Better Housing Improves People’s Lives—Health Benefits Should Be Seen as a Bonus

Population health

Many advocates hope that addressing the quality of housing, or similar social needs, can help us escape from the quagmire of ever-increasing US health care costs combined with stagnating health outcomes. Their logic is simple: low-income populations often live in dismal conditions that are associated with disease, and treating those diseases drives up health system costs. Improving housing quality might both prevent disease and reduce spending. The premise is attractive to cost-conscious health payers, who hope to achieve the Triple Aim of better care, better health, and lower costs as well as to budget-strapped housing authorities, who hope to recoup the costs of renovations with health care financing. Yet addressing housing quality is fundamentally different from prescribing medicine. By the standards used to evaluate medical interventions, housing improvements will rarely appear successful. That’s because the standards are misapplied in this context: better housing confers many benefits, and health outcomes are only a small piece of the very real value of improved housing to residents. As we move toward health in all policies, we need to make sure we don’t assume that health is all.

The strategy of improving housing to generate better health is grounded in the robust correlation between the quality of housing and health.1 Improving the quality of housing is, indeed, an ideal intervention in the case of the few housing risk factors that are analogous to monogenetic risks in medicine: think of lead poisoning or falls from high-rise windows. In these cases, the health problem is almost always present when the housing problem is present, and almost always absent when the housing problem is absent. Removal of lead-based paint is the best way to prevent lead poisoning, and installation of window guards is the most effective way to reduce injuries associated with falls from windows.

Most common housing quality issues, such as inadequate heating, pests, and mold, address problems that behave more like multifactorial genetic conditions, where the presence or absence of a housing condition is not dispositive. Because most housing-quality-related health problems have multiple causes, intervening in housing alone will not fully eliminate these problems. People live in substandard housing because they can’t afford anything better, or because of discrimination, or because they need to live near a job to avoid long commutes, or because they have high expenses for other needs. All of these motivations directly affect health. Housing interventions to address these problems are also typically multifaceted—repairing leaks, adding insulation, improving heating—and an intervention strategy will likely impact many aspects of health and well-being.

These challenges mean that the causal evidence supporting the use of specific housing interventions to improve specific dimensions of health is surprisingly weak. A recent careful and comprehensive review of the evidence linking housing quality to health demonstrates that improving heating and reducing mold alone can affect some health outcomes (though these are often self-reported measures of symptoms such as colds or wheezing), but finds little evidence that eliminating dust or pests or making alterations to reduce falls, without further medical intervention, improve health.2,3 The magnitude of the positive effects in the comprehensive review may be optimistic, because many of the reviewed studies targeted renovations to those already identified as having health risks, magnifying their measured effects compared to less-targeted interventions. Finally, each reviewed study took place in a particular locality, with particular environmental conditions, so their generalizability to areas with different types of housing and weather conditions may be limited. Nonetheless, I can use this literature to provide a first-pass assessment of how effective housing interventions might be in addressing health alone, from the perspective of either a housing authority or a health care provider.

I apply estimates of the prevalence of health conditions in units with poor housing quality (drawn from the control arms of studies included in the reviews) to the prevalence of these housing conditions in the United States (drawn from the 2015 American Housing Survey) to estimate how many cases of health problems likely exist in units with poor housing quality (Table 1). Many people live in low-quality housing. In an analysis that adjusts for overlap of housing conditions among units, I find that close to 13 million housing units in the nation have either inadequate heating or mold. Based on the studies, it’s plausible that people living in these units experience at least 7 million adverse health outcomes each year. That’s a big potential target.

Table 1. Prevalence of Housing Problems and Housing-Related Health Problems and Size of Interventions Needed to Detect Renovation Effects

Housing ProblemNumber of Units Susceptible to Housing Problem
(Millions of Units)
Rate of Health Problems Per Unit (Control Arm of Trials at Baseline)Expected Number of Health Problems Among Residents of Affected Units
(Millions of People)
Number of Units with Housing Problem Needed to Treat in Order to Detect an Effect on Health Relative to BaselineNumber of Patients with Identified Health Condition and Housing Problem Needed to Treat in Order to Achieve Benefit
Inadequate heat9.7Wheezing (Past three months): 34.1%33.327013
Colds and flu: 61.7%36.01539
Poor mental health: 7.8%30.881337
High blood pressure: 20.6%42.069124
Mold4.3Colds and flu (Rhino conjunctivitis): 38.0%51.6415
Wheeze limits speech: 17.2%50.717513

Source: Sample size calculations using calculator at http://clincalc.com/stats/samplesize.aspx. Number needed to treat computed as 1/(Control Event Rate – Treatment Event Rate).

Suppose a housing authority decided to focus its attention on the goal of improving residents’ health by renovating deficient units, anticipating that the costs of renovations could be recouped through lower health costs. How many units with a specific deficiency would it need to rehabilitate to detect a statistically meaningful effect on a particular health problem? I can compute this using data from the reviewed studies. I calculate the number of units that would need to be renovated using a standard power analysis based on a one-arm intervention trial (with a comparison group of the existing population). The fourth column of Table 1 presents these estimates.

If housing interventions are initially targeted based only on the presence of housing quality deficiencies (not considering health conditions), detecting health effects is likely to be quite challenging. I estimate, based on the magnitudes reported in the literature, that a housing authority would need to improve the heating in about 270 under-heated units to detect any significant effects of this intervention on a proximate symptom such as a self-report of wheezing. That is a conservative estimate, because it’s based on health outcomes measured in studies that targeted interventions to populations with rates of health problems about twice as high as is typical in low-income populations. If the housing authority didn’t target renovations to units where someone had symptoms already, about 2.5 to 3 times as many units would need to be renovated to discern an effect. Still more units would need to be renovated to detect effects for costly events such as emergency room visits, which occur much less frequently than self-reported episodes of wheezing. These small effect sizes mean that linking housing renovations to clinical outcomes in the hope of recouping the cost of renovations through verifiably lower health care costs will be beyond the scope of most local housing authorities.

An alternative approach might be to tie addressing housing quality to the clinical care system. In this model, providers would identify potential housing-related health problems in clinical care, follow-up to determine whether a housing deficiency is present, and then work with a housing-focused organization to repair specific problems in specific units. The final column in Table 1 reports analyses, based on the reviewed studies, of the number of units of housing with people suffering from a given health problem that would need to be renovated to improve outcomes for one patient. For the outcomes in this table, these figures range from about 5 to 37 units. These figures would be much higher if I adjusted for weak study targeting and outcome severity as described above. Put simply, many people whose units are renovated would not have developed housing-related diseases; some will develop those diseases despite the renovations.

While these number-needed-to-treat figures may seem large (many housing units need to be renovated to observe a health effect), they are well within the range of many commonly used clinical interventions. From the perspective of a health system, however, dispensing a prescription is a lot less trouble than ascertaining housing deficiencies, building coalitions that can address them, and implementing the necessary renovations. And, unlike the case with conventional clinical treatment, limiting the receipt of a housing intervention to those most likely to experience a health benefit raises serious logistical and ethical problems. Health systems that seek to improve health outcomes through housing renovations are likely to find the benefit to their bottom line to not be worth the cost.

There’s a basic logical error at work here: better housing is not, and should not be thought of, as primarily a health intervention—housing is so much more! While a pill confers, at best, nothing but a health benefit, we value the comfort and quality of our homes for many reasons, with health being only one factor. The predictably disappointing health outcomes of housing renovation interventions could have the perverse effect of leading us to underinvest in social determinants that generate a great deal of benefit across both health and other dimensions of well-being. Rats, inadequate heat, and mold substantially diminish people’s lives whether or not they cause direct harm to their health. In addressing social determinants of health, any cost savings to the health care system should not act as a barrier, but rather should be viewed as a bonus.

 

Acknowledgment: I thank Sarah Holder and Dane Gambrell for wonderful research assistance. This research was supported by a Polices for Action grant from the Robert Wood Johnson Foundation.

References

  1. Thomson H, Thomas S, Sellstrom E, Petticrew M. Housing improvements for health and associated socio-economic outcomes. Cochrane Database of Systematic Reviews. 2013;2(2):CD008657.
  2. Thomson H, Thomas S, Sellstrom E, Petticrew M. Housing improvements for health and associated socio-economic outcomes. London: Sons; 2013.
  3. Sauni R, Uitti J, Jauhiainen M, Kreiss K, Sigsgaard T, Verbeek JH. Remediating buildings damaged by dampness and mould for preventing or reducing respiratory tract symptoms, infections and asthma. Evidence‐Based Child Health: A Cochrane Review Journal. 2013;8(3):944-1000.
  4. Platt S, Mitchell R, Walker J, Hopton J, Petticrew M, Corbett J. The Scottish Executive Central Heating Programme: assessing impacts on health. Research Findings 239. Edinburgh: Scottish Executive: Social Research Development Department; 2007.
  5. Burr ML, Matthews IP, Arthur RA, et al. Effects on patients with asthma of eradicating visible indoor mould: a randomised controlled trial. Thorax. 2007;62(9):767-772.

 


Citation:
Glied S. Better Housing Improves People’s Lives—Health Benefits Should Be Seen as a Bonus. Milbank Quarterly Opinion. November 12, 2020. https://doi.org/10.1599/mqop.2020.1112


About the Author

Sherry Glied was named dean of New York University’s Robert F. Wagner Graduate School of Public Service in 2013. From 1989-2013, she was professor of health policy and management at Columbia University’s Mailman School of Public Health. She was chair of the Department of Health Policy and Management from 1998-2009. On June 22, 2010, Glied was confirmed by the US Senate as assistant secretary for planning and evaluation at the Department of Health and Human Services, and served in that capacity from July 2010 through August 2012. She had previously served as senior economist for health care and labor market policy on the President’s Council of Economic Advisers in 1992-1993, under Presidents Bush and Clinton, and participated in the Clinton Health Care Task Force. She has been elected to the National Academy of Medicine, the National Academy of Social Insurance, and served as a member of the Commission on Evidence-Based Policymaking. Glied’s principal areas of research are in health policy reform and mental health care policy. Her book on health care reform, Chronic Condition, was published by Harvard University Press in January 1998. Her book with Richard Frank, Better But Not Well: Mental Health Policy in the US since 1950, was published by The Johns Hopkins University Press in 2006. She is co-editor, with Peter C. Smith, of The Oxford Handbook of Health Economics, which was published by the Oxford University Press in 2011. Glied holds a BA in economics from Yale University, an MA in economics from the University of Toronto, and a PhD in economics from Harvard University.

See Full Bio