Abstract
Should Medicaid pay for supportive housing for homeless persons? After describing current limits on how states can use Medicaid funds to support housing, this article considers whether justice requires treating Medicaid recipients residing in nursing homes and Medicaid recipients needing supportive housing similarly.
Introduction
Achieving safe and stable housing presents a profound and ongoing challenge for many people living in poverty. These challenges include only being able to afford housing that falls below minimum standards for safety and habitability and regularly facing the possibility of eviction. In some communities, available housing is so expensive that paying rent consumes an untenable portion of monthly income, leaving inadequate funds for other necessities. For a growing number, the challenges of maintaining housing prove too much, and they become homeless. 1
Little imagination is required to understand how housing-related challenges that are part of daily life for many poor people can influence their physical and mental health. We do not have to depend on our imagination, however. Research provides a solid basis for treating housing as a social determinant of health. 2 Homelessness in particular is both a cause of and result of poor health. Being without housing may lead to new health problems for a person and make it harder to treat existing illness. Living on the streets increases the risk of physical and sexual victimization. In addition, research indicates that a majority of homeless persons experience some kind of substance use or significant mental health issues. 3
Increased attention to the health impacts of inadequate, insecure, and unaffordable housing has prompted some — including public health experts, physicians, and sociologists studying housing — to urge that housing issues, and homelessness in particular, be addressed as a health problem. Two articles published in JAMA in October 2017 illustrate this approach, using language that medicalizes the housing challenges associated with poverty. Megan Sandel, a Boston physician, and Matthew Desmond, the author of the widely acclaimed book Evicted, wrote bluntly: “Housing is similar to a drug prescription.” 4 Mitchell Katz, the long-time Director of the Los Angeles Department of Health Services, took the next logical step, connecting housing as a health issue to sources of funding for health care. He wrote: “If Medicaid were to pay for supportive housing for persons with chronic medical or psychiatric illness, eligibility would be determined by medical necessity, just as is the case for a nursing home.” 5
This article considers whether Medicaid should pay for supportive housing for some recipients. First, it briefly describes how state Medicaid programs already are permitted to use federal Medicaid funds in support of recipients' shelter needs in three different contexts: (1) the mandatory nursing home benefit; (2) expanded home and community-based services (HCBS) programs for recipients needing long-term support and services (LTSS); and (3) initiatives to provide permanent supportive housing for homeless persons. After describing the ways that states currently are permitted to use federal Medicaid funds to pay for housing, the article interrogates the simultaneous permissibility of states' paying for room and board for Medicaid recipients residing in nursing homes and impermissibility of paying rent for other Medicaid recipients needing supportive housing. Using the concept of bundled payments, I argue that justice requires treating these populations similarly.
Medicaid's Existing Funding for Housing
The Nursing Home Benefit
Medicaid, the federal-state program providing health coverage for low-income persons, is the single largest payer for nursing home care in the United States. In 2015 it spent nearly $55 billion on nursing facilities. 6 Unlike Medicare, which covers nursing home care only for a limited time following a patient's qualifying hospital stay, 7 Medicaid can cover extended and indefinite stays in nursing homes for adult recipients. This difference has historical roots in federal-state cooperative programs that paid nursing homes to provide care to poor and “medically needy” elderly persons and that predated the legislative birth in 1965 of the unlikely twins, Medicare and Medicaid. 8 More than a half century later, Medicaid remains the primary vehicle for public support of nursing home care.
While permitting some variation in what state Medicaid programs cover as part of their nursing home benefit, the federal government requires states to cover nursing and related services, room and board, specialized rehabilitative services, medically related social services, and a range of other services. 9 By contrast, states can charge Medicaid beneficiaries for “extras” that do not relate directly to the recipient's health. Examples include items such as private rooms, specially prepared food, cosmetic and grooming items, personal clothing, and televisions. 10 Within these parameters, states can develop their own rules about what their programs will actually cover.
Of the 1.4 million people residing in nursing homes in 2015, sixty-two percent relied on Medicaid to pay for their care. 11 Although Medicaid is a means-tested program, with eligibility limited to persons with low incomes, many Medicaid recipients residing in nursing homes may have come from well-heeled backgrounds. Most states permit people to “spend down” their income and assets to the point where they meet Medicaid's financial eligibility criteria. 12 Because of this pathway to eligibility, Medicaid is sometimes referred to as a long-term care insurance policy where the deductible is your life savings. Moreover, the ability to spend down assets to qualify for Medicaid provides a long-term care safety net for persons of any socioeconomic class whose “medical costs overwhelm their income.” 13 This safety net gives more affluent persons a personal interest in Medicaid and its coverage of nursing home care. 14
Coverage of Home and Community-Based Services
Medicaid's coverage of nursing home care, however, is just one aspect of its coverage of a broader category of services referred to as long-term services and supports (LTSS), 15 which includes both institutional care and care provided in the community. While Medicaid's absolute spending on nursing facilities has grown steadily, more Medicaid beneficiaries are receiving long-term care services in their homes or community-based settings in the past few decades. In 2011, 80% of non-elderly Medicaid recipients with disabilities who were receiving some kind of LTSS lived in the community, as did 50% of elderly recipients with disabilities. 16 This increase has resulted from the federal government's creation of a proliferation of Medicaid waiver authorities and funding streams — aptly described as a “bewildering array of programs” 17 — to encourage home and community-based services (HCBS). As of 2013 at least nine different paths existed to permit state Medicaid programs to cover an enhanced set of services — including personal-attendant care, specialized therapies, and care coordination — to enable people with disabilities to live in the community. 18 These programs have eroded Medicaid's historical bias toward institutional care. That bias arose partly from the Medicaid statute itself, which makes mandatory states' coverage of nursing home services, while treating as optional most services that fall under the umbrella of HCBS. 19
Additional stimulus for states to rebalance their spending on LTSS towards community-based care came from the Supreme Court's decision in Olmstead v. L.C. ex rel. Zimring. 20 Interpreting Title II of the Americans with Disabilities Act (ADA) as prohibiting unjustified institutionalization of people with disabilities, that case established a community integration mandate for state-funded services, like Medicaid. Olmstead signaled to states that a failure to provide services to disabled persons in the community could violate the ADA. As a result, states increasingly have had both the financial means (Medicaid funding) and the legal motive (Olmstead) to provide services and supports to their disabled citizens in the community.
Shifting Medicaid spending away from nursing home care and towards HCBS, however, has laid bare the question of what exactly Medicaid will (and should) pay for. As part of various HCBS programs, states can use federal Medicaid dollars for a wide range of services including personal attendant services, habilitation services, transportation support, case management, and respite services. Recognizing that stable housing is critical to the effectiveness of LTSS delivered in community settings, in 2015 the Centers for Medicare and Medicaid Services (CMS) issued an Informational Bulletin clarifying the circumstances in which Medicaid pays for housing-related activities and services, including services to help recipients transition from an institutional setting to housing in the community and, once there, to succeed as a tenant. 21
By contrast, Medicaid will not pay rent for persons receiving Medicaid-funded HCBS. Instead, disabled Medicaid recipients seeking community placements must rely on other programs to help them find affordable housing. Local governments and housing agencies may disburse a mix of local, state, and federal (non-Medicaid) dollars to housing providers; tax credits for developers may spur the building of low-income units; and local non-profits often support housing initiatives. 22
Despite states' ability to use Medicaid funding for HCBS, the demand for community-based living services often outstrips the availability of funded services. Lengthy waiting lists for those services exist in many states. 23 Moreover, political, financial, and logistical factors impede continued progress in advancing the community integration of Medicaid recipients with disabilities. Recent proposals to fundamentally restructure Medicaid and slash its federal funding pose a serious threat to states' support of HCBS, 24 and the nursing home lobby has a vested interest in preventing further shifting of Medicaid spending away from institutional care. On a personal level, however, the immediate challenges for a Medicaid recipient seeking to transition out of (or to avoid going into) a nursing home include finding a place to live that is both affordable and accessible. Without affordable housing or financial assistance, many Medicaid recipients have little choice but to live in institutions in order to receive the services they need. Thus, it is partly for want of resources to pay for housing that many disabled Medicaid recipients must endure life in the isolating and segregating environment of a nursing home, where Medicaid will pay for their shelter needs. 25
Supportive Housing for Homeless Persons
The third scenario in which state Medicaid programs may provide funding to help shelter low-income persons is supportive housing for homeless persons. The mechanisms by which state Medicaid programs can provide that support (e.g., reimbursing housing transition services like assisting with the housing search and application processes and identifying resources to cover security deposits or moving costs and tenancy-sustaining services like coaching on meeting lease obligations and linking tenants with community resources to avoid eviction, but not rent) are similar to those discussed in the previous section, 26 but the aims are different. Funding housing-related services in conjunction with Medicaid-funded HCBS seeks to move (or keep) disabled people out of institutions in order to advance their dignity and autonomy and avoid their isolation and segregation. By contrast, funding housing-related services in conjunction with supportive housing for homeless persons seeks to move people off the streets and into stable housing in order to improve their health outcomes — and perhaps decrease spending.
For some time, states and local governments have run programs aimed at providing permanent supportive housing to persons who are chronically homeless. Many of these initiatives align with principles of the “Housing First” model, which operates from the premise that stable housing is a precondition to persons' ability to benefit fully from medical, mental health, or addiction services. 27 Thus, rather than placing homeless persons in temporary or transitional housing until those persons address their mental illness or substance addiction, these programs supply permanent housing first. 28 That housing is accompanied by a full range of social and medical supports to address other needs of the homeless person, but those services are not a condition of receiving housing.
More recently, supportive housing programs have begun to partner with state Medicaid agencies. 29 Many homeless persons are covered by or eligible for Medicaid, especially in the states that have implemented the Affordable Care Act's Medicaid expansion. From the perspective of housing agencies, tapping into Medicaid to pay for physical and mental health services and housing-related services frees up housing funds to pay for core housing expenses like rent subsidies and developing low-income housing projects. From the perspective of state Medicaid agencies, devoting resources to support housing for homeless Medicaid recipients reflects a belief that housing homeless persons improves their health outcomes and potentially decreases their health spending.
That belief is based on research suggesting that supportive housing for chronically homeless persons may reduce state spending — including health care spending — on them. 30 For example, researchers who assessed the impact of supportive housing for one small cohort of highly vulnerable homeless persons in Oregon found that per capita spending on Medicaid recipients in the group declined by $8,724 in the first year following move-in. 31 The researchers found that major contributors to the reduced spending were lower costs for inpatient hospital care and emergency department usage. Participants' usage of primary care and outpatient behavioral health care, by contrast increased. 32 Other studies have produced similar results, 33 with some examining a broader range of potential cost savings flowing from investments in supportive housing, including reduced spending for emergency shelters and prisons. 34
While encouraging, the evidence regarding permanent supportive housing's production of cost savings is incomplete, 35 and some researchers have cautioned that higher-quality trials of Housing First interventions have not shown net cost savings. 36 Not surprisingly, the return on investment in supportive housing appears to vary with the intensity of participants' support needs. For example, reductions in spending may be greater when persons who regularly require inpatient psychiatric care or frequently use emergency department services are moved into permanent supportive housing. 37
We need more and better research to assess whether public investments in Housing First interventions result in savings in health care spending and other budgets. Existing evidence, however, indicates that supportive housing helps decrease unmet health needs and improve health outcomes for homeless persons, 38 reinforcing housing's importance as a determinant of health. Since federal policy permits states to use Medicaid funding to pay for housing-related activities (though not rent) integral to Housing First programs, Medicaid has already assumed a role in addressing a social determinant of health. The balance of this article considers challenges to Medicaid programs' current ability to address homelessness effectively and suggests an argument for permitting the use of Medicaid funds for direct housing costs.
We need more and better research to assess whether public investments in Housing First interventions result in savings in health care spending and other budgets. Existing evidence, however, indicates that supportive housing helps decrease unmet health needs and improve health outcomes for homeless persons, reinforcing housing's importance as a determinant of health.
Medicaid Partnerships and Supportive Housing
Because of prohibitions on using federal Medicaid dollars for rent or housing capital costs, state Medicaid programs seeking to address the health needs of homeless recipients must partner with other public or private entities who can supply funding for those costs. The result is an array of promising supportive housing collaborations taking various forms. But these collaborations face several challenges to their ability to provide stable housing for homeless persons. These challenges cluster around a shortage of affordable housing in many communities, 39 difficulties in securing stable funding streams to assist residents in paying rent and other expenses, and barriers to effective coordination. 40
Much of the financial support for Housing First projects comes by way of grants from the federal Department of Housing and Urban Development (HUD), often supplemented by funding from state or local governments. Growing need for supportive housing, however, has outpaced budgeted funds. 41 Indeed, federal budget sequestration requirements imposed in 2011 led to spending cuts at HUD, and Trump administration proposals would further cut funding for housing assistance. 42 As a consequence, government housing programs may be unreliable and inadequate partners for state Medicaid directors seeking to participate in Housing First initiatives involving Medicaid recipients. Instead, private partners — typically philanthropies or health-related non-profits — have in many instances stepped in to help create or pay for housing in many projects. 43 Beyond issues of funding scarcity, constraints imposed by different funding sources may make it difficult for collaborating agencies to “braid” funding streams, and legal constraints regarding data privacy complicate efforts to coordinate data collection and program evaluation. 44
Of course, state Medicaid programs also face internal funding challenges to devoting resources to supportive housing for homeless Medicaid recipients. With Medicaid spending overall mushrooming, competing demands for Medicaid funds are multiple and pressing. As a political matter, deciding to allocate funds to housing-related activities for homeless persons requires either additional taxpayer support, reducing funds for other optional benefits (e.g., physical therapy, dental services, or hospice care), or persuasive evidence that investments in supportive housing are cost effective.
Should Medicaid Pay Rent for Homeless Recipients?
Even in a budget-constrained environment, one can argue for expanding Medicaid's support for housing homeless persons. Once we acknowledge the legitimacy of using Medicaid funds to address social determinants that negatively affect the health of homeless recipients, the logic behind expansion is straightforward. Why not loosen existing limits on the use of Medicaid funds to permit direct payments of rent, at least in some cases?
One objection to using Medicaid funds to pay rent for homeless Medicaid recipients arises from legitimate concerns with “mission creep.” Paying rent for homeless persons (or providing capital to build supportive housing) arguably falls outside of Medicaid's statutory purpose, which is to provide “medical assistance” for low-income Americans. From this perspective, if Medicaid is permitted to cover payments for housing, what wouldn't Medicaid potentially cover? Numerous social and environmental factors affect poor persons' health. Should Medicaid thus pay for nutritious food, improved schools, neighborhood safety, and job training? An objection based on statutory purpose has some force, but does not seem conclusive. Recent Medicaid waivers approved by the current administration construe the statute's purpose broadly. Paying for housing for homeless Medicaid recipients so that they can receive effective physical and mental health services seems more closely connected to Medicaid's statutory purpose than does requiring recipients to work or risk losing Medicaid eligibility. 45
Moreover, providing housing as a means of meeting the health needs of homeless persons is no further removed from Medicaid's purpose than is providing for room and board for nursing home patients. In the case of nursing homes, Medicaid pays nursing homes to provide a bundle of required services, including nursing services, food and nutrition services (including three meals per day), an activities program, personal hygiene items, along with a room and maintenance of the facility. 46 For Medicaid recipients who, because of physical or cognitive impairments or frailty, are unable to live independently, Medicaid pays for a bundle of services permitting them to live in a setting where their health needs can be met. In short, Medicaid pays for the bundled nursing home benefit (which includes non-medical services and items, including housing) so that recipients needing that level of care are situated where they can receive effective health care.
Recognizing Medicaid's nursing home coverage as a bundled benefit suggests the logic of similarly conceiving of Medicaid's payment for supportive housing as a bundled benefit. Persons with chronic mental illness or substance use disorders may need an array of services (including non-medical services and housing) so that they are situated where they can receive effective health care. While homeless persons may be physically capable of performing activities of daily living, in many cases mental illness or addiction renders them unable to achieve stable housing without supports. Research indicates that housing chronically homeless persons improves health outcomes and may reduce public spending on those persons. 47
Thus, having Medicaid pay housing costs when needed for homeless recipients to receive effective health care makes sense. Admittedly, some Medicaid recipients who experience short-term homelessness as a result of financial or family dislocations may not fall into this category; Medicaid may be able to meet their health needs without providing them with housing. Thus, I do not argue that justice necessarily requires that Medicaid pay for housing all recipients who are homeless. But for persons for whom the evidence shows that providing housing supports improves health outcomes, the bundled benefit argument is strong. Public health insurance programs have considered the use of bundled payments to promote coordinated care and achieve cost savings for interventions ranging from hip replacement to cardiac surgery; why shouldn't the concept apply to interventions for persons with chronic mental illness and substance use disorders?
Indeed, denying bundled coverage (including housing) for Medicaid recipients with mental impairments (like mental illness or substance use disorders) while providing bundled coverage (including housing) for recipients with physical and cognitive impairments (like frailty or dementia) represents a serious injustice. The continuing stigma attached to mental illness and addiction, along with the historical origins of the nursing home benefit, explain why this injustice persists. But once recognized, the injustice should not be accepted. If the federal government maintains that Medicaid dollars cannot pay housing costs associated with providing effective health care to persons with mental illness and substance use disorders, then neither should Medicaid pay housing costs associated with providing effective health care to persons with physical or cognitive impairments who reside in nursing homes. In other words, if the federal government refuses to bundle coverage to include housing costs in the former case, justice arguably requires it to unbundle coverage in the latter case. Unbundling coverage of nursing home care and shifting financial responsibility for the housing component either to the resident or another government program would be politically unpalatable, however. Politicians are unlikely to take on powerful voting blocs like senior citizens and affluent whites, many of whom have come to depend on the availability of Medicaid coverage as a backstop in the event they need an extended period of nursing home care.
Commitments to humaneness and cost-effectiveness in public spending buttress the justice-based argument for bundling housing costs into the coverage provided to homeless Medicaid recipients for whom housing is medically necessary to achieve improved health. We already devote substantial public and private funding to paying to shelter persons who are unable to maintain stable housing. Most obvious is spending on temporary or emergency shelters for homeless persons, but the costs of emergency department services and inpatient hospitalization for repeated acute health episodes suffered by chronically homeless persons are large. 48 Moreover, as a growing number of localities pass measures criminalizing homelessness, the costs of jailing homeless persons must also be considered. 49 Current funding streams exist in silos, producing a range of vested interests in the status quo, so implementing a health-oriented, holistic approach to spending on homeless persons' shelter needs will not be easy. But both justice and Medicaid's purpose of improving health access for poor persons demand that we take up the challenge.
Footnotes
The author has no conflicts to disclose.