[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure.  Adjusted Percentage Point Differences in Past-Year Asthma-Related Emergency Department Visits for Children Experiencing an Asthma Attack Until Entry Into a US Department of Housing and Urban Development Program
Adjusted Percentage Point Differences in Past-Year Asthma-Related Emergency Department Visits for Children Experiencing an Asthma Attack Until Entry Into a US Department of Housing and Urban Development Program

The sample includes children from the 1999 to 2001 and 2004 to 2012 National Health Interview Survey who reported an asthma attack. Percentage point differences come from the incremental differences implied by a multivariable logistic regression. The year before entry (−1) is the omitted reference category. Year 0 is the first year of entry. The error bars indicate 95% CIs.

Table 1.  Characteristics of Respondents by HUD Statusa
Characteristics of Respondents by HUD Statusa
Table 2.  Percentage Point Differences in Asthma Outcomes for Children Currently Participating in a HUD Rental Assistance Program vs Those Waiting to Enter a Programa
Percentage Point Differences in Asthma Outcomes for Children Currently Participating in a HUD Rental Assistance Program vs Those Waiting to Enter a Programa
1.
Azar  A.The Root of the Problem: America’s Social Determinants of Health. Washington, DC: Department of Health and Human Services; November 2018.
2.
Alley  DE, Asomugha  CN, Conway  PH, Sanghavi  DM.  Accountable health communities: addressing social needs through Medicare and Medicaid.  N Engl J Med. 2016;374(1):8-11. doi:10.1056/NEJMp1512532PubMedGoogle ScholarCrossref
3.
Council on Community Pediatrics.  Poverty and child health in the United States.  Pediatrics. 2016;137(4):e20160339. doi:10.1542/peds.2016-0339PubMedGoogle Scholar
4.
Kaiser Permanente. Announcing $200M impact investment to address housing crisis. May 2018. https://share.kaiserpermanente.org/article/kaiser-permanente-announces-200-million-impact-investment-partners-with-u-s-mayors-and-ceos-to-address-housing-stability/. Accessed June 26, 2019.
5.
Council on Community Pediatrics.  Providing care for children and adolescents facing homelessness and housing insecurity.  Pediatrics. 2013;131(6):1206-1210. doi:10.1542/peds.2013-0645PubMedGoogle ScholarCrossref
6.
Leventhal  T, Newmann  S.  Housing and child development.  Child Youth Serv Rev. 2010;32:1165-1174. doi:10.1016/j.childyouth.2010.03.008Google ScholarCrossref
7.
Watson  NE, Steffen  BL, Martin  M, Vandenbroucke  DA.  Worst Case Housing Needs: Report to Congress. Washington, DC: US Department of Housing and Urban Development; 2017.
8.
Slopen  N, Fenelon  A, Newman  S, Boudreaux  M.  Housing assistance and child health: a systematic review.  Pediatrics. 2018;141(6):1-14. doi:10.1542/peds.2017-2742PubMedGoogle ScholarCrossref
9.
Center for Budget and Policy Priorities.  Federal Rental Assistance. Washington, DC: Center for Budget and Policy Priorities; 2017.
10.
Forston  JG, Sanbonmatsu  L.  Child health and neighborhood conditions: results from a randomized housing voucher experiment.  J Hum Resour. 2010;45(4):840-864.Google Scholar
11.
Jacob  BA, Ludwig  J, Miller  DL.  The effects of housing and neighborhood conditions on child mortality.  J Health Econ. 2013;32(1):195-206. doi:10.1016/j.jhealeco.2012.10.008PubMedGoogle ScholarCrossref
12.
Kessler  RC, Duncan  GJ, Gennetian  LA,  et al.  Associations of housing mobility interventions for children in high-poverty neighborhoods with subsequent mental disorders during adolescence.  JAMA. 2014;311(9):937-948. doi:10.1001/jama.2014.607PubMedGoogle ScholarCrossref
13.
Schmidt  NM, Lincoln  AK, Nguyen  QC, Acevedo-Garcia  D, Osypuk  TL.  Examining mediators of housing mobility on adolescent asthma: results from a housing voucher experiment.  Soc Sci Med. 2014;107:136-144. doi:10.1016/j.socscimed.2014.02.020PubMedGoogle ScholarCrossref
14.
Zahran  HS, Bailey  CM, Damon  SA, Garbe  PL, Breysse  PN.  Vital signs: asthma in children—United States, 2001-2016.  MMWR Morb Mortal Wkly Rep. 2018;67(5):149-155. doi:10.15585/mmwr.mm6705e1PubMedGoogle ScholarCrossref
15.
National Heart, Lung, and Blood Institute.  Expert Panel Report 3: Guidelines for the Diagnosis and Management of Asthma. Bethesda, MD: National Heart, Lung, and Blood Institute; 2007.
16.
Popkin  SJ, Katz  B, Cunningham  MK, Brown  KD, Gustafson  J, Turner  MA.  A Decade of HOPE IV: Research Findings and Policy Challenges. Washington, DC: The Urban Institute; 2004.
17.
Finkel  M, Buron  L. Study on Section 8 Voucher Success Rates: Quantitative Study of Success Rates In Metropolitan Areas. Vol 1. Cambridge, MA: Abt Associates; 2001.
18.
Smith  R, Naparstek  A, Popkin  S,  et al.  Housing Choice for HOPE VI Relocatees. Washington, DC: The Urban Institute; 2002.
19.
Pearson  WS, Goates  SA, Harrykissoon  SD, Miller  SA.  State-based Medicaid costs for pediatric asthma emergency department visits.  Prev Chronic Dis. 2014;11:E108. doi:10.5888/pcd11.140139PubMedGoogle Scholar
20.
Lantz  PM.  The medicalization of population health: who will stay upstream?  Milbank Q. 2019;97(1):36-39. doi:10.1111/1468-0009.12363PubMedGoogle ScholarCrossref
21.
Lloyd  PC, Helms  VE.  NCHS-HUD Linked Data: Analytic Considerations and Guidelines. Hyattsville, MD: National Center for Health Statistics; 2019.
22.
Parsons  VL, Moriarity  C, Jonas  K,  et al.  Design and Estimation for the National Health Interview Survey, 2006-2015. Hyattsville, MD: National Center for Health Statistics; 2014.
23.
Blewett  LA, Rivera Drew  JA, Griffin  R, King  ML, Williams  KCW.  IPUMS Health Surveys: National Health Interview Survey, Version 6.3. Minneapolis, MN: IPUMS; 2018.
24.
US Department of Housing and Urban Development. Assisted housing: national and local. https://www.huduser.gov/portal/datasets/assthsg.html. Accessed August 1, 2019.
25.
Fenelon  A, Mayne  P, Simon  AE,  et al.  Housing assistance programs and adult health in the United States.  Am J Public Health. 2017;107(4):571-578. doi:10.2105/AJPH.2016.303649PubMedGoogle ScholarCrossref
26.
Fenelon  A, Slopen  N, Boudreaux  M, Newman  SJ.  The impact of housing assistance on the mental health of children in the United States.  J Health Soc Behav. 2018;59(3):447-463. doi:10.1177/0022146518792286PubMedGoogle ScholarCrossref
27.
Simon  AE, Fenelon  A, Helms  V, Lloyd  PC, Rossen  LM.  HUD housing assistance associated with lower uninsurance rates and unmet medical need.  Health Aff (Millwood). 2017;36(6):1016-1023. doi:10.1377/hlthaff.2016.1152PubMedGoogle ScholarCrossref
28.
Nath  JB, Hsia  RY.  Children’s emergency department use for asthma, 2001-2010.  Acad Pediatr. 2015;15(2):225-230. doi:10.1016/j.acap.2014.10.011PubMedGoogle ScholarCrossref
29.
Akinbami  LJ, Moorman  JE, Simon  AE, Schoendorf  KC.  Trends in racial disparities for asthma outcomes among children 0 to 17 years, 2001-2010.  J Allergy Clin Immunol. 2014;134(3):547-553.e5. doi:10.1016/j.jaci.2014.05.037PubMedGoogle ScholarCrossref
30.
Norton  EC, Dowd  BE, Maciejewski  ML.  Marginal effects—quantifying the effect of changes in risk factors in logistic regression models.  JAMA. 2019;321(13):1304-1305. doi:10.1001/jama.2019.1954PubMedGoogle ScholarCrossref
31.
Jacob  BA, Kapustin  M, Ludwig  J.  The impact of housing assistance on child outcomes: evidence from a randomized housing lottery.  Q J Econ. 2015;130(1):465-506. doi:10.1093/qje/qju030Google ScholarCrossref
32.
Clampet-Lundquist  S.  “Everyone had your back”: social ties, perceived safety, and public housing relocation.  City Community. 2010;(1):87-108. doi:10.1111/j.1540-6040.2009.01304.xGoogle Scholar
33.
Keene  DE, Geronimus  AT.  Community-based support among African American public housing residents.  J Urban Health. 2011;88(1):41-53. doi:10.1007/s11524-010-9511-zPubMedGoogle ScholarCrossref
34.
Chen  E, Chim  LS, Strunk  RC, Miller  GE.  The role of the social environment in children and adolescents with asthma.  Am J Respir Crit Care Med. 2007;176(7):644-649. doi:10.1164/rccm.200610-1473OCPubMedGoogle ScholarCrossref
35.
Peterson-Sweeney  K.  The relationship of household routines to morbidity outcomes in childhood asthma.  J Spec Pediatr Nurs. 2009;14(1):59-69. doi:10.1111/j.1744-6155.2008.00175.xPubMedGoogle ScholarCrossref
36.
Gubits  D, Shinn  M, Wood  M,  et al.  Family Options Study: 3-Year Impacts of Housing and Services Interventions for Homeless Families. Washington, DC: US Dept of Housing and Urban Development; 2015.
37.
Newman  S, Holupka  CS.  The effects of assisted housing on child well-being.  Am J Community Psychol. 2017;60(1-2):66-78. doi:10.1002/ajcp.12100PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Views 573
    Citations 0
    Original Investigation
    March 9, 2020

    Association of Childhood Asthma With Federal Rental Assistance

    Author Affiliations
    • 1Department of Health Policy and Management, University of Maryland School of Public Health, College Park
    • 2Department of Sociology and Criminology, Penn State University School of Public Policy, State College, Pennsylvania
    • 3Department of Epidemiology and Biostatistics, University of Maryland School of Public Health, College Park
    • 4Department of Health Policy and Management, Johns Hopkins University Bloomberg School of Public Health, Baltimore, Maryland
    JAMA Pediatr. Published online March 9, 2020. doi:10.1001/jamapediatrics.2019.6242
    Key Points

    Question  Is participation in a federal rental assistance program associated with asthma outcomes in children?

    Findings  In this survey study of 2992 children compared those currently participating in a rental assistance program with those waiting to enter such a program, participation in a rental assistance program was associated with a reduction in emergency department use among children with an asthma attack in the past year, but no significant changes were found in asthma attack or asthma diagnoses.

    Meaning  Rental assistance may be associated with reduced use of emergency department services for the treatment of asthma.

    Abstract

    Importance  Millions of low-income children in the United States reside in substandard or unaffordable housing. Relieving these burdens may be associated with changes in asthma outcomes.

    Objectives  To examine whether participation in the US Department of Housing and Urban Development’s (HUD) rental assistance programs is associated with childhood asthma outcomes and to examine whether associations varied by program type (public housing, multifamily housing, or housing choice vouchers).

    Design, Setting, and Participants  This survey study used data from the nationally representative National Health Interview Survey linked to administrative housing assistance records from January 1, 1999, to December 31, 2014. A total of 2992 children aged 0 to 17 years who were currently receiving rental assistance or would enter a rental assistance program within 2 years of survey interview were included. Data analysis was performed from January 15, 2018, to August 31, 2019.

    Exposures  Participation in rental assistance provided by HUD.

    Main Outcomes and Measures  Ever been diagnosed with asthma, 12-month history of asthma attack, and 12-month history of visiting an emergency department for the treatment of asthma among program participants vs those waiting to enter a program. Overall participation was examined, and participation in public or multifamily housing was compared with participation in housing choice vouchers.

    Results  This study included 2992 children who were currently participating in a HUD program or would enter a program within 2 years. Among children with an asthma attack in the past year, participation in a rental assistance program was associated with a reduced use of emergency departments for asthma of 18.2 percentage points (95% CI, −29.7 to −6.6 percentage points). Associations were only found after entrance into a program, suggesting that they were not confounded by time-varying factors. Statistically significant results were found for participation in public or multifamily housing (percentage point change, −36.6; 95% CI, −54.8 to −18.4) but not housing choice vouchers (percentage point change, −7.2; 95% CI, −24.6 to 10.3). No statistically significant evidence of changes in asthma attacks was found (percentage point change, −2.7; 95% CI, −12.3 to 7.0 percentage points). Results for asthma diagnosis were smaller and only significant at the 10% level (−4.3; 95% CI, −8.8 to 0.2 percentage points).

    Conclusions and Relevance  Among children with a recent asthma attack, rental assistance was associated with less emergency department use. These results may have important implications for the well-being of low-income families and health care system costs.

    Introduction

    Policy makers and health care practitioners are increasingly focused on leveraging health system resources in cross-sector collaborations that aim to mitigate socially determined health conditions.1-4 Such efforts highlight the need for credible evidence on which social services are associated with improved health and reduced health care costs. Rental assistance has garnered particular attention, especially for promoting child health.4,5 Unaffordable, inadequate, and unstable housing may adversely affect child health by increasing stress and environmental exposures and by creating barriers to health care, education, physical activity, and social networks.6 Nearly 3 million low-income families with children pay more than 50% of family income in rent or live in substandard housing.7

    The US Department of Housing and Urban Development’s (HUD’s) rental assistance programs may be associated with improved child health.8 HUD assists approximately 4 million children through 3 programs: public housing (publicly owned developments in which all units are subsidized), multifamily housing (developments in which at least some units are subsidized), and housing choice vouchers (which enable renting in the private market). Benefits ensure that households pay approximately 30% of their income in rent, and all assisted units must pass annual quality inspections.9

    The association of HUD’s rental assistance programs with child health have been relatively understudied because of insufficient data and the difficulty of overcoming unobserved selection.8 Several studies10,11 have examined the effects of changes in program type. For example, the Moving to Opportunity (MTO) trial, set in 5 US cities, randomized existing public housing residents to remain in public housing (the control group), to a traditional voucher, or to a restricted voucher that could be used only in a low-poverty neighborhood. Results for childhood health were mixed. For example, although many health outcomes were unaffected by treatment assignment, adolescent boys in the 2 voucher groups had worse mental health outcomes than did the control group.10,12 Adolescent girls had the opposite pattern. Children in the MTO trial assigned to the voucher conditions had worse asthma outcomes than did those in the control group.10,13

    In this observational study, which accounted for time-invariant selection into assistance, we examined the association of rental assistance with the prevalence of childhood asthma and the use of emergency departments (EDs) for asthma. Asthma is a chronic condition that affects 1 in 12 children and disproportionately affects children from lower-income households.14 The condition can be managed by reducing environmental irritants and through care plans that require high levels of patient engagement.15

    Measuring whether associations differ across program type is important given longstanding interest in promoting residence in low-poverty neighborhoods through mobility-based assistance (ie, vouchers) vs place-based programs (public and multifamily housing).16 Although the MTO trial suggested that public housing is associated with better asthma outcomes than voucher assistance, the mechanisms are not well understood and may differ by sex.13 Voucher holders face additional pressures not experienced by public housing residents. For example, searching for a private unit that accepts vouchers can be stressful and often results in participants settling for units that do not meet their needs.17,18 To our knowledge, no study that credibly accounts for unobserved selection has examined program differences in a national sample. We examined the association of rental assistance with the prevalence of childhood asthma and the use of emergency departments (EDs) for asthma.

    A frequent argument in favor of expanding rental assistance is that it will improve health and reduce publicly financed health system costs. Indeed, asthma-related ED visits cost Medicaid/Children’s Health Insurance Program approximately $272 million per year or $433 for an average visit.19 The motivation for providing rental assistance clearly extends beyond the health system and housing initiatives driven by the health system risk overmedicalizing social problems.20 Nonetheless, developing evidence about how housing policy affects health and health system resources would inform cross-sector collaborations focused on the social determinants of health.

    Methods
    Data

    This survey study used data from the National Center for Health Statistics, which links longitudinal administrative records of HUD participation with cross-sectional data from the National Health Interview Survey (NHIS).21 The administrative records indicated the dates of program participation from January 1, 1999, to December 31, 2014. Data analysis was performed from January 15, 2018, to August 31, 2019. We used a deidentified secondary data set. Consent was managed by the National Center for Health Statistics. The University of Maryland Institutional Review Board reviewed and approved the study.

    The NHIS is a nationally representative survey that collects information on the sociodemographic, health, and health care characteristics of the US civilian noninstitutional population.22 The in-person survey generally achieves an approximately 70% response rate. We limited our analysis to the sample child files, which include detailed information about 1 randomly selected child (aged 0-17 years) per household. We used NHIS variables that are disseminated by the Minnesota Population Center.23

    HUD administrative records were linked to NHIS respondents who consented to linking and provided US Social Security numbers and other identifying characteristics required for linkage. Social Security numbers were collected only for sample child cases in 1999 to 2001 and 2004 to 2012. Within those years, only a portion of respondents consented to linking or provided complete Social Security numbers.21 We created sample weights to ensure that children eligible to be linked were representative of the US population (eAppendix in the Supplement). Our final analytic sample consisted of 2992 children from 1999 to 2001 and 2004 to 2012 that were participating in a HUD program at interview or would enter a program within 2 years (eFigure 1 in the Supplement).

    Study Design

    A challenge in measuring the association of rental assistance with child asthma outcomes is that program participation is voluntary, and unobserved participation factors could also be associated with child health. To address that challenge, we compared outcomes in children currently participating in rental assistance programs (the treatment group) with those in a similar group of children who were not participating when outcomes were measured but would enter such programs within 2 years (the comparison group). The 2-year restriction is aligned with the mean wait-list length and was considered to be suitable for the remitting and relapsing nature of asthma. However, we examined sensitivity to longer waiting periods. Our design accounted for unobserved time-invariant factors associated with a family’s decision to participate and reflects the association of the treatment (rental assistance) for those who selected to receive the treatment.

    Rental assistance is not an entitlement. As a result, participants must first wait for a program to accept applications and then spend a mean of 2 years on a wait-list between when they apply and when they receive assistance.24 However, there is substantial variability in wait-list lengths across local public housing authorities. We did not directly observe participation on a formal wait-list. Instead, the wait-list group included those on a formal list and those who had not yet applied. As such, we refer to the group yet to enter housing as the quasi–wait-list group. Previous studies25-27 have used this method to examine the association of rental assistance with adult health, adult health care access, and child mental health.

    Subgroups

    We examined whether differences varied by program type (housing choice vouchers vs multifamily or public housing). Differences across subgroups were determined by interactions of HUD status with program indicators (eTable 1 to eTable 4 in the Supplement). Multifamily and public housing differ substantively, but we combined them to enhance statistical power and because they share the similar feature of being a nonportable place-based benefit, unlike housing choice vouchers.

    Outcomes

    We examined 3 dichotomous outcomes reported by a knowledgeable household adult: (1) ever been diagnosed with asthma, (2) asthma attack in the 12 months preceding interview among those ever diagnosed, and (3) asthma-related ED visit in the past 12 months. The ED outcome included urgent care centers, and we examined ED rates for the population with a diagnosis and the population experiencing an asthma attack in the past year. Before 2010, the ED question was asked only of children who had an asthma attack, and in subsequent years, it was asked of all children with a diagnosis. We edited the asthma attack variable for consistency so that all those reporting ED use were coded as having an asthma attack in every year. The edit affected 1.2% of all patients coded as having an asthma attack. In accordance with previous work,28 we considered an ED visit to be an adverse outcome. eTable 5 in the Supplement reports results for asthma attack and ED use in the general population. Although results were similar, we preferred the conditional outcomes because they focus on children at risk for the events of interest.29

    Statistical Analysis

    We used logistic regression to implement the quasi–wait-list comparisons and report the mean incremental differences (percentage point differences) estimated by the models with 95% CIs.30 For each outcome, we examined 4 models. The first included only the HUD status variable. The second included controls for demographic characteristics: age, race/ethnicity, sex, census region, and survey year. The third added individual socioeconomic- and community-level factors that could potentially mediate the association between rental assistance and the outcomes: family poverty level, parental work status, parental marriage status, parental educational attainment, county-by-year air pollution (measured using ozone and particulate matter <2.5 μm), county-by-year rates of violent and property crime, and a series of census tract-by-year indexes that measured the racial/ethnic, economic, and housing characteristics of neighborhoods (eAppendix in the Supplement). The final model added individual health care access variables: health insurance coverage, usual source of care, and financial and nonfinancial barriers to care (eg, difficulty accessing care because of transportation). All models accounted for the complex sample design. Missing data were excluded using case-wise deletion except for poverty level, which was measured using imputed family income provided by the National Center for Health Statistics. Two-sided significance testing using a threshold of P ≤ .05 was used. Analyses were conducted at a federal statistical research data center using Stata software, version 14 (StataCorp).

    Sensitivity Analysis

    A potential threat to the validity of our study was omitted variables that varied over time and were associated with the timing of HUD entry. To investigate that issue, we estimated similar models to those described above except that we replaced the 0 or 1 HUD status variable with variables that measured years relative to program entry (ranging from 3 years earlier to ≥3 years after entry). These models used a longer wait-list period to better measure trends leading to program entry. If outcome trends were observed before entry, it would suggest that the comparison of interest was being affected by a time-varying process other than entry itself. However, if outcomes changed only after entry, it would support the validity of our design. We also examined whether the comparisons of interest were sensitive to alternative wait-list lengths.

    Results
    Sample

    A total of 2992 children who were currently participating in a HUD program or would enter a program within 2 years participated in the study. Table 1 gives selected characteristics of children in the quasi–wait-list and current participation groups. Members of the quasi–wait-list group were younger (mean [SD] age, 6.8 [3.1] vs 8.2 [3.2] years) and less likely to reside in the Midwest (168 [weighted 17.1%] vs 470 [weighted 23.2%]). We did not find evidence that the 2 groups differed by sex, race/ethnicity, or parental educational attainment. eTable 2 in the Supplement provides comparisons across a larger set of variables and comparisons with low-income children never participating in a HUD program.

    Main Results

    In the quasi–wait-list group, 24.1% of children had been diagnosed with asthma, 45.6% of those with an asthma diagnosis had an asthma attack in the past 12 months, and 60.7% of those with an asthma attack visited the ED for asthma (Table 2). These rates were higher than rates in the general low-income population (eTable 3 in the Supplement). Unadjusted logistic regression did not reveal a statistically significant difference in asthma diagnosis (0.9 percentage points; 95% CI, −5.1 to 3.2 percentage points) or past-year asthma attack (−2.0 percentage points; 95% CI, −11.7 to 7.7 percentage points) for current HUD participants compared with those waiting to enter HUD programs. Emergency department use, conditional on having an asthma attack, was 15.3 percentage points (95% CI, −29.0 to −1.6 percentage points) lower among current participants compared with the quasi–wait-list group. This finding represents a 25% relative reduction.

    Table 2 gives results from models that include control variables. After demographic controls were included, the size of the difference for ED use for those having an asthma attack was −18.9 percentage points (95% CI, −31.5 to −6.2 percentage points). This estimate was not substantively affected by the inclusion of the individual socioeconomic, community, or health care access variables. The difference for asthma attack was approximately the same magnitude in every model and never statistically significant. Differences in asthma diagnosis were larger when including demographic controls (−4.1 percentage points; 95% CI, −8.5 to 0.2 percentage points). The point estimates and 95% CIs in the asthma diagnosis model were substantively unchanged in the fully adjusted model. eTable 6 in the Supplement compares logistic regression results found in Table 2 with results from a linear probability model.

    Subgroup Analyses

    The percentage point difference between current participants in housing choice voucher programs and their quasi–wait-list counterparts was −7.2 (95% CI, −24.6 to 10.3 percentage points). The difference for those in multifamily or public housing was −36.6 percentage points (95% CI, −54.8 to −18.4 percentage points; P = .02 for interaction). eTables 7 in the Supplement give program-specific results for those with a diagnosis of asthma.

    Sensitivity Analysis

    The Figure displays results from a fully adjusted logistic regression model that replaced the 0 or 1 HUD status indicator with time relative to HUD entry. The percentage point differences plotted in the Figure were in comparison with 1 year before program entry. The outcome was ED visit, and we estimated differences for those with an asthma attack (eFigure 2 in the Supplement gives results for those with an asthma diagnosis). The Figure shows no discernable trend in ED use in the years before entry. However, after HUD entry, ED use was consistently lower among current HUD participants by 10 to 18 percentage points, and the association was larger after at least 2 years of participation. Only the point estimate for 3 or more years of participation was statistically significant (−18.0 percentage points; 95% CI, −34.5 to −1.0 percentage points). The shape of the coefficients suggests that differences in ED use were not associated with preexisting trends leading to HUD entry. The eAppendix in the Supplement shows that using a longer wait-list period to define the comparison group led to moderately different results but not enough to alter conclusions (eTables 8-10 in the Supplement).

    Discussion

    In this study, we investigated whether participation in a federal rental assistance program was associated with childhood asthma outcomes. We found that among children receiving rental assistance who had an asthma attack, there was a 15.3– to 18.2–percentage point reduction in ED use for asthma compared with children waiting to enter a program. This finding translates into a 25% to 30% relative reduction.

    Differences in ED use were larger for participants in multifamily or public housing but smaller and not significant among those with housing choice vouchers. This finding aligns with evidence from the MTO trial, which found that providing vouchers to public housing residents was associated with worse asthma outcomes.10,13 Similarly, a study31 of a Chicago housing lottery found no association with ED use when low-income families in private housing were provided a voucher. The reasons why place-based housing might be beneficial in ways that vouchers are not are not well understood.13 The presence of richer social networks is one avenue worth investigation.32,33 Voucher recipients may also face additional challenges associated with finding a private market unit that meets their needs.16,17

    Our estimates were not sensitive to a large set of controls. Reductions in ED use were not explained by neighborhood characteristics or better access to primary care. Our study could not determine whether lower ED use was associated with differential asthma attack severity, unobserved differences in access to an ED or alternative sites of care, or better management of symptoms outside the ED. Given that adherence to asthma treatment can be disrupted by social stressors and enhanced by regular household routines,34,35 it is possible that the stability that rental assistance provides could enable families to better manage asthma symptoms when they occur.

    We did not find evidence that rental assistance was associated with reduced asthma attacks, which could suggest that rental assistance may not be associated with improved underlying health. Such results are consistent with experimental evidence from the Family Options Study36 and evidence from the Panel Study of Income Dynamics,37 which suggests that rental assistance has no association with parent-reported child health. However, the CIs in the asthma attack models were wide, and we cannot rule out clinically meaningful increases or decreases in the incidence of asthma attack, and we did not observe asthma attack severity.

    We also observed differences in ever being diagnosed with asthma, which were significant at the 10% level in the fully adjusted model. That difference was unexpected given that current HUD assistance program recipients were previously exposed to a wait period and thus might be expected to have the same or an increased risk of a past diagnosis. It is possible that in some cases the ED visit was the child’s first asthma-related medical encounter, in which case it would follow that the group that was less likely to visit the ED would also have a lower prevalence of ever being diagnosed. Another explanation is differential response error. Emergency department visits are salient events that could improve recall of past asthma diagnoses. The higher rate of ED use in the quasi–wait-list group might have been associated with improved reporting compared with current participants.

    Limitations

    This study had a number of limitations. We cannot definitively exclude the possibility that the changes we observed were associated with factors other than entry into a HUD program. However, the design accounted for time-invariant selection by comparing current HUD program participants with those who would enter such programs in the future. We also did not observe trends leading to HUD entry, suggesting that our results are not associated with time-varying factors, such as changes in a family’s economic status or regression to the mean. Our study period extended back to 1999 and might not be generalizable to current HUD programs. For example, the size and composition of the HUD program population change with changes in housing markets, the broader economy, and program rules and regulations. Furthermore, the relative benefit of rental assistance is a function of the affordability of housing.

    Conclusions

    Our findings may have important implications for policy makers and health care practitioners who are considering how to leverage health care system resources in cross-sector collaborations that address the social determinants of health. Our estimates suggest that rental assistance may be associated with health care system savings. However, the magnitude remains uncertain given that we did not observe the number of avoided ED visits or typical treatment intensity.

    Back to top
    Article Information

    Accepted for Publication: October 16, 2019.

    Corresponding Author: Michel Boudreaux, PhD, Department of Health Policy and Management, University of Maryland School of Public Health, 4200 Valley Dr, College Park, MD 20742 (mhb@umd.edu).

    Published Online: March 9, 2020. doi:10.1001/jamapediatrics.2019.6242

    Author Contributions: Dr Boudreaux had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: All authors.

    Acquisition, analysis, or interpretation of data: All authors.

    Drafting of the manuscript: Boudreaux.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: Boudreaux, Fenelon, Slopen.

    Obtained funding: Boudreaux, Fenelon, Slopen.

    Administrative, technical, or material support: Boudreaux, Fenelon.

    Supervision: Boudreaux.

    Conflict of Interest Disclosures: None reported.

    Funding/Support: This study was supported by a grant from The Eunice Kennedy Shriver National Institute of Child Health and Human Development.

    Role of the Funder/Sponsor: The funding organization had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    References
    1.
    Azar  A.The Root of the Problem: America’s Social Determinants of Health. Washington, DC: Department of Health and Human Services; November 2018.
    2.
    Alley  DE, Asomugha  CN, Conway  PH, Sanghavi  DM.  Accountable health communities: addressing social needs through Medicare and Medicaid.  N Engl J Med. 2016;374(1):8-11. doi:10.1056/NEJMp1512532PubMedGoogle ScholarCrossref
    3.
    Council on Community Pediatrics.  Poverty and child health in the United States.  Pediatrics. 2016;137(4):e20160339. doi:10.1542/peds.2016-0339PubMedGoogle Scholar
    4.
    Kaiser Permanente. Announcing $200M impact investment to address housing crisis. May 2018. https://share.kaiserpermanente.org/article/kaiser-permanente-announces-200-million-impact-investment-partners-with-u-s-mayors-and-ceos-to-address-housing-stability/. Accessed June 26, 2019.
    5.
    Council on Community Pediatrics.  Providing care for children and adolescents facing homelessness and housing insecurity.  Pediatrics. 2013;131(6):1206-1210. doi:10.1542/peds.2013-0645PubMedGoogle ScholarCrossref
    6.
    Leventhal  T, Newmann  S.  Housing and child development.  Child Youth Serv Rev. 2010;32:1165-1174. doi:10.1016/j.childyouth.2010.03.008Google ScholarCrossref
    7.
    Watson  NE, Steffen  BL, Martin  M, Vandenbroucke  DA.  Worst Case Housing Needs: Report to Congress. Washington, DC: US Department of Housing and Urban Development; 2017.
    8.
    Slopen  N, Fenelon  A, Newman  S, Boudreaux  M.  Housing assistance and child health: a systematic review.  Pediatrics. 2018;141(6):1-14. doi:10.1542/peds.2017-2742PubMedGoogle ScholarCrossref
    9.
    Center for Budget and Policy Priorities.  Federal Rental Assistance. Washington, DC: Center for Budget and Policy Priorities; 2017.
    10.
    Forston  JG, Sanbonmatsu  L.  Child health and neighborhood conditions: results from a randomized housing voucher experiment.  J Hum Resour. 2010;45(4):840-864.Google Scholar
    11.
    Jacob  BA, Ludwig  J, Miller  DL.  The effects of housing and neighborhood conditions on child mortality.  J Health Econ. 2013;32(1):195-206. doi:10.1016/j.jhealeco.2012.10.008PubMedGoogle ScholarCrossref
    12.
    Kessler  RC, Duncan  GJ, Gennetian  LA,  et al.  Associations of housing mobility interventions for children in high-poverty neighborhoods with subsequent mental disorders during adolescence.  JAMA. 2014;311(9):937-948. doi:10.1001/jama.2014.607PubMedGoogle ScholarCrossref
    13.
    Schmidt  NM, Lincoln  AK, Nguyen  QC, Acevedo-Garcia  D, Osypuk  TL.  Examining mediators of housing mobility on adolescent asthma: results from a housing voucher experiment.  Soc Sci Med. 2014;107:136-144. doi:10.1016/j.socscimed.2014.02.020PubMedGoogle ScholarCrossref
    14.
    Zahran  HS, Bailey  CM, Damon  SA, Garbe  PL, Breysse  PN.  Vital signs: asthma in children—United States, 2001-2016.  MMWR Morb Mortal Wkly Rep. 2018;67(5):149-155. doi:10.15585/mmwr.mm6705e1PubMedGoogle ScholarCrossref
    15.
    National Heart, Lung, and Blood Institute.  Expert Panel Report 3: Guidelines for the Diagnosis and Management of Asthma. Bethesda, MD: National Heart, Lung, and Blood Institute; 2007.
    16.
    Popkin  SJ, Katz  B, Cunningham  MK, Brown  KD, Gustafson  J, Turner  MA.  A Decade of HOPE IV: Research Findings and Policy Challenges. Washington, DC: The Urban Institute; 2004.
    17.
    Finkel  M, Buron  L. Study on Section 8 Voucher Success Rates: Quantitative Study of Success Rates In Metropolitan Areas. Vol 1. Cambridge, MA: Abt Associates; 2001.
    18.
    Smith  R, Naparstek  A, Popkin  S,  et al.  Housing Choice for HOPE VI Relocatees. Washington, DC: The Urban Institute; 2002.
    19.
    Pearson  WS, Goates  SA, Harrykissoon  SD, Miller  SA.  State-based Medicaid costs for pediatric asthma emergency department visits.  Prev Chronic Dis. 2014;11:E108. doi:10.5888/pcd11.140139PubMedGoogle Scholar
    20.
    Lantz  PM.  The medicalization of population health: who will stay upstream?  Milbank Q. 2019;97(1):36-39. doi:10.1111/1468-0009.12363PubMedGoogle ScholarCrossref
    21.
    Lloyd  PC, Helms  VE.  NCHS-HUD Linked Data: Analytic Considerations and Guidelines. Hyattsville, MD: National Center for Health Statistics; 2019.
    22.
    Parsons  VL, Moriarity  C, Jonas  K,  et al.  Design and Estimation for the National Health Interview Survey, 2006-2015. Hyattsville, MD: National Center for Health Statistics; 2014.
    23.
    Blewett  LA, Rivera Drew  JA, Griffin  R, King  ML, Williams  KCW.  IPUMS Health Surveys: National Health Interview Survey, Version 6.3. Minneapolis, MN: IPUMS; 2018.
    24.
    US Department of Housing and Urban Development. Assisted housing: national and local. https://www.huduser.gov/portal/datasets/assthsg.html. Accessed August 1, 2019.
    25.
    Fenelon  A, Mayne  P, Simon  AE,  et al.  Housing assistance programs and adult health in the United States.  Am J Public Health. 2017;107(4):571-578. doi:10.2105/AJPH.2016.303649PubMedGoogle ScholarCrossref
    26.
    Fenelon  A, Slopen  N, Boudreaux  M, Newman  SJ.  The impact of housing assistance on the mental health of children in the United States.  J Health Soc Behav. 2018;59(3):447-463. doi:10.1177/0022146518792286PubMedGoogle ScholarCrossref
    27.
    Simon  AE, Fenelon  A, Helms  V, Lloyd  PC, Rossen  LM.  HUD housing assistance associated with lower uninsurance rates and unmet medical need.  Health Aff (Millwood). 2017;36(6):1016-1023. doi:10.1377/hlthaff.2016.1152PubMedGoogle ScholarCrossref
    28.
    Nath  JB, Hsia  RY.  Children’s emergency department use for asthma, 2001-2010.  Acad Pediatr. 2015;15(2):225-230. doi:10.1016/j.acap.2014.10.011PubMedGoogle ScholarCrossref
    29.
    Akinbami  LJ, Moorman  JE, Simon  AE, Schoendorf  KC.  Trends in racial disparities for asthma outcomes among children 0 to 17 years, 2001-2010.  J Allergy Clin Immunol. 2014;134(3):547-553.e5. doi:10.1016/j.jaci.2014.05.037PubMedGoogle ScholarCrossref
    30.
    Norton  EC, Dowd  BE, Maciejewski  ML.  Marginal effects—quantifying the effect of changes in risk factors in logistic regression models.  JAMA. 2019;321(13):1304-1305. doi:10.1001/jama.2019.1954PubMedGoogle ScholarCrossref
    31.
    Jacob  BA, Kapustin  M, Ludwig  J.  The impact of housing assistance on child outcomes: evidence from a randomized housing lottery.  Q J Econ. 2015;130(1):465-506. doi:10.1093/qje/qju030Google ScholarCrossref
    32.
    Clampet-Lundquist  S.  “Everyone had your back”: social ties, perceived safety, and public housing relocation.  City Community. 2010;(1):87-108. doi:10.1111/j.1540-6040.2009.01304.xGoogle Scholar
    33.
    Keene  DE, Geronimus  AT.  Community-based support among African American public housing residents.  J Urban Health. 2011;88(1):41-53. doi:10.1007/s11524-010-9511-zPubMedGoogle ScholarCrossref
    34.
    Chen  E, Chim  LS, Strunk  RC, Miller  GE.  The role of the social environment in children and adolescents with asthma.  Am J Respir Crit Care Med. 2007;176(7):644-649. doi:10.1164/rccm.200610-1473OCPubMedGoogle ScholarCrossref
    35.
    Peterson-Sweeney  K.  The relationship of household routines to morbidity outcomes in childhood asthma.  J Spec Pediatr Nurs. 2009;14(1):59-69. doi:10.1111/j.1744-6155.2008.00175.xPubMedGoogle ScholarCrossref
    36.
    Gubits  D, Shinn  M, Wood  M,  et al.  Family Options Study: 3-Year Impacts of Housing and Services Interventions for Homeless Families. Washington, DC: US Dept of Housing and Urban Development; 2015.
    37.
    Newman  S, Holupka  CS.  The effects of assisted housing on child well-being.  Am J Community Psychol. 2017;60(1-2):66-78. doi:10.1002/ajcp.12100PubMedGoogle ScholarCrossref
    ×