OBJECTIVE: To describe selected demographics; maternal health behaviors before, during, and after pregnancy; and infant health outcomes among homeless women in the United States.
PATIENTS AND METHODS: Analyses are based on Pregnancy Risk Assessment Monitoring System data from 31 participating states from 2000 to 2007 that were linked to birth certificate data, which contain demographic and medical information collected through the state's vital records system. Responses were then weighted to be representative of all women who gave birth in each state during that year. Assessment of these data takes into account the complex sampling designs employed by the states.
RESULTS: Four percent of women reported homelessness within 12 months before pregnancy, with the highest percentage in Illinois, followed by Oregon and Washington. Homeless women were younger, unmarried, uninsured, less educated, less likely to initiate and sustain breastfeeding, and had less prenatal care and well-visits. They were also more likely to be black, Hispanic, smoke cigarettes, be underweight or have class III obesity, and not take preconception multivitamins. Infants had lower birth weights, a longer hospital stay, and were more likely to receive neonatal intensive care.
CONCLUSIONS: In this study we provide additional insight into homelessness in the perinatal period and provide information on ways to direct interventions aimed at improving the health of homeless mothers and infants. Additional research is needed to determine factors that influence pregnancy weight gain and infant feeding practices among homeless women and how this affects infant health.
WHAT'S KNOWN ON THIS SUBJECT:
Homeless pregnant women tend to be younger, less educated, less likely to get prenatal care, and have a higher likelihood of relying on government assistance and delivering premature, low-birth weight infants.
WHAT THIS STUDY ADDS:
This study provides insight about the effects of homelessness on maternal and infant health in the United States. Homeless women had less preconception supplement intake, less breastfeeding initiation and duration, and a higher likelihood of being underweight or obese.
More than 640 000 individuals experience homelessness on any given night in the United States.1 Recent estimates indicate that 37% of these are families, which is an increase of 4% from 2008 to 2009.1 Also, more than half of homelessness occurs in 6 states across the nation, including 3 western states (California, Nevada, and Washington) and 3 southern states (Georgia, Florida, and Texas).1 In one study involving data from the Pregnancy Risk Assessment Monitoring System (PRAMS), it was found that women of lower socioeconomic status were 13.1 times more likely to experience homelessness before or during pregnancy than those of higher socioeconomic status.2 Another study reported that 11.4% of childbearing women in an eastern metropolitan area had experienced homelessness within a 7-year period surrounding their infants' birth.3
Pregnancy is a critical time period since many health outcomes for the child are strongly influenced by maternal health behaviors. Homelessness, in general, presents greater challenges for individuals to engage in healthy behaviors, such as optimal dietary intake and access to adequate medical care.4,–,8 However, little is known about maternal health behaviors both during pregnancy and in the postpartum period, and the related infant health outcomes among homeless women in the United States. In some studies it has been shown that experiencing homelessness during pregnancy increases the likelihood of having inadequate prenatal care and delivering infants with low birth weights, prematurity, or developmental disability disorders.4,9 However, these studies were based on small samples sizes or small geographical regions in the United States, thus providing a limited understanding of the effect of homelessness on pregnancy in this nation. Thus, the purpose of this study is to describe (1) selected demographic characteristics according to geographical region, (2) maternal health behaviors before, during, and after pregnancy, and (3) infant health outcomes among homeless women in the United States. Results will be compared with nonhomeless women.
PATIENTS AND METHODS
This study was based on PRAMS data from 31 participating states/cities from 2000 to 2007. PRAMS is an ongoing, state-specific surveillance system that collects data about health behaviors before, during, and after pregnancy from mothers who delivered a live-born infant 2 to 4 months before survey administration.10 Analyses included data from states/cities with at least a 70% response rate for each individual year and among women who reported being homeless sometime during the 12 months before their infant's delivery. Approximately 1300 to 3400 women complete the survey each year within each participating state/city. In the PRAMS data set there were 11 310 (4%) homeless women, 257 566 (95%) not homeless, and 3983 (1%) without information about homeless status. Those with missing information about homeless status were excluded from the current study.
Within each participating PRAMS state/city, a systematic sample of 100 to 250 new mothers is taken every month from mothers who recently gave birth, from a frame of eligible birth certificates. States typically oversample low weight births and stratify by the mother's race and ethnicity. Three weighting variables are provided in the PRAMS data file: a sampling weight; a nonresponse weight; and a noncoverage weight. These weights are described in detail elsewhere.10 Multiplying together the sampling, nonresponse, and noncoverage components of the weights produces the analysis weight, which is interpreted as the number of women like themselves in the population that each respondent represents.
The mother's responses are linked to selected variables from birth certificate data, which contain demographic and medical information collected through the state's vital records system. Responses are then weighted to be representative of all women who gave birth in each state during that year. Analysis of the PRAMS data involved statistical software that takes into account the complex sampling designs that the states employ. Specifically, survey procedures in SAS were used to compute and assess estimates that take into account the complex survey sample design, which involved stratification, clustering, and unequal weighting. In the current study, the weighted number of nonhomeless women was 10 229 730, and of homeless women was 441 528.
Independent and Outcome Variables
Homeless status was based on the following question: “This question is about things that may have happened during the 12 months before your new infant was born … you were homeless.” Demographic variables include maternal age, race, ethnicity, education, and marital status; number of children in household; geographic region; whether government aid had been received in the previous 12 months; whether the mother had health insurance; and whether she received Medicaid. Maternal health behaviors included smoking status, alcohol drinking status, prepregnancy BMI, use of prenatal vitamins, number of prenatal care visits, breastfeeding duration, infant sleep position, and well-baby checkups. Although maternal illicit drug use during pregnancy has been associated with adverse infant health outcomes,11 this variable was not included in our analyses because these data are not collected on the PRAMS survey. Infant health outcomes included gestational age, infant birth weight, infant length of time in the hospital, and whether the infant attended the NICU.
To determine breastfeeding duration, the following PRAMS survey question was used: “Are you still breastfeeding or feeding pumped milk to your new infant?” If women responded “yes,” then duration of breastfeeding was estimated by subtracting the infant's date of birth from the date the PRAMS survey was filled out by the participant. If women responded “no,” then duration of breastfeeding was determined from the following PRAMS survey question: “How many weeks or months did you breastfeed or pump milk to feed your infant?” Previous research has shown good reliability and validity of maternal recall about breastfeeding initiation and duration, especially if recalled within 3 years after their infant's birth.12 BMI (kg/m2) was derived from self-reported height and weight questions: “Just before you got pregnant, how much did you weigh?” and “How tall are you without shoes?” Because BMI was based on self-reported heights and weights, and research has shown that women in the age range 15 to 44 tend to underestimate their weight, BMI is likely underestimated.13,14 BMI-based weight classifications are underweight (BMI < 18.5), normal weight (BMI 18.5–24.9), overweight (BMI 25–29.9), class I and II obesity (BMI 30–39.9); and class III obesity (BMI ≥ 40).
Frequency distributions were used to describe the prevalence of selected variables according to homeless status. The Rao-Scott χ2 was used to assess significant difference in distributions of selected variables by homeless status. Multiple logistic regression was used to assess the odds of being homeless according to selected variables, adjusting for maternal age, race, ethnicity, and region. Multiple regression was also used to assess the simultaneous effects of homeless status and other selected variables on birth weight (g) and number of prenatal care visits. Poststratification weights were used to adjust for nonresponse and noncoverage of households and for deriving representative population-based estimates in the United States. Statistical significance was based on the 0.05 level. Statistical analyses were derived from SAS 9.2 (SAS Institute, Inc, Cary, NC).
The percentage of homeless women with a live birth is presented in Fig 1 for each PRAMS city/state within 4 regions of the United States. The highest percentage of pregnant women who had been homeless within the past 12 months was in Illinois, followed by Oregon and Washington. New Jersey, Florida, Colorado, and New Mexico also had relatively high percentages of pregnant women who were homeless. Overall, 4% of PRAMS participants identified themselves as homeless sometime during the previous 12 months.
Sociodemographic characteristics of PRAMS participants are presented in Table 1 according to homeless status. Homeless compared with nonhomeless pregnant women were significantly more likely to be younger, black, Hispanic, reside in the Midwest or West, have less than a high school education, be single, receive government aid, not have health insurance, have Medicaid, smoke cigarettes, be underweight or have class III obesity, and not take preconception multivitamins.
Selected maternal health behaviors (before, during, and after pregnancy) and infant health outcomes are presented in Table 2 according to homeless status. Homeless women compared with nonhomeless women were significantly less likely to have a prenatal visit during the first trimester, breastfeed their child, and have a well-baby checkup. In addition, infants of homeless women had significantly longer stays in the hospital, were more likely to require intensive care, and had lower duration of breastfeeding.
Gestational age was similar between homeless and nonhomeless women, ∼38.6 weeks, but mean birth weight was significantly lower for homeless women (3242 g [7.1 lb] versus 3311 g [7.3 lb], P < .001), and the number of prenatal care visits was significantly lower for homeless women (10.0 vs 11.6, P < .001), after adjusting for maternal age, race, ethnicity, and region. The prevalence of very low birth weight infants (< 1500 g [3.3 lbs]) was significantly greater for homeless women (1.6% vs 1.3%, P < .001), and the prevalence of low birth weight infants (1500–2499 g [5.5 lbs]) was also significantly greater in homeless women (7.1 vs 5.9, P < .001). Regression models for birth weight and number of prenatal care visits are presented in Table 3 for selected independent variables. Birth weight was significantly lower for homeless women, those not receiving prenatal care, those not having well-baby checkups, those with younger maternal age, among nonCaucasians and Hispanics, and in the Southern and Western regions of the United States. In addition, the number of prenatal care visits was significantly lower among homeless women, those not having well-baby checkups, those with younger maternal age, nonCaucasians and Hispanics, and in the Western region of the United States.
Homelessness in our study was associated with women being black, Hispanic, unmarried, without health insurance/Medicaid, receiving government aid, and not taking a multivitamin supplement before pregnancy. Other studies have similarly reported that homeless pregnant women were less educated, younger, unmarried, but more likely to be black and use government assistance programs.1,4,8 In addition, in our study we found that homelessness among pregnant women was most common in the Midwest (Illinois) and Pacific Northwest (Oregon and Washington) regions. Because previous studies have evaluated homelessness during pregnancy in selected regions of the United States, the current study offers a more comprehensive portrayal of the scope of this public health problem nationwide and provides information that may be useful in directing interventions and effectively allocating resources to improve the health of homeless mothers and infants.
Experiencing homelessness at some point before delivering can pose challenges for women in engaging in healthy behaviors. Our study demonstrated that inadequate use of a multivitamin supplement before conception is a common phenomenon, which behavior may impede healthy fetal growth and lead to impaired nutritional status for the pregnant woman. It is well-established that inadequate folic acid within the first 28 days of pregnancy increases the risk of infants born with neural tube defects (NTDs).15 However, the extent of NTDs in infants born to the homeless women in our sample is not known because specific information about NTDs is not collected on the PRAMS survey. It is likely that women in our study had limited funds to purchase multivitamin supplements, were less educated to the importance of vitamins/minerals during pregnancy, and had more pressing needs and stresses related to their homeless condition.6,16,17 Because of health concerns about nutrient inadequacy commonly observed among those experiencing homelessness, a recent innovative intervention was conducted in France to offer homeless men with a nutrient-rich chocolate-flavored food packet (Vitapoche) to improve overall health.18 Such an intervention to supplement homeless childbearing age women in the United States may provide an effective vehicle to promote healthier maternal and infant health outcomes.
Homeless pregnant women in this study were less likely to have prenatal care in the first trimester of pregnancy, and had less prenatal care visits overall, which has been consistently observed in other studies involving homeless pregnant women.6,8 Inadequate prenatal care has been associated with adverse infant health outcomes, including low-birth weight, preterm delivery, small-for-gestational age, and infant mortality.4,9,19 In a consistent manner, the current results showed that homeless women were more likely than nonhomeless women to have low-birth weight infants and to have infants that had a longer hospital stay, spent some time in the NICU, and received fewer well-baby checkups after birth. Furthermore, compared with nonhomeless women, those who experienced homelessness were less likely to initiate breastfeeding and had shorter duration of breastfeeding past 8 weeks, which may further compromise an infant's health.
The results reveal that better initiatives are needed to encourage breastfeeding among homeless populations. Human breast milk provides optimal nutrition for an infant's health and may provide an economic benefit to homeless mothers.20,21 Because homeless women in our study were less likely to breastfeed longer than 8 weeks, most likely they relied on alternative food sources, such as infant formula, which may pose food safety challenges because of a limited ability to sterilize water and infant bottles and inadequate storage facilities associated with the homeless condition.16 Homeless mothers, especially those living on the streets, may have to resort to inappropriate means of preparing infant formula because they may be unable to heat water or sterilize utensils. Future research should evaluate infant feeding practices among homeless mothers and how they affect overall health of the infant. It is also possible that homeless women had to cease breastfeeding as they returned to the workforce.22 Policy interventions should be developed to enforce time allocation and space for women to pump breast milk in an employment setting as mandated by 2010 health care legislation.23
Our findings indicated that homeless women were more likely to be either underweight or have class III obesity compared with their nonhomeless counterparts. Previous studies in which anthropometric data among homeless individuals have been evaluated have revealed mixed results related to underweight and overweight status.24,25 However, in more recent analyses the hunger-obesity paradox has been suggested as fairly prevalent among homeless individuals.16,26,–,28 It is unclear from our study what factors contributed to homeless women's likelihood of having inadequate or excess weight status before pregnancy. In some research it has been suggested that the environment in which a person experiences homelessness may affect weight and overall health status.5 Because our study could not differentiate between residencies, it is possible that residency location (ie, homeless shelter, living on the streets or in a car) could affect daily caloric intake and ultimately influence weight gain or loss. In previous research it has been suggested that the homeless condition can impede access to healthy food choices because of poor food quality and inability to find economically priced nutrient-dense foods in surrounding neighborhood food stores, inadequate space to store and cook food, and shelter rules that encourage the consumption of high-calorie foods.5,16,29,–,31 Thus it is possible that poor nutritional intake, either inadequate or excess calories, could offer an explanation for the underweight and obesity status observed among the homeless women in our sample.
It is also unclear how prepregnancy weight may have influenced weight gain during pregnancy and its effect on the infant health outcomes observed in our study. Because women from our study were less likely to have health insurance or Medicaid and were less likely to access prenatal care, it is probable that they did not obtain adequate information about appropriate weight gain during pregnancy, which may be responsible for the negative infant health outcomes after delivery. This suggests the need for future research to determine the effect of homelessness on pregnancy weight gain and the subsequent immediate and long-term health outcomes for the mother and infant.
This study has certain limitations. Although the PRAMS survey asked whether the mother had been homeless sometime during the 12 months before the delivery of their child, it did not ask about the duration of homelessness and whether they were currently homeless. Duration would directly relate to the adverse effects of homelessness and duration and current homeless status could influence recall. In addition to recall bias potentially influencing the results, lack of survey anonymity could have biased responses. Residency of women at the time of survey administration was also unknown, so in this study we were unable to determine if they were homeless and living in a shelter environment or living with friends or family. However, because the survey is administered only to women with a known mailing address or home telephone number, the data likely underestimate homelessness among pregnant women. It may also under-represent women who were homeless for a longer duration.
Homelessness in the United States presents a situation that can adversely affect maternal health behaviors during pregnancy and subsequent infant health outcomes. Homelessness among women in this study was associated with inadequate prenatal care, less preconception multivitamin use, fewer well-baby checkups, less initiation and duration of breastfeeding, and higher likelihood of being underweight or having class III obesity. Infants born to mothers experiencing homelessness had a lower birth weight, longer hospital stays, and higher likelihood of being in the NICU after delivery. Future research should focus on gaining a better understanding of the impact of homelessness on other related health behaviors, such as factors influencing pregnancy weight gain and infant feeding practices, and how these behaviors impact maternal health status and an infant's well-being.
The PRAMS Working Group comprises Albert Woolbright, PhD (Alabama); Kathy Perham-Hester, MS, MPH (Alaska); Mary McGehee, PhD (Arkansas); Alyson Shupe, PhD (Colorado); George Yocher, MS (Delaware); Marie Bailey, MA, MSW, MPH (Florida); Carol Hoban, PhD, MS, MPH (Georgia); Mark Eshima, MA (Hawaii); Theresa Sandidge, MA (Illinois); Joan Wightkin (Louisiana); Tom Patenaude (Maine); Diana Cheng, MD (Maryland); Hafsatou Diop, MD, MPH (Massachusetts); Violanda Grigorescu, MD, MSPH (Michigan); Judy Punyko, PhD, MPH (Minnesota); Marilyn Jones, MEd (Mississippi); Venkata Garikapaty, MSc, MS, PhD, MPH (Missouri); JoAnn Dotson (Montana); Brenda Coufal (Nebraska); Lakota Kruse, MD (New Jersey); Eirian Coronado, MPH (New Mexico); Anne Radigan-Garcia (New York); Candace Mulready-Ward, MPH (New York, NY); Paul Buescher, PhD (North Carolina); Sandra Anseth (North Dakota); Connie Geidenberger (Ohio); Alicia Lincoln, MSW, MSPH (Oklahoma); Kenneth Rosenberg, MD (Oregon); Tony Norwood (Pennsylvania); Sam Viner-Brown, PhD (Rhode Island); Mike Smith (South Carolina); Christine Rinki, MPH (South Dakota); Kate Sullivan, PhD (Texas); David Law, PhD (Tennessee); Laurie Baksh (Utah); Peggy Brozicevic (Vermont); Marilyn Wenner (Virginia); Linda Lohdefinck (Washington); Melissa Baker, MA (West Virginia); Katherine Kvale, PhD (Wisconsin); Angi Crotsenberg (Wyoming); and the Centers for Disease Control and Prevention PRAMS Team, Applied Sciences Branch, Division of Reproductive Health.
- Accepted May 5, 2011.
- Address correspondence to Rickelle Richards, PhD, MPH, RD, Assistant Professor, Department of Nutrition, Dietetics, and Food Science, Brigham Young University, S-233 ESC, Provo, UT 84602. E-mail:
Dr Richards made substantial contributions to the conception and design of the study and the acquisition and interpretation of the data used, in drafting and revising the manuscript, and approved the final version submitted to Pediatrics; Dr Merrill made substantial contributions to the conception and design of the study and the acquisition, analysis, and interpretation of the data and in drafting and revising the article, and approved the final version submitted to Pediatrics; and Ms Baksh made substantial contributions to the conception and design of the study and the acquisition of the data, in reviewing and revising the article critically, and approved the final version submitted to Pediatrics.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
- PRAMS —
- Pregnancy Risk Assessment Monitoring System
- NTD —
- neural tube defect
United States Interagency Council on Homelessness. Opening doors: federal strategic plan to prevent and end homelessness. 2010. Available at: www.usich.gov/pdf/openingdoors_2010_fsppreventendhomeless.pdf. Accessed August 19, 2010
- Little M,
- Shah R,
- Vermeulen MJ,
- Gorman A,
- Dzendoleta D,
- Ray JG
Centers for Disease Control and Prevention. Detailed PRAMS methodology. Available at: www.cdc.gov/PRAMS/methodology.htm. Accessed September 4, 2010
- Pitkin RM
- Bartick M,
- Reinhold A
Public Law 111–148. The Patient Protection and Affordable Care Act. March 23, 2010. Available at: www.gpo.gov/fdsys/pkg/PLAW-111publ148/pdf/PLAW-111publ148.pdf. Accessed April 18, 2011
- Copyright © 2011 by the American Academy of Pediatrics