Is Everything Health Care? The Overblown Social Determinants of Health

The Biden administration’s recently published The U.S. Playbook to Address Social Determinants of Health cites extensive research to argue that increasing public spending on housing, education, food, and neighborhood environments will often pay for itself by reducing health care costs. This justification has been used to expand the ability of states, insurers, and nonprofit organizations to claim Medicaid and Medicare funds for providing nonmedical services.

These nonmedical factors are called social determinants of health (SDOH), but the bulk of SDOH research is substandard and does not appropriately disentangle causation from correlation. Much of it fails to include a control group in assessing the effectiveness of nonmedical interventions, nor does it consider their cost. The best-designed experiments with randomized controlled trials—the gold standard of social science research—typically find that SDOH expenditures have weak effects on health and few offsetting savings.

SDOH justifications have become so popular because they allow states, nonprofit groups, and other social policy advocates to tap into the much larger pool of federal funding that is allocated to health care. But health care costs will not be greatly reduced by increasing spending on social services; instead, it will tend to inflate expenses. Nor should social policies be structured according to their incidental impacts on health, as Americans who most need public assistance are often those whose health has the weakest prospects of recovery.

In November 2023, the Biden White House released the 53-page U.S. Playbook to Address Social Determinants of Health and declared: “Improving health and well-being across America requires addressing the social circumstances and related environmental hazards and exposures that improve health outcomes.” It accepted the Department of Health and Human Services’ definition of SDOH as “the conditions in the environments where people are born, live, learn, work, play, worship, and age that affect a wide range of health, functioning, and quality-of-life outcomes and risks.”[1]

The same month as the Playbook’s release, federal regulations revising Medicare payments to physicians declared that “around 50 percent of an individual’s health is directly related to SDOH.”[2] Others have suggested that 40% to 90% of health outcomes are attributable to social, behavioral, or economic factors.[3]

These claims are based on some stark facts. Low-income Americans have higher rates of disability, anxiety, heart disease, stroke, diabetes, and other chronic conditions, and they are more subject to obesity, substance abuse, physical strain, and environmental pollutants.[4] From 2001 to 2014, the life expectancy of the richest 1% of Americans averaged 15 years longer than that of the poorest 1%.[5] Furthermore, poorer social classes have worse health outcomes even when they receive the same access to medical care.[6]

Decades ago, sociologists Bruce G. Link and Jo Phelan argued that socioeconomic status and social supports are the “fundamental causes of disease” because they “embody access to important resources, affect multiple disease outcomes through multiple mechanisms, and consequently maintain an association with disease even when intervening mechanisms change.”[7]

The World Health Organization embraced this assessment, declaring that “based on their respective social status, individuals experience differences in exposure and vulnerability to health-compromising conditions.”[8] In 2008, it argued that “to address health inequities, and inequitable conditions of daily living, it is necessary to address inequities . . . in the way society is organized”—namely, by tackling “the Inequitable Distribution of Power, Money, and Resources.”[9]

Progressive scholars have long argued that increasing public spending on social services could prevent and reduce health care costs. Whereas U.S. spending on health care well exceeded the level in other developed countries (16% vs. 11% of GDP, respectively), its spending on social services was less (16% vs. 20%).[10]

The Biden administration’s Playbook laments that “funds associated with ‘health’ are too often walled off from investments in improving SDOH.”[11] Other SDOH advocates cite a “wrong pockets problem,” in which one institution makes investments in improving a social condition but the benefits accrue to another. Returns on investment may be split between public and private sectors, or between public agencies.[12] For instance, infrastructure spending by states could yield savings to the federal government in reduced Medicare costs.

Scholars have conjectured a vast array of potential social determinants of health, ranging across the entirety of domestic policy responsibilities (see the text box).

Spending on health care benefits dwarfs all other social welfare activities in the U.S., and classifying activities as health care greatly increases the federal funding available to such programs. While the federal government spent $616 billion on Medicaid in 2023, it provided only $22 billion to Temporary Assistance for Needy Families (TANF), a federally funded, state-run welfare program.[13] Advocates for social policy programs are therefore usually eager to tap into the largest pot of available resources by highlighting any associated health impacts.

For example, although the Department of Housing and Urban Development may be a more appropriate agency to tackle issues around homelessness, according to a senior policy director for the National Health Care for the Homeless Council, “more and more states are desperate to find any help, and that’s why they’re turning to Medicaid because they’re not getting real assistance from HUD.”[14]

As a result of such actions, the federal government has added various forms of funding for SDOH to its major health care programs over recent years.

Medicaid

Medicaid’s matching fund structure (which provides up to $9 in federal funds for every $1 that states spend on the program, without limit) offers states the most lucrative terms of federal support for their expenditures. Wherever possible, states have reclassified expenditures, from law enforcement to education, under Medicaid, to gain federal matching funds.[15]

Medicaid offers several opportunities for states to claim federal funding for SDOH:

• Managed Care Benefits in Lieu of Services (ILOS)[16] 
  
  States can claim federal Medicaid funds for nonmedical benefits by requiring Managed Care Organizations (MCOs) to provide additional medically appropriate social services, so long as costs do not exceed 5% of managed care expenses.[17] Funds have been used to finance meals, asthma remediation, sobering centers, post-hospitalization services, substance abuse treatment, caregiver respite, home modification, personal care, food, rent, and temporary housing.[18] Massachusetts adjusts payments to MCOs according to “neighborhood stress” and expects plans to assist with housing, substance abuse, and disability.[19]

• Section 1115 Demonstration Waivers[20] 
  
  By obtaining “demonstration”waivers from standard restrictions on the use of Medicaid funds, states can claim federal matching funds to pay for nonstandard benefits such as housing, aid to caregivers, home modification, utilities, or food. Expenditures must not increase net costs to the federal government or exceed predetermined caps, and spending on nonmedical benefits must not exceed 3% of total Medicaid spending. To qualify, states must exceed thresholds in paying providers of core medical services and maintain existing levels of other spending on social services.[21]

• Section 1915 Home and Community-Based Services (HCBS) Waivers[22] 
  
  States can claim HCBS waivers to allow them to claim Medicaid funds for nonmedical services to support people who need a nursing home level of care to remain living at home. They can be used to fund case management, post-hospitalization transition, rehousing expenses, caregiver respite, nonmedical transportation, home-delivered meals, medically necessary home modifications, and employment support for Medicaid beneficiaries with special needs. Federal funds under this section may not be used to pay for services of “general utility,” such as long-term “room and board” outside of specific institutions.[23]

• Section 1945 Health Homes[24] 
  
  States can claim federal funds for comprehensive care coordination, care transition supports, and referral to social services.

• State Plan Authority[25] 
  
  Without special permission through a waiver, states can claim federal funds for rehabilitative services for physically or mentally disabled beneficiaries. These may include case-management services to assist beneficiaries access medical, social, educational, and other social services.

• Program of All-Inclusive Care for the Elderly (PACE)[26] 
  
  PACE consolidates payments for Medicare and Medicaid benefits to cover long-term care services for dually eligible beneficiaries needing nursing facility care but who are able to live in the community. Funds can be used to finance care teams delivering comprehensive medical, long-term care, and social services, including meals, social work services, and transportation.

States can also use various Medicaid funding streams in concert for instance, by combining Section 1115 and Section 1915 waivers to get around restrictions particular to each. Although the use of Medicaid funds to pay rent has traditionally been prohibited, Arizona has done so since October 2022, while Oregon has recently required MCOs to provide it. Additionally, seven states now claim federal Medicaid funds for purposes of gun-violence prevention, using State Plan Authority or Section 1115 waivers, with violence prevention activities being billed as “mental health services.”[27]

CHIP Health Service Initiatives

The Children’s Health Insurance Program (CHIP) provides federal matching funds for states to provide health care benefits to children and pregnant women in low-income households. States can use up to 10% of CHIP funds to improve the health of those eligible for the program with public health services such as lead abatement, home visits to reduce asthma triggers, emergency food relief, and violence prevention.

Medicare

Medicare is a fully federally financed and administered health care benefit for elderly and disabled Americans. Medicare can directly pay only for services that are “reasonable and necessary” to the diagnosis or treatment of illness or injury, or to improve bodily functioning.[28] But indirect payments for SDOH activities have gradually been established.

• Basic Fee-for-Service Payments
  
  Payment for SDOH risk assessment is provided as an add-on to an annual wellness visit, hospital discharge, or behavioral health visit.[29] The Biden administration’s FY 2024 budget proposes paying for nutrition counseling and community health workers to deliver screening, prevention, and care navigation services.[30]

• Value-Based Payment Innovations
  
  The CMS Innovation Center is authorized to develop alternative payment arrangements for Medicare services. It has established capitated (i.e., per patient), bundled, and shared savings payment arrangements, which can be used to fund SDOH expenditures if they reduce covered medical expenses.[31] Risk-based contracts with providers may include specific rewards for social needs assessments and other SDOH practices. The Accountable Health Communities Model provides social-need screenings, referrals to community services, and alignment of clinical and community resources for 28 geographic areas.[32]

• Medicare Advantage
  
  Medicare beneficiaries can opt to receive their Medicare benefits administered by private insurers through Medicare Advantage (MA). MA plans can offer supplemental benefits to pay for nonmedical SDOH expenditures.[33] The passage of the CHRONIC Care Act in 2018 broadened the definition of permitted “health-related” spending on supplemental benefits (to include payment for services that ameliorate chronic conditions or reduce avoidable health care utilization, such as food, transportation, pest control, counseling, HVAC, education, and home modification) and allowed plans to reserve these for subsets of high-cost beneficiaries.[34]

Hospital Tax-Exemption Conditions

Two-thirds of hospitals in the U.S. operate as nonprofit institutions. To maintain that tax-exempt status, hospitals must provide “community benefits,” and they can count “community building activities” (such as support for improvements to nutrition, housing, education, environment, and economy of the local area) as satisfying the requirements.[35] In 2019, U.S. hospitals spent $800 million on “community building activities.”[36]

Many studies identify correlations between household living circumstances and health. Some find that health improvements follow the delivery of social services, and in rare circumstances, this may even save money.

Housing is the most often cited beneficial impact of SDOH. Its advocates argue that good housing provides psychological stability, safety from toxins, and low neighborhood crime, and it does not strain family budgets in a way that leaves inadequate funds for food or medical care. They note that water leaks, dirty carpets, pest infections, and poor climate control systems are correlated with asthma and adverse cardiovascular events among the elderly.[37] The elderly who received supportive services to help them live at home saw a lower risk of falls, improvements in physical health, and lower rates of depression, compared with other seniors.[38] Families who were given housing vouchers to move to more affluent neighborhoods subsequently enjoyed lower obesity and diabetes levels.[39]

Studies offer evidence that improvements to ventilation, bedding, and dust mitigation reduced asthma symptoms, hospitalizations, and days of school missed by children—with reductions in medical costs exceeding the costs of interventions, in some cases.[40] Advocates advanced similar evidence for the cost-effectiveness of removing lead paint from homes of children from low-income families.[41]

Homelessness is correlated with higher rates of hospitalization, HIV infection, substance abuse, untreated mental illness, and tuberculosis.[42] Homeless HIV/AIDS patients who were given housing had higher medication adherence and survival rates than other studies found for similar populations.[43] Policy experiments providing housing to alcoholics and the mentally ill have been found to reduce the use of hospital services, rehab services, psychiatric hospitals, and incarceration.[44] In some cases, it has been suggested that reductions in health care costs exceeded the cost of housing the most medically at-risk homeless.[45]

Nutrition is also commonly cited as an SDOH. Insecure access to food is correlated with chronic illness, mortality, hospitalizations, obesity, anxiety, depression, and child behavior problems.[46] Most Americans’ diets lack sufficient fruits and vegetables, while three-quarters fail to get the recommended level of exercise.[47] Schools have been able to improve this through concerted interventions, while most recipients of nutrition assistance under the Older Americans Act “indicated the program had improved their health.”[48]

The receipt of food stamp benefits is correlated with lower rates of rehospitalization among Medicaid beneficiaries, while prenatal participation in the Women, Infants, and Children (WIC) nutrition aid program is associated with higher birth weights and reduced fetal death.[49] Recently hospitalized patients suffering diabetes, heart failure, or chronic obstructive pulmonary disease (COPD) who were provided home-delivered meals subsequently saw a significant reduction in their medical expenses.[50] Medicaid beneficiaries receiving home-delivered medically tailored meals saw better health outcomes, fewer hospitalizations, and reduced medical costs relative to similar enrollees—possibly exceeding the cost of food assistance.[51]

Transportation is also widely identified as an SDOH. The Biden administration argues that providing free transportation to scheduled medical appointments “has been shown to be cost-effective by increasing preventive and outpatient care and decreasing more expensive care.”[52] A Federal Transit Administration report claims that providing transportation saves on prenatal care, asthma, congestive heart failure, and diabetes.[53] The administration has also argued that aid for public transportation further improves health by reducing road traffic accidents and encouraging people to walk more.[54]

The structure of neighborhoods more broadly affects health through levels of pollution, access to amenities, and the degree to which the community supports or threatens its residents.[55] A federal review of “evidence-based strategies” concluded that “health and wellness are shaped by and within overarching systems, including structural racism, ableism, homophobia, and transphobia; and broad neighborhood and community structures.”[56] Public-health agencies have attempted to reduce crime by providing anger-management counseling, domestic-violence screening, drug-abuse treatment, and alternative activities for gang members.[57]

The White House’s SDOH Playbook also notes that levels of educational attainment are “associated with health outcomes.”[58] Early childhood education may improve health “via access to health screenings, health care, improved nutrition, or other health-promoting activities.”[59] In particular, studies have suggested that pre-K education of disadvantaged children reduced obesity, blood pressure, and metabolic syndrome in adulthood.[60]

Money might suffice to avert some health problems. Low-income disabled families receiving randomized cash payments saw better physical health and emotional well-being.[61] Health improvements were similarly registered for recipients of the Low-Income Home Energy Assistance Program, parents receiving the Earned Income Tax Credit, and seniors receiving Supplemental Security Income.[62]

These returns on preventive investments may be even greater if they are targeted toward “super-utilizers” of medical services.[63] Several studies have concluded that substantial reductions in hospitalization and overall savings may be generated by providing case management, housing support, and substance-abuse treatment to the seriously mentally ill and homeless.[64] In 2010, the 14% of Medicare beneficiaries with six or more chronic medical conditions accounted for 46% of the program’s costs—at an average of $32,658 that year.[65] Overall, 1.7% of hospital emergency-department patients made 10 or more visits per year, accounting for 12.3% of total ED usage—a cohort of patients who mostly had substance-abuse and mental-illness problems.[66]

The provision of some social services may be more able to generate net savings in those cases where taxpayers are already on the hook for long-term care as well as medical expenses. Medicaid, for example, has long sought to encourage the provision of home and community-based services as a cheaper and more appealing substitute for nursing-home care.[67] Home visits by a team comprising nurses, occupational therapists, nurses, and handymen might help make homes more supportive of individuals’ living needs—improving their ability to undertake basic tasks (such as bathing or cooking), reducing depression, and averting hospitalization and nursing-home expenses.[68]

Shoddy Evidence

Despite the seeming benefits, however, the bulk of SDOH scholarship merely identifies correlation, without demonstrating causal effects by rigorously eliminating confounding factors.

The Biden White House’s SDOH Playbook boasts of an “evidence-based” attitude, buttressed by a catalog of citations from social science journals,[69] but SDOH research is highly entangled with advocacy, subject to widespread conflicts of interest, and often overtly politically motivated. Overblown claims are frequently recited uncritically, or exaggerated, and sweeping policy recommendations are justified by reference to the volume rather than the quality of scholarship.

According to the second sentence of “A Playbook for State Medicaid Agencies,” commissioned by a federally supported partnership between foundations, public-health agencies, health care systems, Medicaid managed-care organizations, and think tanks: “Research shows that . . . SDOHs . . . are responsible for 80 percent of health outcomes.”[70]

The first source cited in the Medicaid playbook for this claim is a study of county-level variation in health outcomes that attributes 20% of differences to “clinical care.” To reach this conclusion, the Medicaid playbook bundles together all other factors reported by the study (30% “health behavior,” like tobacco use, diet, and exercise; 40% “social and economic factors,” like education and employment; and 10% “physical environment,” like housing, transit, and air and water quality) as socially determined.[71]

In using this study, the Medicaid playbook does not ponder the enormous difference between “health outcomes” and “county-level variation in health outcomes.” It should be no surprise that clinical care accounts for relatively little of the variation in health outcomes between counties, since entitlements, insurance benefits, and medical procedures are largely standardized across counties, whereas population attributes are not. Nor does it consider interaction effects between the factors considered, the appropriateness of categorizing personal behavior as entirely socially determined, or the fact that the influence of genetics and other important factors were entirely neglected. Yet,“evidence-based” reports advocating greater attention to SDOHs rarely fail to lead with a reference to this or some other similarly misleading (and misinterpreted) statistic.

The SDOH concept lumps together access to home help services, exposure to pollution, vulnerability to crime, personal lifestyles, cultural pressures, and economic opportunities—factors that arise for radically different reasons. To imagine that these elements are each merely manifestations of underlying socioeconomic status is to neglect the complexity of causes associated with each.

In practice, it is hard to disentangle causal factors behind health outcomes. That income is correlated with health does not prove that low-income status is necessarily responsible for ill health. A Duke University assessment of SDOH research warned that “evidence often is generated through less rigorous study designs, which makes it harder to conclusively rule out confounding factors.”[72] For instance, smoking is much more common among poorer social groups and accounts for much of the disparity in health outcomes, while genetics and education similarly contribute to levels of income and health.[73]

The SDOH literature is dominated by sloppily designed observational studies that do not seriously attempt to disentangle causation from correlation or to control for obvious sources of bias. Few observational assessments of SDOH policies employ methods of causal inference needed to control for unobserved confounders (e.g., natural experiments, regression discontinuity designs, instrumental variables), which would be necessary for a study to be published in the leading peer-reviewed economics journals.

Study results are published even though they lack statistical significance; often, they are cited even though effects are weak.[74] Most published studies focus on process measures (such as the number of patients screened) rather than outcomes (such as health, utilization, and cost).[75] This is important because an intervention that has a significant impact on birth weight, for instance, may have little measurable impact on health, such as pediatric hospitalizations.[76]

When studies assess interventions by analyzing a variety of different dependent variables, the administration’s 2022 review of the literature cites significant results but fails to mention numerous associated null findings.[77] A more independent review noted, quite reasonably, that “positive effects documented in the published literature should be considered an upper bound on our understanding of positive impact.”[78]

Furthermore, widely cited SDOH policy experiments typically fail to randomly assign treatments and frequently do not even include appropriate control groups. Many report only effects on health, without attempting to quantify costs. A systematic review of SDOH housing demonstrations in the Journal of the American Medical Association, for example, warned that even randomized clinical trials “had a high risk of bias owing to nonblinding of participants and personnel.”[79]

Weak Effects

Despite the claims by advocates of using medical resources for nonmedical interventions, SDOH expenditures typically have weak effects on health.

A correlation between social circumstances and health outcomes does not mean that changes in social circumstances would necessarily alter health. Where socioeconomic disadvantages are large enough that they substantially affect health, they are seldom cheap to remedy. Sherry Glied of New York University notes that “because most housing-quality-related health problems have multiple causes, intervening in housing alone will not fully eliminate these problems.”[80] Even though numerous experiments have provided housing to the homeless, there is no substantial evidence of resulting improvements in health.[81]

Nor does a broader increase in income, by itself, greatly affect health. A recent study of a natural experiment, by which infants born to pregnant mothers received higher cash payments from the federal government during the Covid-19 pandemic, found that the infants did not enjoy significantly or meaningfully better health.[82]

A review of the impact of social policy experiments on health concluded that “improvements in social circumstances produced by the intervention must occur early in life or be of sufficient intensity or duration to overcome the effect of adverse economic circumstances accumulated over the life course.”[83] This is a standard that few SDOH interventions meet. The evidence for health benefits to babies from food aid targeted at low-income pregnant women appears to be robust,[84] but a study of nutrition counseling found that it was less effective at reducing obesity among kids from disadvantaged backgrounds.[85]

“Upstream” public spending to improve the prosperity and welfare of disadvantaged communities is likely to have a largely incidental impact on health and to be less effective than “downstream” expenditures that directly address medical needs. Early 20th-century improvements in sanitation and sewage had a substantial health impact, but these were undertaken primarily for the purpose of improving health. Few nonmedical interventions offer such a bang for the buck or are as focused on yielding improvements in health. Targeted programs to remove lead from housing might be a rare exception.[86]

Expenditures on filling basic human needs are also likely to be subject to diminishing returns. At the level of severe deprivation, such as during a famine, increasing the availability of food will obviously yield major benefits to health, but in the U.S., the most substantial health problems are associated with an excess, rather than a deficiency, of food.

A study of county-level variation in the implementation of the food stamp program estimated that only 16% of the value of food stamps served to increase the consumption of food, while a randomized trial of the federal school breakfast program found that it increased obesity.[87] Although the complete absence of fresh foods can lead to scurvy, costly interventions to increase the supply of rapidly perishable foods to neighborhoods (where demand for them is low) are likely to have negligible effects on health. In general, significant health effects of modest changes to nutrition are hard to identify.[88]

Many studies of SDOH interventions are satisfied by merely demonstrating that they improved health without inquiring into the cost or cost-effectiveness of doing so.[89] Others have found that while some policy experiments were associated with declining use of services from hospital emergency departments, they did not significantly reduce overall health care utilization or spending levels.[90] 

Slim Savings

Everyone wants a free lunch, but few significant expansions of social services pay for themselves.

Even when savings may be demonstrated in certain cases and circumstances, that does not mean that they would have similar effects more broadly. While providing some social services to the cohort of seriously ill patients with the greatest likelihood of rehospitalization might be able to reduce medical costs, housing all the homeless would clearly cost far more than it would save. If broadly provided, such aid would also end up paying for people who are already housed. Similarly, although it might be cheaper for Medicaid to provide home- and community-based services than nursing home care to individual beneficiaries, doing so might increase expenditures overall due to its appeal to those who previously would have opted against entering a nursing home.[91]

The degree to which SDOH interventions are cost-effective therefore depends on how well targeted they are. This is often overlooked by SDOH advocates, who cite the cost-effectiveness of narrowly targeted interventions to justify their broad expansion to cohorts for which similar effects are entirely implausible.

Studies often boast of reduced medical spending after SDOH interventions, but health care spending levels among a group of high utilizers (such as those recently hospitalized) selected for treatment should be expected to decline over time without interventions due to the regression of utilization levels to the broader population mean.[92] Remarkably few demonstrations include an appropriate randomly selected control group to address this statistical challenge.

Some studies have inflated savings estimates by counting reduced expenditures that accrue to private insurers.[93] Others have counted reduced medical expenditures in terms of hospital “charges” or list prices, which are more than three times the costs actually incurred.[94] Some experiments have found that SDOH interventions shifted care between inpatient and outpatient settings without reducing overall costs.[95]

Experiments that have demonstrated offsetting savings have sometimes bundled medical interventions (i.e., case management) and nonmedical ones (e.g., housing), such that their relative effectiveness in reducing the use of medical services is unclear.[96] One study demonstrated the value of housing in improving medication adherence among homeless HIV/AIDS patients by comparison with nonintervention without considering alternative methods of promoting medication adherence that may be more cost-effective.[97]

The potential savings to be had from targeting “super-utilizers” of medical care is also often overestimated. The fact that 50% of Medicaid spending is concentrated among 5% of the population does not mean that the 5% are easily identified ex ante.[98] Information on medical diagnoses and expenditures from the previous year explains only 17% of the variation in health care spending.[99] 

The best-designed studies have found negligible savings from the costliest SDOH interventions.[100] MIT economists who randomly assigned “super-utilizer” patients with medically and socially complex needs to a highly touted care-transition program of social workers found that this did not significantly reduce their rate of hospital readmissions relative to a control group.[101] A RAND Corporation assessment of a more substantial intervention, which provided $30,540 in subsidized housing, meals, transport, and case management to a group of high utilizers, found that their medical utilization and expenditures decreased by less ($21,418) than those of a comparison group ($25,273).[102]

More modest interventions might sometimes be more cost-effective, but this tends to be because they depend more on behavioral modifications than expenditures—and intended beneficiaries might be resistant to these types of interventions. This is well recognized in the case of diet and exercise, but this is also true when it comes to improving housing quality by cleaning to remove dust and mold. Similarly, although teaching kids about healthy eating may cost little, nutrition has long been part of the K–12 curriculum, to limited effect. The challenge of remedying problematic behavior is even greater when an externality is involved, such as crime, cultures of substance abuse, and other actions with adverse effects on neighborhood quality.

Ultimately, social policies should not be pursued for the sake of their incidental impact on health.

Basic services are valuable for their own sake. Health outcomes, in fact, represent only a small part of the value of improved housing to residents. As Sherry Glied notes: “Rats, inadequate heat, and mold substantially diminish people’s lives whether or not they cause direct harm to their health.”[103] 

There are many good reasons to want to remedy the shortage of affordable housing, but the problem is unlikely to be solved by framing it as a “health equity” challenge to be addressed by health insurers. It would place resources and authority where they cannot be leveraged to tackle the underlying problems and political challenges. For instance, as long as the supply of housing is fixed by restrictions on development, housing subsidies from the health care system would likely serve only to inflate the price of housing and shift shortages from one set of residents to another.[104] Nor does the establishment of public housing have a great track record of improving the safety of neighborhoods. Many policy interventions fail to achieve their primary objectives, and so are unlikely to yield substantial second-order benefits to health.

Although public aid might help some people move away from others whose behavior is toxic, this will likely only redistribute the burden to others unless crime and antisocial behavior are tackled at the source. Whatever a “public health approach” might do to help reduce violent crime is at best, auxiliary to the need for law enforcement to catch, remove, or deter perpetrators.[105]

Social services are best provided by social work agencies, not doctors or hospitals. Hospitals interact with most patients only sporadically and are unlikely to have a good understanding of their specific nonmedical needs.[106] Community residents with the greatest unmet social needs are likely to be those most disconnected from the health-care system and least likely to have well-established primary-care relationships.

Because hospitals and healthcare systems are “bigger, stronger, and politically well-connected,” they will tend to dominate partnerships with social workers and distort the effective provision of social services.[107] Hospitals in poor neighborhoods, where most unmet social needs exist, often struggle to finance medical care. If policymakers decide to place the responsibility for addressing SDOHs on these hospitals, it will likely further strain the funds needed to care for the uninsured.[108] 

Shoehorning social work into medical practice burdens clinicians with extraneous tasks for which they are poorly suited and leaves less time for them to treat patients.[109] The use of Annual Wellness Visits for SDOH assessments, for example, turns physicians into data-gatherers for a poorly designed social survey in which information is skewed by entanglement with provider payment incentives.[110] It is not clear why Medicare is paying physicians $19 per patient to collect data on the prevalence of transportation security when the Centers for Medicare and Medicaid Services has little capacity to improve safety on the New York City subway or any other mode of transportation.[111] Nor are diagnoses of, for example, “excessive transportation time” particularly informative.[112]

Giving physicians authority to write “prescriptions” for subsidized housing places them in the awkward position as gatekeepers for substantial resources. It is enough of a challenge to deliver medical care cost-effectively; medical providers do not want to be made responsible for solving all other social ills.

The medical profession gains its authority from the ability of patients to trust that physicians’ advice serves only the objective good of their health, with advice rooted in medical science. Although patients might be willing to accept physicians being judgmental about their diet, exercise, and hygiene practices, making medical practitioners responsible for remedying social determinants of ill health necessarily pushes them into more ambiguous and political contentious territory. If the public views medical advice as a dubious justification for the allocation of scarce resources and questionable stances in cultural disputes, the perceived legitimacy of the medical profession is likely to suffer.

Much SDOH scholarship appears to be a methodologically sloppy attempt to claim the objective superiority of lifestyle preferences characteristic of urban academic liberals. Farmers’ markets, mass transit, and bike paths each have their charms, but there is little scientific basis for the SDOH literature’s focus on them at the exclusion of traditionally red-state pursuits such as hunting, fishing, or golf. Policies to promote marriage, control recreational drugs, and prosecute low-level, “broken window” crimes might each substantially improve the health of communities, but references to them are nowhere to be found in the SDOH literature.

SDOH studies also often neglect second-order effects of public policies. An insistence on improving the quality of basic goods may come at the cost of reduced access. For instance, the elimination of poor-quality, single-room-occupancy housing has led to an increase in the overall homelessness rate.[113]

Although foreclosures and evictions are associated with adverse health outcomes in the short run, abolishing the ability of private owners to profit from investment in housing would inflict much greater harm in the long run.[114] Although individual beneficiaries might be better off if they are given the essentials of life for free, the burden of paying for such aid must be considered. A payout or benefit will do little to improve health outcomes if the only effect of public expenditures is to crowd out private payment for the same services.

Viewing health as socially determined downplays personal agency, and although positive externalities exist, health is a relatively internalized good. The pain of a heart attack, the benefit of heart medications, and the burden of adjusting one’s lifestyle to prevent future incidents accrue largely to individual patients.[115] Many supposed SDOHs, such as income, housing, food, or education, relate to goods that can be internalized. Insofar as lack of money is an obstacle to health, its greatest adverse impact is likely to be as an impediment to access purely medical services.

While medical services are often required to demonstrate clinical effectiveness in placebo-controlled trials, SDOH interventions serve loosely defined objectives such as promoting “food insecurity” or reducing “housing instability.”[116] Whereas the degree to which hospitals provide free or discounted medical care to the uninsured can be quantified, the open-endedness of SDOH concepts allows hospitals to claim tax credit for almost any spending classified as “community building activities.” The appropriateness and effectiveness of SDOH expenditures is therefore harder to ascertain.

Making hospital systems responsible for the broader health and social welfare of local neighborhoods entrenches their market power over those communities. It is difficult to reconcile such power with vigorous competition over the provision of narrowly defined medical services.

Social programs established through ad hoc waivers are also unlikely to allocate resources where they are most needed. The provision of services “based on individual assessments of need, rather than a one-size-fits-all approach,” as recommended by CMS, invites the arbitrary and politicized distribution of benefits.[117] As SDOH funding is often distributed through “nonprofit” partner organizations, the absence of accountability according to clearly measurable objectives affords enormous discretion for fraud, waste, and abuse.

The correlation between socioeconomic conditions and health outcomes is increasingly cited to justify expanded public spending on social services.

Of course, health and other human goods are intertwined. Any policy or development of circumstances that substantially improves living standards is likely to improve health, but this does not mean that health is greatly improved by increasing spending on social services. Nor does it mean that spending on social services should be guided by the associated impact on health. In fact, the spillover benefits to health from expenditures on social services are slim. Nor are the essential foundations of a prosperous society (“peace, easy taxes, and a tolerable administration of justice”) best secured by a single-minded focus on public health.[118]

Rather, the SDOH push reflects the great skew of social welfare expenditures toward health-care programs and the desire of other social policy interests to tap into them. In particular, the abundance of federal matching funds for Medicaid rewards states for spending on health-care services (or services redefined as health care) at the expense of other social policy services. This incentive for states to put a lower priority on nonmedical services can more completely be remedied by capping federal subsidies for Medicaid than it can by expanding aid to all other possible social welfare expenditures.

Rather than expanding health care programs to pay for more nonmedical services, policymakers should:

• Resist pressure to expand the capacity of states to claim federal Medicaid funding for purposes beyond the program’s core covered benefits.
• Not establish additional payments in Medicare for health care providers to address SDOH beyond incentives inherent in medically risk-adjusted payments to Medicare Advantage plans.
• Oppose efforts to make insurers, hospitals, or other medical providers responsible for addressing SDOH with tax or payment incentives.

Even if a social welfare program is a well-intentioned and wise idea, that does not make it health-care. Health care costs will not be greatly reduced by expanding the meaning of health-care to cover every social service; nor would doing so distribute nonmedical assistance to those who need it most.

Please see Endnotes in PDF

Photo: SDI Productions / E+ via Getty Images