). The percentage of respondents who reported putting their babies to sleep on their back was highest in Washington (42.9%) and Alaska (40.8%) and lowest in Georgia (24.5%), Florida (25.4%), and South Carolina (25.8%). In most states, respondents usually put their babies to sleep on their side.

The percentage of black mothers who put their babies to sleep on their stomach was 11%-54% higher than that for white mothers; the percentages ranged from 22.5% in Washington to 42.1% in Florida among black mothers, and from 16.1% in Maine to 30.5% in Oklahoma among white mothers. For American Indians in two states (Washington and Oklahoma), 16.0% and 33.9% of respondents, respectively, reported usually putting their babies to sleep on their stomach. The comparable percentage for Alaska Natives was 23.5% in Alaska.

The median age of infants in Oklahoma (132 days) was at least 1 month older than that in all other states except New York (103 days) and South Carolina (117 days). Median infant age in Washington and Maine, where the prevalence of the stomach sleeping position was lowest, was 98 days and 87 days, respectively.

Pregnancy Risk Assessment Monitoring System Working Group. Div of Reproductive Health, National Center for Chronic Disease Prevention and Health Promotion, CDC.

Usual infant sleeping position is monitored periodically to assess the success of efforts to encourage mothers and other caretakers to place babies on their back for sleeping. During 1992-1996, placement on the stomach declined from 70% to 24%, and placement on the back increased from 13% to 35%.⁶ In the PRAMS survey, state-specific prevalence of the stomach sleeping position in 1996 exceeded the national average in five states and was lower than the national average in four states. The variation observed among states may result from differences in infant age at the time the mother responded to the questionaire, the rate of decline since 1992, or the distribution of factors (i.e., maternal age, education, parity, and exposure to health-promotion messages) related to the choice of infant position.

Infants aged ≥16 weeks were more likely to be placed on their stomach than were infants in younger age groups.⁶ However, the relation between the state percentages of babies put to sleep on their stomach and median infant age when mothers responded to the questionnaire was not always consistent. Differences in the rate of decline by state may result from variations in the intensity and effectiveness of efforts to encourage back sleeping through the "Back to Sleep" campaign and other efforts. However, differences in the rate of decline cannot be assessed because state-specific data are not available before 1996. Additional analysis is required to determine whether socioeconomic status, access to health care, or advice by health-care providers in addition to other predictors of infant position are related to the state or race differences found in this report.

The higher rate of stomach sleeping among blacks than whites is consistent with the twofold higher rate reported nationally in a previous study (22% versus 43%).⁶ The rate for Alaska Natives was similar to the national average but still was higher than that for whites in Alaska. In Washington, the rate for American Indians was comparable to that for whites (16.0% and 16.7%, respectively) and is the lowest rate for any racial group in the 10 states. In comparison, in Oklahoma the rate for American Indians was the same as that for blacks (33.9%). These findings suggest that infant sleep positioning practices vary within groups of American Indians and may explain the unequal risk for SIDS found among American Indians.⁷

The findings in this report are subject to at least three limitations. First, PRAMS does not collect information from adoptive mothers or birth mothers who put their infants up for adoption, no longer care for their infants, or are nonresidents of the states in which they gave birth. Second, misclassification of sleep position may have occurred because mothers had difficulty recalling or assigning the sleep position they used most of the time. Because the question solicits only one response, mothers who selected multiple responses to the question were not included in the analysis. Finally, the survey did not include other sleep-related questions such as stability of the initial sleep position during the night and changes in positioning with increasing infant age. Infant age at the time of the mother's response varied by state and may explain why some mothers whose infants were older reported using a stomach position.

Despite these limitations, the findings in this report provide useful data that states can use as a baseline to measure progress toward the national goal of the "Back to Sleep" campaign to reduce the percentage of infants put to sleep on their stomach to ≤10% by 2000.⁴ The 38% decline in SIDS during 1992-1996 in the United States is associated with the substantial declines observed in the percentage of infants put to sleep on their stomach.^2,8

Innovative communication strategies and outreach programs are needed to educate all persons who care for infants, particularly blacks and certain American Indian populations, to reduce the proportion of babies placed to sleep on their stomach. These risk-reduction strategies must consider cultural and other barriers to adopting the recommended infant sleeping position and/or the appropriateness of the health-education message for high-risk groups. In designing outreach programs to promote the recommended infant sleeping position, public health officials also should consider factors that influence a caregiver's behavior, such as advice given by a health-care provider, mother's observation of a newborn's health-care provider, experience with previous children, or presence of a grandmother in the home.^6,8- 9 Decreasing the difference in SIDS rates in high-risk populations will require new educational efforts and the identification and modification of the risk factors that contribute to the disparity in mortality.

MMWR. 1998;47:825-830

1 table omitted

IN THE United States, breast cancer is the most commonly diagnosed malignancy among women and the second leading cause of cancer death.¹ Lack of health insurance coverage often is an important financial barrier to seeking preventive health care such as mammography screenings.^2,3 To assess mammography use and the impact of insurance status on mammography use, state-specific proportions of women aged ≥40 years who reported receiving a mammogram during the preceding 2 years by insurance status were derived using data from the Behavioral Risk Factor Surveillance System (BRFSS) for 1991-1992 and 1996-1997. This report describes the results of this analysis, which indicate that the percentage of women reporting having had a screening mammogram during the previous 2 years increased, but women with insurance were substantially more likely than women without insurance to have had a mammogram.

Forty-six states and the District of Columbia (DC) participated in BRFSS surveys during 1991-1992 and 1996-1997.* Using a multistage sampling design and random-digit dialing, each state conducted monthly telephone interviews sampling noninstitutionalized adults (aged ≥18 years).^3,4 Annual data were weighted to the age, sex, and race distribution of each state's adult population using 1994 census or intercensal estimates. Female respondents aged ≥40 years were asked, "Have you ever had a mammogram?" If the respondent answered "yes," she was asked, "How long has it been since your last mammogram?" and "Was it part of a routine checkup, or was it because of a breast problem other than cancer, or was it because you had already had breast cancer?" Respondents also were asked, "Do you have any kind of health care coverage, including health insurance, prepaid plans such as HMOs, or government plans such as Medicare?" In the 1996 and 1997 surveys, respondents who answered "no" were asked to reconsider the question. For consistency between the surveys, this analysis categorized respondents in 1996 and 1997 who first answered "no" to the insurance question as uninsured, even if they answered "yes" when asked again; the increase in the percentage of persons insured based on "yes" responses on reconsideration of the question was <2%.

Almost all women aged ≥65 years have Medicare coverage.³ However, the aggregated results for all women aged ≥40 years are presented because this format is consistent with prior analyses of trends in mammography coverage using data from the BRFSS and national objectives for breast cancer screening.^5,6 To compensate for the potential affects of the resulting differences in age distributions between insured and uninsured women, estimates were age-adjusted to the age distribution of women in the 1994 BRFSS sample for participating states.

The overall pooled age-adjusted proportion of women with insurance who reported having had a mammogram was 65.2% in 1991-1992 and 70.9% in 1996-1997; the proportion of women without insurance who reported having had a mammogram was 39.6% in 1991-1992 and 46.2% in 1996-1997. In each of the 46 states and DC in both 1991-1992 and 1996-1997, the prevalence of self-reported screening mammography use within the previous 2 years was higher among insured women than among uninsured women; uninsured women represented approximately 9% of the sample in 1996-1997.

Among insured women, from 1991-1992 to 1996-1997, the age-adjusted proportion aged ≥40 years who reported having had a mammogram during the preceding 2 years increased in 43 states. Increases in 26 states were statistically significant; the largest absolute increases in mammography use were in Mississippi (from 51.4% to 65.3%) and Alaska (from 63.9% to 76.4%). Mammography use decreased in three states (Minnesota, Vermont, and Washington), and DC, but the changes were not statistically significant.

Among uninsured women, mammography use increased in 33 states; the increase was significant in six. The largest absolute increases were 31.0% in Alaska (from 33.8% to 64.8%) and 23.9% in New Jersey (from 23.7% to 47.6%). Although there were decreases in 14 states, the only statistically significant decrease was in New Hampshire (from 51.1% to 32.4%; p=0.047).

J Cook, MBA, Alabama; P Owen, Alaska; B Bender, MBA, Arizona; J Senner, PhD, Arkansas; B Davis, PhD, California; M Leff, MSPH, Colorado; M Adams, MPH, Connecticut; F Breukelman, Delaware; C Mitchell, District of Columbia; S Hoecherl, Florida; L Martin, MS, Georgia; A Onaka, PhD, Hawaii; J Aydelotte, Idaho; B Steiner, MS, Illinois; K Horvath, Indiana; A Wineski, Iowa; M Perry, Kansas; K Asher, Kentucky; R Jiles, PhD, Louisiana; D Maines, Maine; A Weinstein, MA, Maryland; D Brooks, MPH, Massachusetts; H McGee, MPH, Michigan; N Salem, PhD, Minnesota; D Johnson, Mississippi; T Murayi, PhD, Missouri; P Feigley, PhD, Montana; M Metroka, Nebraska; E DeJan, MPH, Nevada; L Powers, MA, New Hampshire; G Boeselager, MS, New Jersey; W Honey, MPH, New Mexico; T Melnik, DrPH, New York; K Passaro, PhD, North Carolina; J Kaske, MPH, North Dakota; P Pullen, Ohio; N Hann, MPH, Oklahoma; J Grant-Worley, MS, Oregon; L Mann, Pennsylvania; J Hesser, PhD, Rhode Island; D Shepard, South Carolina; M Gildemaster, South Dakota; D Ridings, Tennessee; K Condon, Texas; R Giles, Utah; C Roe, MS, Vermont; L Redman, MPH, Virginia; K Wynkoop-Simmons, PhD, Washington; F King, West Virginia; P Imm, MS, Wisconsin; M Futa, MA, Wyoming. Epidemiology and Health Svcs Research Br, Div of Cancer Prevention and Control, and Health Care and Aging Studies Br, Div of Adult and Community Health, National Center for Chronic Disease Prevention and Health Promotion, CDC.

The findings in this report indicate that the percentage of women reporting having had a screening mammogram in the previous 2 years has increased over time, and this increase has been observed among both insured and uninsured women. However, women without insurance continue to be substantially less likely than women with insurance to have this procedure. These results underscore the importance of public health activities to increase access to breast and cervical cancer screening services for women who are medically underserved.⁷ If breast cancer mortality is to continue to decrease, then access to mammography for all women, particularly the uninsured, must be enhanced.⁸

The findings in this report are subject to at least three limitations. First, because the BRFSS is a telephone survey, women living in a household without a telephone (5% of U.S. households) are excluded.⁹ Second, the survey's self-reported data may not be consistent with reports of mammography use from medical records. However, studies comparing self-reports with medical records found that the error in self-reporting mammography use is not substantial enough to explain the differences seen in the analyses described in this report.¹⁰ Finally, the response rates within the BRFSS have dropped from 84.1% and 82.9% in 1991 and 1992, respectively, to 77.9% and 76.8% in 1996 and 1997, respectively. Because respondents may differ from nonrespondents, this increase in nonresponse could portend greater bias in later samples.

This study indicates that lack of health insurance decreases the likelihood that a woman will receive a mammogram. This is an important finding given the efforts being made to reduce breast cancer mortality in this country, where a substantial proportion of women lack health insurance. The demonstrated efficacy of regular breast cancer screening with mammography suggests that efforts such as CDC's National Breast and Cervical Cancer Early Detection Program, a comprehensive nationwide program administered through state health departments and American Indian/Alaskan Native tribal organizations, could facilitate the early detection of breast cancer in underserved women.

References 10 available.

*Arkansas, Kansas, Nevada, and Wyoming did not participate.

MMWR. 1998;47:761-762

IN OCTOBER 1997, the Injury Control and Emergency Health Services Section of the American Public Health Association (APHA) conducted a survey of all 50 states, the District of Columbia (DC), and Puerto Rico to assess the availability of external cause-of-injury data in statewide hospital discharge data systems (HDDS), hospital emergency department data systems (HEDDS), and other ambulatory care data systems. The report on the findings of the analysis, How States are Collecting and Using Cause of Injury Data,¹ includes recommendations for improving the quality and availability of statewide injury-related data for injury-prevention activities.

The findings in the survey indicated that (1) 36 states and DC routinely collect external cause-of-injury data in their HDDS, and 23 of the states have laws or mandates requiring external cause-of-injury coding; and (2) 11 states have developed the capacity to provide external cause-of-injury data on injury-related visits in their statewide HEDDS, and nine of those states have laws or mandates requiring external cause-of-injury coding. A coordinated effort among states is needed to develop standard methods for collecting, coding, analyzing, and presenting injury-related data from statewide data systems. Timely dissemination of uniform, population-based injury morbidity data to hospital administrators, public health professionals, and policy makers will enhance their usefulness for injury-prevention efforts.

This survey was funded by the APHA through a mini-grant to the Trauma Foundation at San Francisco General Hospital and was conducted in partnership with CDC's National Center for Injury Prevention and Control (NCIPC) and National Center for Health Statistics. A copy of the report is available from the Office of Statistics and Programming, NCIPC, telephone (770) 488-4656, e-mail jmc1@cdc.gov, or from the Trauma Foundation site on the World-Wide Web, http://www.traumafdn.org/injuries/apha4.html.

References 1 available.