TABLE 1.

Overview of the included studies evaluating homelessness or housing instability and pregnancy outcomes

Study no.	Title	Author, y	Study design	Population	Sample size	Outcomes	Results	Quality of the evidence (GRADE)
Exposure: maternal homelessness
1	Association between homelessness and hospital readmissions––an analysis of 3 large states	Khatana et al,21 2020	Retrospective cohort	Recently postpartum individuals in Florida, Massachusetts, and New York from 2010 to 2015	N=92,720	Hospitalization owing to complications of pregnancy, childbirth, and the puerperium within 30-d and 90-d postpartum	• 30-d readmission rates for which the cause of index hospitalization was complications of pregnancy, childbirth, and the puerperium were 0.2% points lower (P=.01) in the homeless group.  • Similar trends were seen for 90-d readmission rates.	Low
2	Characteristics of mothers and infants living in homeless shelters and public housing in New York City	Reilly et al,22 2019	Cross-sectional	Individuals and newborns in New York City from 2008 to 2014	N=41,076; 3228 births matched to homeless shelter addresses 37,848 births matched to stable housing	Low birthweight (<2500 g), preterm birth (GA <37 wk), NICU admission, breastfeeding following delivery, infant death within 1 y, infant discharge status	• Neonates born to mothers living in a shelter were more likely to have a low birthweight (13.3% vs 12.0%; P=.02), be preterm (14.5% vs 12.8%; P<.0001), be admitted to the NICU (14.8% vs 12.6%; P<.0001), less likely to exclusively breastfeed (77.4% vs 80.8%; P<.001), and be discharged home at the same time as the mother (73.4% vs 80.1%; P<.0001)when compared with mothers not living in shelters.  • There was no significant association between maternal homelessness and infant death within 1 y (1.1% vs 0.8%; P=.07).	Very low
3	Effects of maternal homelessness, supplemental nutrition programs, and prenatal PM (2.5) on birthweight	Rhee et al,23 2019	Retrospective cohort study	Individuals and children enrolled in Boston-based Children’s HealthWatch cohort from 2007 to 2015	N=3366; 524 pregnant individuals who are homeless and 2842 pregnant individuals who are not homeless	Infant birthweight (g)	Maternal homelessness during pregnancy was associated with a 56 g lower birth-weight (95% CI, −97.8 g to −13.7 g).	Low
4	Health behaviors and infant health outcomes in homeless pregnant women in the United States	Richards et al,2 2011	Retrospective cohort study	Recently postpartum individuals participating in PRAMS 31 states/cities from 2000 to 2007	N=10,671,258; 441,528 individuals who are homeless and 10,229,730 individuals who are not homeless	Prenatal visits in first trimester, GA at delivery, infant birthweight (g), infant length of time in hospital, NICU admission, breastfeeding initiation after delivery, breastfeeding duration	• Pregnant individuals who are homeless were less likely to have a prenatal visit during the first trimester (aOR, 2.0; 95% CI, 1.9–2.2) and to breastfeed their infant (aOR, 1.4; 95% CI, 1.2–1.6).  • Pregnant individuals who are homeless were less likely to not have an NICU admission (aOR, 0.8; 95% CI, 0.7–0.8).  • Mean birthweight was lower for individuals who are homeless than for individuals who are not homeless (3242 g vs 3311 g; P<.001).	Low
5	Homelessness contributes to pregnancy complications	Clark et al,18 2019	Retrospective cohort study	Pregnant individuals with Medicaid coverage in Massachusetts from January 2008 to June 2015	N=9124; 4379 pregnant individuals who were homeless and 4745 pregnant individuals who were not homeless	Hypertension complicating pregnancy, iron deficiency and other anemia, polyhydramnios, hemorrhage during pregnancy, early or threatened labor, other complications of birth affecting mother	• Individuals who are homeless had significantly higher odds of pregnancy-related conditions, such as hypertension complicating pregnancy (aOR, 1.5; 95% CI, 1.3–1.6), iron deficiency and other anemia (aOR, 1.3; 95% CI, 1.2–1.4), polyhydramnios (aOR, 1.7; 95% CI, 1.6–1.9), hemorrhage (aOR, 1.9; 95% CI, 1.7–2.0), early or threatened labor (aOR, 1.9; 95% CI, 1.8–2.1), and other complications of birth (aOR, 2.6; 95% CI, 2.4–2.8).	Low
6	Homelessness during pregnancy: a unique, time-dependent risk factor of birth outcomes	Cutts et al,1 2015	Cross-sectional	<2 y postpartum individuals recounting experiences during pregnancy from 5 US cities from 2009 to 2011	N=9666; 580 individuals with any prenatal homelessness and 9086 consistently housed individuals	Preterm birth (GA <37 wk), low birthweight (<2500 g)	• Individuals with any antenatal homelessness had greater odds of preterm birth (aOR, 1.43; 95% CI, 1.1–1.8). There was no association between prenatal homelessness and birthweight (aOR, 1.24; 95% CI, 0.98–1.56).	Very low
7	Prenatal maternal stress and physical abuse among homeless women and infant health outcomes in the United States	Merrill et al,24 2011	Retrospective cohort	Recently postpartum individuals participating in PRAMS 31 states/cities from 2000 to 2007	N=10,586,614; 434,534 individuals who are homeless and 10,152,080 individuals who are not homeless	Prenatal care as early as wanted, preterm birth (<37 wk GA), birthweight	• Individuals who are not homeless had higher odds of prenatal care early as wanted (aOR, 1.76; 95% CI, 1.63–1.90) and lower odds of preterm labor (aOR, 0.73; 95% CI, 0.68–0.79) than in individuals who are homeless.  • Birthweight among infants of individuals who are homeless was an average of 17.37 g (standard error, 1.01; P<.001) lower than for infants of individuals who are not homeless.	Low
8	Severity of homelessness and adverse birth outcomes	Stein et al,26 2000	Cross-sectional	Individuals of reproductive age in Los Angeles who reported a live birth within the last 3 y	Total N is unknown; 974 homeless individuals compared with national norms	Low birthweight (<2500 g), birthweight (continuous), preterm birth (<37 wk GA), GA at delivery	• Almost 17% of the sample reported neonates who weighed <2500 g at birth as opposed to the national average of 6%.	Very low
9	The reproductive experience of women living in hotels for the homeless in New York City	Chavkin et al,271987	Retrospective cohort	Singleton births in New York City from 1982 to 1984	N=255,206; 401 individuals living in homeless hotels, 13,247 individuals living in low-income housing, and 241,558 individuals citywide	Prenatal visits, infant birthweight (g), infant mortality	• Individuals living in hotels had significantly fewer prenatal visits than the low-income housing group and citywide group.  • There was an estimated reduction of 125 g in the neonatal birthweight associated with hotel residence and a reduction of 48 g in the neonatal birthweight associated with low-income housing project residence (P<.001).  • The hotel group had a greater relative risk than the low-income housing project and citywide groups (respective RR, 1.4; P=.01; RR, 2.07; P<.01). The hotel residents had a 2.5-fold reduction in the likelihood of getting prenatal care than the low-income housing group and a 4.12-fold reduction in the likelihood than the citywide population.	Very low
	Exposure: maternal housing instability
10	Associations between unstable housing, obstetrical outcomes, and perinatal healthcare utilization	Pantell et al,28 2019	Retrospective cohort	Births from singleton pregnancies in California from 2007–2012	N=5588; 2794 births to people with unstable housing code with propensity-matched controls	Preterm birth (GA<37 wk), GA at birth, preterm labor, preeclampsia, chorioamnionitis, placental abruption, small for gestational age, long length of stay, maternal ED visit within 3 mo and 1 y after delivery, maternal readmission within 3 mo and 1 y after delivery	• Individuals with unstable housing had greater odds of preterm birth (aOR, 1.2; 95% CI, 1.0–1.4) and preterm labor (aOR, 1.4; 95% CI, 1.2–1.6). No relationship with any other obstetrical outcome was found.  • Individuals with unstable housing were more likely to have a long length of stay after childbirth (OR, 1.6; 95% CI, 1.4—1.8), an emergency department visit within 3 months (OR, 2.4; 95% CI, 2.1–2.8) and 1 y (OR, 2.7; 95% CI, 2.4–3.0,) after delivery.  • Individuals with unstable housing were more likely to have a readmission within 3 mo (OR, 2.7; 95% CI, 2.2–3.4) and 1 y after delivery (OR, 2.6; 95% CI, 2.2–3.0).	Moderate
11	Housing instability and birthweight among young urban mothers	Carrion et al,29 2015	Retrospective cohort study of a cluster randomized controlled trial	Second trimester pregnant individuals between 14 and 21 y old at community hospitals and health centers in New York City	N=613; 175 with housing instability and 438 with stable housing	Infant birthweight (g)	• Housing instability was significantly associated with lower birthweight.  • On average, infants of housing stable individuals weighed 3155.96 g (SD, ±532.69), whereas infants of housing instable individuals weighed 3028.17 g (SD, ±641.18).	Low
12	Maternity shelter care for adolescents: its effect on incidence of low birthweight	LaGuardia et al,30 1989	Prospective cohort	Adolescents (19 y and younger) who delivered at the New York Hospital-Cornell Medical Center in New York City from 1984 to 1986	N=225; 112 individuals residing in a shelter; 113 individuals not residing in a shelter (ie, housed) individuals	Low birthweight (<2500 g), preterm birth (<37 wk), maternal ICU admission, average number of days in hospital per person with preterm labor	• Sheltered individuals had lower rates of low birthweight neonates and preterm birth than the control group.  • The sheltered group spent a greater number of days in hospital after preterm labor (11.4% vs 5.8%; P<.025).	Low
13	Predicting preterm birth among women screened by North Carolina’s Pregnancy Medical Home Program	Tucker et al,31 2015	Retrospective cohort	Individuals participating in North Carolina’s Pregnancy Medical Home Program screened between 6–24 wk gestational age from 2011 to 2012	N=15,427; 956 individuals with unstable housing; 14,471 individuals with stable housing	Preterm birth (<37 wk GA)	• Individuals with unsafe or unstable housing have greater odds of preterm birth than in individuals with stable housing (aOR, 1.26; 95% CI, 1.04–1.53).	Very low
14	Severe housing insecurity during pregnancy: association with adverse birth and infant outcomes	Leifheit et al,25 2020	Retrospective cohort	Mother-infant dyads enrolled in the Fragile Families and Child Wellbeing Study in 20 large US cities from 1998 to 2000	N=3428; 55 severely housing insecure mother-infant dyads; 3373 mother-infant dyads who were not experiencing severe housing insecurity	Low birthweight (<2500 g), preterm birth (<37 wk GA), NICU admission	• Individuals experiencing severe housing insecurity during pregnancy had 1.73 times higher risk of having low birthweight neonates and/or preterm birth and 1.64 times higher risk for NICU admission than individuals who did not experience severe housing insecurity (95% CI, 1.28–2.32; 95% CI, 1.17–2.31, respectively).	Low

aOR, adjusted odds ratio; CI, confidence interval; GA, gestational age; ICU, intensive care unit; PRAMS, Pregnancy Risk Assessment Monitoring System; RR, relative risk; SD, standard deviation.