Abstract
Introduction.
Obesity affects four in ten US adults. One of the most prevalent health-related social risk factors in the US is housing instability, which is also associated with cardiovascular health outcomes, including obesity. The objective of this research brief is to examine the association between housing instability with obesity status among a representative sample of insured adults across seven integrated health systems.
Methods.
Kaiser Permanente National Social Needs Survey used a multistage stratified sampling framework to administer a cross-sectional survey across seven integrated health systems (administered Jan.-Sept. 2020). Survey data were linked with electronic health records (EHR). Housing instability was categorized into levels of risk: 1) ‘No Risk’; 2) ‘Moderate Risk’; and 3) ‘Severe Risk’. Based on established BMI thresholds, obesity and severe obesity served as the primary outcome variables. In 2023, weighted multivariable logistic regression accounted for the complex sampling design and response probability and controlled for covariates.
Results.
The analytic cohort comprised 6,397 adults. Unadjusted weighted prevalence of obesity and severe obesity was 31.1% and 5.3%, respectively; and 15.5% reported housing instability. Adjusted regression models showed that the odds of severe obesity was nearly double among adults exposed to severe housing instability (Adjusted OR=1.93; 95% CI 1.14–3.26). Other BMI categories were not associated with housing instability.
Conclusions.
Among a representative cohort of insured adults, this study suggested increasing levels of housing instability are associated with increasing levels of obesity. Future research should further explore the temporal, longitudinal, and independent association of housing instability with obesity.
Keywords: obesity, weight status, housing stability, social determinants of health, adult
Introduction
Obesity prevalence has risen dramatically over past decades and affects four in ten US adults. Obesity and severe obesity are associated with serious cardiovascular outcomes and vary across social and demographic factors.1 One of the most prevalent health-related social risk factors is housing instability – affecting one in three US households. This has been exacerbated by rising housing costs and a deficit of affordable homes. Housing instability is also associated with numerous poor health outcomes, including indicators of cardiovascular health.2,3
Emerging evidence generally suggests an adverse association between housing instability and weight-related outcomes using area-based social risk measures.3,4 Proposed mechanisms explaining the association between higher rates of housing instability and obesity are complex and may act through: 1) direct physiologic pathways to influence weight outcomes (i.e., chronic stress and mental health pathway); 2) indirect pathways by which increased housing cost limits remaining household income available for other resources (i.e., financial strain and food insecurity); and 3) indirect pathway whereby housing cost burden leads to forced relocation to poorer quality housing and disadvantaged neighborhoods (i.e., displacement and distribution).3 Individuals experiencing severe housing instability likely encounter additional socioeconomic difficulties, exacerbating this relationship.
Examining housing instability as a key social determinant of health in the context of increasing obesity rates is of paramount importance because it will help elucidate social contributors to the obesity epidemic and inform on actionable ways to reduce health disparities. However, to the authors’ knowledge, no studies have examined these associations at the individual level among a racially and ethnically and geographically diverse sample of insured adults. Hence, this study examined the association of housing instability with obesity and severe obesity among a representative sample of insured adults with access to health care across seven integrated healthcare systems.
Methods
A cross-sectional design was employed across seven regional markets (Colorado, Georgia, Hawaii, Mid-Atlantic States, Northern California, Northwest, Southern California) of Kaiser Permanente (KP) – an integrated U.S. health system serving ~8.6 million members. The KP National Social Needs Survey was administered between January-September 2020 and used a multistage stratified sampling framework proportional to the age and sex distributions across each region; stratified by potential for experiencing social risk.5 Survey data were linked with electronic health records (EHR).
Housing instability was assessed with four validated questions asking about housing experiences in the past year (i.e., ability to pay mortgage/rent on time; number of places lived in past year; steady place to sleep; current living situation).6–8 Responses were categorized into one of three mutually exclusive levels of risk for experiencing housing instability: 1) ‘No Risk’; 2) ‘Moderate Risk’; and 3) ‘Severe Risk’ (Appendix Table 1).
The height and weight measurement closest to survey completion between July 2018-September 2020 was used to derive body mass index (BMI=kg/m2) and weight status categories. Obesity and severe obesity served as the primary outcome variables (see Table 1 for definitions).1
Table 1.
Weighted Baseline Sample Characteristics (n= 6,397)
| Sample Characteristics | Weighted Percent (95% CI)a |
|---|---|
| Age group | |
| 18–34 years | 23.4 (21.5 – 25.3) |
| 35–49 years | 23.2 (21.4 – 25.2) |
| 50–64 years | 28.6 (26.6 – 30.6) |
| 65+ years | 24.8 (23.1 – 26.7) |
| Sex | |
| Female | 55.3 (53.1 – 57.5) |
| Male | 44.7 (42.5 – 46.9) |
| Race and Ethnicity | |
| Hispanic, Latino/a or Spanish origin | 26.1 (24.2 – 28.1) |
| White/Caucasian (non-Hispanic) | 44.0 (41.9 – 46.1) |
| Black or African American (non-Hispanic) | 9.3 (8.1 – 10.5) |
| Asian / Pacific Islander (non-Hispanic) | 17.5 (15.8 – 19.3) |
| Other race selected, including multi-racial (non-Hispanic) | 3.2 (2.4 – 4.1) |
| Education Level | |
| Less than High School | 5.7 (4.8 – 6.7) |
| High school graduate or GED | 15.8 (14.4 – 17.4) |
| Some college or 2-year degree | 31.3 (29.3 – 33.4) |
| 4-year college graduate | 24.4 (22.6 – 26.4) |
| More than a 4-year college degree | 22.7 (20.9 – 24.6) |
| Household Income | |
| Less than 35,000 | 16.8 (15.3 – 18.4) |
| 35,000 to less than 75,000 | 26.0 (24.2 – 27.9) |
| 75,000 or greater | 41.9 (39.8 – 44.1) |
| Missing self-report income | 15.3 (13.7 – 16.9) |
| Insurance Type | |
| Medicare, Medicaid, Special Program | 26.8 (25.0 – 28.6) |
| Commercial | 73.2 (71.4 – 75.0) |
| Self-rated physical health | |
| Excellent / Very Good | 48.2 (46.0 – 50.4) |
| Good | 37.4 (35.3 – 39.6) |
| Fair / Poor | 14.4 (12.9 – 15.9) |
| Self-rated mental health | |
| Excellent / Very Good | 56.7 (54.6 – 58.9) |
| Good | 29.3 (27.3 – 31.3) |
| Fair / Poor | 14.0 (12.6 – 15.5) |
| Housing Stability Indicatorsb | |
| Unable to pay mortgage/ rent | |
| Yes | 10.1 (8.9 – 11.5) |
| No | 89.9 (88.5 – 91.1) |
| Places lived | |
| One | 89.8 (88.4 – 91.1) |
| Two | 8.7 (7.5 – 10.0) |
| Three or more | 1.5 (1.0 – 2.1) |
| No stable place to sleep | |
| Yes | 1.8 (1.3 – 2.5) |
| No | 98.2 (97.5 – 98.7) |
| Current living situation | |
| Have a steady place to live | 93.5 (92.4 – 94.5) |
| Have a steady place to live but are worried about losing it in the future | 6.1 (5.1 – 7.2) |
| Do not have a steady place to live | 0.4 (0.2 – 0.7) |
| Housing Instability Risk Categories | |
| Severe Risk of Housing Instabilityc | 12.0 (10.7 – 13.5) |
| Moderate Risk of Housing Instabilityd | 3.5 (2.75 – 4.31) |
| No Risk of Housing Instabilitye | 84.5 (83.0 – 86.0) |
| Weight Status Categoriesf | |
| Underweight (BMI <18.5 kg/m2) | 1.3 (0.84 – 1.91) |
| Healthy Weight (BMI = 18.5 - <25.0 kg/m2) | 25.9 (24.0 – 27.8) |
| Overweight (BMI = 25.0 - <30.0 kg/m2) | 36.5 (34.4 – 38.6) |
| Obese, Class I (BMI = 30.0 - <35.0 kg/m2) | 22.4 (20.6 – 24.3) |
| Obese, Class II (BMI = 35.0 - <40.0 kg/m2) | 8.7 (7.5 – 10.0) |
| Severe Obesity (Class III, BMI ≥40.0 kg/m2) | 5.3 (4.4 – 6.2) |
Sum of Weights=5,366,813; Sample Eligible for Analysis=6,397
Housing Instability Questions (Response Options): #1: In the past 12 months, was there a time when you were not able to pay the mortgage or rent on time? (Yes, No): #2: In the past 12 months, how many places have you lived? (One, Two, Three or more); #3: In the past 12 months, was there a time when you did not have a steady place to sleep or slept in a shelter? (Yes, No); #4: What is your living situation today? (You have a steady place to live; you have a steady place to live but are worried about losing it in the future; you do not have a steady place to live)
Severe Risk for Housing Instability: Respondents that indicated they do not have a steady place to live and/or screened positive to two or more exposures, i.e., [[(Places lived = ‘Three or more’) or (No stable place to sleep = “Yes’)] and [Living Situation = ‘You have a steady place to live but are worried about losing it in the future’ or ‘You do not have a steady place to live’]] or [Living Situation = ‘You do not have a steady place to live’]
Moderate Risk for Housing Instability: Screening positive to at least one housing stability question and severity not elevated to severe risk for housing instability, i.e., [Unable to pay = ‘Yes’] and/or [Places lived = ‘Three or more’] and/or [No stable place to sleep = “Yes’] and/or [Living Situation = ‘You have a steady place to live but are worried about losing it in the future’]
No Risk for Housing Instability: Screening negative for all housing instability questions, i.e., [Unable to pay = ‘No’] and [Places lived = ‘One’ or ‘Two’] and [No stable place to sleep = “No’] and [Living Situation = ‘Have a steady place to live’]
Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared (BMI=kg/m2); with implausible BMI values (BMI <15th percentile; or BMI >99th percentile) excluded (<1% of measurements). Stated BMI thresholds for weight status categories applicable for all race and ethnicity groups except for Asian and Pacific Islander populations for which a lower threshold was applied based on World Health Organization recommendations and prior literature (i.e., Overweight: BMI≥23.0 kg/m2; Class I Obesity: BMI≥ 27.5 kg/m2; Class II Obesity: BMI ≥32.5 kg/m2; Severe Obesity: BMI ≥ 37.5 kg/m2).
All analyses were conducted between January 2021-June 2023. Descriptive statistics and bivariate associations were examined for the analytic sample (Rao-Scott chi-square). Weighted multivariable logistic regression accounted for the complex sampling design and response probability and controlled for covariates. Model covariates included age, sex, and insurance type (EHR-derived); and race and ethnicity, household income, education, physical health, mental health (survey). This quality improvement project was deemed not to be human subjects research by the KP Interregional Institutional Review Board.
Results
The eligible study cohort included 10,226 adult survey respondents (23% response rate). Survey respondents were excluded if they were 1) enrolled <18 months prior to survey completion (period aligning with survey recall period, allowing for 6-month potential disruptions in care due to COVID-19 pandemic) (n=1,947); and/or 2) missing outcome (n=912) or exposures/covariate values (n=885, one regional market) (Appendix Table 2). The analytic cohort comprised 6,397 adults with at least 18 months of health plan enrollment prior to survey completion and complete data for the variables of interest. Table 1 shows that the distribution of females was slightly higher than males (i.e., 55.3% vs. 44.7%, respectively) and approximately half of the sample was less than 50 years old. Respondents were racially and ethnically diverse and nearly half had received a 4-year college degree or more.
Approximately 15.5% respondents demonstrated moderate or severe housing instability (3.4% at moderate risk and 12.0% at severe risk). The mean BMI of respondents was 28.5kg/m2 (not shown). Unadjusted weighted prevalence of obesity was 36.4%; with 22.4% categorized as Class I, 8.7% as Class II, and 5.3% as Class III (i.e., severe obesity). The unadjusted distribution of weight status categories varied significantly across housing instability risk categories (Figure 1, Rao-Scott Chi-Square, p<0.0001). Among adults experiencing severe housing instability, 44.8% of adults had obesity and 11.3% had severe obesity. The proportion of adults with obesity and severe obesity was lower among those reporting no risk (35.3% and 4.4%, respectively) or moderate risk (33.0% and 5.2%, respectively) for housing instability.
Figure 1.
Unadjusted Proportion of Adults with Obesity by Housing Instability Severity Categories, Weighted Sample (presented as row percentages) a, b
a Sum of Weights=5,366,813; Sample Eligible for Analysis=6,397
b Rao-Scott Chi-Square, p<.0001
Weighted logistic regression models adjusting for sociodemographic characteristics and health status showed that the odds of severe obesity were nearly double among adults exposed to severe housing instability (Adjusted OR=1.93; 95%CI 1.14–3.26) compared to those that did not report any housing instability (Table 2). A relationship with severe obesity was not observed among those reporting moderate risk for housing instability (Adjusted OR=0.98, 95%CI 0.45–2.16). The adjusted odds of obesity did not statistically vary across categories of housing instability (Moderate Risk: Adjusted OR=0.80, 95%CI 0.50–1.30; Severe Risk: Adjusted OR=1.11, 95%CI 0.82–1.50).
Table 2.
Adjusted odds ratios (OR) of obesity and severe obesity by level of housing instability risk.
| Weighted Logistic Regression Models | Housing Instability Severity | ||
|---|---|---|---|
| No Risk (ref) | Moderate Risk | Severe Risk | |
| Obesitya | |||
| Unadjusted OR (95% CI) | 1.00 | 0.9 (0.56 – 1.45) | 1.49 (1.14 – 1.94) |
| Adjusted OR (95% CI)c | 1.00 | 0.80 (0.50 – 1.30) | 1.11 (0.82 – 1.50) |
| Severe Obesityb | |||
| Unadjusted OR (95% CI) | 1.00 | 1.19 (0.55 – 2.54) | 2.76 (1.77 – 4.29) |
| Adjusted OR (95% CI)c | 1.00 | 0.98 (0.45 – 2.16) | 1.93 (1.14 – 3.26) |
Obesity defined as BMI≥30.0kg/m2 for all race and ethnicity groups except for Asian and Pacific Islander (API) populations for which a lower threshold was applied based on World Health Organization recommendations and prior literature (API Obesity: BMI≥ 27.5 kg/m2)
Severe obesity defined as BMI≥40.0kg/m2 for all race and ethnicity groups except for Asian and Pacific Islander API) populations for which a lower threshold was applied based on World Health Organization recommendations and prior literature (API Severe Obesity: BMI ≥ 37.5 kg/m2)
Models adjusted for age group, sex, race and ethnicity, education level, household income, insurance type, and self-reported physical and mental health status, as well as health plan region and sample design weights.
Notes: Secondary analysis (results not shown) explored the potential impact of the COVID-19 pandemic on the associations between housing instability and weight status outcomes before and after the start of the pandemic using an interaction term between housing instability and COVID-19 timing (pre-COVID vs COVID time period) in the adjusted model. No significant difference between housing instability and weight status outcomes were observed by pre-COVID vs. COVID time period (p-value for interaction >0.05).
CI = confidence interval; OR = odds ratio
Boldface indicates statistical significance (p<0.05)
Discussion
Among a representative cohort of insured adults across seven U.S. integrated healthcare systems, a significant association between housing instability severity and weight status was observed. Respondents who experienced severe housing instability were nearly twice as likely to have severe obesity, independent of age, sex, race/ethnicity, socioeconomic factors, and health status. An association between housing instability and the remaining weight status categories was not observed.
While prior observational research is limited, this study’s findings align with recent studies including a narrative review examining the relationship between housing instability and cardiometabolic health.3,4,9–12 With respect to obesity, they found that housing instability was generally associated with a higher prevalence of overweight and obesity.3 At the county-level, for example, one study reported a 37% increase in the odds of obesity with each percentage point increase in the proportion of household income spent on housing cost.12 Taken together, this study and others suggested increasing levels of housing instability are associated with increasing levels of obesity.
Limitations
First, the cross-sectional design limits the ability to determine causality or direction of these relationships. Second, BMI does not directly assess body fat and loses precision when examined at the individual-level, especially across sociodemographic groups. Lastly, there may be unmeasured potential confounders that account for the observed associations (e.g., lifestyle behaviors).
Conclusions
To the author’s knowledge, this is one of the first efforts to investigate the connection between housing instability and obesity among a representative cohort of geographically diverse adults with access to health care. Future longitudinal studies should consider BMI alongside other biometric and cardiometabolic risk factors when examining these relationships among insured and uninsured populations. In conclusion, these findings underscore the need for future studies to investigate the underlying mechanisms of the relationship between housing instability and obesity to better understand how best to optimize cardiovascular health and inform future public health and policy interventions.
Supplementary Material
Acknowledgements.
Funding for this project was provided from the Social Needs Network for Evaluation and Translation (SONNET), a national Kaiser Permanente (KP) Office of Community Health program that seeks to improve health of KP members and the communities they live in by developing and implementing new, scientifically-driven strategies to shape social health practice and policy. SONNET is supported by KP’s Office of Community Health. The sponsor/funder had no role in the study design; collection, analysis, and interpretation of findings; writing the research brief; and/or the decision to submit the report for publication. This quality improvement project was deemed not to be human subjects research by the KP Interregional Institutional Review Board. The findings presented in this paper are that of the authors and does not reflect the official policy of Kaiser Permanente
The authors declare that there are no conflicts of interest associated with this manuscript. None of the authors have professional or financial affiliations that could be perceived to influence the content of the manuscript.
Dr. Clennin is also supported by funding unrelated to this project from the American Heart Association (938627). Dr. Vupputuri is also supported by funding unrelated to this project from the NIH (HL165376 and AG066956). Dr. Daugherty is also supported by funding unrelated to this project from the NHLBI (HL164106 and HL168504). The remaining authors including Dr. Schootman, Dr. Brown, Ms. Reifler, and Ms. Goodman have no financial disclosures to report.
References
- 1.Hales CM, Carroll MD, Fryar CD, Ogden CL. Prevalence of obesity and severe obesity among adults: United States, 2017–2018. NCHS Data Brief, no 360. Hyattsville, MD: National Center for Health Statistics. 2020. [PubMed] [Google Scholar]
- 2.Harvard Joint Center for Housing Studies. The State of the Nation’s Housing 2020. Harvard University Report. 2020. Accessed October 2023. Available at: https://www.jchs.harvard.edu/state-nations-housing-2020. [Google Scholar]
- 3.Gu KD, Faulkner KC & Thorndike AN. Housing instability and cardiometabolic health in the United States: a narrative review of the literature. BMC Public Health. 2023; 23(1):1–27. 10.1186/s12889-023-15875-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Lotfata A, Tomal M. Exploring Housing Determinants of Obesity Prevalence Using Multiscale Geographically Weighted Regression in Chicago, Illinois. Prof Geogr. 2023;75(3):335–44. 10.1080/00330124.2022.2111692. [DOI] [Google Scholar]
- 5.Brown MC, Lewis CC, Wellman RD, Haugen KL, Bain C, Ramaprasan A, DiJulio BS, Shah AR. 2022 Kaiser Permanente National Social Health Survey Final Report. Kaiser Permanente Social Health Network for Evaluation and Translation (SONNET). 2023. Accessed September 2023. Available at: https://www.kpwashingtonresearch.org/application/files/9916/9833/8109/KP-National-Social-Health-Survey_2022_Quant-Results_Final-Report.pdf [Google Scholar]
- 6.Sandel M, Sheward R, Ettinger de Cuba S, et al. Unstable housing and caregiver and child health in renter families. Pediatrics. 2018;141(2). 10.1542/peds.2017-2199. [DOI] [PubMed] [Google Scholar]
- 7.Billioux A, Verlander K, Anthony S, Alley D. Standardized screening for health-related social needs in clinical settings: the accountable health communities screening tool. Discussion paper. NAM Perspect, Washington DC. 2017; 1–9. Available at: https://nam.edu/wp-content/uploads/2017/05/Standardized-Screening-for-Health-Related-Social-Needsin-Clinical-Settings.pdf. [Google Scholar]
- 8.Weir RC, Proser M, Jester M, Li V, Hood-Ronick CM, Gurewich D. Collecting social determinants of health data in the clinical setting: findings from national PRAPARE implementation. J Health Care Poor Underserved. 2020;31(2):1018–35. 10.1353/hpu.2020.0075. [DOI] [PubMed] [Google Scholar]
- 9.Ludwig J, Sanbonmatsu L, Gennetian L, Adam E, Duncan GJ, Katz LF, et al. Neighborhoods, obesity, and diabetes–a randomized social experiment. N Engl J Med. 2011;365(16):1509–19. 10.1056/NEJMsa1103216. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Datar A, Shier V, Braboy A, Jimenez-Ortiz M, Hernandez A, King SE, Liu Y. Assessing impacts of redeveloping public housing communities on obesity in low-income minority residents: Rationale, study design, and baseline data from the Watts Neighborhood Health Study. Contemp Clin Trials Commun. 2022;25:100879. 10.1016/j.conctc.2021.100879. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bowen DJ, Quintiliani M, Bhosrekar SG, et al. Changing the housing environment to reduce obesity in public housing residents: a cluster randomized trial. BMC Public Health. 2019;18:883. 10.1186/s12889-018-5777-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Rodgers J, Briesacher BA, Wallace RB, Kawachi I, Baum CF, Kim D. County-level housing affordability in relation to risk factors for cardiovascular disease among middle-aged adults: The National Longitudinal Survey of Youths 1979. Health Place. 2019;59:102194. 10.1016/j.healthplace.2019.102194 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.