Beyond covariates: an intersectional framework for housing conditions, disability, and health equity

1 Introduction

Housing conditions are among the most consequential social determinants of health. The physical quality of a dwelling, the security of tenure, and the affordability of housing each shape the environments in which people sleep, breathe, manage chronic conditions, and age, influencing health outcomes across the life course in ways that clinical care alone cannot address. The relationship between housing and health is well established in public health scholarship (1, 2), yet the field has largely examined it through single analytical lenses, asking how income poverty drives exposure to substandard housing, how racial residential segregation concentrates environmental hazards, or how disability creates barriers to accessible accommodation. What remains largely unexplored is how these dimensions intersect and how race, income, and disability interact with housing conditions to produce differential health outcomes that no single-factor analysis can fully explain.

The theoretical case for an intersectional approach draws on three complementary traditions. Crenshaw's foundational work on intersectionality established that systems of oppression, including racism, classism, and ableism, do not operate independently but interact to produce qualitatively distinct forms of disadvantage (3). Applied to housing and health, this means that a low-income Black person with a disability occupies a structural position shaped by the convergence of racial residential segregation, economic marginalisation, and the systematic inaccessibility of housing markets, and the health consequences of that position cannot be predicted from any single dimension alone. Krieger's ecosocial theory provides a complementary biological mechanism which is bodies literally incorporate into their biological structure the social conditions in which people live, leaving measurable traces in allostatic load, inflammatory markers, and neurological development (4). Psychosocial stress, chronic exposure to indoor environmental hazards, and the physiological wear of housing insecurity all contribute to this embodied burden, and the magnitude of these effects is shaped by the intersection of social positions rather than by any single factor. The capability approach associated with Sen and Nussbaum adds a third dimension, framing housing not merely as a physical condition but as a resource that enables or constrains people's fundamental capacities to achieve health and wellbeing (5). This framing is particularly important for integrating disability into the analysis, since disability scholarship has long argued that health outcomes for disabled people depend critically on whether their living environments enable or restrict functioning.

Despite the theoretical coherence of an intersectional approach, the housing-health literature remains largely organised around single social dimensions. Reviews of housing and mental health rarely disaggregate by disability. Reviews of racial disparities in housing conditions rarely examine how income moderates those disparities. Disability scholarship on housing and health rarely centres race or ethnicity. The result is a fragmented evidence base that cannot answer the question that matters most for policy and intervention design, namely who is most vulnerable to the health consequences of poor housing, and why. This fragmentation also has practical costs. Housing programmes, disability policy, and health equity initiatives are typically designed and implemented in silos, each addressing one dimension of disadvantage while remaining blind to the compounded vulnerabilities of those at the intersection of multiple disadvantaged positions. Without an intersectional evidence base, these programmes risk being calibrated for an average population that does not reflect those most in need, and may inadvertently widen disparities for those at the most disadvantaged intersections.

The housing-health relationship has a substantial earlier evidence base. Shaw's foundational review established the range of pathways through which housing conditions affect health, positioning housing as a core social determinant operating through direct physical exposures, psychosocial mechanisms, and neighbourhood-level processes (6). More recently, Anyanwu and Beyer developed a conceptual model integrating housing disparities into the environmental justice framework, documenting how historical redlining, exclusionary zoning, and segregation produce the spatial concentration of poor housing and environmental risk that disproportionately affects minoritised communities in the United States (7). Both contributions are important, but neither addresses the specific question this paper pursues. Shaw's review treats the housing-health relationship as broadly applicable across the population without examining how it differs across social groups. Anyanwu and Beyer's model treats race and income as structural context but does not position them as intersecting social positions that modify the health consequences of housing exposure, and disability is absent from their framework entirely. Neither paper addresses housing accessibility as a distinct exposure dimension relevant to disabled people, nor the bidirectionality of the disability-housing relationship. The IHCHE framework described below builds on these foundations while extending them in three directions: positioning race, income, and disability as intersecting modifiers rather than structural backdrop; introducing housing accessibility alongside physical conditions and security as a distinct and analytically necessary dimension; and accommodating the bidirectionality of disability through a framework architecture that treats it simultaneously as a social position and a health outcome.

A hybrid narrative review of the literature published between 2019 and 2026 is presented here to address this gap. The review is structured around a new conceptual framework, the Intersectional Housing Conditions and Health Equity (IHCHE) framework, developed to provide a theoretically grounded analytical structure for understanding how housing conditions produce differential health outcomes across intersecting social positions. The framework is presented before the literature is examined through its lens, so that the synthesis can be evaluated against an explicit analytical standard rather than organised purely descriptively. A structured gaps analysis and a research agenda specifying the study designs needed to fill them follow the synthesis. The aim is not to synthesise a settled body of consistent findings, as the literature is not yet at that stage, but to map an emerging field, establish what is known and what is not, and provide the conceptual and methodological foundations for the research that should follow.

2 Theoretical framework: the intersectional housing conditions and health equity (IHCHE) framework

The IHCHE framework brings together the social determinants of health framework, intersectionality theory, and the capability approach into a four-level analytical structure. The Dahlgren-Whitehead layered model positions housing as a living and working condition mediating between broader structural forces and individual health outcomes (1). Intersectionality theory (3) and Krieger's ecosocial elaboration (4) provide the basis for arguing that race, income, and disability interact multiplicatively rather than additively in shaping both exposure to housing conditions and the health consequences of those conditions. What multiplicative means here is that the health disadvantage experienced at the intersection of all three positions is greater than the sum of the disadvantages associated with each position separately. The capability approach (5) frames housing accessibility as a distinct and underappreciated dimension of housing that directly constrains or enables health for people with disabilities. The framework is presented in Figure 1.

2.1 Level 1: structural determinants

Housing inequality is produced by identifiable upstream forces. Historical policies such as redlining, exclusionary zoning, restrictive covenants, and discriminatory lending concentrated substandard housing stock in racially and economically marginalised communities, and the effects persist to the present day. Bryant-Stephens and colleagues document how this legacy is not merely historical. The housing stock in many minoritised communities remains structurally deficient as a direct legacy of disinvestment (8). Lemire and colleagues demonstrated that this structural disadvantage extends to institutional responses. Boston's code enforcement system responds more slowly and less effectively to housing complaints in predominantly Black and Hispanic neighbourhoods, with racial disparities in mold enforcement actually widening over the study period from 2011 to 2021. (9)

Contemporary structural racism operates through housing markets and regulatory systems as well as through historical legacy. Racial disparities in maintenance delays, documented by Schmidt using the American Housing Survey, reflect both structural racism in rental markets and individual discrimination, with different pathways operating for Black and Hispanic renters (10). United States federal welfare and disability policy interact with housing policy to shape who can access stable, affordable, and accessible housing. Trivedi and Brucker's analysis of United States federal rental housing assistance for disability households found that the programme reduced cost burden but did not address housing quality or accessibility gaps, indicating that affordability-focused structural interventions are insufficient when structural deficiencies and accessibility barriers are not also addressed (11). Income inequality, itself shaped by structural racism and ableism through labour market discrimination, constitutes the economic substrate that determines who can exercise market power in housing. Racism produces income inequality which produces housing disadvantage, and these pathways interact differently for people with disabilities, who face the compounded structural exclusions of ableism alongside racial and economic marginalisation. Treating any one of these forces in isolation therefore understates the disadvantage experienced at their intersection.

2.2 Level 2: housing conditions (the exposure)

The framework distinguishes three dimensions of housing conditions that the literature tends to conflate but which have distinct health implications and different relationships with the intersecting social positions at Level 3. Keeping them separate matters because the pathways through which each damages health are different, the populations most affected by each are not identical, and interventions that address one dimension may have no effect on another.

Physical conditions refers to the material quality of the dwelling, encompassing dampness, mold, overcrowding, indoor air quality, structural safety, heating adequacy, and exposure to toxins including lead and pest allergens. These physical hazards produce health outcomes through direct biological pathways, with indoor allergens triggering asthma exacerbations, lead exposure impairing neurodevelopment, and cold homes worsening cardiovascular and respiratory conditions. Housing security refers to the stability and predictability of occupancy, including eviction risk, affordability relative to income, homelessness, and frequency of residential moves. These produce health outcomes primarily through psychosocial pathways, including chronic stress, disrupted social networks, barriers to healthcare continuity, and the neurodevelopmental consequences of residential unpredictability in childhood (12). Housing accessibility refers to the degree to which the dwelling enables functioning for people with disabilities, covering physical adaptations, proximity to services, sensory environment, and the degree to which the built environment constrains or enables independence. This dimension is largely absent from existing housing and health research, which means the literature has systematically neglected the housing exposure most directly relevant to disabled people.

Wang and colleagues demonstrate the value of treating these dimensions separately. Their National Health and Aging Trends Study analysis modelled housing affordability, physical quality, and neighbourhood quality as distinct exposures and found substantially different outcome profiles. Affordability problems predicted dementia and disability as well as frailty and mortality, while physical quality predicted frailty, disability, and mortality but not dementia (13). Had these dimensions been collapsed into a single housing measure, these differences in outcome profile would have been invisible.

2.3 Level 3: intersecting social positions (the modifiers)

Race and ethnicity, income and socioeconomic status, and disability status are positioned at Level 3 not as simple risk factors to be controlled for in statistical models, but as social positions that shape both who is exposed to which housing conditions and how severe the health consequences of that exposure are. The distinction matters because controlling for race or income as a covariate treats these dimensions as statistical nuisances to be adjusted away, whereas intersectional analysis asks whether the housing-health relationship operates differently for people occupying different social positions, and why. The central claim is that these positions interact multiplicatively. A low-income disabled person of colour does not simply experience the sum of three separate disadvantages. They occupy a distinct structural position whose health implications cannot be predicted from any single dimension.

Disability occupies a particularly important and theoretically complex position here because of its bidirectionality. Disability is both a social position that shapes exposure to poor housing conditions, through the inaccessibility of housing markets, the inadequacy of housing adaptations, and the structural ableism of built environments, and a health outcome that poor housing conditions can produce or worsen. Murphy and Grundy document this bidirectionality empirically, showing that housing tenure predicts disability in longitudinal analyses while disability simultaneously constrains housing tenure options (14). This means that disability cannot be treated simply as either a cause or an effect in housing-health research. It is both, simultaneously, and any framework that does not accommodate this risks misspecifying the relationship.

The IHCHE framework focuses on three intersecting social positions, but it is important to acknowledge that other social dimensions can also modify the disability-housing-health relationship. Social support and social capital represent the most consequential of these. A person with a disability who has reliable informal support networks may be better able to navigate inaccessible housing, access community services, and buffer the psychosocial consequences of housing insecurity. Lucas and Chopik demonstrated using longitudinal panel data that post-event social support moderates declines in life satisfaction following the onset of a disabling condition, suggesting that social resources can partially offset the health consequences of adverse circumstances, including poor housing (15). Agha and colleagues showed that housing stability predicts acquisition of social capital among people with histories of homelessness, confirming that the relationship runs in both directions, i.e., social support shapes how people experience housing conditions, and housing conditions shape the social support available to them (16). The IHCHE framework does not formally incorporate social support at Level 3, both because doing so would substantially increase the complexity of the diagram and because the empirical evidence on social support as a modifier of the specific disability-housing-health intersection across race and income dimensions is not yet sufficient to characterise how it operates. The framework is, however, extensible such that social support and social capital can be incorporated as additional Level 3 modifiers in future applications, and their omission from the current version is acknowledged as a limitation in Section 7.

2.4 Level 4: health outcomes

Health outcomes in this framework are organised along two dimensions because conflating them obscures important differences in how housing conditions do their damage. The first dimension is health domain, covering physical health outcomes such as respiratory disease, cardiovascular disease, infectious disease, and mortality; mental health outcomes such as depression, anxiety, psychosocial stress, and wellbeing; cognitive health including dementia, cognitive decline, and neurodevelopment; and functional capacity including frailty, disability as an outcome, and activities of daily living. The second dimension is pathway, distinguishing direct pathways from physical housing conditions to health through biological exposure from indirect pathways through psychosocial stress, healthcare access barriers, health behaviour changes, and capability constraints. A damp dwelling damages respiratory health through a direct biological pathway; housing insecurity damages mental health through chronic stress and disrupted social networks. Both matter, but they call for different interventions.

The framework makes a specific argument about cognitive outcomes that distinguishes it from prior housing and health frameworks. Emerging evidence suggests that poor housing affordability predicts dementia onset through chronic stress and allostatic load pathways independent of other socioeconomic factors (13, 17), and this cognitive pathway has not yet been examined across the intersection of race, income, and disability. Given that Black and Hispanic Americans carry substantially higher dementia risk and also face substantially higher rates of housing disadvantage, the absence of intersectional research on this question represents a consequential gap. Health outcomes, particularly disability and chronic illness, can also reinforce the structural conditions that perpetuate housing disadvantage through their effects on employment, income, and policy eligibility, creating a feedback loop back to Level 1 that single-time-point studies cannot capture.

3 Methods

A hybrid narrative review design is used here, combining the interpretive flexibility of a narrative review with transparent search documentation and a structured evidence mapping approach that draws on scoping review methodology. This design is appropriate because the field is characterised by heterogeneous study designs and diverse outcome measures, and because the research question is better served by evidence mapping and gap identification than by formal synthesis of effect estimates, which would require a degree of methodological homogeneity the literature does not yet possess.

3.1 Search strategy

Web of Science and Dimensions were searched covering the period January 2019 to May 2026. Three complementary search strategies were used, each combining housing conditions, health outcomes, and subsets of the intersecting social positions of interest. Running the three searches separately rather than as a single combined string was a deliberate choice as it allowed the breadth of each partial literature to be mapped before assessing how much overlap existed at the fully intersectional level.

3.1.1 Search 1: full intersectional search (housing, health, race, income, disability)

“housing condition\” OR “housing quality” OR “substandard housing” OR “housing insecurity” OR “housing instability” OR “residential condition\” OR “housing precarity”) AND (“health outcome\” OR “health disparity” OR “health disparities” OR “mental health” OR “physical health” OR morbidity OR wellbeing) AND (race OR ethnicity OR “racial minority” OR minoritized OR Black OR Hispanic OR Latino OR Indigenous OR “ethnic minority”) AND (income OR poverty OR “socioeconomic status” OR deprivation OR “low-income”) AND (disability OR disabilities OR “functional limitation\” OR impairment OR “disabled people” OR “chronic illness”

3.1.2 Search 2: housing, health, race and income (without disability)

As above but omitting the disability domain terms. This search established the broader race-income-housing-health evidence base as a baseline.

3.1.3 Search 3: housing, health and disability (without race and income)

As Search 1 with housing, health, and disability domain terms but omitting race and income terms. This search mapped the disability-housing-health literature independently.

Web of Science searches were conducted using plain text export format (full record) and Dimensions searches used RIS format. Records were imported into R 4.6.0 and deduplicated using the ASySD package (18). The three searches across both databases yielded 375 records prior to deduplication, leaving 280 unique records for review.

3.2 Record selection

All 280 deduplicated records were reviewed against three core criteria that follow directly from the review question. A study had to measure housing conditions in some concrete way, not merely note housing as a background demographic; it had to report at least one health outcome; and at least one of the three intersecting social positions (race/ethnicity, income, or disability) had to feature in the study design. Studies conducted entirely in institutional settings such as hospitals or care homes were excluded because the housing conditions there are not comparable to community living and would introduce a different exposure entirely. Studies where the primary outcome was economic rather than health-related were also excluded. Studies engaging with all three intersecting factors were prioritised in the synthesis, but studies addressing two of the three were included to characterise the partial evidence base and make the gaps visible. Of the 280 records reviewed, 48 were excluded because housing was noted only as a background demographic without a concrete housing condition measure, 22 were excluded because the study was conducted entirely in institutional settings, and 19 were excluded because the primary outcome was economic rather than health-related, yielding 191 records included in the synthesis.

3.3 Data extraction and synthesis

For each included study, structured data extraction recorded the housing variable(s) measured, health outcome(s) reported, whether and how each intersecting social position featured (as a primary variable, subgroup, or covariate), study design, country, and key findings. Studies were then positioned within the four levels of the IHCHE framework at the level most relevant to their primary contribution.

The synthesis draws selectively on the 191 included records, with studies chosen for explicit discussion on the basis of their direct contribution to the framework levels and the arguments being developed. Studies informing the synthesis but not explicitly discussed in the text are not individually cited, consistent with narrative review conventions.

4 Literature synthesis

The synthesis below is organised by the four levels of the IHCHE framework, asking at each level what the literature shows, how the intersecting social positions of race, income, and disability feature, and where evidence is present or absent. Section 5 provides the structured gaps analysis.

4.1 Level 1: evidence on structural determinants

The most developed evidence at this level concerns structural racism and its legacy in producing racially differentiated housing conditions, though studies that directly measure or model structural factors remain relatively few. Bryant-Stephens and colleagues review how redlining, displacement, and exclusionary zoning produced the structurally deficient housing stock that characterises many minoritised communities, creating indoor environments rife with asthma triggers that drive racial disparities in childhood asthma (8). Lemire and colleagues add the critical finding that structural racism operates not only through historical housing investment patterns but through contemporary institutional responses. Boston's code enforcement system responds more slowly and less effectively in predominantly Black and Hispanic neighbourhoods, with disparities in mold enforcement actually widening rather than closing over a decade (9).

Schmidt distinguishes between two mechanisms producing racial disparities in housing maintenance delays i.e., structural racism in rental markets, which shapes the experiences of both Black and Hispanic renters, and individual landlord discrimination, which operates more strongly for Black renters than Hispanic renters (10). Anyanwu and Beyer's conceptual review integrates housing instability, including eviction, into a broader environmental justice framework, positioning historical redlining as a central driver of the contemporary housing-health inequity landscape (7). Together, these papers establish that structural racism is not merely a background context for housing-health research but an active and ongoing mechanism that continues to produce and deepen inequities.

Evidence on welfare and disability policy as structural determinants is thinner but present. Trivedi and Brucker's analysis demonstrates that the structural design of United States federal rental housing assistance for disability households addresses affordability but not quality or accessibility gaps (11). Cheetham and colleagues document how the structural redesign of the United Kingdom welfare system through Universal Credit drove housing insecurity among claimants with complex needs and disabilities, with clear mental health consequences, providing a quasi-natural experiment in how welfare policy restructuring becomes a structural determinant of housing-health inequity (19).

4.2 Level 2: evidence on housing conditions as exposures

4.2.1 Physical housing conditions

The strongest evidence on physical housing conditions concerns respiratory health, where the literature engages explicitly with race and income as modifying factors. Bryant-Stephens and colleagues establish that the physical housing conditions driving paediatric asthma, including indoor allergens, mold, pest infestations, secondhand smoke, and poor ventilation, are disproportionately concentrated in urban low-income minority renting environments as a direct legacy of structural racism (8). Patti and colleagues provide clinical guidance for addressing this problem, noting that unhealthy housing most frequently occurs in urban low-income renting environments and proposing multi-disciplinary approaches to exposure reduction (20). Liu and colleagues document that caregivers of children with asthma in Washington DC subsidised housing frequently encounter multiple simultaneous triggers, with structural barriers tied to geography, race, and socioeconomic status impeding healthy housing services (21).

Beyond respiratory health, physical housing conditions predict mental health through multiple pathways. The systematic review by Singh and colleagues, the methodologically strongest contribution to this question, reviewed 12 longitudinal studies of housing disadvantage and mental health and found consistent positive associations across diverse measures of physical housing conditions, with follow-up periods ranging from within one year to 34 years (22). Park and colleagues extend this finding to adults with disabilities, showing that structural unsafety and inadequate ventilation predict worsening depressive symptoms in longitudinal analysis, with the effect less pronounced for those living near community services (23).

Cheng and colleagues illustrate how physical conditions interact with disability and ageing, using data from 30,632 older adults in India to show dose-response associations between housing quality, measured across sanitation, water access, cooking fuel, and electricity, and fall risk, with poor sanitation, inadequate water access, solid fuel use, and indoor air pollutants all independently predicting falls (24). Since falls are a leading cause of disability and mortality among older adults, this work demonstrates a concrete pathway through which physical housing conditions produce disability as an outcome.

4.2.2 Housing security

Housing insecurity, encompassing homelessness, eviction risk, rent arrears, affordability stress, and frequent moves, is the most frequently studied housing dimension in the 2019 to 2026 literature, partly as a consequence of the COVID-19 pandemic generating intense research interest in the health consequences of economic disruption and eviction. Wang and colleagues provide the most methodologically rigorous evidence for physical health outcomes, using a prospective cohort design to show that poor housing affordability independently predicts frailty, disability, dementia, and mortality among older Americans, and that poor physical quality predicts frailty, disability, and mortality (13). Rao and colleagues found that county-level eviction rates were associated with higher all-cause mortality, with the strongest associations in counties with the highest proportions of Black and women residents, confirming that the eviction-health relationship is not race-neutral (25).

Renson and colleagues used the Future of Families and Child Wellbeing Study to show that between 45% and 71% of children in large United States cities experienced at least one housing insecurity event by age 15, with Black and Hispanic children more likely than White children to experience a first event (26). Foster and colleagues document the neurodevelopmental pathways through which residential instability affects children's mental health, showing that residential unpredictability, a specific and underappreciated dimension of housing insecurity, disrupts neural circuitry supporting cognitive and emotional development (12).

4.2.3 Housing accessibility

Housing accessibility as a distinct dimension is the most poorly evidenced of the three in this literature. Studies of people with disabilities and housing conditions rarely measure accessibility specifically, using instead general housing quality or tenure as proxies. Murphy and Grundy document that social renters (households renting from local authorities or registered housing associations, as distinct from private renters or owner-occupiers) have dramatically higher rates of disability and health-related economic inactivity than owner-occupiers, but the analysis does not distinguish between the effects of housing quality, affordability, and accessibility on these outcomes (14). Heyman and colleagues note that disabled parents face higher rates of unaffordable housing and renting but do not measure whether the housing stock is accessible to their disability-related needs (27). This is a major evidence gap that the IHCHE framework makes visible.

4.3 Level 3: evidence on intersecting social positions

4.3.1 Race and ethnicity

Race and ethnicity are the most extensively studied of the three intersecting factors, and the literature consistently documents racial disparities in both exposure to poor housing conditions and health consequences. Nakphong and colleagues document high prevalence of multiple forms of housing instability among low-income urban Black young adults, with unaffordable and overcrowded housing associated with four times higher odds of depression compared with more stably housed peers (28). Oláh and colleagues find that Roma adults living in segregated settlements have more than double the rate of high mental morbidity risk compared with the general Hungarian population, with settlement type rather than ethnic identity the primary predictor, a finding that warns against conflating racial or ethnic group membership with the housing conditions that actually produce health disadvantage (29).

The asthma literature provides the most developed evidence on how race and income jointly shape housing-related health outcomes. Racial disparities in asthma triggers prevalence are driven by the intersection of racial residential segregation, determining where people live, and income, determining what housing quality they can afford. Lemire and colleagues show these disparities are compounded by institutional responses that are themselves racially differentiated (9). Race and income are not independent modifiers of the housing-health relationship; they interact through structural pathways that produce outcomes neither would generate alone.

4.3.2 Income and socioeconomic status

Bundy and colleagues provide the most direct finding at this level, showing that social determinants including housing instability completely mediated the Black-White gap in premature mortality, which suggests that income and socioeconomic factors, through their operation on housing and other conditions, account for much of the observed racial health gap (30). Wang shows that the adverse health consequences of neighbourhood housing instability, spanning mental health, cardiovascular outcomes, and mortality, vary by income level as well as race, with lower-income neighbourhoods showing more persistent and severe effects than higher-income neighbourhoods facing comparable levels of housing instability (31).

Graff and colleagues show that Black and Hispanic households are more likely than White households to be energy insecure even after accounting for deficient housing conditions and energy burdens, suggesting additional unmeasured structural mechanisms related to racial discrimination (32). Ballesteros-Arjona and colleagues find in their scoping review that energy poverty's health effects are consistently worse for vulnerable groups across dimensions of inequality, though disaggregated analysis by the intersection of race and income is largely absent from the underlying studies (33).

4.3.3 Disability status

Disability is the least analytically developed of the three intersecting factors, and its role in the housing-health relationship is the most complex because of the bidirectionality noted in the framework. Murphy and Grundy provide the longitudinal evidence with the longest follow-up, showing that housing tenure predicts disability in United Kingdom adults aged 40 to 69, with social renters (households in publicly subsidised rental accommodation) showing dramatically higher rates of disability and illness than owner-occupiers, and projections indicating that these rates will continue to diverge through 2030 (14). Disability is also produced and exacerbated by poor housing conditions through multiple pathways. Cheng and colleagues show that poor physical housing quality predicts falls in older adults, a leading cause of functional decline and disability (24). Park and colleagues show that structural housing deficiencies predict worsening depressive symptoms in adults already living with disabilities, an effect that diminishes when community services are nearby, suggesting that neighbourhood resources can partially buffer the damage poor housing does to mental health (23). Wang and colleagues show that both housing affordability and quality problems predict the onset of frailty and disability in older Americans (13). Taken together, these studies confirm that disability and housing conditions are bound together in a relationship that runs in both directions.

4.4 Level 4: evidence on health outcomes

4.4.1 Mental health

The systematic review by Singh and colleagues is the methodological foundation for the mental health evidence base, establishing through 12 longitudinal studies that housing disadvantage across multiple dimensions is causally antecedent to poor mental health (22). Nakphong and colleagues and Opia and colleagues extend this to low-income Black populations specifically, finding that multidimensional housing instability compounds mental health risk in ways that affordability or crowding alone cannot explain (28, 34). Westbrook documents how displacement pressure, rather than displacement itself, becomes a chronic embodied health burden for Hispanic and Latinx immigrant communities (35). Oláh and colleagues extend this evidence to Roma populations in Hungary, finding that settlement type predicts mental morbidity more strongly than ethnic identity itself, a finding that shows why housing conditions must be measured directly rather than inferred from racial or ethnic group membership (29).

The COVID-19 pandemic generated a cluster of studies demonstrating that housing insecurity predicted mental health deterioration across multiple racial and ethnic groups, with consistently stronger effects for racial and ethnic minorities and for lower-income households. Kim, using United States Household Pulse Survey data, found that financial hardship closely linked to housing insecurity was associated with nearly three-fold higher risks of anxiety and depressive symptoms (36). Alhomsi and colleagues found dose-response associations between financial hardship and psychological distress across all racial and ethnic groups, with Spanish-speaking Latino and Native Hawaiian and Pacific Islander adults experiencing the highest rates of hardship (37).

4.4.2 Physical health outcomes

The respiratory health literature is the most developed physical health evidence base linking housing conditions to differential outcomes by race and income. Asthma in children provides the paradigmatic case, with well-documented associations between indoor physical housing conditions and asthma morbidity, explicit engagement with racial and socioeconomic disparities as drivers, and emerging evidence on the role of institutional failure in compounding disparities (8, 9, 20, 21). Cardiovascular outcomes are increasingly studied alongside respiratory ones. Wang and colleagues and Rao and colleagues document associations between housing instability and outcomes including stroke and mortality, with racial variation in these associations that mirrors the patterns found in the asthma literature (13, 25, 38).

4.4.3 Cognitive outcomes and older adults

Wang and colleagues provide strong evidence on housing and cognition, showing that poor housing affordability predicts dementia onset among older Americans after adjustment for a range of socioeconomic and health covariates, suggesting a pathway through chronic psychosocial stress and allostatic load (13). Baek and Yoon replicate this association for multidimensional housing precarity and cognitive function among older Koreans (17). Neither study examines the intersection of race, income, and disability with housing and cognitive outcomes, which matters because Black and Hispanic Americans carry substantially higher dementia risk and also face substantially higher rates of housing disadvantage. Whether those two facts are causally connected through housing conditions is a question the literature cannot currently answer.

4.4.4 The COVID-19 pandemic as natural experiment

The COVID-19 pandemic constitutes an important natural experiment in housing-health inequity. Multiple systematic reviews and population-level studies document that housing conditions, including overcrowding, poor quality, and housing insecurity, mediated and amplified the differential COVID-19 health burden falling on racial and ethnic minorities and lower-income populations (39, 40). Parolin and Lee identify housing quality and neighbourhood conditions as one of four mechanisms through which poverty and racial discrimination exacerbated COVID-19 health consequences (41). Li and colleagues’ quasi-experimental study of emergency rental assistance provides the strongest causal evidence that housing policy directly affected COVID-19 mortality rates at the county level (42).

5 Gaps analysis

Measuring the literature against the IHCHE framework reveals five categories of gap, each reflecting a different way in which existing research falls short of what is needed to understand how housing conditions produce differential health outcomes across intersecting social positions.

5.1 Intersectional analytical gap

The most fundamental gap is the near-absence of studies that formally model the interaction between race, income, and disability in relation to housing conditions and health outcomes simultaneously. The overwhelming majority of studies include these factors as covariates, which adjusts for them statistically but does not test whether the housing-health relationship operates differently across the populations they define. Kohlbeck and colleagues come closest, finding that the association between housing mobility and adolescent suicidality shows disparities by race, disability status, and sexual orientation simultaneously (43). Brokamp and colleagues use causal mediation analysis to show that place-based exposures including housing conditions fully mediate racial inequities in paediatric hospitalisation, a methodological model for the field (44). A review of all 280 records found no study that formally tests the three-way interaction between race, income, and disability as modifiers of the housing-health relationship. Search 1, the fully intersectional search requiring all five concept domains simultaneously, returned 21 records before deduplication. All 21 met the keyword criteria but none contained formal intersectional modelling. They included the terms race, income, and disability as descriptive background characteristics rather than as analytical variables whose interactions were tested. This confirms that the absence of intersectional evidence is an empirical finding of the search, not an artefact of how the searches were constructed.

5.2 Disability intersection gap

Disability is substantially underrepresented as an analytical focus in the housing-health literature. The two studies that most directly engage with disability as a primary analytical focus alongside housing conditions, namely Park and colleagues (23) and Murphy and Grundy (14), are from Korea and the United Kingdom respectively, and neither focuses on the intersection with race or income. This matters because disabled people of colour and disabled people in poverty face compounded structural disadvantages that studies treating disability in isolation cannot capture, and because the bidirectionality of the disability-housing relationship means that failing to account for race and income when studying disability and housing also risks confounding.

5.3 Housing accessibility measurement gap

The accessibility dimension of housing conditions identified in the IHCHE framework is almost entirely absent from the empirical literature. Studies consistently measure physical conditions and increasingly measure housing security, but housing accessibility, i.e., the degree to which the dwelling enables functioning for people with disabilities, is rarely measured as a distinct exposure. Since accessibility is the dimension most directly relevant to the health of disabled people, its absence from measurement means that the housing dimension with the greatest potential to explain disability-related health disparities has almost no evidence base, and that evaluations of housing modification programmes for disabled people rarely disaggregate outcomes by race or income.

The absence of validated multidimensional housing assessment tools suitable for population health research is a central reason accessibility has been so poorly measured. Existing instruments for assessing housing accessibility were largely developed for clinical home modification decisions or for single-country applications, and not validated for use across racial and income subgroups in large-scale population health research. The predominance of physical condition and tenure measures in the reviewed literature therefore, reflects in part the availability of validated instruments for these dimensions rather than purely theoretical choices by researchers, which is a path dependency in measurement that the field needs to actively disrupt.

5.4 Cognitive and longitudinal outcomes gap

Black Americans and Hispanic Americans carry substantially higher dementia risk than White Americans, and they also face substantially higher rates of housing-related disadvantage. Whether those two facts are causally connected through housing conditions, and whether the relationship is moderated by income or disability status, remains unanswered. Longitudinal evidence more broadly is thinner than cross-sectional evidence, which limits what can be said about cumulative and lagged effects. Housing conditions accumulated over decades may produce health consequences that short-follow-up studies cannot detect.

5.5 Intervention evidence gap

The intervention and policy evidence base is thin, particularly for population subgroups defined by the intersection of race, income, and disability. The Housing First literature finds improvements across health domains when homeless individuals are provided permanent housing, but most evidence focuses on chronically homeless men, leaving outcomes for racial minority populations, families, women, older adults, and disabled individuals largely unexamined (45). Li and colleagues’ rental assistance study shows that housing policy can reduce mortality, but does not report heterogeneous effects by race or disability (42). Without evidence on whether and how intervention effects vary across intersectional groups, there is no basis for targeting programmes to those most in need or for detecting when an intervention that works for one group may be less effective or even harmful for another.

6 Research agenda

The gaps identified above translate into five research priorities. Each priority is described in terms of the study design most likely to address it, so that the agenda specifies what should be done rather than merely what is needed.

6.1 Studies modelling intersectional effects

The field needs studies that formally test interactions between race, income, and disability as modifiers of the housing-health relationship, rather than controlling for these factors. Controlling for race or disability in a regression model asks whether average housing effects persist after adjusting for group membership; interaction testing asks whether the effect is larger or smaller in some groups than others and why. Prospective cohort studies with adequate representation of racial and ethnic minority participants, disabled people, and low-income households, and with sufficient sample sizes to power interaction tests, are needed. Analysis should pre-specify interaction terms for race-by-income, race-by-disability, income-by-disability, and ideally the three-way interaction. Multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA) is an intersectional quantitative method that models how health outcomes are distributed across intersectional strata rather than averaging across them (46, 47). It was developed specifically to address the limitations of conventional regression approaches to intersectionality and should be adopted more widely in this field.

6.2 Multidimensional housing measurement

Studies that measure only one dimension of housing conditions cannot determine whether their findings generalise to other dimensions, which makes cumulative evidence difficult to build. Future studies should measure physical conditions, housing security, and housing accessibility within a single study, enabling examination of their independent and joint effects on health outcomes. At present no validated tool exists that does all three simultaneously for use in population health research. Developing and testing such a tool is a prerequisite for the kind of multidimensional evidence the field needs. Developing and validating such a tool, i.e., one that integrates physical condition assessments with measures of housing security and accessibility, and that is suitable for administration across diverse racial and socioeconomic groups in large-scale population studies, is a prerequisite for the kind of multidimensional evidence the field needs. This represents a distinct research and methodological development priority.

6.3 Disability-centred housing and health research

Most existing research treats disability as something to adjust for rather than something to understand. What the field needs instead are studies that ask what the housing-health relationship looks like specifically for disabled people, and how it differs across disabled people who also face racial or economic disadvantage. Community-based participatory research that engages disabled communities of colour in defining research questions, developing measures, and interpreting findings is particularly important given the historical exclusion of these communities from research design. Mixed-methods studies that combine quantitative measurement of housing conditions and health outcomes with qualitative documentation of lived experience would capture both the scale and the mechanisms of the disability-housing-health relationship across different intersectional positions.

6.4 Quasi-experimental and longitudinal designs

Establishing causation in this field requires more than cross-sectional observational studies, which cannot rule out the possibility that health status, rather than housing conditions, determines housing outcomes. Changes in eviction moratorium laws, housing voucher programme expansions, welfare benefit redesigns, and disability housing assistance reforms all create natural experiments in which some populations experience policy changes before others, enabling quasi-experimental estimation of causal effects for specific intersectional groups. Li and colleagues' study of emergency rental assistance demonstrates what this kind of analysis can show (42). Longer follow-up studies are also needed, particularly for cognitive outcomes where the pathway from housing insecurity to dementia likely operates over decades.

6.5 Intervention studies with intersectional outcome reporting

Evaluations of housing interventions, including Housing First programmes, asthma home visiting, housing remediation, accessible housing modification, and rental assistance, should be designed from the outset with sufficient diversity of participants to allow health outcomes to be disaggregated by race, income, and disability status, and ideally to test whether intervention effects differ across intersectional groups. New intervention studies should be co-designed with communities at the intersection of the most disadvantaged positions to ensure that interventions address the specific barriers and needs those communities face. An intervention designed without this input risks being calibrated to the needs of more advantaged subgroups within the target population.

7 Discussion

The IHCHE framework developed here makes an argument that the housing-health literature has not yet operationalised. Race, income, and disability interact to produce patterns of housing-related health disadvantage that cannot be read off from any single dimension, and understanding those patterns requires both a different theoretical lens and a different set of analytical tools. The framework's four levels organise a heterogeneous literature into a coherent structure and, in doing so, make the gaps visible in a way that a purely descriptive synthesis cannot. Most distinctively, the tripartite typology of housing conditions, separating physical conditions, security, and accessibility, reveals that the housing dimension most relevant to disabled people's health has barely been measured. The bidirectional treatment of disability, as both a social position that shapes housing exposure and a health outcome that housing conditions can produce, is similarly absent from existing frameworks.

The synthesis confirms that the housing-health literature has grown substantially since 2019, but what it has not yet done is generate intersectional evidence. Studies that include race, income, and disability as covariates adjust for these factors rather than learning from them, and the architecture of the field remains one in which each factor is studied in relative isolation. The evidence base therefore describes three partially overlapping literatures rather than a single integrated one. This is not a critique of individual studies, many of which are methodologically strong within their scope, but an argument about what the field has collectively prioritised and what a different set of priorities would produce.

The policy implications of the existing evidence are clear even without waiting for intersectional studies. Lemire and colleagues' evidence that racial disparities in the enforcement of housing code violations for asthma triggers are widening despite over a decade of healthy housing initiatives confirms that generic housing programmes without intersectional accountability mechanisms are unlikely to close disparate health outcomes (9). Trivedi and Brucker's finding that United States federal rental assistance does not address health or quality gaps for disability households argues for redesigning housing assistance with accessibility and quality, not just affordability, as explicit goals (11). The structural determinants evidence confirms that these problems are products of identifiable policy choices and can therefore be reversed by different policy choices, provided those choices are informed by evidence about who bears the greatest burden and why.

Several limitations of the IHCHE framework and the current review warrant acknowledgement. First, the framework diagram is an analytical stylization rather than a complete causal map. Some pathways run in directions not fully represented. For example, Level 2 housing conditions can influence Level 1 structural determinants through advocacy, collective action, and policy feedback. Housing accessibility campaigns by disabled people's organisations have historically shaped welfare policy and building regulations, and this upward pathway is absent from the current diagram. Second, social support and social capital represent the most consequential omissions from Level 3. Social networks can buffer the health consequences of inaccessible housing for disabled people and can simultaneously be depleted by housing instability, but they are not formally incorporated as modifiers in the current framework. Third, the literature reviewed is predominantly from the United States and United Kingdom, which limits the immediate applicability of the framework to health systems and housing markets with very different welfare state configurations, tenure structures, and anti-discrimination protections. Fourth, the framework treats race, income, and disability as analytically equivalent modifiers at Level 3, whereas their relative influence on the housing-health relationship is likely to vary across health outcome domains and housing condition types. Future empirical work should test whether the framework's analytical equivalence assumption holds or whether some modifiers operate more powerfully in specific domains.

8 Conclusion

Housing conditions shape health through multiple pathways across the life course, and this relationship is not socially neutral. Race, income, and disability interact to determine both who is exposed to poor housing conditions and how severe the health consequences of that exposure are. The IHCHE framework provides the theoretical structure for investigating these interactions systematically, and the literature from 2019 to 2026 confirms the importance of each factor individually while making clear that the fully intersectional evidence needed to understand how they combine has not yet been generated. The field is emerging, the gaps are identifiable, and the research designs needed to fill them are feasible. What remains is to do the work.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Arin Balalian, Texas A&M Agriculture, Food and Nutrition Evidence Center, United States

Reviewed by: Catherine Ipsen, University of Montana, United States

Steven Haworth, University of Kansas Medical Center, United States