Association of Antenatal Housing Instability with Perinatal Care Utilization and Outcomes

Abstract

Background:

Social determinants of health are important contributors to maternal and child health outcomes. Limited existing research examines the relationship between housing instability during pregnancy and perinatal care utilization. Our objective was to evaluate whether antenatal housing instability is associated with differences in perinatal care utilization and outcomes.

Materials and Methods:

Participants who were surveyed during their postpartum hospitalization were considered to have experienced housing instability if they answered affirmatively to at least one of six screening items. The primary outcome was adequacy of prenatal care measured by the Adequacy of Prenatal Care Utilization index. Maternal, neonatal, and postpartum outcomes, including utilization and breastfeeding, were also collected as secondary outcomes. Multivariable logistic regression models were adjusted for sociodemographic and clinical covariates.

Results:

In this cohort (N = 490), 11.2% (N = 55) experienced housing instability during pregnancy. Participants with unstable housing were more likely to have inadequate prenatal care (17.3% vs. 3.9%; odds ratio [OR] 5.11, 95% confidence interval [CI] 2.15–12.14, p < 0.001), but findings were not significant after adjustment (aOR 1.72, 95% CI 0.55–5.41, p = 0.35). Similarly, postpartum visit attendance was lower for individuals with unstable housing (79.6% vs. 91.2%), but there was no difference in the odds of the postpartum visit attendance after adjustment (OR 0.69, 95% CI 0.29–1.66, p = 0.14).

Conclusions:

There were no statistically significant association with the maternal, neonatal, and other postpartum secondary outcomes. Housing instability appears to be a risk marker that is related to other social determinants of health. Given the range of housing instability experiences, future research must account for specific types and degrees of housing instability and their potential perinatal consequences.

Introduction

Social determinants of health significantly influence both maternal and child health. The complex interactions of social determinants—including issues such as poverty, food access, violence, housing, transportation, and many others—in the context of obstetrical care and outcomes are beginning to be understood and considered points of potential interventions.¹ One important social determinant is housing status.

A recent systematic review identified experiences with episodes of homelessness and housing instability as significantly associated with perinatal outcomes such as low birth weight, preterm birth, neonatal intensive care unit admission, and delivery complications.² Most included studies examined only unhoused status, which is independently associated with preterm birth and low birth weight even after accounting for health care utilization and demographic factors.^3,4 Pregnancy also correlates with increased risk of experiencing homelessness.⁵ However, an episode of homelessness is the most extreme end of the housing status spectrum; the relationship between housing instability and maternal-child health in the perinatal period has been less well studied.

Traditionally, investigations into the relationship between housing and health have categorized participants dichotomously as either homeless or housed. In the wake of the 2008 foreclosure crisis, however, recognition of housing status as a continuum of conditions led to a growing body of literature examining the relationship between housing instability and a variety of health outcomes.⁶ The US Department of Health and Human Services defines housing instability as including at least one of the following conditions: (a) high housing costs (>30% of monthly household income), (b) poor housing quality, (c) unstable neighborhoods, (d) overcrowding, or (e) homelessness.⁷

Other studies examining the relationship between housing instability and health outcomes have used heterogeneous definitions, including greater than two moves in 1 year, falling behind on rent or utility payments, and experiencing eviction or foreclosure.^8

–11 In the general population, housing instability is associated with decreased access to health care, as well as negative health outcomes. For example, a 2006 study found that unstable housing was independently associated with lack of a usual source of care, postponing needed medical care and medications, and higher rates of emergency department visits and hospitalizations.¹² A 2012 study found that housing instability was independently associated with poor or fair self-reported health, major depressive disorder, and anxiety.⁶

Although housing instability has been associated with poor access to care in the general adult population, its association with utilization and adequacy of prenatal care has not been studied. Prior research has demonstrated increased emergency visit utilization and decreased preventive care utilization among patients with housing instability; we therefore hypothesized that housing instability may be associated with decreased prenatal care utilization.⁶ Thus, our objective was to assess whether there is a relationship between antenatal experiences of housing instability assessed by direct patient survey and adequacy of prenatal care, as well as secondary maternal, neonatal, and postpartum outcomes.

Materials and Methods

In this prospective observational study, a cohort of 500 postpartum patients was recruited between June 2018 and February 2020 from a single tertiary care facility in Chicago, IL, during admission to the postpartum floor. Patients were eligible for enrollment if they were 18 years of age or greater, were English-speaking, had a singleton live birth, and had at least one prenatal visit with one of the hospital's ambulatory practices using the same electronic health record. The study was approved by the Northwestern University Institutional Review Board and all participants provided written informed consent.

The target sample size was calculated to detect a twofold increased likelihood of inadequate prenatal care with a power of 0.80 and an alpha of 0.05, estimating a 10%–20% prevalence of housing instability and a 15% prevalence of inadequate prenatal care in the unexposed group based on prior literature.^6,8,13,14 Potentially eligible participants were identified by screening patients admitted to the postpartum floor and then approaching them in their hospital room regarding participation in the study. Enrolled participants were provided a $10 gift card.

Exposure was determined by surveying participants regarding their housing over the past year (Supplementary Appendix S1). Surveys were multiple choice and conducted in English on a touch screen tablet using REDCap electronic data capture tools hosted at Northwestern University Feinberg School of Medicine.^15,16 Participants were given the option to self-complete the survey with study personnel remaining in the room for question clarification and technical assistance, or to have the survey questions read aloud by study personnel and answer verbally. The survey was conducted in English and took 10–20 minutes to complete. We defined experiencing housing instability as reporting one or more of the following in the year before delivery: Two or more moves, greater than 1 month behind on rent or utilities payments, moving in with friends or family for financial reasons, experiencing eviction or foreclosure, and/or experiencing an episode of homelessness.

Due to the lack of a validated scale or uniform definition of housing instability in the published literature, these criteria were selected to achieve a comprehensive definition of housing instability, combining more limited definitions used in prior research.^{3,7
–9,11,12,17} Outcome data were obtained through the same survey and medical record review. Abstraction of data from the medical record and surveys was completed in May of 2020 for the primary analysis, and August of 2020 for the post hoc postpartum analysis.

The primary outcome was the adequacy of prenatal care utilization index (APNCU), as described by Kotelchuck.¹⁸ The APNCU incorporates the timing of initiation of prenatal care and number of visits attended compared to the expected number of visits based on gestational age of entry. Inadequate prenatal care by the APNCU has been associated with increased risk of low birth weight and preterm delivery.^13,18
–20

We selected maternal, neonatal, and care utilization outcomes of interest for the secondary analysis based on a plausible relationship with housing instability based on prior research.^8,12,14 Secondary maternal outcomes were gestational excessive weight gain, cesarean delivery, and hypertensive disorders of pregnancy; neonatal outcomes were preterm birth, large for gestational age status, and neonatal intensive care unit (NICU) admission. Postpartum care utilization outcomes included the following within 12 weeks after delivery: attendance at a comprehensive postpartum clinic visit, emergency department visit, and hospital readmission. Clinical postpartum outcomes were continuation of breastfeeding and postpartum weight change.

Gestational weight gain above recommendations was defined using the Institute of Medicine's 2009 criteria.²¹ Mode of delivery was defined as delivery type documented in the delivery note (cesarean vs. vaginal birth). Participants were considered to have a hypertensive disorder of pregnancy if a diagnosis of gestational hypertension, preeclampsia, or superimposed preeclampsia was made antenatally or during the admission for delivery. Preterm birth was defined as delivery before 37 weeks 0 days gestation. Large for gestational age birth weight was defined as birth weight greater than the 90th percentile for gestational age by World Health Organization (WHO) criteria. NICU admission was defined as admission to the neonatal intensive care unit at the time of delivery.

Postpartum visit attendance was defined as attendance at a comprehensive visit involving discussion of breastfeeding, contraception, mood, and overall maternal health and well-being at 4–12 weeks after delivery.²² Emergency department visits or hospital readmissions were defined as any visit or readmission after being discharged from the hospital within 30 days. Postpartum weight change was defined as the difference between weight at postpartum visit and weight at delivery in kilograms. Breastfeeding was defined as exclusive breastfeeding or breastfeeding with formula supplements, as assessed at the postpartum visit. All primary and secondary outcomes were abstracted from the medical chart by study staff.

Descriptive statistics were calculated for all variables of interest. Categorical variables were summarized with the use of counts and percentages, and continuous variables with means and standard deviations or median and interquartile range, as appropriate. Potential sociodemographic confounders were selected a priori from existing literature and included age, participant-reported race/ethnicity as a proxy for the experience of structural and interpersonal racism, insurance type, education, relationship status, and employment status for the primary outcome.² Maternal and neonatal secondary outcomes were adjusted for age, race, insurance type, education, relationship status, employment, parity, and history of preterm birth. Postpartum outcomes, which were analyzed separately, were adjusted for age and insurance type due to the lack of association of other variables with the exposure in the post hoc sample. We also conducted a sensitivity analysis, which included race in models for the postpartum outcomes.

The multivariable logistic regression with housing instability as outcome and potential confounders as predictors was used to create individual propensity scores, which are predicted probabilities of being housing unstable based on the covariates included for primary outcomes and secondary maternal and neonatal outcomes, respectively. Due to the close association of housing instability with other social determinants of health, propensity scoring was chosen over traditional logistic regression to avoid overcorrection by controlling for traditional sociodemographic covariates, which may limit causal inference due to increased confounding when presence or absence of each covariate in any individual participant is collapsed.^23,24

To test linearity between propensity score and log odds, we added an extra variable: The natural logarithm of the propensity score multiplied by the propensity score. If this extra variable is significant, then it violates the assumption of linearity. In our analysis, there were only four models that violated the linearity assumption. For those models, we performed additional models adjusting square root transformation of the propensity score. When we compared the results for models with propensity score and models with square root of propensity score, the conclusion stayed the same; therefore, our final analysis adjusted for the original propensity score without any transformation.

The primary analyses were performed using logistic regression. The unadjusted and adjusted for propensity score model results are presented. A separate multiple logistic regression was then performed for secondary postpartum outcomes and adjusted for age and insurance directly without calculating propensity score. A similar statistical strategy was then conducted to stratify by participant-reported race/ethnicity (minoritized, e.g., self-reported race/ethnicity as LatinX or Hispanic, Black, and other, vs. nonminoritized, e.g., self-reported race as White). This exploratory post hoc analysis was completed to assess whether the social construct of minoritized race/ethnicity acted as an effect modifier for the postpartum outcomes, rather than treating it as a confounder due to potentially overlapping causal pathways. All p-values presented were for two-sided tests. p-values <0.05 were considered significant. All statistical analyses were performed in SAS version 9.4.

Results

The final analyzed cohort included 490 participants. Ten enrolled participants were excluded due to inability to calculate the primary outcome. The prevalence of housing instability in our cohort was 11.2% (N = 55).

Participants experienced a wide range of types of housing instability (Table 1). Being behind on utilities payments was the most common marker of housing instability, with 56.4% of patients who screened positive reporting it (N = 31). Moving in with friends or family for financial reasons was the second most common (N = 17, 30.9%), followed by being behind on rent (N = 14, 25.5%). The more severe forms of housing instability, including eviction, foreclosure, and episode of homelessness, were the least common, with only three participants reporting these experiences.

Table 1.

Types of Housing Instability Reported in the Past Year

	N	Percentage of housing unstable (N = 55)	Percentage of total cohort (N = 490)
Two or more moves	9	16.4	1.8
Behind on rent	14	25.5	2.9
Behind on utilities	31	56.4	6.3
Moved in with friends or family	17	30.9	3.5
Episode of homelessness	1	1.8	0.2
Eviction or foreclosure	2	3.6	0.4

Participants may answer yes to more than one item.

Participants experiencing antenatal housing instability differed significantly from the analyzed sample in all measured sociodemographic and clinical characteristics (Table 2). A higher proportion of participants experiencing housing instability identified as Black (45.5%) and Hispanic/LatinX (23.6%) than in the stably housed population (16.1% and 13.3%, respectively). A higher proportion of participants experiencing housing instability had Medicaid insurance (67.3% vs. 12.7%), and a lower proportion were married or had a long-term partner (47.3% vs. 90.1%) or reported at least a college education (30.9% vs. 81.6%). Participants experiencing housing instability were less likely to be nulliparous (35.2% vs. 53.1%), and were more likely to have a history of spontaneous preterm birth (13.0% vs. 3.4%), compared to participants with stable housing.

Table 2.

Cohort Characteristics

		All (N = 490)	No housing instability (N = 435)	Housing instability (N = 55)
Age^a, years		34 (31, 36)	34 (31, 37)	29 (27, 34)
Nulliparous^b		250 (51.1)	231 (53.1)	19 (35.2)
Race/Ethnicity^c	Non-Hispanic White	248 (50.6)	235 (54.0)	13 (23.6)
	Non-Hispanic Black	95 (19.4)	70 (16.1)	25 (45.5)
	Hispanic/LatinX	71 (14.5)	58 (13.3)	13 (23.6)
	Other	76 (15.5)	72 (16.6)	4 (7.3)
Education	Some college or less	118 (24.1)	80 (18.4)	38 (69.1)
Education	College and above	372 (75.9)	355 (81.6)	17 (30.9)
Relationship^d	Unpartnered	72 (14.7)	43 (9.9)	29 (52.7)
Relationship^d	Partnered	417 (85.3)	391 (90.1)	26 (47.3)
Employment	Yes	395 (80.6)	364 (83.7)	31 (56.4)
Employment	No/Unknown	95 (19.4)	71 (16.3)	24 (43.6)
Medicaid^e	Yes	92 (18.9)	55 (12.7)	37 (67.3)
Medicaid^e	No	395 (81.1)	377 (87.3)	18 (32.7)
History of preterm delivery^f	Yes	22 (4.5)	15 (3.4)	7 (13.0)
History of preterm delivery^f	No	467 (95.5)	420 (6.6)	47. (87.0)

Data displayed as N (%) or median (quartile 1–3).

19 missing.

One missing.

p-values (Age, p < 0.001; nulliparous, p = 0.013; race/ethnicity, p < 0.001; education, p < 0.001; relationship, <0.001; employment, p < 0.001; Medicaid, p < 0.001; History of preterm delivery, p = 0.006).

One missing.

Three missing.

One missing.

The frequency of inadequate prenatal care among participants experiencing housing instability was 17.3%, and among participants reporting stable housing was 3.9% (Table 3). Before adjusting for propensity scoring, participants experiencing housing instability had significantly higher odds of inadequate prenatal care (odds ratio [OR] 5.11, 95% confidence interval [CI] 2.15–12.14, p < 0.001). Following adjustment for propensity score, the odds of inadequate prenatal care were no longer statistically significant (aOR 1.72, 95% CI 0.55–5.41, p = 0.35).

Table 3.

Proportions of Participants Experiencing Housing Instability and No Housing Instability Receiving Inadequate Prenatal Care and Secondary Outcomes with Odds Ratios

Outcome	Housing instability (N = 55)	No housing instability (N = 435)	OR (95% CI)	aOR^a (95% CI)	p-value^b
Primary outcome
Inadequate prenatal care^c	17.3%	3.9%	5.11 (2.15–12.16)	1.72 (0.55–5.41)	0.35
Secondary maternal outcomes
Excessive weight gain^d	65.9%	59.1%	1.33 (0.68–2.63)	0.91(0.41–2.02)	0.81
Cesarean delivery^e	33.3%	32.2%	1.05 (0.58–1.92)	0.99(0.48–2.06)	0.98
Hypertensive disorders of pregnancy^f	16.7%	15.0%	1.14 (0.53–2.44)	0.95(0.36–2.48)	0.92
Secondary neonatal outcomes
Preterm birth^g	10.9%	10.3%	1.06 (0.43–2.62)	1.18 (0.4–3.54)	0.76
LGA infant^h	19.2%	10.2%	2.10 (0.98–4.48)	2.30 (0.83–6.38)	0.11
NICU admissionⁱ	9.4%	11.8%	0.78 (0.3–2.04)	0.56 (0.18–1.75)	0.32
Secondary postpartum outcomes
Postpartum visit^j	79.6%	91.2%	0.38 (0.18–0.79)	0.69 (0.29–1.66)	0.41
Emergency department visit^k	14.0	9.4%	1.57 (0.66–3.72)	1.22 (0.42–3.52)	0.71
Readmission^l	6.0%	2.9%	2.14 (0.58–7.87)	1.36 (0.29–6.46)	0.70
Postpartum weight change (kg)^m	−9.5 ± 4.0	−9.1 ± 4.2	−0.44 (−1.83,0.95)	−0.64 (−2.3–1.0)	0.44
Breastfeedingⁿ	71.4%	82.8%	0.52 (0.25–1.06)	0.86 (0.37–2.01)	0.72

Sample size of N = 490.

For primary outcome, odds ratios adjusted for propensity scoring using the following covariates: Age, race/ethnicity, education, marital status, employment status, Medicaid; for secondary maternal and neonatal outcomes, odds ratios adjusted for propensity scoring using the following covariates: age, race/ethnicity, education, marital status, employment status, Medicaid, nulliparity and prior spontaneous preterm birth; for secondary postpartum outcomes, odds ratios adjusted for age and Medicaid.

p-value after adjustment.

Defined as Inadequate care on Adequacy of Prenatal Care Index (APNCU), five missing.

Weight gain exceeding Institute of Medicine guidelines based on prepregnancy body mass index, comparing subjects with excessive weight gain and subjects with normal weight gain. Subjects below suggested weight gain were excluded.

Four missing.

Includes diagnosis of gestational hypertension, preeclampsia, or superimposed preeclampsia at the time of delivery or during inpatient care after delivery, eight missing.

Birth at <37 weeks gestational age.

Large for gestational age birth weight for gestational age by WHO score, six missing.

Six missing.

Postpartum visit: A comprehensive visit involving discussion of breastfeeding, contraception, mood, and overall maternal health and well-being at 4–12 weeks after delivery, two missing.

25 missing.

Only included subjects who went to postpartum visit. Postpartum weight change: weight at postpartum visit—weight at delivery in kilograms, 33 missing.

Only included subjects who went to postpartum visit. Breastfeeding: exclusive breastfeeding or breastfeeding with formula supplements, as assessed at the postpartum visit, seven missing.

aOR, adjusted odds ratio; CI, confidence interval; LGA, large for gestational age; NICU, neonatal intensive care unit; WHO, World Health Organization.

Secondary maternal, neonatal, and postpartum outcomes are shown in Table 3. Among other maternal outcomes, there was no statistically significant difference in gestational weight gain, mode of delivery, or hypertensive disorders of pregnancy. Among neonatal outcomes, there was no statistically significant difference in preterm birth, large for gestational age birth weight, or NICU admission. Although the frequency of postpartum visit attendance was lower for individuals with unstable housing (79.6% vs. 91.2%, OR 0.38, 95% CI 0.18–0.79), there was no difference in the odds of the postpartum visit attendance after adjusting for age and insurance status (aOR 0.69, 95% CI 0.29–1.66). The remaining postpartum outcomes did not differ by housing status. Findings did not differ in the sensitivity analysis, including race in models (Supplementary Table S1).

Discussion

Principal findings

Antenatal housing instability was common in this diverse urban cohort, and included a wide range of experiences ranging from being behind on rent and utility payments to experiencing an episode of homelessness. In the primary analysis, before adjusting for sociodemographic confounders, participants experiencing antenatal housing instability had approximately five times the odds of receiving inadequate prenatal care compared with participants with no housing instability. After adjusting for sociodemographic covariates, however, this difference no longer reached statistical significance. Antenatal housing instability is therefore a risk marker associated with other social determinants of health that increase the likelihood of receiving inadequate prenatal care, but it is not independently associated with inadequate prenatal care in our study population. Antenatal housing instability was not an independent predictor of any other analyzed maternal, neonatal, or postpartum outcome.

Results in context

A 2021 systematic review from DiTosto et al. identified 14 studies examining the relationship between housing status and obstetric outcomes.² One study of 623 young pregnant participants from a predominantly low-income sample in New York City demonstrated a lower average birth weight among participants who reported two or more moves in the past year.⁸ A more recent study examined the relationship between housing instability, obstetric outcomes, and perinatal care utilization by using a large population database with exposure defined by an International Classification of Diseases (ICD) 9 code for lack of housing or inadequate housing.¹⁴

Pantell et al. found that housing instability was associated with adverse outcomes, including preterm delivery, long length of stay, emergency department visits, and hospital readmission. The mechanism through which housing instability influences obstetric outcomes is likely complex. It may include the accumulation of maternal life stressful events, which has been associated with increasing risk for preterm delivery, low birth weight, and other adverse child health outcomes in prior studies.^25
–27 Unstable housing may also determine exposure to harmful environmental factors such as air pollutants and heavy metals, or influence a mother's social environment and family- and neighborhood-level factors.^28,29

The frequency of housing instability in this analyzed cohort was 11.0%, and the population of participants experiencing housing instability differed significantly from those with stable housing in the year before delivery. In published literature, the rate of housing instability in the peripartum ranges widely, from <0.01% in population studies using ICD-9 codes to as high as 28.5% in surveyed populations, likely due to heterogeneity in the definition of housing instability and in the demographics of the study population.^2,8,19 Our study is unique in its direct measurement of housing instability by participant report in a socio-demographically diverse urban birth cohort, increasing its generalizability to demographically similar populations. Participants experiencing housing instability were more likely to be publicly insured, of minority race/ethnicity, younger, multiparous, single, and unemployed, and have less than a college education. These attributes are consistent with findings of prior studies in both obstetric and nonobstetric populations.^8,9,11,12

Clinical implications

Although we did not find an independent association between housing instability and inadequate prenatal care or other secondary outcomes after adjusting for covariates, our data do support housing instability as a marker for obstetric patients at risk for receiving inadequate prenatal care, perhaps due to a confluence of social factors. The accumulation of social stressors over a lifetime is theorized to widen racial and socioeconomic disparities in preterm birth and other outcomes.^30,31 In addition, social conditions that occur concurrently with the experience of housing instability, such as restricted access to care due to public insurance status or transportation barriers, also contribute to prenatal care inadequacy.³² Intervention to address housing instability in the antenatal period may therefore be one of a complement of approaches to reduce “weathering” over the life course and improve outcomes.

In contrast with prior studies where inadequate prenatal care was analyzed as a covariate in the relationship between housing instability and perinatal outcomes, our study does not support an independent association between housing instability and inadequate prenatal care, or analyzed secondary outcomes when adjusting for sociodemographic and clinical covariates.^2,8,14

We hypothesize that these contradictory findings may be explained by differing definitions of housing instability. Pantell et al. acknowledge that the use of ICD 9 codes may underestimate the prevalence of housing instability and select for cases where instability is linked with more severe or detectable hardships. The prevalence of housing instability in our population was much higher than in Pantell et al., suggesting that our broader definition captured a larger proportion of obstetric patients with housing-related concerns.¹³ These findings taken together may suggest that not all forms of housing instability are equivalent in their relationship with care utilization or birth outcomes.

Research implications

Our research demonstrates that housing instability is closely intertwined with other social determinants of health. While existing research identifies housing status as a unique and sometimes independent driver of care utilization and health outcomes, further research must account for variation in types and degrees of housing instability to draw conclusions about the relationship between housing and perinatal health and health care. Our study was not designed or sufficiently powered to assess outcomes of each subgroup of participants experiencing housing instability. Use of a standardized and comprehensive definition or validated scale for housing instability and a larger sample size in future research is essential for distilling its distinct relationship with other covariates and outcomes.

Further work is required to understand how best to support pregnant individuals who experience housing instability. Interventions targeting housing instability in pregnant patients may affect care utilization and outcomes through a variety of mediating variables even in the absence of an independent association. Future work may therefore include quantitative and qualitative analyses of housing-focused interventions, for example Boston's “Healthy Start in Housing” initiative, which expedited subsidized housing vouchers for pregnant individuals at risk of homelessness, and their impact on health care utilization and health outcomes.³³

Strengths and limitations

Previous studies of housing instability during pregnancy have been limited by reliance on medical record documentation or responses to one or two housing-related questions as part of a larger survey. In this cohort analysis, housing instability was defined by direct participant survey regarding several varieties of housing instability, leading to a robustly selected exposure group. In addition, our recruitment strategy minimized selection bias by recruiting from a population of postpartum patients at a large urban hospital serving a socioeconomically diverse population, thereby identifying participants experiencing housing instability, who may not have been otherwise screened for housing concerns during pregnancy.

Several important limitations to this study exist. The small sample size prevented the inclusion of several important covariates, including substance use, intimate partner violence, and other medical comorbidities. Recruitment from the postpartum floors may have led to oversampling of patients with longer hospital stays, including those who delivered through cesarean, or who had medical or social complications delaying discharge. Due to the timing of control variable measurement, the study may have controlled for mediators, which could weaken focal associations and may partly explain our findings in the adjusted models.

Although our definition of housing instability was intentionally broad and the sample was large, another limitation is the relatively smaller proportion of participants who experienced housing instability. The most common indicators of housing instability in our sample were being behind on utility bills, moving in with friends/family, and being behind on rent. Given the small total number of participants experiencing housing instability, and that the majority reported more mild forms of housing instability, a smaller effect size than twofold increase in inadequate prenatal care may not have been detected due to insufficient power.

Moreover, while we believe our novel definition of housing instability using six different screening items is a major strength of the study, it may also introduce a limitation in the interpretation and application of results by grouping participants with milder and more severe forms of housing instability. While homelessness is a well-defined risk factor for adverse perinatal outcomes, further research with a larger sample size would be required to determine whether other specific forms of housing instability such as eviction or crowding are associated with these outcomes.

Conclusion

In this cohort analysis, housing instability in the year before delivery was common and reflected a wide variety of experiences, including being behind on rent and utilities payments, frequent moves, crowding, and homelessness. Housing instability was not independently associated with maternal, neonatal, or postpartum outcomes. Participants experiencing housing instability did have higher odds of inadequate prenatal care in our unadjusted model, suggesting that housing instability may be one of a confluence of social determinants associated with inadequate prenatal care utilization. Future research in the impact of housing on perinatal health must account for the diversity of types and severity of housing instability to inform policies and interventions to improve care utilization and outcomes.

Footnotes

Acknowledgment

The authors would like to thank the Northwestern University Clinical and Translational Sciences (NUCATS) Institute for their support of this project.

Authors' Contributions

M.Z.: Conceptualization (lead); investigation (equal); writing—original draft (lead); and writing—editing and revising (supporting). E.E.: Conceptualization (supporting); investigation (equal); writing—original draft (supporting); and writing—editing and revising (lead). A.O.: Data curation (lead); investigation (equal); and writing—editing and revising (supporting). C.Y.: Formal analysis (lead). L.M.Y.: Supervision (lead); conceptualization (supporting); and writing—editing and revising (supporting).

Author Disclosure Statement

No competing financial interests exist.

Funding Information

Resident research grant from the Northwestern University Feinberg School of Medicine Department of Obstetrics and Gynecology and National Institutes of Health's National Center for Advancing Translational Sciences, Grant No. UL1TR001422.

Supplementary Material

Supplementary Table S1

Supplementary Appendix S1

References

Sufrin

, Davidson

, Markenson

. Importance of social determinants of health and cultural awareness in the delivery of reproductive health care. Obstet Gynecol, 2018; 131(1):E43–E48; doi: 10.1097/AOG.0000000000002459

DiTosto

, Holder

, Soyemi

, et al. Housing instability and adverse perinatal outcomes: A systematic review. Am J Obstet Gynecol MFM, 2021; 3(6):100477; doi: 10.1016/j.ajogmf.2021.100477

Cutts

, Coleman

, Black

, et al. Homelessness during pregnancy: A unique, time-dependent risk factor of birth outcomes. Matern Child Health J, 2015; 19(6):1276–1283; doi: 10.1007/s10995-014-1633-6

Stein

, Lu

, Gelberg

. Severity of homelessness and adverse birth outcomes. Health Psychol, 2000; 19(6):524–534; doi: 10.1037/0278-6133.19.6.524

Weitzman

. Pregnancy and childbirth: Risk factors for homelessness?. Fam Plann Perspect, 1989; 21(4):175–178; doi: 10.2307/2135809

Burgard

, Seefeldt

, Zelner

. Housing instability and health: Findings from the Michigan Recession and Recovery Study. Soc Sci Med, 2012; 75(12):2215–2224; doi: 10.1016/j.socscimed.2012.08.02

Johnson

, Meckstroth

. Ancillary Services to Support Welfare-to-Work. Mathematica Policy Research; 1998. Available from: https://ideas.repec.org/p/mpr/mprres/0aa15be44f8b4caeb7be1a2fea30a372.html [Last accessed: June 5, 2020 ].

Carrion

, Earnshaw

, Kershaw

, et al. Housing instability and birth weight among young urban mothers. J Urban Health, 2015; 92(1):1–9; doi: 10.1007/s11524-014-9913-4

Suglia

, Duarte

, Sandel

. Housing quality, housing instability, and maternal mental health. J Urban Health, 2011; 88(6):1105–1116; doi: 10.1007/s11524-011-9587-0

10.

Pollack

, Lynch

. Health status of people undergoing foreclosure in the Philadelphia region. Am J Public Health, 2009; 99(10):1833–1839; doi: 10.2105/AJPH.2009.161380

11.

Cutts

, Meyers

, Black

, et al. US Housing insecurity and the health of very young children. Am J Public Health, 2011; 101(8):1508–1514; doi: 10.2105/AJPH.2011.300139

12.

Kushel

, Gupta

, Gee

, et al. Housing instability and food insecurity as barriers to health care among low-income Americans. J Gen Intern Med, 2006; 21(1):71–77; doi: 10.1111/j.1525-1497.2005.00278.x

13.

Heaman

, Newburn-Cook

, Green

, et al. Inadequate prenatal care and its association with adverse pregnancy outcomes: A comparison of indices. BMC Pregnancy Childbirth, 2008; 8:15; doi: 10.1186/1471-2393-8-15

14.

Pantell

, Baer

, Torres

, et al. Associations between unstable housing, obstetric outcomes, and perinatal health care utilization. Am J Obstet Gynecol MFM, 2019; 1(4):100053; doi: 10.1016/j.ajogmf.2019.100053

15.

Harris

, Taylor

, Thielke

, et al. Research electronic data capture (REDCap)—A metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform, 2009; 42(2):377–381; doi: 10.1016/j.jbi.2008.08.010

16.

Harris

, Taylor

, Minor

, et al. The REDCap consortium: Building an international community of software platform partners. J Biomed Inform, 2019; 95:103208; doi: 10.1016/j.jbi.2019.103208

17.

Frederick

, Chwalek

, Hughes

, et al. How stable is stable? defining and measuring housing stability. J Community Psychol, 2014; 42(8):964–979; doi: 10.1002/jcop.21665

18.

Kotelchuck

. An evaluation of the Kessner Adequacy of Prenatal Care Index and a proposed Adequacy of Prenatal Care Utilization Index. Am J Public Health, 1994; 84(9):1414–1420; doi: 10.2105/ajph.84.9.1414

19.

Alexander

, Kotelchuck

. Assessing the role and effectiveness of prenatal care: History, challenges, and directions for future research. Public Health Rep, 2001; 116(4):306–316; doi: 10.1093/phr/116.4.306

20.

Krueger

, Scholl

. Adequacy of prenatal care and pregnancy outcome. J Am Osteopath Assoc, 2000; 100(8):485–492; doi: 10.7556/jaoa.2000.100.8.485

21.

Institute of Medicine (US) Committee on Nutritional Status During Pregnancy and Lactation. Nutrition During Pregnancy: Part I Weight Gain. Washington, DC, USA: National Academies Press; 1990. Available from: https://pubmed.ncbi.nlm.nih.gov/25144018/ [Last accessed: June 12, 2021 ].

22.

American College of Obstetricians and Gynecologists. Committee Opinion no. 736: Optimizing Postpartum Care. Obstet Gynecol, 2018; 131(5):e140–e150; doi: 10.1097/AOG.0000000000002633

23.

Haukoos

, Lewis

. The propensity score. JAMA, 2015; 314(15):1637–1638; doi: 10.1001/jama.2015.13480

24.

Arbogast

, Ray

. Performance of disease risk scores, propensity scores, and traditional multivariable outcome regression in the presence of multiple confounders. Am J Epidemiol, 2011; 174(5):613–620; doi: 10.1093/aje/kwr143

25.

Witt

, Cheng

, Wisk

, et al. Maternal stressful life events prior to conception and the impact on infant birth weight in the United States. Am J Public Health, 2014; 104(Suppl 1):S81–S89; doi: 10.2105/AJPH.2013.301544

26.

Witt

, Cheng

, Wisk

, et al. Preterm birth in the United States: The impact of stressful life events prior to conception and maternal age. Am J Public Health, 2014; 104(Suppl 1):S73–S80; doi: 10.2105/AJPH.2013.301688

27.

Robinson

, Mattes

, Oddy

, et al. Prenatal stress and risk of behavioral morbidity from age 2 to 14 years: The influence of the number, type, and timing of stressful life events. Dev Psychopathol, 2011; 23(2):507–520; doi: 10.1017/S0954579411000241

28.

Miranda

, Maxson

, Edwards

. Environmental contributions to disparities in pregnancy outcomes. Epidemiol Rev, 2009; 31:67–83; doi: 10.1093/epirev/mxp011

29.

Roberts

. Neighborhood social environments and the distribution of low birthweight in Chicago. Am J Public Health, 1997; 87(4):597–603; doi: 10.2105/ajph.87.4.597

30.

, Halfon

. Racial and ethnic disparities in birth outcomes: A life-course perspective. Matern Child Health J, 2003; 7(1):13–30; doi: 10.1023/a:1022537516969

31.

Blumenshine

, Egerter

, Barclay

, et al. Socioeconomic disparities in adverse birth outcomes: A systematic review. Am J Prev Med, 2010; 39(3):263–272; doi: 10.1016/j.amepre.2010.05.012

32.

Taylor

, Liu

, Howell

. Insurance differences in preventive care use and adverse birth outcomes among pregnant women in a medicaid nonexpansion state: A retrospective cohort study. J Womens Health, 2020; 29(1):29–37; doi: 10.1089/jwh.2019.7658

33.

Feinberg

, Trejo

, Sullivan

, et al. Healthy start in housing: A case study of a public health and housing partnership to improve birth outcomes. Cityscape, 2014; 16(1):141–164.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.02 MB

0.00 MB

0.01 MB