Eviction and Household Health and Hardships in Families With Very Young Children

Researchers conducted household-level surveys with caregivers and audited medical records from February 2011 through December 2019 in pediatric emergency departments (ED) (Baltimore, MD; Boston, MA; Little Rock, AR; and Philadelphia, PA) and primary care clinics (Minneapolis, MN and Baltimore, MD). Institutional review board approval was obtained for all sites and renewed annually. Study eligibility criteria included caregiver: consent, lived in the child’s household, residency in the study site’s state, and ability to speak English, Spanish, or Somali (in Minneapolis only), and the child’s age <48 months. Of those screened, 34 729 caregivers (88.6%) were eligible, and 31 749 completed the interview. The sample was restricted to unique renter households (homeowners excluded n = 4811) and nonmissing eviction history data (missing n = 497) with a final sample of 26 441.

The survey included questions about the caregivers’ race and ethnicity, country of birth, marital status, highest level of education, age, and employment status; number of children in household; and children’s insurance type, breastfeeding history, and birth weight. Self-reported household income data were collected starting in 2013 with a question about monthly income from all sources, excluding noncash benefits, in categorical ranges of $1000. The birth mothers’ race, a social construct, and ethnicity were self-identified using questions from the US Census and combined into the following categories: Hispanic, all races; Black non-Hispanic; White non-Hispanic; other or multiple races non-Hispanic (composed of caregivers identifying multiple races or groups too small in this sample to analyze independently including Asian, Native American). Child age, sex, weight, and length and height on the visit date were abstracted from medical records.

Eviction history is based on a validated measure (Milwaukee Area Renters Study) defined by answers to the following question: “An eviction is when your landlord or a government or bank official forces you to move when you don’t want to. In the past 5 years, have you ever been evicted?”¹  The question identifies any eviction experience (any versus none). If eviction was affirmed, caregivers were asked about an eviction order in court and a ruling in the landlord’s favor to determine a formal eviction. Caregivers were divided into 3 groups based on their 5-year history: no eviction, eviction history–formal, eviction history–informal.

Child and caregiver health: Caregivers evaluated their own and their child’s health on the basis of the validated question from the third National Health and Nutrition Examination Survey. Responses were dichotomized into fair or poor versus good or excellent.

Child hospitalizations: The child’s lifetime history of hospitalizations excluding birth reported by the caregiver (any versus none).¹⁷ 

Developmental Risk: Children’s developmental risk was defined by 1 or more concerns on the parent’s evaluation of developmental status, approved by the American Academy of Pediatrics for child developmental screening tools for this age group.¹⁸  Only children ≥4 months were included in developmental risk analysis because of improved parent’s evaluation of developmental status specificity and sensitivity for those ages.¹⁸ 

Hospital admission from the ED: For ED sites only, admission to inpatient care on the visit date was abstracted from the medical record.

Child anthropometrics: The child’s length and height and weight on the visit date were used to calculate weight-for-age and weight-for-length or height percentiles on the basis of World Health Organization or Centers for Disease Control and Prevention standards. Risk of underweight was defined as weight-for-age <5^th percentile or weight-for-length or height <10^th percentile.¹⁹  Obesity risk was defined as weight-for-age ≥ 90^th percentile.²⁰ 

Maternal depressive symptoms: Endorsement of 2 of 3 questions on a validated maternal depression screener²¹  was used to determine evidence of depressive symptoms.

Four measures of housing-related hardship were evaluated: behind on rent in the previous year, 2 or more moves in the previous year, homelessness in the child’s lifetime,²²  and current homelessness defined as living in a shelter, hotel or motel, or having no steady place to sleep.

Food insecurity: The 18-question US Household Food Security Survey Module asks about the previous 12 months.²³  Household food insecurity was defined as endorsement (sometimes or often true) of 3 or more of 10 household questions and child food insecurity as endorsement of 2 or more of 8 child-specific questions.

Energy insecurity: Families were classified as energy insecure if any of 4 validated questions were affirmed²⁴ : threatened or actual utility shutoff, unheated or uncooled days, or heating with a cooking stove.

Health care hardship: Health care hardship was defined by: forgone health care, in which a household member needed care or prescriptions that were not received because of cost and/or health cost sacrifices; or struggling with payment of basic needs including food, housing, or utilities as a consequence of medical expenses.

Child care constraints: Child care constraints were defined if caregivers reported challenges obtaining child care made it difficult to work or attend school.

χ² analyses or Student’s t test were used to describe family characteristics, stratified by 3-category eviction history (no eviction, formal, or informal) and combined eviction history (no eviction versus either formal or informal). An initial set of multivariable logistic regression analyses examined informal and formal eviction separately and compared associations across eviction type on child, caregiver, and hardship outcomes. If associations of informal and formal eviction were similar using multivariable logistic regression, we planned to rerun the analysis with a combined eviction history variable to increase precision. Adjusted multivariable logistic regression analyses were fit to evaluate the association between eviction history groups and outcomes. Covariates were chosen based on significant association with eviction history or health outcomes or on previous research demonstrating associations, including well-documented findings of racism and racial inequities as a significant contributor to increased risk of eviction and poor health outcomes.²  Analyses were adjusted for study site, caregiver: age, place of birth (United States–born or foreign-born), race and ethnicity (as a proxy for experienced racism), marital status, employment, education; and child: insurance type, age, breastfeeding history; and number of children in the household. For child health outcomes, low birth weight (LBW) was included as a covariate on the basis of evidence linking LBW with adverse child health outcomes. No history of eviction was the referent group for all analyses. Sensitivity analyses adjusting for (1) household income (2) housing-related hardships and (3) limiting the sample to children with public or no insurance were performed to examine whether these factors were drivers of observed relationships.

For all models, we report adjusted odds ratios (aOR) and corresponding 95% confidence intervals (95% confidence interval [CI]). All analyses were conducted using 2-sided tests and a significance level of 0.05. Statistical analyses were performed by using SAS software (version 9.4; SAS Institute, NC, USA).

Of the 26 441 caregivers, 92.5% (24 457) were birth mothers. 3.9% reported an eviction in the previous 5 years (Table 1). More caregivers reported formal than informal evictions (57.0% vs 43.0%). Compared to the no eviction group, a greater proportion of caregivers in the formal and informal eviction group were Black, non-Hispanic, United States–born, older and had an education beyond high school. Of caregivers with an eviction history, whether informal or formal, the greatest proportion came from Baltimore, Philadelphia, and Boston. Children in the informal and formal eviction versus no eviction group had higher rates of public insurance (93.6%, 95.2% vs 90.8%, respectively) and lower rates of breastfeeding history (63.3%, 59.5% vs 65.9%, respectively). Rates of current housing subsidy receipt (17.3%) were lowest in the formal eviction versus informal and no eviction groups (26.0% and 21.3%, respectively). Notably, a greater proportion of both the informal and formal eviction versus no eviction group reported having subsidized housing rescinded in the past 2 years (5.0%, 5.9% vs 1.0%). Rates of LBW did not differ between groups.

There were higher rates of fair or poor child and caregiver health, child developmental risk, and maternal depressive symptoms in both the informal and formal eviction versus no eviction group and no differences between groups in child weight or health care utilization rates. All household hardships rates were higher in the informal and formal eviction versus no eviction group. (Table 2)

Treating informal and formal eviction as separate exposures showed little difference in the associations of informal and formal eviction compared with no eviction for child health, caregiver health, and hardship outcomes (Table 3). Two outcomes differed by type of eviction. Admission from the ED showed greater odds among the formal compared to informal eviction group. Household food insecurity had greater odds among the informal compared to the formal eviction group. Informal and formal eviction groups were combined for subsequent analyses.

In adjusted analyses (Table 4), compared to the no eviction group, the eviction group children were more likely to be in fair or poor health (aOR: 1.43, [95% CI: 1.17–1.73]), at developmental risk (aOR: 1.55, [95% CI: 1.32–1.82)]) and to have been admitted from the ED (aOR: 1.24 [95% CI: 1.01–1.53]). Groups did not differ in odds of lifetime hospitalizations, risk of underweight, and obesity risk.

Eviction group caregivers were more likely to be in fair or poor health (aOR: 1.96 [95% CI: 1.72–2.25]) and to report depressive symptoms (aOR: 2.71 [95% CI: 2.37–3.11]), than no eviction group caregivers.

Compared to the no eviction group, the eviction group was more likely to have multiple moves (aOR: 4.01, [95% CI: 3.33–4.82]), be behind on rent in the past year (aOR: 3.83, [95% CI: 3.33–4.41]), have homelessness in the child’s lifetime (aOR: 5.48 [95% CI: 4.63–6.49]), and be homeless currently (aOR: 3.66 [95% CI: 2.93–4.57]). The eviction group was more likely to experience all material hardships measured: household food insecurity (aOR: 3.07 [95% CI: 2.69–3.50]), child food insecurity (aOR: 2.75 [95% CI: 2.36–3.20]), energy insecurity (aOR: 2.18 [95% CI: 1.90–2.50]), health care hardship (aOR: 2.43 [95% CI: 2.11–2.78]), and child care constraints (aOR: 2.63 [95% CI: 2.30–2.99]).

Sensitivity analyses controlling for household income did not change the significance of the health and hardship associations but attenuated the odds (Table 5). Exceptions were fair or poor child health and admission from the ED, which did not achieve significance after adjusting for household income (aOR: 1.23 [95% CI: 0.99–1.53]) and (aOR: 1.23 [95% CI: 0.98–1.56]), respectively). Adjusting for housing-related hardships, odds ratios for all health and other hardships were attenuated but remained significant with the exception of admission from the ED (aOR: 1.23 [95% CI: 0.99–1.53]) (Table 5). Limiting the sample to children with public or no insurance to ensure a more uniformly low-income sample resulted in highly consistent findings with no changes in significance and nearly identical odds ratios for many outcomes. (Table 5)

This study found a 5 year history of formal or informal eviction was associated with increased odds of young children’s developmental risk and hospital admission from the ED (acute health concerns requiring inpatient care), child and caregiver fair or poor health, housing-related hardships and 4 household hardship dimensions: food, utilities, health care, and child care. Additionally, there were few differences in outcomes differentiated by formal versus informal evictions. Sensitivity analyses accounting for income and other housing hardships likewise demonstrated few differences. These results are consistent with a small but growing body of evidence on associations between eviction history and adverse health and hardship conditions among young children.^6,7  Longitudinal data has also documented eviction’s association with health and material hardship among families with young children, finding poor child and caregiver health, and elevated risks for both maternal depression and a composite measure of material hardships for 2 years posteviction.⁶  Our study expands these findings and includes specific household hardship dimensions among families with very young children, useful for targeting policy solutions. Evictions during this sensitive developmental period place children at risk for lasting harm.^5,7 

This study also includes informal evictions, which are often excluded from other studies. Similar associations between health outcomes and formal and informal evictions suggest a shared detrimental effect, although possibly through different mechanisms. Informal evictions, like formal evictions, increase parental stress and can lead to residential displacement resulting in loss of social connections, school, and other support systems for families and children. Unlike like formal evictions, which involve an arduous and often lengthy legal process, informal evictions may cause sudden displacement that is catastrophic for families, leading them to accept whatever housing is available, regardless of location or quality. Although informal evictions are not recorded in court, most landlords request previous landlord references, which may have a detrimental effect on future rental opportunities for families.

Food insecurity, housing instability, energy insecurity, and health care hardship are independently associated with adverse health outcomes among adults and children.^22,24–27  Emerging data show that child care constraints are associated with worse child and adult health.²⁸  Regardless of the temporality of hardships and eviction, these associations undermine children and families’ health and wellbeing.

Eviction rates in this sample (3.9%) were approximately double the national average of formal evictions among renters (2.3% in 2016), again revealing that households with children are at elevated eviction risk.^2,29,30  There are many drivers of eviction rates, including economic downturns, income shocks, and neighborhood gentrification, which is independently associated with poor child mental health outcomes.³¹  Previous research also demonstrates adverse health conditions may contribute to increased risk of eviction¹⁵  among families with young children. Despite city, state, and federal eviction moratoria during the coronavirus disease 2019 pandemic, varying widely by state and locality, an estimated 1 out of 5 families with children were at risk for eviction in May 2021.^32,33 

To advance health equity, eviction’s root causes, including discrimination against children, systemic racism, xenophobia, wage and wealth disparities by race and sex, and health inequities, must be addressed.^3,34–37  This sample had complex variations in maternal racial and ethnic and nativity based on eviction history. Although the current study does not focus on racial and ethnic and nativity differences, future research should specifically and intersectionally address eviction rates by race and ethnicity, nativity, and associated health inequities. Research differentiating relationships of state and local policies with eviction incidence and inequities and the bidirectional nature of evictions and health can identify protective policies for family health. Research examining protective effects of housing and other assistance programs against eviction would be informative.^38,39 

Formal evictions remain on record for ∼5 years and can have ramifications on housing access, employment, and credit scores, perhaps partially explaining associations with material hardships in this and other studies.^6,40  A report from the National Academies of Science, Engineering, and Medicine highlights the coronavirus disease 2019 pandemic’s impact on evictions and inequities, providing recommendations for addressing the current eviction crisis.³⁶  Policies that provide tenant protections, increase housing affordability, and promote racial and socioeconomic equity vary widely by state or locality and are critical to mitigate the effect and decrease the eviction incidence.⁴¹  Policy solutions for formal evictions include procedural changes requiring legal representation and, when needed, interpreters for tenants to decrease landlord-tenant power imbalance. Preventing informal evictions may require skilled community-based advocacy to reduce tensions between landlords and tenants. Upstream prevention may include increasing access to rental assistance and affordable housing, especially for families with infants and young children, such as permanently authorizing an emergency assistance fund to pay rent arrears and other income-boosting policies. Recognizing the bidirectional nature of eviction and adverse health outcomes, increasing access to health care services, and improving health-related social needs that support optimal health are important. Expanding Medicaid helps to cover medical expenses, provides access to care that could ameliorate illness, and has been shown to reduce the number and rate of evictions as well as eviction filings.^15,37,38  Providing assistance with health care costs may also present the opportunity for household spending to be distributed to other necessities, such as food, rent, utilities, and child care.¹⁵ 

Although policy solutions are necessary for reducing evictions, clinical and public health practice provides an opportunity to address underlying causes of evictions and mitigate their harms.⁴²  Health care providers could screen patients and families for housing-related hardship.^5,22,43  Partnering with community organizations and connecting patients at risk with resources including housing-specific case management, financial assistance, and affordable housing may prevent evictions and improve health.^36,44  Responses could vary depending on the specific hardships identified, thus screening for and connecting patients to resources to address and alleviate food, utility, child care, and health care costs are also critical.

This study’s strengths include its focus on a large, geographically and racially and ethnically diverse sample of families with young children and information about informal evictions, often missing in eviction research because of a reliance on court filings. This study population is not nationally representative but rather a sentinel sample, primarily composed of families with low incomes accessing urban hospitals with a high proportion of both caregivers of color and immigrant caregivers, compared to national statistics. The sentinel sample is both a strength and a limitation as a dynamic form of data collection designed to signal early trends and identify and monitor policy effects and disease burdens before they become widely prevalent. Although limited in generalizability, it helps identify emerging health impacts promptly so that timely interventions can be developed.⁴⁵  The cross-sectional study design reflects association, not causation, within unknown eviction timing. Caregivers were asked about evictions in the previous 5 years, whereas most hardship questions use a 1 year retrospective period. Therefore, eviction’s proximity to interview date, hardship report, and youngest children’s lifetime is unknown. Some families may have experienced >1 eviction, but multiple incidents of eviction, which is rare in a 5 year window, were not captured in this study. Future studies may consider exploring health outcomes among those who have had multiple evictions compared to 1 or none. Regardless, maternal hardships experienced prenatally have implications for child health^46,47  and, conversely, child health can precipitate evictions.¹⁵  As a result, a relentless cycle can occur between poor health and evictions. Additionally, many variables in this study were self-reported and subject to reporting bias. The eviction history question is validated to identify formal and informal evictions common among low-income renters missed by court record data,¹  and self-reported health and material hardship measures used have been validated in national surveys.^17–24,47  Finally, there may be other unmeasured confounders, such as exposure to discrimination, neighborhood location, and city-specific tenant protections.

Considering evidence on the health and socioeconomic impact of evictions, this research elucidates associations between family history of eviction and infant and toddler health and development. Although eviction’s causes and consequences may be complex and varied, findings suggest reduction of eviction incidence, formal and informal, may address health disparities and the needs of young families. Policymakers, community organizations, and health professionals have important roles in designing evidence-based policy solutions to reduce evictions and improve opportunities for families to meet their basic needs.

We thank Nayab Ahmad, BA, Yasmeen Alsaif, and Mikalia Jackson, BS, for their assistance in the preparation of this manuscript. We also thank the families who shared their time and information with us.

An earlier version of this analysis was presented at the Pediatric Academic Societies meeting in Baltimore in April 2019.

aOR

adjusted odds ratio

CI

confidence interval

ED

emergency department

LBW

low birth weight