CDC—The Universal Declaration of Human Rights recognizes “adequate housing” as a basic human right.
Housing is among the most important social determinants of health. Homeownership is an essential pathway to economic security, social mobility, stability, generational wealth, and healthy life. Financial assets such as owning a house or land passed down from one generation to the next within families is referred to as generational wealth. Recent research conducted by the National Realtor Association reported that, on average, a homeowner gained $139,134 in equity in the last 5 years, and $218,505 in the last 10 years. Homeownership is also key to narrowing racial inequity gaps in income and wealth. Wealth and income are strongly associated with increased illness and death rates and reduced life expectancy. Poor health also leads to reduced and inadequate income, creating a health–poverty trap.
The relationship between adequate housing and health has been well established. Housing can affect health through multiple mechanisms. For example, homeowners have a stable place to live, and residential stability leads to improved health. In the US, most services — such as access to health care, schools, and financial, social, and municipal services — are linked to residential status and residential history. These SDOH also affect health and health outcomes. Owning a residence also contributes to better internal house conditions and affordability, which can impact health status.
Extensive research in recent decades has solidified the association between housing and health outcomes, prompting a paradigm shift that considers housing as an integral component of overall health. Advancing beyond age-old hypotheses to explore a new domain, the relationship between homeownership and health outcomes is necessary for contemporary public health. Homeownership, symbolizing stable and sustainable housing, offers a unique lens to enhance our understanding of community health.
Understanding the connection between homeownership and chronic health conditions is imperative in the field of public health. Recent decades have underscored the critical role of housing in overall health outcomes, yet a substantial knowledge gap exists concerning how homeownership, as a distinct housing aspect, influences chronic health conditions. This understanding holds vital implications for several reasons. Chronic health conditions are the leading cause of illness and death, responsible for 90% of the $4.1 trillion in annual health care expenditures in the US, necessitating their prioritization in public health efforts.
Furthermore, homeownership extends beyond a mere housing arrangement, encompassing stability, community ties, social networks, and financial security — factors that can affect the risk and management of chronic health conditions. By exploring this under-researched area, we can identify interventions and policies to leverage homeownership to mitigate chronic health issues and enhance overall well-being. This study aims to establish homeownership as a critical determinant of health and social well-being, particularly in predicting the prevalence of major chronic health conditions in the US.
The observed associations between homeownership and chronic health conditions highlight the importance of housing as a social determinant of health. These findings are consistent with the existing literature that emphasizes the multifaceted impact of housing and housing type on health outcomes. Our study findings align with prior research indicating that housing and homeownership contribute to improved economic security, generational wealth, and social stability.
Homeownership, as a means of wealth accumulation, can provide people and families with resources that positively influence health and well-being (3,5,11,17). The significant associations between homeownership and chronic health conditions remained robust even after adjusting for various demographic factors. This suggests that homeownership’s influence on health outcomes transcends individual characteristics, highlighting its unique role in shaping health disparities. The observed effect could be attributed to factors such as differences in housing quality, neighborhood environments, and access to health care services based on homeownership status (3,10,11). Homeowners often experience greater housing stability, which has been linked to better health outcomes (3,8,18). Moreover, homeowners may have more control over their living conditions, leading to improved internal housing conditions that positively impact health (18,19).
Methods
We analyzed 2020 data from the Behavioral Risk Factor Surveillance System (BRFSS; N = 401,958). Details of the BRFSS cross-sectional survey, methods, sampling, data collection, and weights applied to calculate population estimates can be found at www.cdc.gov/brfss/index.html. Briefly, BRFSS is the largest telephone survey in the US, collecting self-reported prevalence data on chronic health conditions, risk behaviors, and preventive services use from a representative sample of adults aged 18 years or older. Primarily, the survey uses a stratified random sampling design. BRFSS data use 2 types of weights that account for the survey design and population characteristics, and they are calculated on the basis of each geographic stratum, number of telephones within sampled household, and number of adults aged 18 years or older living in the residence. The survey uses iterative proportional fitting (ie, raking) to adjust estimates for demographic differences between the sample and reference populations. Within the scope of this article and based on data availability, we examined the 6 leading chronic health conditions in the US with the highest disease burden and economic impact. Outcome variables were self-reported ever diabetes, asthma, cancer (other than skin), angina/coronary heart disease (CHD), stroke, and kidney disease. The exposure variable was homeownership.
We defined and used analytic variables based on the 2020 BRFSS questionnaires. The homeownership variable, derived from Core Section 8, queried respondents on owning or renting their homes; response options were “own,” “rent,” “other arrangements,” and “don’t know/not sure/refused.” For analytical precision, the “other arrangements” category was combined with “rent” due to similar housing statuses. Responses in the “don’t know/not sure/refused” category for both homeownership and chronic health conditions were treated as missing values in all analyses. Chronic health conditions, obtained from Core Section 6, used a general question format: “Has a doctor, nurse, or other health professional ever told you had [condition]?” The conditions were angina or CHD, diabetes, stroke, cancer other than skin, asthma, and kidney disease. Response options for chronic health conditions, excluding “ever diabetes,” were yes, no, don’t know/not sure/refused. For “ever diabetes,” response options were yes, “prediabetic or borderline,” “yes, during pregnancy,” no, and don’t know/not sure/refused. To streamline analysis and ensure comparability, yes, “prediabetic or borderline,” and “yes, during pregnancy” were combined based on the assumption of their shared indication of altered glucose metabolism and similar risk profiles.
All analyses were performed in SAS 9.4 (SAS Institute, Inc). According to the BRFSS analysis manual, we calculated weighted estimates (population proportions and odds ratios [ORs]) by supplying appropriate strata, cluster, and weight information. SAS has specialized procedures for sample survey data that could incorporate design factors and the understanding of population characteristics. For descriptive statistics and sample distribution by sociodemographic characteristics we used PROC SURVEYFREQ. This procedure yields n-way frequency and cross tabulation tables for population totals and population proportions. We used PROC SURVEYLOGISTIC to construct our logistic regression models. PROC SURVEYLOGISTIC integrates complex survey sample designs, stratification, clustering, and unequal weighting and offers fitting a broad class of binary response models in the general form of g (π) = α + xβ. The associations between homeownership and demographic characteristics were also examined by using logistic regression. To illustrate the influence of the study design on this association, both unweighted and weighted ORs were computed.
We used a nested model approach to construct our adjusted logistic regression models. Initially, we began with just 2 demographic variables and systematically incorporated additional variables. We employed a sequential adjustment, progressing from age and sex in Model 1 to various sociodemographic variables in Model 5, enabling a nuanced exploration of how diverse demographic factors collectively influence homeownership likelihood. The use of multiple models enhanced the thorough exploration of the interplay between demographic factors and homeownership. This iterative process resulted in our final adjusted model, which encompassed a comprehensive set of covariates: sex, age, race and ethnicity, education, income, marital status, and employment. To evaluate the effect of these variable additions on model fit, we leveraged the likelihood ratio test, a conventional statistical technique for assessing the significance of model improvements within nested models. For adjusted models, certain categories of marital status, education, and employment were collapsed and combined to simplify analysis while preserving relevant characteristics within each group. In regression analyses the choice of referent groups aimed to consistently present ORs above 1, emphasizing increased odds for an easy and intuitive interpretation of study results. We present findings from descriptive and regression analyses and report population proportions, population-level ORs, and corresponding 95% CIs.
Top
Results
Most of the sample population (66.8%) reported living in their own residences, 26.9% were renting, and the remaining 6.3% were residing under “other arrangements” (Table 1). Individuals aged 65 years or older accounted for 21.9% of the sample, the largest proportion. The population was roughly equal between female (51.3%) and male (48.7%) respondents. The study population was predominantly non-Hispanic White (61.8%), followed by Hispanic (17.8%), Black (11.8%), and Asian (5.6%). Most of the study population was married (50.3%) or had never married (24.8%). The largest proportion of the study population attended some college or graduated from college or technical school (59.8%), followed by high school graduates (27.8%) and those who did not complete high school (12.5%). The largest proportion of respondents in the employment category were those employed for wages (46.8%), followed by retired people (19.6%). The largest proportion of respondents in the income category were those who earned $50,000 or more (53.3%), and the remaining proportions varied.
The odds of homeownership increased significantly with advancing age (Table 2). People aged 65 years or older had the highest odds of homeownership (OR = 17.45; 95% CI, 16.21–18.79). Female respondents had a modestly elevated odds of homeownership compared with male respondents (OR = 1.03; 95% CI, 1.00–1.06). Non-Hispanic White people had substantially higher odds of homeownership (OR = 3.34; 95% CI, 3.18–3.52) compared with Hispanic people, while various other racial and ethnic groups exhibited comparatively lower odds. Married people had the highest odds of homeownership (OR = 10.16; 95% CI, 9.71–10.64). Education and income also showed strong positive associations with homeownership. Respondents who graduated from college or technical school had higher odds of homeownership compared with those who did not graduate high school (OR = 4.13; 95% CI, 3.89–4.38), and respondents in higher income brackets had notably higher odds than those who earned less than $15,000 per year. Respondents who had an annual household income of $50,000 or more were 7.8 times more likely to own a home than those earning less than $15,000 annually.
Among people aged 18 to 44 years, homeowners exhibited a significantly higher prevalence of cancer other than skin (1.6% vs 1.2%; P = .002), and lower prevalence of asthma (13.9% vs 16.6%; P <.001) (Table 3). The prevalence of other conditions was comparable between the 2 groups. Among respondents aged 45 to 64 years, homeowners had a significantly lower prevalence of angina/CHD (3.8% vs 6.0%), ever diabetes (16.3% vs 25.4%), asthma (12.4% vs 17.4%), stroke (2.7% vs 6.2%), and kidney disease (2.6% vs 5.1%) (all P <.001). Among people aged 65 to 80 years, homeowners displayed lower prevalence rates of all chronic health conditions than renters, except for cancer other than skin.
After adjusting for age and sex (Model 1), the odds of several chronic health conditions were significantly higher for renters in comparison with homeowners (Table 4). These significant associations remained in Model 2, which was adjusted for age, sex, and race and ethnicity. In Models 3 (adjusted for age, sex, race and ethnicity, education, and marital status) and 4 (adjusted for same variables as Model 3, plus employment and income), the associations remained significant for several conditions. In Model 4, renters had 1.14 (95% CI, 1.03–1.27) higher odds of angina/CHD and 1.15 (95% CI, 1.08–1.23) higher odds of ever diabetes compared with homeowners. Renters also had 1.18 (95% CI, 1.11–1.22) higher odds of asthma and 1.34 (95% CI, 1.19–1.52) higher odds of stroke in comparison to homeowners. Similarly, renters had 1.38 higher odds of kidney disease compared with homeowners (95% CI, 1.22–1.56).
