Pregnancy and the responsibilities of caring for children create additional barriers for low-income women and families seeking adequate and affordable housing, yet few studies have assessed the relative risk or prevalence of homelessness among pregnant or parenting women in the United States.1 The economic, social, and psychological dislocations associated with housing instability and homelessness may have particularly serious negative consequences for the health and well-being of pregnant and parenting women and their children,2–4 highlighting the need to document the number of women exposed to such conditions. Enumerating homeless populations is fraught with both difficulty and controversy5; however, effective public health planning and advocacy are problematic in the absence of a better understanding of the scope of the homelessness problem among pregnant and parenting women.
In light of the need for relevant data, we sought to estimate the prevalence of homelessness among childbearing women living in a single, large urban area within the continental United States.
METHODS
Prevalence of homelessness was assessed based on public shelter use and birth record data, which were available for a defined cohort of Philadelphia, Pa, resident childbearing women. Shelter use data are routinely collected as a function of the city of Philadelphia’s centralized public shelter intake system, and use of such data to obtain estimates of homeless prevalence has been discussed in previous publications.5–7 For this study, electronic shelter intake records pertaining to any official request for emergency shelter occurring from January 1, 1990, through December 31, 1998, were linked to electronic birth records for all women older than 18 years who gave birth in 1993 or 1994. For women who delivered more than 1 live infant, only 1 randomly selected delivery record was chosen as the index birth. Women for whom an official record of a request for emergency shelter existed, in either the 3-year period preceding or the 4-year period following the index birth, were defined as having had a homeless episode. Homeless episode prevalence rates during this 7-year period were then calculated for the entire 1993 to 1994 cohort, and for selected subgroups of the cohort, based on demographic and other characteristics available in the birth record.
RESULTS
The results are presented in Table 1 ▶. Of the total 44 430 women in the cohort, 11.4% had a documented homeless episode in either the 3 years before or the 4 years following the index birth. Prevalence rates varied dramatically by race/ethnicity, previous live births, and maternal education. More than 20% of all African American adult women had a documented homeless episode during the 7-year period, compared with much lower rates for White (1.2%), Asian (0.9%), and Hispanic (4.8%) women. For each racial/ethnic group, the rates increased with greater parity, and within each racial/ethnic and parity group, the rates tended to be higher for those with fewer compared with more years of education. Nearly one third (32.3%) of the African American women with 3 or more previous live births and almost half (45.2%) of all African American women with 4 or more previous live births had homeless episodes during the 7-year period surrounding the index birth. Prevalence rates were substantially higher for African American women within each parity and education subgroup of the cohort. Although African American women represented only about 50% of the women in the study cohort, they accounted for more than 90% of all homeless episodes that occurred between 1990 and 1998.
TABLE 1—
Seven-Year Homeless Episode Prevalence Rates for Adult Resident Childbearing Women, by Race/Ethnicity, Parity, and Education
| African American (n = 22 457) | White (n = 15 558) | Asian (n = 2090) | Hispanica (n = 4195) | Allb (N = 44 430) | |
| Parity = 0 | 9.8% (648) | 0.6% (41) | 0.5% (4) | 2.5% (32) | 4.7% (727) |
| Education, y | |||||
| 0–11 | 24.1% (213) | 3.3% (19) | 0.6% (1) | 3.4% (14) | 12.2% (247) |
| ≥ 12 | 7.6% (435) | 0.4% (22) | 0.4% (3) | 2.0% (18) | 3.6% (480) |
| Parity = 1 | 16.1% (1064) | 0.8% (43) | 0.7% (5) | 3.6% (47) | 8.5% (1162) |
| Education, y | |||||
| 0–11 | 31.6% (427) | 3.0% (19) | 0.0% (0) | 6.6% (32) | 18.3% (478) |
| ≥ 12 | 12.1% (637) | 0.5% (24) | 0.9% (5) | 1.8% (15) | 6.2% (684) |
| Parity = 2 | 24.8% (1105) | 1.5% (37) | 0.8% (2) | 5.0% (41) | 14.8% (1191) |
| Education, y | |||||
| 0–11 | 40.5% (498) | 3.9% (16) | 0.0% (2) | 6.0% (24) | 25.3% (540) |
| ≥ 12 | 18.9% (607) | 1.0% (21) | 2.3% (0) | 4.0% (17) | 11.0% (651) |
| Parity = 3 | 32.3% (776) | 3.1% (28) | 2.2% (3) | 9.3% (39) | 21.9% (849) |
| Education, y | |||||
| 0–11 | 41.4% (345) | 6.3% (13) | 3.4% (2) | 10.0% (25) | 28.4% (385) |
| ≥ 12 | 27.4% (431) | 2.2% (15) | 1.3% (1) | 8.3% (14) | 18.4% (464) |
| Parity ≥ 4 | 45.2% (1068) | 6.2% (31) | 4.0% (5) | 11.5% (41) | 34.2% (1149) |
| Education, y | |||||
| 0–11 | 53.0% (546) | 10.7% (18) | 3.4% (2) | 12.5% (30) | 39.8% (597) |
| ≥ 12 | 39.3% (522) | 3.9% (13) | 4.4% (3) | 9.5% (11) | 29.7% (552) |
| All | 20.8% (4661) | 1.2% (180) | 0.9% (19) | 4.8% (200) | 11.4% (5078) |
Note. Numbers in parentheses refer to number of women in each subcategory with a documented homeless episode during the 7-year period.
aFigures refer to women who were identified as being of Hispanic origin on the birth record, regardless of race; those who were identified as being of Hispanic origin were not included in preceding columns even if race was also specified.
bFigures refer to all women regardless of race/ethnicity and include some women for whom race/ethnicity was other than African American, White, Asian, or Hispanic or was not stated on the birth record.
DISCUSSION
Access to adequate and affordable housing has significant implications for the health of individuals and the health of populations in several different and important ways.8 The negative consequences associated with lack of access to adequate and affordable housing for women and children may be particularly severe for the individuals involved and have particularly important implications from the standpoint of public health. The significant finding from this study is that such a large number of urban (and overwhelmingly African American and Hispanic) childbearing women are exposed to the disruptions and dislocations associated with becoming homeless at some point during their reproductive careers. In addition, the risk of this exposure is closely tied to the number of children a woman has had, suggesting that childbearing and child-rearing responsibilities render many women much more vulnerable to the vagaries of lowincome housing markets.
More generally, the findings highlight the need for appropriate health and social service delivery systems that adequately address the multiple and overlapping needs of the numerous childbearing women and their families whose housing situation is extremely precarious and who, in the absence of preventive interventions, will surface in need of basic shelter.
Finally, if homelessness can be viewed as the failure of our society to meet even the most basic needs of families,9 then failure appears to occur on a disturbingly large scale among urban minority women with dependent children. This, in turn, lends support to the argument that the lack of adequate low-income housing, housing instability, and homelessness are structural rather than aberrant social problems, serious enough to be ranked high on the list of priority public health concerns.9,10
Acknowledgments
The authors would like to thank both the city of Philadelphia’s Department of Public Health and the Office of Emergency Shelter Services for their cooperation in conducting this study.
Human Participant Protection No protocol approval was needed for this study. Only secondary data sources from which personal identifiers had been removed were involved.
Contributors D. A. Webb, J. Culhane, and D. Culhane suggested and designed the study. D. A. Webb wrote the first and subsequent drafts. S. Metraux and J. M. Robbins assisted with the data analysis and editing of the brief.
Peer Reviewed
References
- 1.Weitzman B. Pregnancy and childbirth: risk factors for homelessness? Fam Plann Perspect. 1989;21:175–178. [PubMed] [Google Scholar]
- 2.Burg MA. Health problems of sheltered homeless women and their dependent children. Health Soc Work. 1994;19:125–131. [DOI] [PubMed] [Google Scholar]
- 3.Chavkin W, Kristal A, Seabron C, et al. The reproductive experience of women living in hotels for the homeless in New York City. N Y State J Med. 1987;87:10–13. [PubMed] [Google Scholar]
- 4.Killion CM. Special health care needs of homeless pregnant women. Adv Nurs Sci. 1995;18:44–56. [DOI] [PubMed] [Google Scholar]
- 5.Culhane D, Dejowski EF, Ibanez J, et al. Public shelter admission rates in Philadelphia and New York City: the implications of turnover for sheltered population counts. Housing Policy Debate. 1994;5:107–140. [Google Scholar]
- 6.Culhane DP. The organization and utilization of the shelter system in Philadelphia: estimating length of stay and turnover. J Health Soc Policy. 1993;4:55–78. [Google Scholar]
- 7.Metraux S, Culhane D, Raphael S, et al. Assessing homeless population size through the use of emergency and transitional shelter services in 1998: results from the analysis of administrative data from nine US jurisdictions. Public Health Rep. 2001;116:344–352. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Fullilove MT, Fullilove RE. What’s housing got to do with it? Am J Public Health. 2000;90:183–184. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Breakey WR. It’s time for the public health community to declare war on homelessness. Am J Public Health. 1997;87:153–155. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Link BG, Susser E, Stueve A, et al. Lifetime and five-year prevalence of homelessness in the United States. Am J Public Health. 1994;84:1907–1912. [DOI] [PMC free article] [PubMed] [Google Scholar]