Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

Adults with Housing Insecurity Have Worse Access to Primary and Preventive Care

Adults with Housing Insecurity Have Worse Access to Primary and Preventive Care J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. ORIGINAL RESEARCH Adults with Housing Insecurity Have Worse Access to Primary and Preventive Care Patricia Martin, DO, Winston Liaw, MD, MPH, Andrew Bazemore, MD, MPH, Anuradha Jetty, MPH, Stephen Petterson, PhD, and Margot Kushel, MD Objective: Housing insecurity has been linked to high-risk behaviors and chronic disease, although less is known about the pathways leading to poor health. We sought to determine whether housing insecu- rity is associated with access to preventive and primary care. Methods: We conducted weighted univariate, bivariate, and multivariate analyses by using 2011 to 2015 Behavioral Risk factor Surveillance Survey data (N  228,131 adults). The independent variable was housing insecurity derived from the question on worry about paying rent or mortgage. The outcome measures were health services utilization (no usual source of care, no routine checkup in the past 1 year, and delayed medical care due to cost), self-rated health (number of days reported physical, men- tal health not good, and poor overall health), and number of chronic diseases (0, 1, 2 or more). The covariates included age, sex, race/ethnicity, income, level of education, marital status, and number of children in the family. We also adjusted for state fixed effects and survey year. We performed  tests and binary logistic regressions on categorical variables and ran t tests and estimated linear regression models on continuous variables. Multinomial logistic regressions were estimated for the number of chronic diseases. Results: Of the 228,131 adults in the study sample, 28,704 adults reported housing insecurity. We found that those with housing insecurity were more likely to forgo routine check-ups and lack usual sources of care. Low-income individuals, minorities, the unmarried, and middle-aged adults were more likely to report housing insecurity. Conclusion: Housing insecurity is associated with worse access to preventive and primary care. In- terventions to enhance access for these patients should be developed and studied. (J Am Board Fam Med 2019;32:521–530.) Keywords: Behavioral Risk Factor, Chronic Disease, Homeless Persons, Housing, Linear Models, Logistic Models, Mental Health, Multivariate Analysis, Outcomes Assessment, Primary Health Care, Risk-Taking, Social Determi- nants of Health, Surveillance System Social determinants of health (SDH) have a greater organizations have called for the integration of impact on people’s health and longevity than clin- public health and primary care to address SDH, ical care, and addressing these determinants is operationalizing the integration of data has proven viewed as a key strategy for meeting the triple aim difficult, with uncertainty around the types of data 2,3 of lower costs, improved patient experience, and to collect to improve outcomes. improved population health. Although numerous Housing insecurity is likely the SDH with the largest potential impact on population health, if properly addressed. Defining housing insecurity can This article was externally peer reviewed. be challenging as it refers to a spectrum of housing Submitted 15 December 2018; revised 26 March 2019; accepted 31 March 2019. experiences, including homelessness, crowding, high From Unity Health Care, Washington, D.C. (PM); De- partment of Health Systems and Population Health Sci- ences, University of Houston College of Medicine, Houston (WL); Robert Graham Center: Policy Studies in Family Medicine and Primary Care, Washington D.C. (AB, AJ, SP); Funding: none. Center for Vulnerable Populations, University of California, Conflict of interest: none declared. San Francisco/Zuckerberg San Francisco General, San Corresponding author: Patricia Martin, DO, Unity Health Francisco (MK). Care, Washington, D.C. E-mail: pmartin@unityhealthcare.org). doi: 10.3122/jabfm.2019.04.180374 Housing Insecurity and Access to Care 521 J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. housing costs in proportion to income (defined vari- about the relationship between housing insecurity ably as 30% and 50% of household income), and preventive and primary care access. In partic- 4,5 foreclosure, and frequent moves. - ular, having a usual source of care (USOC, or a Despite disagree ment about the definition, housing insecurity affects singular person or facility for navigating most millions of Americans, with over 21 million paying health care needs) is critical to receiving the rec- 30% to 50% and nearly 19 million households paying ommended screening tests and preventive services more than 50% of their income in housing costs. that can delay or prevent the development of The independent association between housing chronic disease. Our objective was to determine insecurity and poor health has been well-docu- how this BRFSS housing insecurity assessment was mented for both adults and children over the past associated with preventive and primary care access, 5,7–10 30 years. - self-rated health, and chronic disease. Nationally, we found a graded asso ciation between degree of housing insecurity and decreased access to care. Adults experiencing homelessness reported higher rates of acute care Methods services, such as emergency department visits, as We obtained data from the 2011 to 2015 BRFSS. well as postponement of needed medical care and The BRFSS is a phone-based survey that collects medications. data regarding health-related risk behaviors, Housing instability and frequent moves affect children in similar ways, resulting in chronic health conditions, and health services uti- 11,12 increased use of acute care services, - lization. It is administered annually to noninstitu- postpone ment of needed care, tionalized adults over the age of 18 in all 50 US and lack of a regular site for 12 17 preventive care. - states, Washington D.C., and 3 US territories. The prevalence of housing inse curity among low-income children (defined as liv- The BRFSS is composed of standard core ques- ing 200% below the federal poverty line or lacking tions, rotating core questions, optional modules, commercial health insurance coverage) ranges from and state-based questions. Standard core questions 11 13 29.5% are asked every year; whereas, rotating core ques- to 46% respectively. There are long-term health outcomes associated with childhood housing tions are asked every other year. States may also insecurity, which include earlier use of illicit select to include optional BRFSS modules as well as 14,15 19 drugs, - additional state-specific questions. increased rates of depression and preg nancy among teenagers, - The Social Context Module is an optional mod- and poor emotional adjust ment. ule that focuses on housing insecurity, food inse- Despite housing insecurity’s impact on health curity, and employment. We included all 23 states outcomes, researchers and policy makers have not that incorporated this module into their BRFSS achieved consensus around its measurement. Sev- questionnaires at any point between 2011 and eral clinically validated screening questions in use 2015. For the states included, the sample popula- by organizations such as Children’s Health Watch tion was comparable to the population not included and the Veteran’s Administration. in the sample (Supplemental Table 1), although the The Centers for Disease Control and Prevention Behavioral sample population was noted to be older, more Risk Factor Surveillance Survey (BRFSS) has in- affluent, and have a lower proportion of Hispanics. cluded a question regarding housing insecurity The survey includes both landline and cell phone since 2009. respondents from 2011 onwards. Response rates Deemed to be written at a 12th grade reading level with adequate precision and clinical for landline and cell phone users were 53.0% and validity, 27.9% (2011), 49.1% and 35.5% (2012), 49.6% and the question asks, “How often in the past 12 months would you say that you were worried or 37.8% (2013), 48.7% and 40.5% (2014), and 47.7% stressed about having enough money to pay your and 39.5% (2015), respectively. rent/mortgage?” - We determined the presence of housing insecu- The question addresses per ceived stress rather than a specific housing-related rity based on the question, “How often in the past metric, such as percent of household income spent 12 months would you say that you were worried or on housing. Perceived stress, however, offers a stressed about having enough money to pay your highly sensitive screening tool with the opportunity rent/mortgage?” Responses included the following: to capture the highest number of individuals at risk always, usually, sometimes, rarely, and never. We for becoming homeless. Furthermore, less is known deemed the affirmative responses “always” and 522 JABFM July–August 2019 Vol. 32 No. 4 http://www.jabfm.org J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. “usually” as markers for housing insecurity and Respondents are asked about several different dichotomized the sample based on the presence or health conditions with the following common absence of housing insecurity. question stem: “Have you ever been told by a doc- The covariates included age, sex, race/ethnicity, tor, nurse, or health care professional that you have income, level of education, marital status, number had any of the following?” Diabetes, hypertension, of children, and health status. We created 6 mutu- coronary artery disease, stroke, chronic obstructive ally exclusive age brackets (18 to 24, 25 to 34, 35 to pulmonary disease/asthma, skin cancer, other can- 44, 45 to 54, 55 to 64, and 65) and 2 mutually cer, arthritis, depression, and chronic kidney dis- exclusive sex self-identifiers. The race/ethnicity ease were among the included conditions. Those categories were white, black, Hispanic, and other. reporting skin cancer or other cancer were grouped We stratified income levels in 4 brackets based on as any cancer. We created 3 chronic disease cate- household income less than $15,000, $15,000 or gories: (1) no chronic conditions, (2) 1 chronic more and less than $25,000, $25,000 or more and condition, and (3) 2 or more chronic conditions. less than $50,000, and $50,000 or more. We clas- sified level of education in 4 mutually exclusive Statistical Analysis groups: less than high school education, high Using Stata 14.0, we conducted univariate and bi- school graduate/general education degree, some variate ( for categorical variables and t tests for college or technical school (1 to 3 years), and 4 continuous variables) analyses and multivariate lin- years or more of college education. Self-reported ear and binary logistic regressions. We used sam- health status was classified as excellent, very good, pling weights and BRFSS survey design variables good, fair, or poor. Finally, to assess family struc- throughout the analysis to obtain a nationally rep- ture, we determined the number of children and resentative sample of undersampled populations. marital status (married, divorced, widowed, sepa- We first compared the demographic characteristics rated, never married, and unmarried with partner). (sex, age, education, race/ethnicity, and poverty) of The dependent variables can be grouped into 3 the respondents in sample and out of sample by categories: (1) health services utilization, (2) self- conducting bivariate analysis using  tests for cat- reported health status, and (3) presence of chronic egorical and t tests for continuous variables. For the medical conditions. For health services utilization, respondents in the sample, we then computed de- we assessed the respondents’ USOC status, defer- scriptive statistics of demographic characteristics ment of medical care due to cost, and prolonged and conducted bivariate analyses by housing inse- time since last medical checkup. We characterized curity. We estimated 3 sets of regression models by absence of a USOC by a “No” response to the using linear regression for continuous outcomes question “Do you have 1 person that you think of as (health status measures reported as number of days) your personal doctor or health care provider?” We and logistic regressions for binary outcomes (hav- characterized deferment of medical care due to cost ing a USOC, avoiding or delaying medical care, as a “Yes” response to the question “Was there a and having a chronic medical condition). In all the time in the past 12 months when you needed to see models, we used housing insecurity as an indepen- a doctor but could not because of cost?” Lastly, we dent variable and patient characteristics as covari- characterized prolonged time since last medical ates. Because our analysis was restricted to a pub- checkup as any response other than “Within the licly available data set, institutional review board past year” to the question “About how long has it approval was neither required nor obtained. been since you last visited a doctor for a routine checkup?” The second and third categories of interest were Results self-reported health and the presence of chronic Of the 228,131 individuals in our sample, 14.3% medical conditions. We assessed self-reported reported housing insecurity (Supplemental Table health by determining the responses to 3 questions 2). Those reporting housing insecurity were that inquire about the specific number of days more likely to be female, have lower incomes, and within the past month that physical, mental, and be black or Hispanic. Of all age groups, middle-age overall health were not good. The presence of adults (35 to 54) were the most likely to experience chronic medical conditions is also by self-report. housing insecurity. As expected, those with a college doi: 10.3122/jabfm.2019.04.180374 Housing Insecurity and Access to Care 523 J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. Table 1. Health Outcomes and Utilization by Housing Insecurity Worry About Paying Rent or Mortgage Yes (n  28,704) No (n  199,427) Self-Reported Health Status and Utilization n% n % P Value Number of Chronic Conditions 0 910 30.8 10,984 43.5 .001 1 919 27.7 8,620 26 .6397 2 or more 1,973 41.5 12,312 30.4 .001 Health services utilization Deferred care due to cost 11,009 42.8 14,899 10.2 .001 Had a routine check-up within past year 18,821 64.0 148,768 73.8 .001 No usual source of care 6,168 29.0 26,682 19.4 .001 Self-reported health (mean number of days in last month) Poor physical health 8.6 3.9 .001 Poor mental health 9.5 2.6 .001 Poor overall health 8.0 3.3 .001 Source: Author analysis of 2011 to 2015 Behavioral Risk Factor Surveillance Survey (BRFSS) data. Between 2011 to 2015, 23 states implemented the Social Context Module that included this question. BRFSS asks about self-reported health by determining the responses to three questions that inquire about the specific number of days within the past month that physical, mental, and overall health were not good. education were less likely to be housing insecure. found that those with housing insecurity were more Those with more than a high school education were likely to have 1 or more chronic conditions. also less likely to be housing insecure, which may reflect individuals living in areas with a high cost-of- living who despite reported income of $50,000 are Discussion still uncomfortable with their disproportionate hous- BRFSS has been widely used to demonstrate the ing costs. Not being married, having larger families, association between housing insecurity and health 9,20 –22 and being in poor health were all associated with outcomes at the single-state level as well as 23–27 housing insecurity. the multistate level although none of these Those with housing insecurity were more likely studies assessed housing insecurity’s impact on pre- to report chronic diseases, lack a USOC, and defer ventive and primary care access. Our findings con- care in the past year due to cost (Table 1). With firm the potent relationship between housing inse- respect to overall health, those with housing inse- curity and poor health and add to the literature by curity had more days in the last month with poor documenting that, among those with housing inse- health. Furthermore, the number of days in poor curity, poor health may be the consequence of mental health was the highest of the 3 health cat- forgoing routine care and lacking a USOC. Fur- egories. thermore, we found that low-income, minority, un- After controlling for covariates, we found that married, middle-aged adults are more likely to re- those with housing insecurity had 3 times higher port housing insecurity. As public health and odds of delaying care due to cost. Housing-insecure primary care collaborate to identify and address respondents also had 35% higher odds of delaying housing insecurity, these findings can provide guid- check-ups and had 19% higher odds of lacking a ance for future screening efforts, such as whether to USOC. In our second regression model, where the use targeted or universal screening for housing in- outcome was number of days in poor health, we security. found that housing insecurity was associated with There are 1-item screening tools already in use 4.7 more days in poor physical health, 6.9 more by other agencies, such as Children’s Health days in poor mental health, and 4.7 more days in Watch (“An eviction is when your landlord or a poor overall health. In our final regression model, government or bank official forces you to move relative to those with no chronic conditions, we when you do not want to. In the past 5 years, have 524 JABFM July–August 2019 Vol. 32 No. 4 http://www.jabfm.org J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. you ever been evicted?”) and the Veterans Admin- community resources and following up with par- istration (“Are you worried or concerned that in the ents has proven beneficial in the pediatric set- 31,32 next 2 months you may not have stable housing that tings. - Among the adult population, pilot stud you own, rent, or stay in as part of a household?”). ies like the WellRx in New Mexico have drawn The populations of interest for these agencies, attention to the important role of community however, are not reflective of the general popula- health workers in helping patients with unmet so- tion, limiting their generalizability. cial needs to follow through and successfully access In 2017, the Center for Medicaid and Medicare community resources. Medical-legal partnerships Services also released the 10-item Accountable also have a role in the primary care setting in Health Communities Screening Tool, which in- addressing unmet social needs that involve the legal cludes 2 items relating to housing insecurity system and have been successful in many patient (“What is your housing situation today?” and center medical home settings. “Think about the place where you live. Do you Although the demand for stable affordable hous- have problems with any of the following? Bug In- ing in the United States far outpaces the limited festation, Mold, Lead Paint or Pipes, Inadequate supply, there is still benefit to screening for SDH in Heat, Oven of Stove Not Working, No or Not primary care. Having information about patients’ Working Smoke Detectors, Water Leaks”). unmet social needs can improve patient care by The first question was adapted from the Protocol for better informing providers, who can engage pa- Responding to and Addressing Patients’ Assets, tients in discussions that would not otherwise oc- 29 34 Risks, and Experiences assessment tool. - cur The sec and advocate on their behalf for interventions ond question was drawn from a screening tool de- at the community level, including enhancing access veloped by Norwalk Community Health Center to to housing voucher programs and programs that capture patients that would benefit from legal ser- prevent evictions and building high-quality, low- vices offered by their medical-legal partnership. income housing within mixed-income neighbor- These items are evidence-based but have not yet hoods. been validated. - Even the briefest SDH screening tool, however, In addition, given the recent de velopment of the Center for Medicaid and Medi- requires time and investment on the part of pro- care Services screening tool, there are no available viders. For providers who are not working directly studies at this time to link these particular housing with a social worker, referrals for community re- insecurity items with health-related outcomes. sources and follow up may fall to the provider. Housing insecurity was omitted as a National Although these efforts may result in improved Academy of Medicine-recommended domain for health outcomes and health care cost savings, it is inclusion in electronic health records as part of unclear how that cost savings would be shared. Meaningful Use because of the lack of standard More work is needed to explore how primary care measure and perceived difficulty in collecting hous- practices could be appropriately compensated for 4,5 ing insecurity data. the increased expenditure of time and staff re- Without a standard housing insecurity metric, there is a need for a validated sources spent in deploying screening for housing screening tool for the general population to begin insecurity. collecting and appropriately acting on this SDH Some limitations to the data source used in our data. Our findings add to the existing literature analyses are worth noting, including sampling that supporting the use of this BRFSS housing insecu- is not nationally representative, reflecting only the rity question. In the absence of systematic assess- states that administered the BRFSS Social Context ment of housing within medicine, this critical SDH Module. As noted earlier, the housing insecurity will continue to be ignored by clinicians and link- question addresses self-reported stress regarding ages between medicine and public health will fail to paying rent or a mortgage, which is subject to develop. survey participant interpretation of both the defi- With appropriate screening, those with housing nition of stress as well as their own stress level. insecurity can be identified and referred for assis- Dependent variables were also based on self-report, tance. The role of primary care providers in the and the survey questions asked about behaviors implementation of interventions to address SDH is over different time periods. For example, the ques- evolving. Thus far, generating referrals for existing tion regarding housing insecurity includes the past doi: 10.3122/jabfm.2019.04.180374 Housing Insecurity and Access to Care 525 J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. care among low-income Americans. J Gen Intern 12 months while the questions regarding self-re- Med 2006;21:71–7. ported health included the past 30 days and the 6. The State of the Nation’s Housing 2016. President questions regarding health care utilization go back and Fellows of Harvard College; 2016. Available as far as 5 years. Finally, BRFSS is a phone-based from: https://www.jchs.harvard.edu/sites/default/files/ survey that excludes institutionalized adults, which jchs_2016_state_of_the_nations_housing_lowres_0.pdf. omits those without phone lines as well as those 7. Reid KW, Vittinghoff E, Kushel MB. Association living in shelters, transitional housing, or outside. It between the level of housing instability, economic is unclear how inclusion of these populations would standing and health care access: a meta-regression. J Health Care Poor Underserved 2008;19:1212–28. impact our findings. The BRFSS question is ori- 8. Burgard SA, Seefeldt KS, Zelner S. Housing in- ented toward those paying rent or a mortgage, so stability and health: Findings from the Michigan institutionalized individuals or individuals living in recession and recovery study. Soc Sci Med 2012; shelters, transitional housing, or outside may not be 75:2215–24. captured as experiencing housing insecurity. 9. Stahre M, VanEenwyk J, Siegel P, Njai R. Housing In summary, housing insecurity is associated insecurity and the association with health outcomes with poor health, chronic disease, and inefficiently and unhealthy behaviors, Washington State, 2011. Prev Chronic Dis 2015;12. delivered care and is should be considered when 10. Pollack CE, Griffin BA, Lynch J. Housing afford- conducting social need assessments in high-risk ability and health among homeowners and renters. clinical settings or for incorporation into commu- Am J Prev Med 2010;39:515–21. nity vital signs. Future research will need to deter- 11. Ma CT, Gee L, Kushel MB. Associations between mine those populations that will most benefit from housing instability and food insecurity with health screening. Clinicians and public health practitio- care access in low-income children. Ambul Pediatr ners will need to identify effective pathways for 2008;8:50 –7. connecting individuals to community resources and 12. Fowler MG, Simpson GA, Schoendorf KC. Families on the move and children’s health care. Pediatrics scale effective community interventions. 1993;91:934 – 40. To see this article online, please go to: http://jabfm.org/content/ 13. Cutts DB, Meyers AF, Black MM, et al. US housing 32/4/521.full. insecurity and the health of very young children. Am J Public Health 2011;101:1508 –14. References 14. DeWit DJ. Frequent childhood geographic reloca- tion: Its impact on drug use initiation and the devel- 1. Bachrach D, Pfister H, Wallis K, Lipson M. Ad- opment of alcohol and other drug-related problems dressing patients’ social needs. 2014. Available from: among adolescents and young adults. Addict Behav http://www.manatt.com/uploadedFiles/Content/5_ 1998;23:623–34. News_and_Events/Newsletters/Medicaid_Update/ 15. Jelleyman T, Spencer N. Residential mobility in Addressing-Patients-Social-Needs-An-Emerging- childhood and health outcomes: a systematic review. Business-Case-for-Provider-Investment.pdf. Accessed J Epidemiol Community Health 2008;62:584 –92. September 24, 2016. 16. Social Needs Screening Toolkit. HealthLeadsUSA.org; 2. Gottlieb LM, Tirozzi KJ, Manchanda R, Burns AR, 2016. Available from: https://nopren.org/wp-content/ Sandel MT. Moving electronic medical records up- uploads/2016/12/Health-Leads-Screening-Toolkit-July- stream. Am J Prev Med 2015;48:215– 8. 2016.pdf. 3. DeVoe JE, Bazemore AW, Cottrell EK, et al. Per- 17. Centers for Disease Control and Prevention (CDC). spectives in primary care: a conceptual framework Behavioral Risk Factor Surveillance System Survey and path for integrating social determinants of Questionnaire. Atlanta, Georgia: U.S. Department health into primary care practice. Ann Fam Med of Health and Human Services, Centers for Disease 2016;14:104 – 8. Control and Prevention, 2009. 4. Committee on the Recommended Social and Behav- 18. DeVoe JE, Fryer GE, Phillips R, Green L. Receipt of ioral Domains and Measures for Electronic Health preventive care among adults: insurance status and usual Records, Board on Population Health and Public source of care. Am J Public Health 2003;93:786–91. Health Practice, Institute of Medicine. Capturing Social and Behavioral Domains and Measures in 19. The Behavioral Risk Factor Surveillance System User Electronic Health Records: Phase 2. Washington, Guide. Behavioral Risk Factor Surveillance Survey; D.C.: National Academies Press; 2014. Available 2013. Available from: https://www.cdc.gov/brfss/data_ from: http://www.nap.edu/catalog/18951. Accessed documentation/pdf/UserguideJune2013.pdf. September 24, 2016. 20. Bossarte R, Blosnich J, Piegari R, Hill L, Kane V. 5. Kushel MB, Gupta R, Gee L, Haas JS. Housing Housing instability and mental distress among US instability and food insecurity as barriers to health veterans. Am J Public Health 2013;103:S211S212. 526 JABFM July–August 2019 Vol. 32 No. 4 http://www.jabfm.org J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. 21. Herrick H, Vito Di Bona MS. The Association of rity, frequent mental distress, and insufficient sleep Housing Stress, Health, Chronic Disease and Health among adults in 12 US States, 2009. Prev Chronic Care Resources: Results from the 2011 North Car- Dis 2014;11:E37. olina BRFSS Survey. Citeseer; 2013. Available from: 28. National Association of Community Health Centers, http://citeseerx.ist.psu.edu/viewdoc/download? Association of Asian Pacific Community Health doi10.1.1.446.2110&reprep1&typepdf. Accessed Organizations, Association OPC, Institute for Al- August 18, 2016. ternative Futures. The Protocol for Responding to 22. Pobutsky AM, Baker KK, Reyes-Salvail F. Investi- and Assessing Patients’ Assets, Risks, and Experi- gating measures of social context on 2 population- ences (PRAPARE) Published 2016. Available based health surveys, Hawaii, 2010 –2012. Prev from: www.nachc.org/prapare. Chronic Dis 2015;12:E221. 29. Nuruzzaman N, Broadwin M, Kourouma K, Olson 23. Charkhchi P, Fazeli Dehkordy S, Carlos RC. Hous- DP. Making the social determinants of health a rou- ing and food insecurity, care access, and health status tine part of medical care. J Health Care Poor Un- among the chronically ill: an analysis of the behav- derserved 2015;26:321–7. ioral risk factor surveillance system. J Gen Intern 30. Garg A, Toy S, Tripodis Y, Silverstein M, Freeman Med 2018;33:644 –50. E. Addressing social determinants of health at well 24. Charters TJ, Harper S, Strumpf EC, Subramanian child care visits: a cluster RCT. Pediatrics 2015;135: SV, Arcaya M, Nandi A. The effect of metropolitan- e296 – e304. area mortgage delinquency on health behaviors, ac- 31. Fazalullasha F, Taras J, Morinis J, et al. From office cess to health services, and self-rated health in the tools to community supports: the need for infra- United States, 2003–2010. Soc Sci Med 2016;161: structure to address the social determinants of health 74 – 82. in paediatric practice. Paediatr Child Health 2014; 25. Schure MB, Katon JG, Wong E, Liu C-F. Food and 19:195–9. housing insecurity and health status among U.S. 32. Page-Reeves J, Kaufman W, Bleecker M, et al. Ad- adults with and without prior military service. SSM dressing social determinants of health in a clinic Popul Health 2016;2:244 – 8. setting: the WellRx pilot in Albuquerque, New Mex- 26. Njai R, Siegel P, Yin S, Liao Y. Prevalence of per- ico. J Am Board Fam Med 2016;29:414 – 8. ceived food and housing security - 15 states, 2013. 33. Sandel M, Hansen M, Kahn R, et al. Medical-legal MMWR Morb Mortal Wkly Rep 2017;66:12–5. partnerships: transforming primary care by address- 27. Liu Y, Njai RS, Greenlund KJ, Chapman DP, Croft ing the legal needs of vulnerable populations. Health JB. Relationships between housing and food insecu- Aff (Millwood) 2010;29:1697–705. doi: 10.3122/jabfm.2019.04.180374 Housing Insecurity and Access to Care 527 J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. Supplemental Table 1. Demographic Characteristics of Respondents, in and Not in Sample Worry About Paying Rent or Mortgage In sample Not in sample (n  228,131) (n  2,151,916) Characteristic N % N % P Value Sex Male 91,362 47.7 880,370 48.7 .2232 Female 136,769 52.3 1271546 51.3 .001 Age 18–24 years 8,676 9.8 114,874 13.2 .001 25–34 years 21,207 15.9 212,688 17.5 .001 35–44 years 28,776 17.6 262,639 16.7 .001 45–54 years 40,465 19.4 368,741 17.9 .001 55–64 years 52,992 17.8 482,443 16.1 .001 65 and over 76,015 19.5 710,531 18.6 .001 Education Less than 12 years 18,415 14.1 183,753 14.9 .001 Completed 12 years 64,801 28.9 621,592 28.4 .0188 Some college or technical school (3 years) 62,426 30.9 582,443 30.4 .0143 College (4 years or more) 82,099 26.1 752,855 25.6 .001 Income Less than 15k 20,692 9.4 220,649 11.3 .001 Greater than 15k and less than 25k 34,117 15.9 323,625 15.3 .001 Greater than 25 - less than 50k 52,042 22.5 474795 21.0 .001 Greater than 50k 91,179 39.4 800,336 37.3 .001 Race Non-Hispanic white 174,985 68.9 1,648,547 63.0 .001 Non-Hispanic black 26,239 17.2 163,025 10.9 .001 Non-Hispanic other 13,076 5.4 134,593 7.7 .001 Hispanic 11,268 7.5 172,007 16.8 .001 Marital status Married 124,472 55.1 1,125,375 50.0 .001 Divorced 31,600 11.1 299,008 10.6 .001 Widowed 30,719 7.1 286,431 6.8 .001 Separated 4,781 2.5 45,466 2.6 .001 Never married 30,718 20.3 322,395 24.6 .001 Unmarried, with partner 5,049 3.6 57,921 4.7 .001 Number of children 0 166,876 63.0 1,570,583 62.1 .001 1 24,293 15.1 231,474 15.3 .2636 2 21,931 13.2 204,959 13.2 .829 3 9,438 5.6 87,057 5.7 .2745 4 or more 4,909 2.8 45,144 3.0 .1448 Family size 1 60,641 13.6 558,283 12.7 .001 2 75,663 27.8 709,384 25.2 .001 3 23,957 13.8 234,100 13.7 .4744 4 20,564 12.9 198,828 13.2 .0341 5 9,742 6.3 94,870 6.9 .001 6 or more 6,139 4.4 63,050 5.4 .001 Continued 528 JABFM July–August 2019 Vol. 32 No. 4 http://www.jabfm.org J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. Supplemental Table 1. Continued Worry About Paying Rent or Mortgage In sample Not in sample (n  228,131) (n  2,151,916) Characteristic N % N % P Value Reported health status Excellent 39,008 17.7 377,213 19.0 .001 Very good 75,266 32.4 693,445 31.5 .001 Good 70,194 31.3 660,926 31.2 .4132 Fair 30,354 13.3 291,987 13.2 .4565 Poor 12,766 5.0 120,394 4.8 .0016 Missing 543 0.3 7,951 0.4 Number of chronic conditions 0 82,846 41.5 821,662 43.0 .001 1 63,801 26.3 601,230 25.8 .001 2 or more 75,064 32.2 661,783 31.2 .001 High cost 25,908 15.3 250,507 15.4 .001 Routine check-up 167,589 72.3 1,558,547 69.4 .001 No USC 32,850 20.9 324,193 22.6 .001 Source: Author analysis of 2011 to 2015 Behavioral Risk Factor Surveillance Survey (BRFSS). Between 2011 to 2015, 23 states implemented the Social Context Module that included this question. doi: 10.3122/jabfm.2019.04.180374 Housing Insecurity and Access to Care 529 J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. Supplemental Table 2. Demographic Characteristics of Respondents by Housing Insecurity Worry About Paying Rent or Mortgage Yes (n  28,704) No (n  199,427) Characteristic n% n % P Value Sex Male 9,996 42.7 81,366 48.6 .001 Female 18,708 57.3 118,061 51.4 Age 18–24 years 1,435 10.4 7,241 9.7 .1346 25–34 years 3,622 18.9 17,585 15.4 .001 35–44 years 4,891 21.3 23,885 16.9 .001 45–54 years 6,845 23.4 33,620 18.7 .001 55–64 years 6,868 16.6 46,124 18.0 .001 65 and over 5,043 9.5 70,972 21.3 .001 Education Less than 12 years 4,423 24.0 13,992 12.2 .001 Completed 12 years 9,983 32.6 54,818 28.2 .001 Some college or technical school (3 years) 8,700 31.1 53,726 30.8 .6321 College (4 years or more) 5,543 12.1 76,556 28.6 .001 Income Less than 15k 7,197 23.9 13,495 6.7 .001 Greater than 15k and less than 25k 7,817 28.0 26,300 13.7 .001 Greater than 25k and less than 50k 6,262 21.1 45,780 22.7 .001 Greater than 50k 4,240 15.1 86,939 43.8 .001 Race Non-Hispanic white 19,071 61.6 155,914 70.2 .001 Non-Hispanic black 4,972 22.5 21,267 16.3 .001 Non-Hispanic other 2,228 6.0 10,848 5.3 .0043 Hispanic 2,059 9.0 9,209 7.2 .001 Marital status Married 11,106 40.3 113,366 57.8 .001 Divorced 6,576 17.4 25,024 9.9 .001 Widowed 3,041 6.0 27,678 7.4 .001 Separated 1,549 5.7 3,232 1.9 .001 Never married 5,334 25.2 25,384 19.4 .001 Unmarried, with partner 985 5.0 4,064 3.4 .001 Number of children 0 18,241 55.1 148,635 64.4 .001 1 4,293 18.6 20,000 14.5 .001 2 3,511 15.4 18,420 12.8 .001 3 1,601 6.8 7,837 5.3 .001 4 or more 969 3.8 3,940 2.7 .001 Reported health status Excellent 2,278 8.7 36,730 19.3 .001 Very good 5,675 20.4 69,591 34.7 .001 Good 9,100 32.8 61,094 31.1 .001 Fair 7,042 24.2 23,312 11.3 .001 Poor 4,490 13.5 8,276 3.5 .001 Source: Author analysis of 2011 to 2015 BRFSS. Between 2011 and 2015, 23 states implemented the Social Context Module that included this question. 530 JABFM July–August 2019 Vol. 32 No. 4 http://www.jabfm.org http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The Journal of the American Board of Family Medicine Unpaywall

Adults with Housing Insecurity Have Worse Access to Primary and Preventive Care

The Journal of the American Board of Family MedicineJul 1, 2019

Loading next page...
 
/lp/unpaywall/adults-with-housing-insecurity-have-worse-access-to-primary-and-kNTNALne8i

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Unpaywall
ISSN
1557-2625
DOI
10.3122/jabfm.2019.04.180374
Publisher site
See Article on Publisher Site

Abstract

J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. ORIGINAL RESEARCH Adults with Housing Insecurity Have Worse Access to Primary and Preventive Care Patricia Martin, DO, Winston Liaw, MD, MPH, Andrew Bazemore, MD, MPH, Anuradha Jetty, MPH, Stephen Petterson, PhD, and Margot Kushel, MD Objective: Housing insecurity has been linked to high-risk behaviors and chronic disease, although less is known about the pathways leading to poor health. We sought to determine whether housing insecu- rity is associated with access to preventive and primary care. Methods: We conducted weighted univariate, bivariate, and multivariate analyses by using 2011 to 2015 Behavioral Risk factor Surveillance Survey data (N  228,131 adults). The independent variable was housing insecurity derived from the question on worry about paying rent or mortgage. The outcome measures were health services utilization (no usual source of care, no routine checkup in the past 1 year, and delayed medical care due to cost), self-rated health (number of days reported physical, men- tal health not good, and poor overall health), and number of chronic diseases (0, 1, 2 or more). The covariates included age, sex, race/ethnicity, income, level of education, marital status, and number of children in the family. We also adjusted for state fixed effects and survey year. We performed  tests and binary logistic regressions on categorical variables and ran t tests and estimated linear regression models on continuous variables. Multinomial logistic regressions were estimated for the number of chronic diseases. Results: Of the 228,131 adults in the study sample, 28,704 adults reported housing insecurity. We found that those with housing insecurity were more likely to forgo routine check-ups and lack usual sources of care. Low-income individuals, minorities, the unmarried, and middle-aged adults were more likely to report housing insecurity. Conclusion: Housing insecurity is associated with worse access to preventive and primary care. In- terventions to enhance access for these patients should be developed and studied. (J Am Board Fam Med 2019;32:521–530.) Keywords: Behavioral Risk Factor, Chronic Disease, Homeless Persons, Housing, Linear Models, Logistic Models, Mental Health, Multivariate Analysis, Outcomes Assessment, Primary Health Care, Risk-Taking, Social Determi- nants of Health, Surveillance System Social determinants of health (SDH) have a greater organizations have called for the integration of impact on people’s health and longevity than clin- public health and primary care to address SDH, ical care, and addressing these determinants is operationalizing the integration of data has proven viewed as a key strategy for meeting the triple aim difficult, with uncertainty around the types of data 2,3 of lower costs, improved patient experience, and to collect to improve outcomes. improved population health. Although numerous Housing insecurity is likely the SDH with the largest potential impact on population health, if properly addressed. Defining housing insecurity can This article was externally peer reviewed. be challenging as it refers to a spectrum of housing Submitted 15 December 2018; revised 26 March 2019; accepted 31 March 2019. experiences, including homelessness, crowding, high From Unity Health Care, Washington, D.C. (PM); De- partment of Health Systems and Population Health Sci- ences, University of Houston College of Medicine, Houston (WL); Robert Graham Center: Policy Studies in Family Medicine and Primary Care, Washington D.C. (AB, AJ, SP); Funding: none. Center for Vulnerable Populations, University of California, Conflict of interest: none declared. San Francisco/Zuckerberg San Francisco General, San Corresponding author: Patricia Martin, DO, Unity Health Francisco (MK). Care, Washington, D.C. E-mail: pmartin@unityhealthcare.org). doi: 10.3122/jabfm.2019.04.180374 Housing Insecurity and Access to Care 521 J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. housing costs in proportion to income (defined vari- about the relationship between housing insecurity ably as 30% and 50% of household income), and preventive and primary care access. In partic- 4,5 foreclosure, and frequent moves. - ular, having a usual source of care (USOC, or a Despite disagree ment about the definition, housing insecurity affects singular person or facility for navigating most millions of Americans, with over 21 million paying health care needs) is critical to receiving the rec- 30% to 50% and nearly 19 million households paying ommended screening tests and preventive services more than 50% of their income in housing costs. that can delay or prevent the development of The independent association between housing chronic disease. Our objective was to determine insecurity and poor health has been well-docu- how this BRFSS housing insecurity assessment was mented for both adults and children over the past associated with preventive and primary care access, 5,7–10 30 years. - self-rated health, and chronic disease. Nationally, we found a graded asso ciation between degree of housing insecurity and decreased access to care. Adults experiencing homelessness reported higher rates of acute care Methods services, such as emergency department visits, as We obtained data from the 2011 to 2015 BRFSS. well as postponement of needed medical care and The BRFSS is a phone-based survey that collects medications. data regarding health-related risk behaviors, Housing instability and frequent moves affect children in similar ways, resulting in chronic health conditions, and health services uti- 11,12 increased use of acute care services, - lization. It is administered annually to noninstitu- postpone ment of needed care, tionalized adults over the age of 18 in all 50 US and lack of a regular site for 12 17 preventive care. - states, Washington D.C., and 3 US territories. The prevalence of housing inse curity among low-income children (defined as liv- The BRFSS is composed of standard core ques- ing 200% below the federal poverty line or lacking tions, rotating core questions, optional modules, commercial health insurance coverage) ranges from and state-based questions. Standard core questions 11 13 29.5% are asked every year; whereas, rotating core ques- to 46% respectively. There are long-term health outcomes associated with childhood housing tions are asked every other year. States may also insecurity, which include earlier use of illicit select to include optional BRFSS modules as well as 14,15 19 drugs, - additional state-specific questions. increased rates of depression and preg nancy among teenagers, - The Social Context Module is an optional mod- and poor emotional adjust ment. ule that focuses on housing insecurity, food inse- Despite housing insecurity’s impact on health curity, and employment. We included all 23 states outcomes, researchers and policy makers have not that incorporated this module into their BRFSS achieved consensus around its measurement. Sev- questionnaires at any point between 2011 and eral clinically validated screening questions in use 2015. For the states included, the sample popula- by organizations such as Children’s Health Watch tion was comparable to the population not included and the Veteran’s Administration. in the sample (Supplemental Table 1), although the The Centers for Disease Control and Prevention Behavioral sample population was noted to be older, more Risk Factor Surveillance Survey (BRFSS) has in- affluent, and have a lower proportion of Hispanics. cluded a question regarding housing insecurity The survey includes both landline and cell phone since 2009. respondents from 2011 onwards. Response rates Deemed to be written at a 12th grade reading level with adequate precision and clinical for landline and cell phone users were 53.0% and validity, 27.9% (2011), 49.1% and 35.5% (2012), 49.6% and the question asks, “How often in the past 12 months would you say that you were worried or 37.8% (2013), 48.7% and 40.5% (2014), and 47.7% stressed about having enough money to pay your and 39.5% (2015), respectively. rent/mortgage?” - We determined the presence of housing insecu- The question addresses per ceived stress rather than a specific housing-related rity based on the question, “How often in the past metric, such as percent of household income spent 12 months would you say that you were worried or on housing. Perceived stress, however, offers a stressed about having enough money to pay your highly sensitive screening tool with the opportunity rent/mortgage?” Responses included the following: to capture the highest number of individuals at risk always, usually, sometimes, rarely, and never. We for becoming homeless. Furthermore, less is known deemed the affirmative responses “always” and 522 JABFM July–August 2019 Vol. 32 No. 4 http://www.jabfm.org J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. “usually” as markers for housing insecurity and Respondents are asked about several different dichotomized the sample based on the presence or health conditions with the following common absence of housing insecurity. question stem: “Have you ever been told by a doc- The covariates included age, sex, race/ethnicity, tor, nurse, or health care professional that you have income, level of education, marital status, number had any of the following?” Diabetes, hypertension, of children, and health status. We created 6 mutu- coronary artery disease, stroke, chronic obstructive ally exclusive age brackets (18 to 24, 25 to 34, 35 to pulmonary disease/asthma, skin cancer, other can- 44, 45 to 54, 55 to 64, and 65) and 2 mutually cer, arthritis, depression, and chronic kidney dis- exclusive sex self-identifiers. The race/ethnicity ease were among the included conditions. Those categories were white, black, Hispanic, and other. reporting skin cancer or other cancer were grouped We stratified income levels in 4 brackets based on as any cancer. We created 3 chronic disease cate- household income less than $15,000, $15,000 or gories: (1) no chronic conditions, (2) 1 chronic more and less than $25,000, $25,000 or more and condition, and (3) 2 or more chronic conditions. less than $50,000, and $50,000 or more. We clas- sified level of education in 4 mutually exclusive Statistical Analysis groups: less than high school education, high Using Stata 14.0, we conducted univariate and bi- school graduate/general education degree, some variate ( for categorical variables and t tests for college or technical school (1 to 3 years), and 4 continuous variables) analyses and multivariate lin- years or more of college education. Self-reported ear and binary logistic regressions. We used sam- health status was classified as excellent, very good, pling weights and BRFSS survey design variables good, fair, or poor. Finally, to assess family struc- throughout the analysis to obtain a nationally rep- ture, we determined the number of children and resentative sample of undersampled populations. marital status (married, divorced, widowed, sepa- We first compared the demographic characteristics rated, never married, and unmarried with partner). (sex, age, education, race/ethnicity, and poverty) of The dependent variables can be grouped into 3 the respondents in sample and out of sample by categories: (1) health services utilization, (2) self- conducting bivariate analysis using  tests for cat- reported health status, and (3) presence of chronic egorical and t tests for continuous variables. For the medical conditions. For health services utilization, respondents in the sample, we then computed de- we assessed the respondents’ USOC status, defer- scriptive statistics of demographic characteristics ment of medical care due to cost, and prolonged and conducted bivariate analyses by housing inse- time since last medical checkup. We characterized curity. We estimated 3 sets of regression models by absence of a USOC by a “No” response to the using linear regression for continuous outcomes question “Do you have 1 person that you think of as (health status measures reported as number of days) your personal doctor or health care provider?” We and logistic regressions for binary outcomes (hav- characterized deferment of medical care due to cost ing a USOC, avoiding or delaying medical care, as a “Yes” response to the question “Was there a and having a chronic medical condition). In all the time in the past 12 months when you needed to see models, we used housing insecurity as an indepen- a doctor but could not because of cost?” Lastly, we dent variable and patient characteristics as covari- characterized prolonged time since last medical ates. Because our analysis was restricted to a pub- checkup as any response other than “Within the licly available data set, institutional review board past year” to the question “About how long has it approval was neither required nor obtained. been since you last visited a doctor for a routine checkup?” The second and third categories of interest were Results self-reported health and the presence of chronic Of the 228,131 individuals in our sample, 14.3% medical conditions. We assessed self-reported reported housing insecurity (Supplemental Table health by determining the responses to 3 questions 2). Those reporting housing insecurity were that inquire about the specific number of days more likely to be female, have lower incomes, and within the past month that physical, mental, and be black or Hispanic. Of all age groups, middle-age overall health were not good. The presence of adults (35 to 54) were the most likely to experience chronic medical conditions is also by self-report. housing insecurity. As expected, those with a college doi: 10.3122/jabfm.2019.04.180374 Housing Insecurity and Access to Care 523 J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. Table 1. Health Outcomes and Utilization by Housing Insecurity Worry About Paying Rent or Mortgage Yes (n  28,704) No (n  199,427) Self-Reported Health Status and Utilization n% n % P Value Number of Chronic Conditions 0 910 30.8 10,984 43.5 .001 1 919 27.7 8,620 26 .6397 2 or more 1,973 41.5 12,312 30.4 .001 Health services utilization Deferred care due to cost 11,009 42.8 14,899 10.2 .001 Had a routine check-up within past year 18,821 64.0 148,768 73.8 .001 No usual source of care 6,168 29.0 26,682 19.4 .001 Self-reported health (mean number of days in last month) Poor physical health 8.6 3.9 .001 Poor mental health 9.5 2.6 .001 Poor overall health 8.0 3.3 .001 Source: Author analysis of 2011 to 2015 Behavioral Risk Factor Surveillance Survey (BRFSS) data. Between 2011 to 2015, 23 states implemented the Social Context Module that included this question. BRFSS asks about self-reported health by determining the responses to three questions that inquire about the specific number of days within the past month that physical, mental, and overall health were not good. education were less likely to be housing insecure. found that those with housing insecurity were more Those with more than a high school education were likely to have 1 or more chronic conditions. also less likely to be housing insecure, which may reflect individuals living in areas with a high cost-of- living who despite reported income of $50,000 are Discussion still uncomfortable with their disproportionate hous- BRFSS has been widely used to demonstrate the ing costs. Not being married, having larger families, association between housing insecurity and health 9,20 –22 and being in poor health were all associated with outcomes at the single-state level as well as 23–27 housing insecurity. the multistate level although none of these Those with housing insecurity were more likely studies assessed housing insecurity’s impact on pre- to report chronic diseases, lack a USOC, and defer ventive and primary care access. Our findings con- care in the past year due to cost (Table 1). With firm the potent relationship between housing inse- respect to overall health, those with housing inse- curity and poor health and add to the literature by curity had more days in the last month with poor documenting that, among those with housing inse- health. Furthermore, the number of days in poor curity, poor health may be the consequence of mental health was the highest of the 3 health cat- forgoing routine care and lacking a USOC. Fur- egories. thermore, we found that low-income, minority, un- After controlling for covariates, we found that married, middle-aged adults are more likely to re- those with housing insecurity had 3 times higher port housing insecurity. As public health and odds of delaying care due to cost. Housing-insecure primary care collaborate to identify and address respondents also had 35% higher odds of delaying housing insecurity, these findings can provide guid- check-ups and had 19% higher odds of lacking a ance for future screening efforts, such as whether to USOC. In our second regression model, where the use targeted or universal screening for housing in- outcome was number of days in poor health, we security. found that housing insecurity was associated with There are 1-item screening tools already in use 4.7 more days in poor physical health, 6.9 more by other agencies, such as Children’s Health days in poor mental health, and 4.7 more days in Watch (“An eviction is when your landlord or a poor overall health. In our final regression model, government or bank official forces you to move relative to those with no chronic conditions, we when you do not want to. In the past 5 years, have 524 JABFM July–August 2019 Vol. 32 No. 4 http://www.jabfm.org J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. you ever been evicted?”) and the Veterans Admin- community resources and following up with par- istration (“Are you worried or concerned that in the ents has proven beneficial in the pediatric set- 31,32 next 2 months you may not have stable housing that tings. - Among the adult population, pilot stud you own, rent, or stay in as part of a household?”). ies like the WellRx in New Mexico have drawn The populations of interest for these agencies, attention to the important role of community however, are not reflective of the general popula- health workers in helping patients with unmet so- tion, limiting their generalizability. cial needs to follow through and successfully access In 2017, the Center for Medicaid and Medicare community resources. Medical-legal partnerships Services also released the 10-item Accountable also have a role in the primary care setting in Health Communities Screening Tool, which in- addressing unmet social needs that involve the legal cludes 2 items relating to housing insecurity system and have been successful in many patient (“What is your housing situation today?” and center medical home settings. “Think about the place where you live. Do you Although the demand for stable affordable hous- have problems with any of the following? Bug In- ing in the United States far outpaces the limited festation, Mold, Lead Paint or Pipes, Inadequate supply, there is still benefit to screening for SDH in Heat, Oven of Stove Not Working, No or Not primary care. Having information about patients’ Working Smoke Detectors, Water Leaks”). unmet social needs can improve patient care by The first question was adapted from the Protocol for better informing providers, who can engage pa- Responding to and Addressing Patients’ Assets, tients in discussions that would not otherwise oc- 29 34 Risks, and Experiences assessment tool. - cur The sec and advocate on their behalf for interventions ond question was drawn from a screening tool de- at the community level, including enhancing access veloped by Norwalk Community Health Center to to housing voucher programs and programs that capture patients that would benefit from legal ser- prevent evictions and building high-quality, low- vices offered by their medical-legal partnership. income housing within mixed-income neighbor- These items are evidence-based but have not yet hoods. been validated. - Even the briefest SDH screening tool, however, In addition, given the recent de velopment of the Center for Medicaid and Medi- requires time and investment on the part of pro- care Services screening tool, there are no available viders. For providers who are not working directly studies at this time to link these particular housing with a social worker, referrals for community re- insecurity items with health-related outcomes. sources and follow up may fall to the provider. Housing insecurity was omitted as a National Although these efforts may result in improved Academy of Medicine-recommended domain for health outcomes and health care cost savings, it is inclusion in electronic health records as part of unclear how that cost savings would be shared. Meaningful Use because of the lack of standard More work is needed to explore how primary care measure and perceived difficulty in collecting hous- practices could be appropriately compensated for 4,5 ing insecurity data. the increased expenditure of time and staff re- Without a standard housing insecurity metric, there is a need for a validated sources spent in deploying screening for housing screening tool for the general population to begin insecurity. collecting and appropriately acting on this SDH Some limitations to the data source used in our data. Our findings add to the existing literature analyses are worth noting, including sampling that supporting the use of this BRFSS housing insecu- is not nationally representative, reflecting only the rity question. In the absence of systematic assess- states that administered the BRFSS Social Context ment of housing within medicine, this critical SDH Module. As noted earlier, the housing insecurity will continue to be ignored by clinicians and link- question addresses self-reported stress regarding ages between medicine and public health will fail to paying rent or a mortgage, which is subject to develop. survey participant interpretation of both the defi- With appropriate screening, those with housing nition of stress as well as their own stress level. insecurity can be identified and referred for assis- Dependent variables were also based on self-report, tance. The role of primary care providers in the and the survey questions asked about behaviors implementation of interventions to address SDH is over different time periods. For example, the ques- evolving. Thus far, generating referrals for existing tion regarding housing insecurity includes the past doi: 10.3122/jabfm.2019.04.180374 Housing Insecurity and Access to Care 525 J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. care among low-income Americans. J Gen Intern 12 months while the questions regarding self-re- Med 2006;21:71–7. ported health included the past 30 days and the 6. The State of the Nation’s Housing 2016. President questions regarding health care utilization go back and Fellows of Harvard College; 2016. Available as far as 5 years. Finally, BRFSS is a phone-based from: https://www.jchs.harvard.edu/sites/default/files/ survey that excludes institutionalized adults, which jchs_2016_state_of_the_nations_housing_lowres_0.pdf. omits those without phone lines as well as those 7. Reid KW, Vittinghoff E, Kushel MB. Association living in shelters, transitional housing, or outside. It between the level of housing instability, economic is unclear how inclusion of these populations would standing and health care access: a meta-regression. J Health Care Poor Underserved 2008;19:1212–28. impact our findings. The BRFSS question is ori- 8. Burgard SA, Seefeldt KS, Zelner S. Housing in- ented toward those paying rent or a mortgage, so stability and health: Findings from the Michigan institutionalized individuals or individuals living in recession and recovery study. Soc Sci Med 2012; shelters, transitional housing, or outside may not be 75:2215–24. captured as experiencing housing insecurity. 9. Stahre M, VanEenwyk J, Siegel P, Njai R. Housing In summary, housing insecurity is associated insecurity and the association with health outcomes with poor health, chronic disease, and inefficiently and unhealthy behaviors, Washington State, 2011. Prev Chronic Dis 2015;12. delivered care and is should be considered when 10. Pollack CE, Griffin BA, Lynch J. Housing afford- conducting social need assessments in high-risk ability and health among homeowners and renters. clinical settings or for incorporation into commu- Am J Prev Med 2010;39:515–21. nity vital signs. Future research will need to deter- 11. Ma CT, Gee L, Kushel MB. Associations between mine those populations that will most benefit from housing instability and food insecurity with health screening. Clinicians and public health practitio- care access in low-income children. Ambul Pediatr ners will need to identify effective pathways for 2008;8:50 –7. connecting individuals to community resources and 12. Fowler MG, Simpson GA, Schoendorf KC. Families on the move and children’s health care. Pediatrics scale effective community interventions. 1993;91:934 – 40. To see this article online, please go to: http://jabfm.org/content/ 13. Cutts DB, Meyers AF, Black MM, et al. US housing 32/4/521.full. insecurity and the health of very young children. Am J Public Health 2011;101:1508 –14. References 14. DeWit DJ. Frequent childhood geographic reloca- tion: Its impact on drug use initiation and the devel- 1. Bachrach D, Pfister H, Wallis K, Lipson M. Ad- opment of alcohol and other drug-related problems dressing patients’ social needs. 2014. Available from: among adolescents and young adults. Addict Behav http://www.manatt.com/uploadedFiles/Content/5_ 1998;23:623–34. News_and_Events/Newsletters/Medicaid_Update/ 15. Jelleyman T, Spencer N. Residential mobility in Addressing-Patients-Social-Needs-An-Emerging- childhood and health outcomes: a systematic review. Business-Case-for-Provider-Investment.pdf. Accessed J Epidemiol Community Health 2008;62:584 –92. September 24, 2016. 16. Social Needs Screening Toolkit. HealthLeadsUSA.org; 2. Gottlieb LM, Tirozzi KJ, Manchanda R, Burns AR, 2016. Available from: https://nopren.org/wp-content/ Sandel MT. Moving electronic medical records up- uploads/2016/12/Health-Leads-Screening-Toolkit-July- stream. Am J Prev Med 2015;48:215– 8. 2016.pdf. 3. DeVoe JE, Bazemore AW, Cottrell EK, et al. Per- 17. Centers for Disease Control and Prevention (CDC). spectives in primary care: a conceptual framework Behavioral Risk Factor Surveillance System Survey and path for integrating social determinants of Questionnaire. Atlanta, Georgia: U.S. Department health into primary care practice. Ann Fam Med of Health and Human Services, Centers for Disease 2016;14:104 – 8. Control and Prevention, 2009. 4. Committee on the Recommended Social and Behav- 18. DeVoe JE, Fryer GE, Phillips R, Green L. Receipt of ioral Domains and Measures for Electronic Health preventive care among adults: insurance status and usual Records, Board on Population Health and Public source of care. Am J Public Health 2003;93:786–91. Health Practice, Institute of Medicine. Capturing Social and Behavioral Domains and Measures in 19. The Behavioral Risk Factor Surveillance System User Electronic Health Records: Phase 2. Washington, Guide. Behavioral Risk Factor Surveillance Survey; D.C.: National Academies Press; 2014. Available 2013. Available from: https://www.cdc.gov/brfss/data_ from: http://www.nap.edu/catalog/18951. Accessed documentation/pdf/UserguideJune2013.pdf. September 24, 2016. 20. Bossarte R, Blosnich J, Piegari R, Hill L, Kane V. 5. Kushel MB, Gupta R, Gee L, Haas JS. Housing Housing instability and mental distress among US instability and food insecurity as barriers to health veterans. Am J Public Health 2013;103:S211S212. 526 JABFM July–August 2019 Vol. 32 No. 4 http://www.jabfm.org J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. 21. Herrick H, Vito Di Bona MS. The Association of rity, frequent mental distress, and insufficient sleep Housing Stress, Health, Chronic Disease and Health among adults in 12 US States, 2009. Prev Chronic Care Resources: Results from the 2011 North Car- Dis 2014;11:E37. olina BRFSS Survey. Citeseer; 2013. Available from: 28. National Association of Community Health Centers, http://citeseerx.ist.psu.edu/viewdoc/download? Association of Asian Pacific Community Health doi10.1.1.446.2110&reprep1&typepdf. Accessed Organizations, Association OPC, Institute for Al- August 18, 2016. ternative Futures. The Protocol for Responding to 22. Pobutsky AM, Baker KK, Reyes-Salvail F. Investi- and Assessing Patients’ Assets, Risks, and Experi- gating measures of social context on 2 population- ences (PRAPARE) Published 2016. Available based health surveys, Hawaii, 2010 –2012. Prev from: www.nachc.org/prapare. Chronic Dis 2015;12:E221. 29. Nuruzzaman N, Broadwin M, Kourouma K, Olson 23. Charkhchi P, Fazeli Dehkordy S, Carlos RC. Hous- DP. Making the social determinants of health a rou- ing and food insecurity, care access, and health status tine part of medical care. J Health Care Poor Un- among the chronically ill: an analysis of the behav- derserved 2015;26:321–7. ioral risk factor surveillance system. J Gen Intern 30. Garg A, Toy S, Tripodis Y, Silverstein M, Freeman Med 2018;33:644 –50. E. Addressing social determinants of health at well 24. Charters TJ, Harper S, Strumpf EC, Subramanian child care visits: a cluster RCT. Pediatrics 2015;135: SV, Arcaya M, Nandi A. The effect of metropolitan- e296 – e304. area mortgage delinquency on health behaviors, ac- 31. Fazalullasha F, Taras J, Morinis J, et al. From office cess to health services, and self-rated health in the tools to community supports: the need for infra- United States, 2003–2010. Soc Sci Med 2016;161: structure to address the social determinants of health 74 – 82. in paediatric practice. Paediatr Child Health 2014; 25. Schure MB, Katon JG, Wong E, Liu C-F. Food and 19:195–9. housing insecurity and health status among U.S. 32. Page-Reeves J, Kaufman W, Bleecker M, et al. Ad- adults with and without prior military service. SSM dressing social determinants of health in a clinic Popul Health 2016;2:244 – 8. setting: the WellRx pilot in Albuquerque, New Mex- 26. Njai R, Siegel P, Yin S, Liao Y. Prevalence of per- ico. J Am Board Fam Med 2016;29:414 – 8. ceived food and housing security - 15 states, 2013. 33. Sandel M, Hansen M, Kahn R, et al. Medical-legal MMWR Morb Mortal Wkly Rep 2017;66:12–5. partnerships: transforming primary care by address- 27. Liu Y, Njai RS, Greenlund KJ, Chapman DP, Croft ing the legal needs of vulnerable populations. Health JB. Relationships between housing and food insecu- Aff (Millwood) 2010;29:1697–705. doi: 10.3122/jabfm.2019.04.180374 Housing Insecurity and Access to Care 527 J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. Supplemental Table 1. Demographic Characteristics of Respondents, in and Not in Sample Worry About Paying Rent or Mortgage In sample Not in sample (n  228,131) (n  2,151,916) Characteristic N % N % P Value Sex Male 91,362 47.7 880,370 48.7 .2232 Female 136,769 52.3 1271546 51.3 .001 Age 18–24 years 8,676 9.8 114,874 13.2 .001 25–34 years 21,207 15.9 212,688 17.5 .001 35–44 years 28,776 17.6 262,639 16.7 .001 45–54 years 40,465 19.4 368,741 17.9 .001 55–64 years 52,992 17.8 482,443 16.1 .001 65 and over 76,015 19.5 710,531 18.6 .001 Education Less than 12 years 18,415 14.1 183,753 14.9 .001 Completed 12 years 64,801 28.9 621,592 28.4 .0188 Some college or technical school (3 years) 62,426 30.9 582,443 30.4 .0143 College (4 years or more) 82,099 26.1 752,855 25.6 .001 Income Less than 15k 20,692 9.4 220,649 11.3 .001 Greater than 15k and less than 25k 34,117 15.9 323,625 15.3 .001 Greater than 25 - less than 50k 52,042 22.5 474795 21.0 .001 Greater than 50k 91,179 39.4 800,336 37.3 .001 Race Non-Hispanic white 174,985 68.9 1,648,547 63.0 .001 Non-Hispanic black 26,239 17.2 163,025 10.9 .001 Non-Hispanic other 13,076 5.4 134,593 7.7 .001 Hispanic 11,268 7.5 172,007 16.8 .001 Marital status Married 124,472 55.1 1,125,375 50.0 .001 Divorced 31,600 11.1 299,008 10.6 .001 Widowed 30,719 7.1 286,431 6.8 .001 Separated 4,781 2.5 45,466 2.6 .001 Never married 30,718 20.3 322,395 24.6 .001 Unmarried, with partner 5,049 3.6 57,921 4.7 .001 Number of children 0 166,876 63.0 1,570,583 62.1 .001 1 24,293 15.1 231,474 15.3 .2636 2 21,931 13.2 204,959 13.2 .829 3 9,438 5.6 87,057 5.7 .2745 4 or more 4,909 2.8 45,144 3.0 .1448 Family size 1 60,641 13.6 558,283 12.7 .001 2 75,663 27.8 709,384 25.2 .001 3 23,957 13.8 234,100 13.7 .4744 4 20,564 12.9 198,828 13.2 .0341 5 9,742 6.3 94,870 6.9 .001 6 or more 6,139 4.4 63,050 5.4 .001 Continued 528 JABFM July–August 2019 Vol. 32 No. 4 http://www.jabfm.org J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. Supplemental Table 1. Continued Worry About Paying Rent or Mortgage In sample Not in sample (n  228,131) (n  2,151,916) Characteristic N % N % P Value Reported health status Excellent 39,008 17.7 377,213 19.0 .001 Very good 75,266 32.4 693,445 31.5 .001 Good 70,194 31.3 660,926 31.2 .4132 Fair 30,354 13.3 291,987 13.2 .4565 Poor 12,766 5.0 120,394 4.8 .0016 Missing 543 0.3 7,951 0.4 Number of chronic conditions 0 82,846 41.5 821,662 43.0 .001 1 63,801 26.3 601,230 25.8 .001 2 or more 75,064 32.2 661,783 31.2 .001 High cost 25,908 15.3 250,507 15.4 .001 Routine check-up 167,589 72.3 1,558,547 69.4 .001 No USC 32,850 20.9 324,193 22.6 .001 Source: Author analysis of 2011 to 2015 Behavioral Risk Factor Surveillance Survey (BRFSS). Between 2011 to 2015, 23 states implemented the Social Context Module that included this question. doi: 10.3122/jabfm.2019.04.180374 Housing Insecurity and Access to Care 529 J Am Board Fam Med: first published as 10.3122/jabfm.2019.04.180374 on 12 July 2019. Downloaded from http://www.jabfm.org/ on 13 December 2021 by guest. Protected by copyright. Supplemental Table 2. Demographic Characteristics of Respondents by Housing Insecurity Worry About Paying Rent or Mortgage Yes (n  28,704) No (n  199,427) Characteristic n% n % P Value Sex Male 9,996 42.7 81,366 48.6 .001 Female 18,708 57.3 118,061 51.4 Age 18–24 years 1,435 10.4 7,241 9.7 .1346 25–34 years 3,622 18.9 17,585 15.4 .001 35–44 years 4,891 21.3 23,885 16.9 .001 45–54 years 6,845 23.4 33,620 18.7 .001 55–64 years 6,868 16.6 46,124 18.0 .001 65 and over 5,043 9.5 70,972 21.3 .001 Education Less than 12 years 4,423 24.0 13,992 12.2 .001 Completed 12 years 9,983 32.6 54,818 28.2 .001 Some college or technical school (3 years) 8,700 31.1 53,726 30.8 .6321 College (4 years or more) 5,543 12.1 76,556 28.6 .001 Income Less than 15k 7,197 23.9 13,495 6.7 .001 Greater than 15k and less than 25k 7,817 28.0 26,300 13.7 .001 Greater than 25k and less than 50k 6,262 21.1 45,780 22.7 .001 Greater than 50k 4,240 15.1 86,939 43.8 .001 Race Non-Hispanic white 19,071 61.6 155,914 70.2 .001 Non-Hispanic black 4,972 22.5 21,267 16.3 .001 Non-Hispanic other 2,228 6.0 10,848 5.3 .0043 Hispanic 2,059 9.0 9,209 7.2 .001 Marital status Married 11,106 40.3 113,366 57.8 .001 Divorced 6,576 17.4 25,024 9.9 .001 Widowed 3,041 6.0 27,678 7.4 .001 Separated 1,549 5.7 3,232 1.9 .001 Never married 5,334 25.2 25,384 19.4 .001 Unmarried, with partner 985 5.0 4,064 3.4 .001 Number of children 0 18,241 55.1 148,635 64.4 .001 1 4,293 18.6 20,000 14.5 .001 2 3,511 15.4 18,420 12.8 .001 3 1,601 6.8 7,837 5.3 .001 4 or more 969 3.8 3,940 2.7 .001 Reported health status Excellent 2,278 8.7 36,730 19.3 .001 Very good 5,675 20.4 69,591 34.7 .001 Good 9,100 32.8 61,094 31.1 .001 Fair 7,042 24.2 23,312 11.3 .001 Poor 4,490 13.5 8,276 3.5 .001 Source: Author analysis of 2011 to 2015 BRFSS. Between 2011 and 2015, 23 states implemented the Social Context Module that included this question. 530 JABFM July–August 2019 Vol. 32 No. 4 http://www.jabfm.org

Journal

The Journal of the American Board of Family MedicineUnpaywall

Published: Jul 1, 2019

There are no references for this article.