Abstract
Background. This study examines whether the relationship between maternal stress or abuse situations and infant birth weight differs between homeless and non-homeless women. Methods. Analyses are based on data from the Pregnancy Risk Assessment Monitoring System (PRAMS), 2002–2007. Results. Homeless women were significantly more likely to experience stressful life events, abusive situations, and poor maternal health than non-homeless women during pregnancy. Birth weight among infants of homeless women was, on average, 17.4 grams lighter than for infants of non-homeless women, after adjusting for maternal age, race, ethnicity, region, education, and marital status. The impact of maternal health, stress, and abuse variables on pregnancy and infant birth weight significantly interacted with homeless status. For example, vaginal bleeding, nausea, kidney/bladder infection, and failure to receive early prenatal care had significantly larger negative impacts on birth weight among homeless women than non-homeless women. Infant birth weight was consistently lower among homeless women, more so when maternal stress and abuse were involved, across all classifications of their prepregnancy weight. Conclusion. Stress and abusive situations among pregnant women have a negative influence on pregnancy-related conditions and infant birth weight. However, this negative influence is even more pronounced among homeless women.
1. Introduction
The scientific literature shows a consistent association between homelessness and stressful life events, especially among women [1–4]. The lower socioeconomic status experienced by many homeless women has been shown to be a major determinant of stressful life events [5]. Research has also shown that pregnant homeless women are exposed to more violence and abuse than their low-income, non-homeless counterparts [6]. Results from one case-control study in Worcester, Massachusetts showed that homeless mothers were more likely to experience severe sexual and physical assault than low-income housed mothers [7]. The same study also showed that homeless mothers had smaller social networks and experienced increased residential instability [7]. In a New Hampshire case-control study comparing homeless mothers with low-income housed mothers, homeless mothers were significantly more likely to report higher levels of stress and depression [8]. Stress and abuse are therefore associated with homelessness in women.
Prenatal maternal stress and physical abuse have been shown to negatively influence pregnancy outcomes [9–13]. Researchers from the University of California found that women with higher self-reported levels of stress and anxiety were significantly more likely to deliver low birth weight and preterm children [10]. A similar prospective study from London found that self-reported depression or “high nerves” in pregnant mothers was associated with low birth weight [11]. Physical violence from an intimate partner has also been linked to antepartum hemorrhage, intrauterine growth restriction, and perinatal death in the fetus [12]. Partner abuse during pregnancy increases the risk of stress and is associated with a host of risk behaviors like smoking or using drugs, which further increase the chance of preterm or low birth weight infants [13]. In addition, women experiencing partner abuse are less likely to obtain prenatal care [13]. Hence, such pregnancy-related outcomes are linked to stress and abuse experienced by the mother.
Research has shown adverse birth outcomes related to homelessness, including low birth weight and premature infants, which are strong risk factors for infant mortality and developmental disabilities [14–16]. However, to our knowledge, no study has considered the modifying effect of homelessness on the relationship between maternal psychological stress and physical abuse and infant birth weight. In the current study, we examine whether the relationship between maternal stress/abuse and infant birth weight differs between homeless and non-homeless women.
2. Methods
Analyses were based on data from 31 areas in the United States participating in the Pregnancy Risk Assessment Monitoring System (PRAMS), 2000–2007. Of 272,859 women who completed the PRAMS survey during these years, approximately 1% did not have information on homeless status. These women were dropped from the current study, leaving 268,876 for assessment.
2.1. PRAMS Weighting Process
Because PRAMS areas typically oversample low weight births and stratify by the mother’s race and ethnicity, the three weighting variables provided in the PRAMS data file (sampling weight, non-response weight, and non-coverage weight) were used. These weights are described in detail elsewhere [17]. The product of the three 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. Weighted estimates were calculated to obtain a representation of all women who gave birth in each state during the specified years. Statistical evaluation of the PRAMS data involved statistical software that takes into account the complex sampling design (i.e., complex survey modules of SAS).
2.2. Study Variables
Homeless status was based on responses to the following: “This question is about things that may have happened during the 12 months before your new baby was bornI was homeless.” Demographic variables considered in this study were maternal age, race, ethnicity, US geographical region, maternal education, and marital status. Pregnancy- and morbidity-related variables include tobacco smoking and alcohol use, previous live birth, previous low birth weight, previous premature birth, prenatal care, preterm labor, vaginal bleeding, nausea, and kidney/bladder infection. Body Mass Index (BMI) is 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?” BMI was classified as underweight (<18.5), normal weight (18.5–24.9), overweight (25–29.9), class I & II obesity (30–39.9), and class III obesity (≥40). Several psychological stress and physical abuse variables obtained in the originally requested data set were considered. Psychological stress variables included the following stressors: a family member’s illness, being divorced or separated, moving to a new address, having a partner lose their job, the survey respondent losing their job, arguing with their partner more than usual, the respondent’s partner not wanting the pregnancy, being unable to pay bills, the survey respondent being in a physical fight, the survey respondent or their partner going to jail, having someone close to the respondent with drug or alcohol problems, and having someone close to them die. Four physical abuse variables were considered: abuse before pregnancy from a partner, abuse before pregnancy from someone other than a partner, abuse during pregnancy from a partner, and abuse during pregnancy from someone other than a partner. Physical abuse was defined on the PRAMS survey as being pushed, hit, slapped, kicked, choked or physically hurt in any other way by their partner.
2.3. Statistical Techniques
Bivariate analyses were performed between homeless status and selected demographic, pregnancy and morbidity, BMI, stress, and abuse variables. Relationships were assessed for significance using the Rao-Scott chi-square test. Average infant birth weight was compared between homeless and non-homeless women using analysis of variance, adjusted for maternal age, race, ethnicity, marital status, education, area, stress, and abuse variables. Logistic regression models were calculated, regressing homeless status on the demographic, pregnancy and morbidity, BMI, stress, and abuse variables. Odds ratios obtained from logistic regression were adjusted for maternal age, race, ethnicity, marital status, education, and geographical region. Poststratification weights, described above, were applied to obtain representative population-based estimates of all homeless women giving birth in the PRAMS geographic areas. Ninety-five percent confidence intervals were calculated, and two-sided tests of significance were used, based on the 0.05 level. Statistical analyses were derived using complex survey modules in Statistical Analysis System (SAS) software, version 9.2 (SAS Institute Inc., Cary, NC, USA, 2007).
3. Results
The estimated number of births in the PRAMS areas for the specified years is presented in Table 1. The percentage of homeless women giving birth varied from 2 to 7. Overall, 4% of women who completed the PRAMS survey were homeless.
Homeless status is presented for selected pregnancy and morbidity variables in Table 2. Homeless women were significantly more likely to have previously given birth, to have a previous low birth weight child or to have a premature birth, to experience preterm labor, to have vaginal bleeding, nausea, and kidney/bladder infection, and to not receive prenatal care as early as wanted.
Homeless status is presented according to selected stressful life events in Table 3. Each stressful event was significantly greater among homeless women. Homeless women were significantly more likely to have an ill family member, to experience divorce or separation, to move to a new address, to lose their job or to have a partner who lost a job, to argue more than usual with a partner, to have a partner who did not want the pregnancy, to be unable to pay bills, to have been in a physical fight, to go to jail or to have a partner who went to jail, to have someone else close to them with drug or alcohol problems, and to have someone close to them that had died.
Homeless status is presented according to selected abuse variables in Table 4. Homeless women were significantly more likely to experience physical abuse. Abuse from a partner or anyone else before or during pregnancy was about 4 to 6 times greater among homeless women.
Birth weight is presented by homeless status in Table 5. Birth weight among infants to homeless women was an average of 17.4 grams lighter than for infants of non-homeless women, after adjusting for maternal age, race, ethnicity, region, education, and marital status. Birth weight according to pregnancy and health variables and stress and abuse variables tend to significantly interact with homeless status (Table 6). For example, infant birth weight was greater among women with a previous live birth, but significantly more so among non-homeless women. A similar result is observed for previous low birth weight, previous premature birth, and preterm labor. Vaginal bleeding, nausea, kidney/bladder infection, and failure to receive early prenatal care had significantly larger negative impacts on birth weight among homeless women compared with non-homeless women. Infant birth weight was consistently lower among homeless women across all classifications of their pre-pregnancy weight. Finally, experiencing stress or abuse among homeless women compared with non-homeless women tended to have a much more negative impact on infant birth weight.
4. Discussion
The current study showed the regional variation in homelessness and confirms other studies in showing that homelessness is related to poorer pregnancy and health outcomes, stressful life events, and abusive conditions [7, 8]. In addition, as in other studies, lower birth weight was associated with homelessness [9].
This study goes beyond previous studies by evaluating how stress and abuse are related to pregnancy and health outcomes based on homeless status. In this study, homelessness modified the effect of maternal pregnancy practices and outcomes, as well as stressful and abusive life events on infant birth weight. The overall lower infant birth weight and preterm birth among homeless women as compared with non-homeless women was also reported in a Canadian prospective cohort study by Little et al. [16]. The modifying impact of homelessness on adverse maternal health and infant birth weight in our study may have been due to the severity of adverse health circumstances reported by homeless women, such as vaginal bleeding, kidney disease, and nausea. Studies have found that homeless women engage more frequently in risky health behaviors, such as smoking, and experience more barriers to prenatal care, including cost and location; such correlations may contribute to the modifying effect of homelessness on poor maternal health and child outcomes [7, 8, 18].
Likewise, the degree of abuse and stress experienced by homeless women may have contributed to findings that homeless women who reported such symptoms were more likely to have low birth weight children than housed women. Studies have shown that homeless women are more likely to experience stress and violence than housed women [6, 8]. Sexual violence or prostitution in particular could have played a role in unwanted pregnancies and poor maternal health behaviors, such as increased smoking or drinking, which in turn could lead to lower birth weight children. In addition, a study by Stein et al. shows that adverse child health outcomes are seen more frequently in mothers whose homeless situation is more prolonged or severe during pregnancy [14]. Due to the effect of homelessness on stress and abuse, it is not surprising that homeless women in our study who reported these factors were more likely to have more serious health problems than housed women who reported similar circumstances.
This research presents new insight into the prevalence of maternal pregnancy and health conditions, stress, and abuse among homeless women in the United States by identifying the modifying effect of homelessness. The results provide public health professionals with pertinent information about this high-risk population and what measures can be undertaken to improve conditions for pregnant homeless women. It can also serve as an impetus for policy-level recommendations regarding public health programs and practices for pregnant women experiencing homelessness within the United States.
The researchers recognize some limitations related to this study. Based on the homelessness question presented on the PRAMS survey, the duration (e.g., number of weeks, months) of homelessness and whether the mother was still experiencing homelessness at the time of survey administration cannot be determined. In addition, the questions related to stress and to abuse failed to measure the duration of experiencing these factors. Some additional limitations include the self-reported and retrospective nature of the PRAMS data and that there was not information on whether homeless women resided in shelters, with family or friends, or on the street.
5. Conclusion
Homeless women are more likely to previously have experienced a low birth weight child, preterm labor, vaginal bleeding, nausea, and kidney/bladder infection, and less likely to have received early prenatal care. In addition, stressful life events and abusive relationships are more common in homeless women. Infant birth weight among homeless women is, on average, 17.4 grams lighter than in non-homeless women. The negative impact of selected pregnancy and morbidity variables (previous low birth weight, premature birth and preterm labor, receipt of early prenatal care, vaginal bleeding, nausea, kidney/bladder infection) and stress or abuse variables on infant birth weight tended to be significantly greater among homeless women.
Declaration of Interest
The authors have nothing to report.
Acknowledgments
PRAMS Working Group: Alabama—Albert Woolbright, PhD; Alaska—Kathy Perham-Hester, MS, MPH; Arkansas—Mary McGehee, PhD; Colorado—Alyson Shupe, PhD; Delaware—George Yocher, MS; Florida—Marie Bailey, MA, MSW, MPH; Georgia—Carol Hoban, Ph.D, MS, MPH; Hawaii—Mark Eshima, MA; Illinois—Theresa Sandidge, MA; Louisiana—Joan Wightkin; Maine—Tom Patenaude; Maryland—Diana Cheng, MD; Massachusetts—Hafsatou Diop, MD, MPH; Michigan—Violanda Grigorescu, MD, MSPH; Minnesota—Judy Punyko, PhD, MPH; Mississippi—Marilyn Jones, M.Ed; Missouri—Venkata Garikapaty, MSc, MS, PhD, MPH; Montana—JoAnn Dotson; Nebraska—Brenda Coufal; New Jersey—Lakota Kruse, MD; New Mexico—Eirian Coronado, MPH; New York State—Anne Radigan-Garcia; New York City—Candace Mulready-Ward, MPH; North Carolina—Paul Buescher, PhD; North Dakota—Sandra Anseth; Ohio—Connie Geidenberger; Oklahoma—Alicia Lincoln, MSW, MSPH; Oregon—Kenneth Rosenberg, MD; Pennsylvania—Tony Norwood; Rhode Island—Sam Viner-Brown, PhD; South Carolina—Mike Smith; South Dakota Tribal—Christine Rinki, MPH; Texas—Kate Sullivan, PhD; Tennessee—David Law, PhD; Utah—Laurie Baksh; Vermont—Peggy Brozicevic; Virginia—Marilyn Wenner; Washington—Linda Lohdefinck; West Virginia—Melissa Baker, MA; Wisconsin—Katherine Kvale, PhD; Wyoming—Angi Crotsenberg; CDC PRAMS Team, Applied Sciences Branch, Division of Reproductive Health.