Chapter 1 | Weaving a Safety Net: Medicaid’s Origins and Evolution

Role of the American Medical Association (AMA)

The AMA played a major role in opposing the establishment of Medicaid and Medicare, primarily due to concerns about federal involvement in healthcare, which they feared would lead to government control of medicine and undermine the autonomy of physicians.

Initially, the AMA’s opposition to federal health insurance began well before the introduction of Medicaid and Medicare. In the 1940s, the AMA spent a substantial amount of money on public relations campaigns to combat President Harry Truman’s proposal for a national health insurance plan. They effectively branded the proposal as a form of socialized medicine, even suggesting it was part of a Communist plot, which contributed to the defeat of Truman’s plan (https://www.semanticscholar.org/paper/1e2a132d1f0997e1b2ab8c3e4a7b88211e93a8e1).

By the 1960s, when Medicare and Medicaid were proposed, the AMA’s opposition continued. The organization was particularly influential in shaping public and political opinion against these programs. The AMA’s resistance was rooted in a fear that these federal programs would lead to increased government control over the conditions and reimbursements for medical services, which they believed would interfere with the practice of medicine and reduce the quality of care provided to patients (https://pubmed.ncbi.nlm.nih.gov/2187917/).

The AMA’s political activities included lobbying efforts and the mobilization of public opinion against the proposed legislation. They were a central force among the interest groups that aligned to oppose the broad, universal coverage initially envisioned for Medicare and Medicaid. Instead, they advocated for a more limited approach that would not disrupt the existing private and employer-based insurance systems (https://pubmed.ncbi.nlm.nih.gov/11617983/).

Fiscal multiplier

Research on the fiscal multiplier of Medicaid provides nuanced insights into its economic impact. The fiscal multiplier of federal Medicaid assistance to states has generally been found to be high during the Great Recession when the multiplier increased to 1.5, indicating a powerful fiscal stimulus to states during that period (https://dx.doi.org/10.24149/wp2112). Further studies have refined these findings, showing that the fiscal multiplier of Medicaid exceeds one when fiscal space is ample, a finding consistently significant across various identification methods and sample analyses (https://dx.doi.org/10.2139/ssrn.3710114).

The variability of the Medicaid fiscal multiplier has also been observed in relation to monetary policy, with values ranging to as large as 2 depending on the monetary offset. This range suggests significant sensitivity of the fiscal multiplier to monetary conditions, analyzed through a decomposition-based approach (https://dx.doi.org/10.24148/wp2020-12). During the COVID-19 pandemic, the fiscal multiplier of Medicaid was influenced by enhanced federal matching funds intended to cover state Medicaid cost increases. The impact of these funds varied among states, depending on their exposure to Medicaid program cost increases (https://dx.doi.org/10.1111/pbaf.12287).

Notably, the Medicaid expansion’s fiscal multiplier was estimated to increase federal spending by 10% in its first year, eventually reaching 27% by 2018. Meanwhile, changes in state funding due to Medicaid expansion were minimal and statistically non-significant, with annual fluctuations remaining under 1% (https://dx.doi.org/10.3386/w26862).

Using Treasury rate discount factors, the Congressional Budget Office has estimated the fiscal effect of continuous Medicaid eligibility was estimated to be a gain of $3,400 for every $1,700 spent, which is about 197% of the initial outlays (https://www.cbo.gov/system/files/2023-10/59231-Medicaid.pdf). Studies from states like Louisiana, Michigan, and Montana suggest that the economic impacts of Medicaid are significant enough to generate tax revenues that match or exceed the costs of Medicaid expansion (https://www.commonwealthfund.org/sites/default/files/2020-05/Ward_impact_Medicaid_expansion_state_budgets_ib_final.pdf). For example, Michigan’s Medicaid expansion is projected to generate sufficient savings and additional tax revenue to offset its costs through at least 2027–28. Additionally, the increase in the federal medical assistance percentage (FMAP) during economic downturns, such as the 2009 Recovery Act, effectively stimulated the economy. Each additional $100,000 of state fiscal relief from increased FMAP boosts employment by 3.8 job-years, suggesting a high fiscal multiplier effect where each dollar spent adds about two dollars to the GDP (https://www.cbpp.org/blog/increasing-federal-medicaid-assistance-provides-effective-economic-stimulus).

Krugman’s arguments on free markets in healthcare

Paul Krugman presents a compelling argument against relying solely on free market principles to address healthcare issues. He challenges the belief held by some Americans that the free market is the only solution to healthcare problems, asserting that this view contradicts both economic theory and overwhelming evidence. His arguments include:

1. Uncertainty and Unpredictability of Healthcare Needs

‍Krugman highlights that healthcare is distinct from other goods and services because individuals cannot predict when or if they will require care. However, when medical care is needed, it can be extremely expensive, such as triple coronary bypass surgery, which most people cannot afford out-of-pocket.‍

2. Asymmetric Information

‍There is a fundamental asymmetry of information in healthcare, where patients lack the expertise to evaluate the quality and appropriateness of the care they receive. This information gap creates opportunities for exploitation and market failure.‍

3. Adverse Selection

‍Krugman explains that in a free market for healthcare insurance, healthy individuals would opt out, leaving only those with higher risks in the insurance pool. This adverse selection would drive up premiums, making insurance unaffordable for many.‍

4. Moral Hazard

‍The presence of insurance can lead to overconsumption of healthcare services, as individuals have less incentive to restrain their use of care when they don’t bear the full cost. This moral hazard problem can drive up healthcare costs.‍

5. Lack of Successful Free Market Examples

‍Krugman asserts that there are no examples of successful healthcare systems based solely on free market principles. While there are various successful healthcare models, such as socialized medicine or single-payer systems, none rely entirely on market forces.‍

6. Economic Theory and Evidence

‍Krugman cites Kenneth Arrow’s influential paper, “Uncertainty and the Welfare Economics of Health Care,” which demonstrated that healthcare cannot be marketed like ordinary goods due to its unique characteristics. Krugman argues that those advocating fora free market solution to healthcare are disregarding both economic theory and overwhelming evidence.

Hypoglycemia at month’s end

A study by Hilary Seligman and colleagues at the University of California examined the relationship between exhaustion of monthly food budgets and risk of hospital admissions for hypoglycemia (low blood sugar) among people with diabetes. Using data on inpatient admissions in California from 2000-2008, the researchers found that admissions for hypoglycemia were more common in low-income populations compared to high-income populations (270 vs. 210 admissions per 1,000,000). Importantly, the risk of admission for hypoglycemia increased by 27% in the last week of the month relative to the first week for low-income populations, but no similar temporal variation was seen in high-income populations. The authors hypothesize that this pattern is driven by exhaustion of food budgets towards the end of the month in low-income households. These findings suggest that unstable access to adequate nutrition, resulting from running out of food money at the end of the month, may be an important contributor to health disparities (https://www.healthaffairs.org/doi/10.1377/hlthaff.2013.0096).

Chapter 2 | The Medicaid Landscape: Structure, Funding, and Eligibility

Federal matching rates

The Medicaid program’s design allows for the federal matching rate to vary based on a state’s per capita income, with poorer states receiving a higher match to support their Medicaid programs. This structure is meant to ensure that states with greater financial needs receive additional support to provide health coverage to low-income populations. The Affordable Care Act (ACA) of 2010 significantly expanded Medicaid eligibility, allowing states to offer coverage to adults underage 65 with incomes up to 138% of the federal poverty level (https://pubmed.ncbi.nlm.nih.gov/22479735). This expansion aimed to increase access to healthcare for millions of Americans who were previously uninsured.

Outcomes of Medicaid expansion

Research has shown that the ACA’s Medicaid expansion has led to several key outcomes, including improvements in contraceptive care (https://doi.org/10.1001/jamanetworkopen.2020.6874) and shifts in eligibility between Medicaid and insurance exchanges as income and family compositions change (https://doi.org/10.1377/hlthaff.2010.1000). Furthermore, studies have reported labor market impacts, demonstrating how increased Medicaid coverage can affect employment among low-income populations (https://doi.org/10.2139/ssrn.2785938).

Medicaid expansion has also been associated with increased coverage for prenatal care and births, leading to improved birth outcomes. Evidence from Florida showed that Medicaid eligibility expansion for pregnant women led to decreased incidences of low-birthweight infants among low-income women without private insurance (https://doi.org/10.2105/AJPH.88.3.371). Similar to other states, this expanded Medicaid coverage ensures that more pregnant women have access to necessary prenatal and postnatal care, potentially leading to healthier outcomes for both mothers and infants (https://doi.org/10.1093/oxfordjournals.aje.a115692). Medicaid expansion statuses are tracked by state online by the Kaiser Family Foundation (https://www.kff.org/affordable-care-act/issue-brief/status-of-state-medicaid-expansion-decisions-interactive-map).

Medicaid coverage of births

Medicaid pays for a significant proportion of all births in the United States, which reflects its crucial role in providing coverage for pregnant women and newborns (https://doi.org/10.1097/EDE.0000000000001462). This coverage has been essential in states that aggressively implemented Medicaid eligibility expansions for pregnant women, resulting in a marked increase in the number of births financed by Medicaid (https://doi.org/10.2307/2991692).

Medicaid Managed Care Organizations

As of 2023, approximately 72% of Medicaid beneficiaries, or 57 million people across 41 states, are enrolled in managed care programs (https://www.gao.gov/products/gao-24-106627). The early years of Medicaid allowed for a fee-for-service model, where beneficiaries could freely choose their healthcare providers. This model was predominant across the United States and reflected a broader commitment to ensuring that low-income populations could access necessary medical services with-out undue restrictions (https://ccf.georgetown.edu/2023/03/21/early-legislative-history-medicaid-managed-care).

However, the landscape of Medicaid began to shift with various legislative and policy changes. The introduction of Medicaid Managed Care in the 1970s marked a significant pivot; the state of California, under Governor Ronald Reagan, initiated pilot projects to test whether prepayment for services could reduce overall Medicaid costs compared to the traditional fee-for-service model.This led to the statewide implementation of Medicaid managed care under the Medi-Cal Reform Act in 1971 (https://ccf.georgetown.edu/2023/03/21/early-legislative-history-medicaid-managed-care).

The evolution of Medicaid Managed Care gained momentum with the Balanced Budget Act of 1997, which allowed states the option to require most beneficiaries to receive services through Managed Care Organizations (MCOs). This shift aimed to control costs and improve care coordination. As of now, managed care is the dominant delivery system in Medicaid, with comprehensive risk-based managed care being implemented by 40 states. These MCOs can limit the network of providers to which enrollees have access, significantly altering the landscape of provider choice that was foundational in Medicaid’s early days (https://ccf.georgetown.edu/2023/03/21/early-legislative-history-medicaid-managed-care).

The National Law Program conducted an in-depth analysis of state incentives and penalties to MCOs (https://healthlaw.org/medicaid-managed-care-accountability-how-states-use-sanctions).

Waiver programs

Oregon

K John McConnell and colleagues at Oregon Health and Science University evaluated the impact of Oregon’s coordinated care organization (CCO) model, an ambitious Medicaid delivery system reform initiated in 2012. CCOs are a type of accountable care organization that receive global budgets to manage the physical, dental, mental health, and broader social service needs for defined Medicaid populations. Key features of the CCO model included global budgets, financial incentives tied to quality metrics, governance boards with multi-stakeholder representation, community-based programs, and integration of physical and behavioral health services. The authors used a difference-in-differences approach comparing Oregon to the neighboring state of Washington before (2011) and after (2013-2014) CCO implementation. They found that two years into the reform, the CCO model was associated with a 7% relative reduction in standardized expenditures across evaluation/management, imaging, procedures, tests, and inpatient facility services compared to Washington. This reduction was driven primarily by decreases in inpatient utilization. Oregon also showed improvements in some quality metrics like avoidable emergency department visits and low-value imaging compared to Washington. The CCO model’s expenditure reductions were concentrated among adults and higher-risk patients. Rural and urban CCOs, as well as those with prior risk-sharing experience, performed similarly (https://pubmed.ncbi.nlm.nih.gov/28264946).

Arkansas

The implementation of work requirements in Arkansas’s Section 1115 Medicaid waiver had significant impacts on health coverage and financial stability among low-income adults. The policy, which was active from June 2018 until April 2019, led to approximately 18,000 adults losing their Medicaid coverage (https://pubmed.ncbi.nlm.nih.gov/32897784). This disenrollment did not result in increased employment among the affected individuals over an eighteen-month follow-up period. Instead, those who lost Medicaid coverage reported substantial financial distress related to healthcare costs: 50% encountered serious problems paying medical debt, 56% delayed care due to costs, and 64% delayed taking medications due to costs (https://pubmed.ncbi.nlm.nih.gov/32897784). Furthermore, the policy was associated with a higher rate of Medicaid disenrollment compared to other states without such requirements, with significant portions of those disenrolled becoming uninsured (https://pubmed.ncbi.nlm.nih.gov/32552024). This suggests that the work requirements mayhave contributed to increased healthcare insecurity among vulnerable populations, including those with chronic conditions who faced higher risks of losing coverage (https://pubmed.ncbi.nlm.nih.gov/32552024). Further details of the Arkansas requirements have been provided by Laura Harker of the Center on Budget and Policy Priorities (https://www.cbpp.org/research/health/pain-but-no-gain-arkansas-failed-medicaid-work-reporting-requirements-should-not-be).

Chapter 3 | The Medicaid Expansion Controversy: Politics, Policy, and Outcomes

Medicaid expansion in North Carolina

The political history of North Carolina’s Medicaid expansion has been detailed by journalist Kiara Brantley-Jones: (https://abcnews.go.com/Politics/north-carolina-governor-hopes-medicaid-expansion-600k-residents/story?id=105853956).

Health outcomes related to Medicaid expansion

1. General Health Improvements:

Several studies have shown that Medicaid expansion leads to better overall health outcomes. For instance, Federally QualifiedHealth Centers (FQHCs) in states that expanded Medicaid saw improved measures for blood pressure and glucose control over a five-year period, particularly for Black and Latinx patients (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8796924/). Additionally, Medicaid expansion is linked to a 9.4 percent reduction in death rates (https://www.nber.org/system/files/working_papers/w26553/w26553.pdf).

2. Specific Health Conditions:

• HIV Outcomes: In Nebraska, Medicaid expansion was associated with better viral suppression among patients with HIV, which is crucial for improving clinical outcomes and reducing viral transmission (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10679050/).
• Perinatal Health: Medicaid expansion has also been linked to improved perinatal health outcomes (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11001670/).
• Asthma: There was an increase in the utilization and expenditures for asthma-related prescription drugs among low-income patients, indicating better management of asthma due to increased access to necessary medications (https://pubmed.ncbi.nlm.nih.gov/34978935/).
• Cancer Treatment: Medicaid expansion has reduced racial disparities in the time to cancer treatment, ensuring more timely access to care for advanced cancer patients across different racial groups (https://www.semanticscholar.org/paper/6184f1bc0f8813a2b21dfb6712394f76352dbe39).

3. Financial Outcomes

• Reduction in Treatment Costs: In Texas, the Medicaid expansion was associated with a 4.2% reduction in treatment costs among patients from states that expanded Medicaid. This was accompanied by a significant decrease in the uninsured rate and an increase in the share of Medicaid patients (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9445472/).
• Hospital Financial Health: The expansion has had mixed effects on hospital finances. While some studies suggest improvements due to increased coverage reducing uncompensated care, others indicate challenges such as increased operating expenses (https://pubmed.ncbi.nlm.nih.gov/36602454).
• Economic Benefits for Health Workers: There is evidence suggesting that Medicaid expansion has positively affected the economic outcomes of healthcare workers, particularly those in lower-wage positions, by improving incomes and increasing the likelihood of receiving benefits like health insurance (https://pubmed.ncbi.nlm.nih.gov/38411645/).

4. Broader Economic and Social Effects

• Financial Security: Medicaid expansion has provided a safety net that reduces the financial impact of healthcare costs on households. For instance, during the COVID-19 pandemic, individuals in expansion states experienced less financial hardship related to healthcare compared to those in non-expansion states (https://pubmed.ncbi.nlm.nih.gov/37801548/).
• Access to Care: Expansion states have seen a sustained reduction in uninsurance rates at FQHCs, which translates to broader access to healthcare services (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8796924/).

Historical arguments against Medicaid expansion, and their associated data

The discourse surrounding Medicaid, particularly its expansion under the Affordable Care Act (ACA), incorporates various arguments regarding its impacts on healthcare access, financial well-being, hospital finances, and the broader economy. The studies listed provide empirical evidence to assess the validity of common critiques against Medicaid expansion:

1. Employment Effects

Critics have argued that Medicaid expansion could discourage employment by making healthcare access less dependent on job status. However, Buchmueller et al. (2020) found no evidence that Medicaid expansion adversely affects employment rates among the unemployed, suggesting that the policy does not discourage job seeking (https://www.nber.org/system/files/working_papers/w26553/w26553.pdf).

2. Financial Well-being

Critics have posited that Medicaid expansion could strain individual finances through increased taxes or reduced healthcare quality. However, Hu et al. (2018) found that the ACA Medicaid expansions significantly improved financial wellbeing by reducing out-of-pocket medical expenses and lowering debt without increased taxes (https://pubmed.ncbi.nlm.nih.gov/30393411/).

3. Hospital Financial Outcomes

The argument that Medicaid expansion could financially harm hospitals, especially due to lower reimbursement rates compared to private insurance, is countered by evidence of heterogeneous effects. While Rhodes et al. (2020) found varied impacts on hospital finances (https://deepblue.lib.umich.edu/bitstream/han-dle/2027.42/153165/coep12428.pdf?sequence=2), Dranove et al.(2016) reported that uncompensated care costs decreased significantly in expansion states, benefiting hospital finances (https://www.healthaffairs.org/doi/10.1377/hlthaff.2015.1344).

4. Fiscal Relief and Economic Impact

Concerns regarding Medicaid expansion’s strain on state budgets are contrasted by findings from Chodorow-Reich et al.(2012) and Levy et al. (2020), which suggest that state fiscal relief during recessions, including Medicaid expansions, can stimulate employment and have positive macroeconomic feedback effects (https://www.aeaweb.org/articles?id=10.1257/pol.4.3.118, https://read.dukeupress.edu/jhppl/article-abstract/45/1/5/140654/Macroeconomic-Feedback-Effects-of-Medicaid?redirectedFrom=fulltext).

The counter-cyclical nature of Medicaid

Medicaid is a counter-cyclical program, meaning that more people become eligible and enroll during economic downturns when jobs and incomes decline. At the same time, states face reduced revenues that make it difficult to fund their share of Medicaid costs. This counter-cyclical nature acts as an automatic stabilizer, providing increased federal funding to states as economic conditions worsen. During the COVID-19 recession, Medicaid enrollment grew sharply as unemployment rose, increasing by 9.4% from February to September 2020 (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8242625). Total Medicaid spending also increased following the pandemic’s onset, driven primarily by this enrollment growth, according to state budget surveys (https://www.kff.org/medicaid/issue-brief/the-impact-of-the-covid-19-recession-on-medicaid-coverage-and-spending).

The Families First Coronavirus Response Act temporarily increased the federal Medicaid matching rate by 6.2 percentage points to provide fiscal relief to states during the public health emergency. This enhanced federal support helped offset state budget shortfalls and maintain Medicaid services when needs were highest. Unlike previous recessions, the COVID-19 downturn required states to provide continuous Medicaid coverage through the end of the public health emergency as a condition of receiving the increased federal match. This prevented states from restricting eligibility or enrollment procedures that could have limited Medicaid’s counter-cyclical impact (https://www.kff.org/medicaid/issue-brief/the-impact-of-the-covid-19-recession-on-medicaid-coverage-and-spending).

Chapter 4 | Beyond Checkboxes: Rethinking Social Needs in Medicaid

Cost of addressing social needs

In a decision analytical microsimulation study, my colleagues and I estimated the cost of implementing evidence-based interventions to address social needs identified in primary care practices. We used nationally representative data from the National Center for Health Statistics (2015-2018) to estimate needs among the civilian U.S. population of 251,406,318 individuals visiting primary care practices. We assessed self-reported social needs in four domains: food insecurity (17.5% prevalence), severe housing insecurity (0.9%), transportation insecurity (2.8%), and community-based care coordination needs (12.7%). The study found that while most individuals with food (95.6%) and housing (78.0%) needs were eligible for federally funded programs, enrollment was low (70.2% and 24.0%,respectively) due to factors such as inadequate program capacity. Eligibility criteria were the primary limiting factor for transportation (26.3% eligible) and care coordination (5.7% eligible) programs. The estimated cost of providing evidence-based interventions averaged $60 per member per month (PMPM), with only 45.8% ($27 PMPM) covered by existing federal funding mechanisms. Subgroup analyses revealed higher costs for populations attributed to federally qualified health centers ($93 PMPM) and non-FQHC practices in high-poverty areas ($77 PMPM) compared to those in lower-poverty areas ($24 PMPM) (https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/2805020).

Food as medicine

The concept of “food as medicine” (FAM) encompasses a range of food-based interventions designed to prevent, manage, and treat various health conditions. The interventions include medically tailored meals, groceries, and produce prescriptions, often combined with nutrition and culinary education. Observational and correlative studies subject to confounding have associated the participation in FAM interventions to food insecurity, diet quality, and related outcomes (https://pubmed.ncbi.nlm.nih.gov/38383100). By contrast, randomized trials subject to less selection bias have had far less impressive results. In specific populations, such as individuals with HIV, FAM interventions did not impact viral suppression but did have an effect on hospitalizations and self-reported mental/physical health (https://pubmed.ncbi.nlm.nih.gov/38696724). Similarly, in patients with diabetes and food insecurity, intensive FAM programs did not significantly improve glycemic control compared to usual care (https://pubmed.ncbi.nlm.nih.gov/38147326).

Supplemental Nutrition Assistance Program (SNAP)

SNAP has been associated with various positive health outcomes across different populations. Research indicates that SNAP participation is linked to reduced mortality rates, particularly from specific causes of death among adults between 40-64 years old (https://pubmed.ncbi.nlm.nih.gov/31682512). Additionally, SNAP eligibility has been shown to decrease the prevalence of diet-related diseases such as diabetes, hypertension, and cardiovascular conditions among older adults, with significant reductions noted particularly among Hispanic and non-Hispanic Black populations (https://pubmed.ncbi.nlm.nih.gov/33491211).For children, SNAP participation correlates with improved health status, reduced developmental risks, and lower rates of food insecurity (https://pubmed.ncbi.nlm.nih.gov/31542130). It also appears to mitigate the negative health impacts associated with food insecurity in older adults, although it does not significantly alter the association between food insecurity and depression (https://pubmed.ncbi.nlm.nih.gov/31678175). Furthermore, SNAP has been associated with reduced emergency department use among low-income children, mediated by improvements in food hardship and health status (https://pubmed.ncbi.nlm.nih.gov/36710646). Among older adults, particularly those dually eligible for Medicare and Medicaid, SNAP participation is linked to fewer inpatient admissions and lower healthcare costs (https://pubmed.ncbi.nlm.nih.gov/34662150).

Bon Secours housing initiative

Emmanuel Drabo and colleagues at Johns Hopkins University conducted a social return on investment (SROI) analysis of Bon Secours Hospital’s Housing for Health affordable housing program in Baltimore, Maryland. Bon Secours owned 801 affordable housing units across 12 properties aimed at addressing social and environ-mental determinants of health in West Baltimore. The SROI analysis evaluated the broader social, environmental, and economic impacts over one year using stakeholder engagement, outcome mapping, and valuation methods. The analysis estimated the program generated between $1.30 and $1.92 in social returns annually for every $1 invested in operating costs. Key benefits included increased housing stability, reduced environmental hazards, increased property values, and reduced use of public services for residents and the community (https://www.healthaffairs.org/doi/abs/10.1377/hlthaff.2020.00998).

Screening and referral

The effectiveness of rapid ‘screening and referral’ strategies for addressing social determinants of health (SDOH) remains uncertain, with mixed outcomes reported across various studies. A systematic review highlighted that while screening and referral programs tend to show positive influences on patient care experience and some aspects of population health, definitive conclusions about their overall impact cannot be drawn due to the high risk of bias in the studies reviewed (https://pubmed.ncbi.nlm.nih.gov/34387188). Furthermore, another systematic review on intersectoral actions, which often include screening and referral strategies, reported moderate to no effect on the social determinants of health and health equity, indicating limited evidence supporting the effectiveness of these strategies in significantly altering health outcomes or equity (https://pubmed.ncbi.nlm.nih.gov/24209299). Additionally, evidence from specific healthcare settings, such as obstetrics and gynecology, also indicates limited benefits of SDOH screening, with inconsistent results on the resolution of social needs and no substantial data on effects on health systems or providers (https://pubmed.ncbi.nlm.nih.gov/37330393). This suggests that while screening for SDOH is increasingly recognized and implemented, its effectiveness in improving health outcomes and addressing health inequities is not robustly supported by current evidence. Further research, particularly studies that are methodologically rigorous and context-specific are needed to better understand and optimize the impact of these strategies.

Referral versus receipt of social services using closed-loop referral systems

Fred Johnson and colleagues at Duke University compared two metrics for measuring the success of connecting patients to social services to address their health-related social needs (HRSNs) usinga social resource connection platform called NCCARE360/Unite Us.

The two metrics analyzed were:

1. Closed Loop Rate (CLR)

The number of patient cases marked as closed divided by the total number of cases, regardless of whether patients actually received services.

2. Successful Connection Rate (SCR)

The ratio of cases where patients received a real benefit/service divided by all managed cases.

The study compared the CLR and SCR across two time periods: Period 1 (Oct 2020 - Mar 2021) when there was funding for community-based organizations (CBOs) to use NCCARE360, and Period 2(Oct 2021 - Mar 2022) without such funding.

The key findings were:

• CLR was similar in both periods (99% vs 93%), indicating successful use of the platform for placing and closing referrals.
• However, SCR was substantially lower in Period 2 (38%) compared to Period 1 (65%), suggesting the funding and support for CBOs in Period 1 led to better outcomes in actually connecting patients to services.
• For the food domain, which received funding throughout, the SCR remained stable across both periods (76% and 71%).

The authors concluded that while CLR represents platform utilization, SCR is a more patient-centered metric reflecting whether services were truly received (https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2817709?adv=000801070736).

Housing First strategies

Several key quantitative results have been observed from studies on the impact of Housing First strategies:

Reduced Healthcare Utilization

• Housing First participants experienced fewer emergency department visits compared to control groups (incidence rate ratio=0.63, 95% CI 0.48 to 0.82; https://jech.bmj.com/content/73/5/379)
• Housing First participants had fewer hospitalizations (incidence rate ratio=0.76, 95% CI 0.70 to 0.83) and less time spent hospitalized (standardized mean difference=-0.14,95% CI -0.41 to 0.14) than control groups (https://jech.bmj.com/content/73/5/379)

Cost Savings (https://nlihc.org/resource/systematic-research-review-finds-benefits-housing-first-programs-us-outweigh-costs)

• The median economic benefit of Housing First programs across U.S. studies was $18,247 per person per year (PPPY)
• For good-quality U.S. studies, the median economic benefit was $33,637 PPPY
• The median benefit-to-cost ratio across all U.S. studies was1.8:1, meaning for every $1 spent on Housing First, there was an economic benefit of $1.80
• For good-quality U.S. studies, the median benefit-to-cost ratio was 1.3:1
• Studies accounting for averted healthcare, judicial system, welfare assistance, and housing assistance costs had a median PPPY benefit of $26,907

Chapter 5 | Care Segregation and Network Inadequacy: Medicaid’s Network Challenges and Corrective Attempts

Medicaid versus Medicare and commercial payment rates

In 2019, Medicaid fee-for-service payments for physician services were nearly 30 percent below Medicare payment levels, with the differential being even larger for primary care physician services. Medicaid payment rates vary widely across states, with Medicaid fee-for-service rates for primary care being less than half the Medicare payment rate in states like Florida, Illinois, Pennsylvania, New York, Rhode Island, and Wisconsin. By contrast, Medicaid rates were at or above Medicare rates in just four states: Alaska, Delaware, Montana, and North Carolina. Regarding hospital payments, data from the Medicaid and CHIP Payment and Access Commission found thatMedicaid fee-for-service inpatient hospital base payments were 22 percent below comparable Medicare rates on average. After accounting for supplemental payments made by many states to hospitals, however, Medicaid inpatient payments averaged 6 percent above Medicare rates, though not all states make such supplemental payments and not all hospitals receive them (https://www.common-wealthfund.org/blog/2022/how-differences-medicaid-medicare-and-commercial-health-insurance-payment-rates-impact).

Impact on Medicaid provider denials on acceptance of patients

Abe Dunn and colleagues at the National Bureau of Economic Research studied a novel dataset of “remittance data” that captures the back-and-forth billing interactions between a large sample of U.S. physicians and many different insurers following 90 million patient visits between 2013-2015. This remittance data provides granular details about the claim denials, reasons for denials, resubmissions, and payments made, enabling the researchers to observe and model the costly billing and bargaining process that occurs after care is provided, what they term the “costs of incomplete payments” (CIP).

To estimate these CIP across insurers and states, the researchers leveraged the observed resubmission decisions to estimate the expected continuation values for each resubmission decision, obtaining maximum likelihood estimates of the costs physicians incur for resubmitting claims.

Their analysis found that CIP are particularly large when billing Medicaid. About 24% of Medicaid claims face a denial on initial submission, compared to only 6.7% for Medicare and 4.1% for commercial insurance. Translating this into revenue losses, they estimate physicians lose 18% of Medicaid revenue to billing problems, versus 4.7% for Medicare and 2.4% for commercial plans.

Crucially, the authors showed that these administrative frictions substantially impact physicians’ willingness to accept Medicaid patients, just as much as Medicaid’s lower reimbursement rates. Using physicians who move across states for identification, they found a one-standard-deviation increase in CIP (around 10 percentage points) reduced a physician’s probability of accepting new Medicaid patients by 0.8 percentage points—larger than the 0.6 percentage point increase from a standard deviation rise in Medicaid fees.

Looking within physician practices that span multiple states, the impact of CIP was even larger: a one standard deviation increase in CIP decreases Medicaid acceptance probability by 1.5 percentage points, while a comparable increase in fees raises it by 2.2 percentage points.

The researchers quantified potential gains from reducing these frictions, finding that decreasing Medicaid prices by 10% while simultaneously cutting the denial probability by 20% could maintain the same Medicaid acceptance level while saving $10 per visit on average (https://www.nber.org/papers/w29010).

Limited managed care networks

Avital Ludomirsky and colleagues at Yale University examined the extent to which Medicaid managed care plan networks may overstate the availability of physicians in the Medicaid program. The researchers used administrative data from four states to assess the level of Medicaid participation among physicians listed in the provider network directories of each managed care plan. They found that about one-third of outpatient primary care and specialist physicians contracted with Medicaid managed care plans in their sample saw fewer than ten Medicaid beneficiaries in a year. Overall, 16.3% of physicians listed in the networks qualified as “ghost” physicians, meaning they saw zero Medicaid beneficiaries over the course of the year. This share varied from 13.4% to 24.9% across the four states. Care was also highly concentrated: 25% of primary care physicians provided 86.2% of the care, and 25% of specialists provided 75% of the care. When ghost and peripheral physicians were excluded, the average ratios of Medicaid beneficiaries to physicians became less favorable, with only 84.7% of counties meeting a common benchmark for primary care access (https://www.healthaffairs.org/doi/abs/10.1377/hlthaff.2021.01747?journalCode=hlthaff).

Phantom networks

Jane Zhu and colleagues at Oregon Health and Science University examined discrepancies between provider listings in Medicaid managed care plan directories and the actual availability of those providers to see Medicaid patients in Oregon. The researchers found that 58.2% of providers listed in the plan directories were “phantom” providers who did not see any Medicaid patients during the study period. This included two-thirds of mental health prescribers, 59.0% of mental health non-prescribers, and 54.0% of primary care providers.

The researchers also found significant disparities when calculating provider-to-enrollee ratios. Ratios based on the plan directories were much higher than those derived from claims data, sometimes by over 5-fold for mental health prescribers. On average, primary care provider networks had higher “realized access,” with 4 6.0% of listed providers actually seeing patients, compared to 32.6% for mental health prescribers and 41.0% for mental health non-prescribers.

There was high variation, however, in realized access across the state’s 15 Medicaid managed care organizations, suggesting differences in administrative capacity and enforcement mechanisms. The researchers concluded that the substantial discrepancies between listed providers and those who actually see Medicaid patients under-mine the usefulness of provider directories for ensuring network adequacy (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9876384/). Zhu and colleagues similarly noted significant variation across states in network adequacy metrics and standards for mental health services. For example, California and Tennessee have detailed time and distance requirements, while Missouri does not appear to have any published standards (https://jamanetwork.com/journals/jama-health-forum/fullarticle/2779945).

Background on the 340B Program

Karen Mulligan and colleagues from the University of Southern California have provided a detailed summary and analysis of the 340B Drug Pricing Program (https://healthpolicy.usc.edu/research/the-340b-drug-pricing-program-background-ongoing-challeng-es-and-recent-developments/). The 340B Program was created in1992 to enable certain healthcare providers, known as “covered entities,” to “stretch scarce federal resources to reach more eligible patients or provide more comprehensive services.” The program requires drug manufacturers to sell outpatient drugs at discounted prices to covered entities, which include certain hospitals, clinics, and other safety-net providers. Participation in the program has grown substantially, with the number of covered entity sites increasing from 8,100 in 2000 to 50,000 in 2020. Estimated discounted purchases through the program have risen from about $4 billion per year in 2007–2009 to $38 billion in 2020.

Mulligan and colleagues discuss several ongoing controversies and challenges facing the 340B program. These include limited program oversight by the Health Resources and Services Administration (HRSA), which administers the program, lack of transparency around how covered entities use the funds generated through 340B discounts, and the rapid growth in participation by contract pharmacies and disproportionate share (DSH) hospitals. HRSA’s regulatory authority over the program is narrow, with the agency only having the ability to issue rules related to ceiling price calculations, manufacturer overcharge penalties, and a dispute resolution process. This has contributed to issues such as the inability to prevent duplicate discounts, where 340B drugs are subject to both the 340B discount and a Medicaid rebate.

The expansion of contract pharmacies, which now account for about 30% of 340B purchases, has increased the risk of drug diversion and made it more difficult to track 340B claims, particularly in Medicaid managed care plans. Similarly, the growing participation of DSH hospitals in the 340B program, which now account for around 80% of 340B sales, has raised concerns that the benefits are not being directed to the vulnerable populations the program was intended to serve. Proposals to address these issues include limiting the number of contract pharmacies, increasing reporting requirements for covered entities, and revisiting the criteria for hospital eligibility.

Kelsey Owsley and colleagues at the University of Arkansas evaluated changes in the provision of typically unprofitable hospital service lines after the initiation of 340B participation. The researchers used a difference-in-differences design to isolate the effect of 340B enrollment, comparing changes in service line offerings before and after participation to changes at non-participating hospitals over the same time period.

The key finding is that while 340B participation was not associated with increases in unprofitable service offerings on average, there was an important difference between public and nonprofit hospitals. Public hospitals that initiated 340B participation increased the number of unprofitable service offerings by over 10%, particularly in areas like inpatient psychiatric care and substance use treatment. In contrast, nonprofit hospitals saw no such increase in unprofitable services and even exhibited a small increase in the provision of highly profitable oncology services (https://jamanetwork.com/journals/jama-health-forum/fullarticle/2818087).

Chapter 6 | Reinvigorating Primary Care, Care Access, and Coordination in Medicaid

Primary care access and mortality

My colleagues and I examined the association between changes in primary care physician supply and population health outcomes across US counties from 2005 to 2015. We used a mixed-effects regression model as the primary analytic approach, which allowed us to account for both within-county and between-county variation in physician supply and health outcomes. We supplemented this with several additional analyses to test the robustness of the findings. First, we conducted instrumental variable analyses, using changes in county-level purchasing power due to the federal Public Service Loan Forgiveness program as an instrument for changes in primary care physician supply. This helped address potential unmeasured confounding. We also performed a “near-far” matching analysis as a robustness check on the instrumental variable approach.

Second, we conducted individual-level survival analyses using claims data linked to mortality, which helped reduce the risk of ecological fallacy. This individual-level analysis examined the association between area-level primary care physician supply and individual life expectancy, as measured by restricted mean survival time. Finally, we performed a falsification test by examining the association between primary care physician supply and mortality due to interpersonal violence, which would not be expected to be influenced by physician supply.

The key findings were:

• Primary care physician supply per capita decreased from 2005 to 2015, particularly in rural areas, despite an overall increase in the total number of primary care physicians.
• In the mixed-effects models, an increase of 10 primary care physicians per 100,000 population was associated with a 51.5-day increase in life expectancy, as well as reductions in cardiovascular, cancer, and respiratory mortality.
• The instrumental variable and individual-level analyses produced similar or larger estimates of the association between primary care physician supply and improved health outcomes.
• The falsification test found no association between primary care physician supply and mortality due to interpersonal violence, as expected.
• Sensitivity analyses including nurse practitioners and physician assistants in the measure of primary care clinician supply produced consistent but less precisely estimated results.

The key quantitative findings were that in the mixed-effects models, an increase of 10 primary care physicians per 100,000 population was associated with a 51.5-day increase in life expectancy, as well as reductions in cardiovascular, cancer, and respiratory mortality of 0.9% to 1.4%. The instrumental variable analysis detected an even larger association, with an 88.9-day increase in life expectancy per 10 additional primary care physicians. The individual-level analysis found a 114.2-day increase in restricted mean survival time per decade of exposure to 10 more primary care physicians per 100,000 population.

In additional subsequent studies, we estimated that increasing the number of primary care physicians to one per 1,500 residents in areas with shortages could boost life expectancy by 56 days and reduce the annual death rate significantly, preventing thousands of deaths annually (https://pubmed.ncbi.nlm.nih.gov/33750188/).

Factors influencing primary care supply

The decline in medical students entering primary care and the challenges faced by those in the field can be quantitatively analyzed through various studies that highlight income disparities, burnout rates, and the impact of systemic biases in reimbursement models.

1. Income Disparities:

A significant factor influencing the career choices of medical students is the income gap between primary care physicians and specialists. A study by Bryan Vaughan and colleagues at Duke University quantitatively estimated the career wealth accumulation across different medical fields, revealing that primary care physicians earn lower incomes and accumulate considerably less wealth over their lifetimes compared to their specialist counter-parts. This wealth gap is substantial, with the study suggesting that narrowing it would require either substantial reductions in specialists’ practice income or increases in primary care physicians’ practice income by more than $100,000 a year (https://pubmed.ncbi.nlm.nih.gov/20439883/).

2. Burnout Rates:

Burnout among primary care physicians is another critical issue, with quantitative studies providing insight into its prevalence and consequences. For instance, a cross-sectional study involving 136 primary care physicians, finding that burnout was significantly associated with higher rates of referrals for diagnostic tests and specialist clinics. This association was quantified with a path analysis using Structural Equation Modeling, which explained a total of 18.1% of the variance in referral rates (P < 0.05; https://pubmed.ncbi.nlm.nih.gov/24148815/).

3. Systemic Biases in Reimbursement:

The reimbursement model for healthcare services, particularly the fee-for-service system, has been criticized for favoring procedural over cognitive services. Arielle Langer at Brigham and Women’s Hospital used the Community Tracking Study (CTS) Physician Survey to analyze income differences across physician specialties, revealing that procedural medical specialties earned 37.5 percent more than family medicine, compared to 15.3 percent for non-procedural medical specialties (https://pubmed.ncbi.nlm.nih.gov/31837254/). This analysis suggests that the differences in physician income and resulting incentives are a direct consequence of the payment structure itself.

Screening time

A 2022 simulation analysis by Porter and colleagues modeled the time required for primary care physicians to deliver recommended preventive, chronic illness, and acute care to a standard panel of 2,500 adults representative of the US population. Applying current clinical practice guidelines across these domains, they estimated PCPs would need a staggering 26.7 hours per day to provide indicated care. This comprised 14.1 hours for preventive services, 7.2hours for chronic disease management, 2.2 hours for acute visits, and 3.2 hours spent on documentation and inbox tasks. Under conservative estimates for visit lengths, this dramatic mismatch highlights a nearly 700% capacity gap between practice recommendations and reality. With optimized team-based care including nurses and medical assistants, the estimated workload was stillexcessive at over 9 hours daily. By quantifying the unchecked growth of primary care responsibilities, this study demonstrates the urgent need for revised standards and workforce structures aligned with achieving quality goals sustainably (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9848034/#:~:text=Times%20were%20re%2Desti-mated%20in,the%20guideline%2Drecommended%20primary%20care).

Pajama time among primary care providers

Two studies published in 2017 and 2023 demonstrate the excessive burdens electronic health records (EHRs) impose on primary care physicians. Analyzing over 140 PCPs in Wisconsin, Arndt et al. found clinicians averaged 5.9 hours daily on EHR tasks: 4.5 hours during clinic plus 1.4 hours after-hours. Nearly half of this workload comprised clerical activities like documentation and billing (https://www.annfammed.org/content/15/5/419.full). A 2023 follow-up analysis by Rotenstein et al. revealed continued strains, with a median 36 minutes per patient visit spent on EHRs. However, they found organizational factors like team collaboration, medication refill support staff, and community health centers associated with modestly lower burdens, suggesting systems-level changes are essential to restore balance. With EHR interactions consuming over half of PCP workdays, these studies quantify the mismatch between administrative expectations and patient care, underscoring the urgent need for revised workflows and infrastructure to prevent burnout and impaired quality (https://pubmed.ncbi.nlm.nih.gov/37991757/).

Physicians fee schedules

A 2013 analysis by Sinsky and Dugdale published in JAMA Internal Medicine demonstrates the substantial Medicare payment gaps disadvantageous to cognitive physician specialties such as primary care. Examining reimbursements for similar physician time requirements, Sinky and Dugdale found colonoscopy and cataract surgeries garnered 368% and 486% higher revenue respectively compared to evaluation and management office visits. With procedure-focused care yielding 3 to 5 times physicians’ hourly rate for cognitive services critical to prevention and chronic disease management, this imbalance propagates systemwide overuse incentives. Quantifying the financial mismatch, the authors argue Medicare’s physician payment schema directly enables the outsized procedural orientation plaguing US healthcare. The study suggests rebalancing compensation and prestige for primary care could curb excess utilization while supporting the comprehensive, continuous relationships most associated with overall population health. Underscoring the need for payment reform centered on value over volume, this analysis highlights the policies distorting care delivery away from cost-effective cognitive services toward revenue-maximizing procedures (https://jama-network.com/journals/jamainternalmedicine/fullarticle/1754364).

Prior authorizations

The impact of prior authorizations (PAs) on patient outcomes varies across different medical specialties and conditions, but the overarching theme from recent research is that PAs can lead to delays in care, which may adversely affect patient outcomes. For instance, in gynecologic oncology, patients experienced over a two-week delay in care when prior authorization occurred (https://pubmed.ncbi.nlm.nih.gov/36244827). In the context of oral anti-cancer drugs, the introduction of a new PA policy was associated with increased odds of prescription discontinuation and delays in medication access (https://pubmed.ncbi.nlm.nih.gov/38086013). Similarly, for infusible medications in rheumatologic conditions, PAs were associated with treatment delays and greater glucocorticoid exposure following a PA request (https://pubmed.ncbi.nlm.nih.gov/31507077).

In the case of diabetes medications, delays in receiving medication due to PAs were associated with less reduction in hemoglobin A1c levels, indicating poorer glycemic control (https://pubmed.ncbi.nlm.nih.gov/38310735). For pediatric patients with inflammatory bowel disease, PAs were linked to prolonged initiation times for biologic therapies and increased healthcare utilization (https://pubmed.ncbi.nlm.nih.gov/35190811). In multiple sclerosis, insurance restrictions including PAs delayed necessary treatments and increased the likelihood of disease activity (https://pubmed.ncbi.nlm.nih.gov/38213675).

Waiting times for obstetric appointments

A 2023 audit study published in the American Journal of Obstetrics and Gynecology by Corbisiero et al. reveals significant disparities in access to OB/GYN subspecialty care based on insurance status. Calling a random national sample of 800 maternal-fetal medicine, gynecologic oncology, urogynecology, and reproductive endocrinology providers twice (once feigning Medicaid, once commercial coverage), they found 44% longer median wait times for Medicaid patients–20.3 days overall. Stratifying by subspecialty, the gap was most pronounced in urogynecology and less so but still significant in maternal-fetal medicine. Documenting harder-to-quantify barriers low-income patients face, this analysis suggests insurance type independently impedes access beyond socioeconomic factors (https://www.ajog.org/article/S0002-9378(23)00147-3/abstract).

Comprehensive Primary Care Plus

The Comprehensive Primary Care Plus (CPC+) program, implemented in 18 regions, was the largest US primary care delivery model tested. It was designed with a two-track system involving enhanced and alternative payments, care delivery requirements, data feedback, learning, and health information technology support. The primary outcome of interest was annualized Medicare Part A and B expenditures per beneficiary per month (PBPM). The CPC+ program was associated with no discernible changes in total expenditures but did result in increases in expenditures including enhanced payments. Secondary outcomes showed that CPC+ was associated with decreases in emergency department visits starting in year 1, and in acute hospitalizations and acute inpatient expenditures in later years. However, CPC+ was not associated with meaningful changes in claims-based quality-of-care measures (https://pubmed.ncbi.nlm.nih.gov/38100460).

Additionally, CPC+ participation was associated with increases in the delivery of annual wellness visits (AWVs) and flu shots, but not other high-value services such as advance care planning, counseling to prevent tobacco use, or depression screening (https://pubmed.ncbi.nlm.nih.gov/37557935). For privately insured patients, CPC+ did not improve spending or quality (https://pubmed.ncbi.nlm.nih.gov/36067428). There was also no evidence that CPC+ increased continuity or decreased fragmentation of care (https://pubmed.ncbi.nlm.nih.gov/35404554). Practices with higher proportions of Black or Latinx Medicare fee-for-service beneficiaries were less likely to participate in CPC+ (https://pubmed.ncbi.nlm.nih.gov/36941423).The CPC initiative, a precursor to CPC+, did not affect primary care physician experience in terms of burnout, control over work, job satisfaction, or likelihood of leaving current practice (https://pubmed.ncbi.nlm.nih.gov/30019124). Medicare beneficiaries with more comprehensive primary care physicians reported better overall care (https://pubmed.ncbi.nlm.nih.gov/36527443). Lastly, despitethe perceived benefits of longitudinal care management (LCM) for high-risk patients, uptake was low due to various challenges including insufficient care manager staffing (https://pubmed.ncbi.nlm.nih.gov/33496929).

Rhode Island’s initiatives

Rhode Island has implemented several strategies to improve primary care spending, focusing on enhancing the quality of care, increasing reimbursement rates, and fostering a more sustainable healthcare workforce. These efforts are part of a broader initiative to reform healthcare within the state, aiming to make healthcare more accessible, affordable, and effective for its residents. The strategies include setting primary care expenditure targets, increasing insurer investment in primary care, and supporting workforce diversity and training.

1. Primary Care Expenditure Targets

Rhode Island established primary care expenditure targets for commercial health insurers over a decade ago, marking a first-in-the-nation approach to explicitly prioritize primary care spending within the health insurance sector (https://ohic.ri.gov/sites/g/files/xkgbur736/files/2023-12/Primary%20Care%20in%20Rhode%20Island%20-%20Current%20Status%20and%20Policy%20Recommendations%20December%202023.pdf). This initiative was part of the state’s Affordability Standards, which aimed to lower healthcare costs and improve quality by encouraging insurers to invest more in primary care providers and services. The goal was to shift the focus towards preventive care and managing chronic conditions more effectively, thereby reducing overall healthcare costs (https://ohic.ri.gov/policy-reform/affordability-standards).

2. Increasing Insurer Investment in Primary Care

The Office of the Health Insurance Commissioner (OHIC) has been instrumental in driving reforms to increase insurer payment for primary care. This includes advocating for increased reimbursements for evaluation and management and other medical services when provided by primary care providers. The aim is to make primary care payment more competitive with other medical specialties and neighboring states, thereby supporting a robust care team of clinicians, medical assistants, and front office staff (https://ohic.ri.gov/sites/g/files/xkgbur736/files/2023-12/Primary%20Care%20in%20Rhode%20Island%20-%20Current%20Status%20and%20Policy%20Recommendations%20December%202023.pdf). The Affordability Standards further encouraged the transformation of primary care practices into Patient-Centered Medical Homes (PCMHs), which are designed to improve care coordination and outcomes for patients, especially those with chronic conditions (https://ohic.ri.gov/policy-reform/affordability-standards).

3. Supporting Workforce Diversity and Training

The Fund for a Healthy Rhode Island (FHRI) grants, managed by the Rhode Island Foundation, support projects focused on retaining a racially, culturally, ethnically, and linguistically diverse health workforce. This initiative recognizes the importance of a diverse healthcare workforce in meeting the needs of Rhode Island’s population. The grants aim to improve primary care access, utilization, and quality by supporting place-based collaborations and addressing health determinants outside the clinical setting (https://rifoundation.org/grant/fund-for-healthy-rhode-island-grants). Additionally, recent legislative proposals aim to provide $2.7 million to primary care practices for enhanced inter-disciplinary clinical training sites, further supporting the development of a skilled and diverse primary care workforce (https://www.wpri.com/news/politics/ri-lawmakers-launch-new-initiative-to-bolster-health-care-access-affordability/).

My colleague Aaron Baum led a study to further examine the impact of the Rhode Island reforms. We published in Health Affairs a research study that leveraged a quasi-experimental difference-in-differences design to quantify the spending impact of Rhode Island’s 2010 insurance cost control policy. Studying 38,001 Rhode Island enrollees over 2007-2016 compared to a matched out-of-state control cohort, the authors found the regulations including price caps and incentivized primary care coordination lowered inflation-adjusted per-member per-quarter fee-for-service spending by $76 (8.1% below 2009 baseline). Over the same period, quarterly non-fee-for-service care management costs rose by $21 per patient. On net, the state’s affordability standards achieved absolute reductions in total expenditure growth while maintaining or improving quality indicators (https://www.healthaffairs.org/doi/10.1377/hlthaff.2018.05164).

Capitated payments

My colleagues and I used a microsimulation model to estimate the financial implications for primary care practices of shifting to team- and non-visit-based care under fee-for-service (FFS), capitated payment, or a mix of the two. The model incorporated data from 969 primary care practices, including staffing, patient utilization, revenue, and costs.

We simulated practice revenues and costs under FFS and a range of capitated payments, before and after substituting team-based services (e.g., nurse visits) and non-visit-based care (e.g., phone/email) for low-complexity in-person physician visits. They defined low-complexity visits as those for chronic conditions without flares or new acute problems, where no tests or prescriptions other than point-of-care diagnostics were needed.

Under FFS, the substitution of team- and non-visit-based services produced financial losses of $42,398 per full-time-equivalent (FTE) physician per year. The costs of additional staff needed to provide these services outweighed the visit revenue gained from seeing more patients in the freed-up slots.

However, under capitated payment, 95% of simulated practices came out ahead financially from the shift to team-based care if more than 63% of their revenue was capitated. With capitation, practices could increase panel sizes and revenue without being constrained by the FFS visit schedule. Losses occurred in 95% of practices if less than 23% of revenue was capitated.

Practices achieved even greater financial gains with the addition of shared savings bonuses based on total medical expenditures. A 0.6% bonus (half the rate seen in ACO pilots) increased net surplus per FTE physician by $59,823 at full capitation. With shared savings, 95% of practices benefited financially from team-based care at a lower 56% capitation rate (https://www.healthaffairs.org/doi/10.1377/hlthaff.2017.0367).

AI scribes

A 2024 analysis by Tierney et al. from Kaiser’s Permanente Medical Group documents early findings from ambient artificial intelligence (AI) clinical scribes deployed in the Permanente Medical Group. Studying over 10,000 physicians across settings and specialties, the authors found 3,442 used the passive transcription tool in at least 100 patient encounters within 10 weeks, with the highest-intensity user enabling it for 1,210 visits. Clinician feedback indicated reduced after-hours documentation and more meaningful visits, complemented by favorable patient impressions of improved interactions. A statistical analysis also revealed associations between AI scribe usage and lower clinician time in records. Though generating high ratings on quality rubrics, the technology required clarifications and alignment to optimize accuracy (https://catalyst.nejm.org/doi/full/10.1056/CAT.23.0404).

Chapter 7 | Bridging Gaps and Building Integrations in Behavioral Health Care

The era of institutionalization

In the early to mid-20th century, the American approach to mental illness and substance use disorders was predominantly characterized by institutionalization. Psychiatric hospitals and asylums, often depicted in cultural works like “One Flew Over the Cuckoo’s Nest,” became symbols of the inhumane treatment of individuals with mental health conditions. These facilities were notorious for overcrowding, poor sanitation, and the use of cruel treatments such as lobotomies, electroshock therapy without anesthesia, and physical restraints (https://smleo.com/2020/02/27/20th-century-psychiatric-hospitals-and-the-lasting-impacts-of-deinstitutionalization/). Albert Deutsch’s “The Shame of the States” provided a scathing exposé of these conditions, highlighting the widespread abuse and neglect within state mental hospitals (https://books.google.com/books/about/The_Shame_of_the_States.html).

The shift towards deinstitutionalization

The 1950s and 1960s marked the beginning of a significant shift towards deinstitutionalization, driven by a confluence of factors. Advances in psychopharmacology introduced new psychiatric medications, such as chlorpromazine, which made outpatient treatment of mental illness more feasible (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10116376/). Legal and policy changes, including the landmark Wyatt v. Stickney case and the Community Mental Health Act of 1963, established rights for patients and laid the groundwork for community-based mental health services (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10116376/). Fiscal pressures also played a role, as the cost of maintaining large psychiatric institutions prompted state governments to explore more cost-effective alternatives (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10116376/).

Challenges of community-based care

While the movement towards deinstitutionalization aimed to improve the lives of those with mental illness or substance use disorders, it also introduced new challenges. The transition to community-based services often resulted in gaps in care and support, as these services lacked adequate funding and resources (https://smleo.com/2020/02/27/20th-century-psychiatric-hospitals-and-the-last-ing-impacts-of-deinstitutionalization/; https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10116376/). Many individuals struggled to find affordable housing and supportive services, contributing to increased rates of homelessness and incarceration among those with untreated conditions (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10116376/). Stigma and discrimination continued to create barriers to accessing care and integrating into the community (https://smleo.com/2020/02/27/20th-century-psychiatric-hospitals-and-the-lasting-impacts-of-deinstitutionalization/).

The role of Medicaid in addressing behavioral health needs

Medicaid has emerged as the largest payer for mental health services in the country, providing a critical safety net for millions of Americans with behavioral health needs. The Mental Health Parity and Addiction Equity Act of 2008 and the SUPPORT Act of 2018 have expanded coverage for mental health and substance use disorders, including medications for opioid addiction (https://ajp.psychiatryonline.org/doi/full/10.1176/appi.ajp-rj.2021.160404). However, the variability in Medicaid coverage across states and the IMD exclusion limit access to inpatient and residential care for many individuals (https://ajp.psy-chiatryonline.org/doi/full/10.1176/appi.ajp-rj.2021.160404).

Certified Community Behavioral Health Clinics (CCBHCs)

CCBHCs aim to provide comprehensive outpatient mental health and substance use disorder services, integrating physical and behavioral healthcare. This model enhances access to a full continuum of services including crisis mental health services, targeted case management, and psychiatric rehabilitation. By connecting behavioral health clinics with local primary care facilities and building coalitions with criminal justice entities and health advocates, CCBHCs ensure a more holistic approach to patient care (https://www.chcf.org/publication/certified-community-behavioral-health-clinics-cal-ex-plained/; https://www.ncsl.org/state-legislatures-news/details/the-value-of-certified-community-behavioral-health-clinics).

A CCBHC in Missouri reported a 66% decrease in requests for crisis intervention services, demonstrating a significant reduction in acute behavioral health crises (https://www.ncsl.org/state-legis-latures-news/details/the-value-of-certified-community-behavior-al-health-clinics). In Oregon, a CCBHC partnership with a local jail estimated a savings of $2.5 million in prison costs, highlighting the cost-effectiveness of integrating behavioral health with community and correctional health services (https://www.ncsl.org/state-legis-latures-news/details/the-value-of-certified-community-behavior-al-health-clinics). Overall, the CCBHC model has been associated with reduced use of high-cost services like emergency departments, which not only improves patient outcomes but also reduces over-all healthcare costs (https://www.ncsl.org/state-legislatures-news/details/the-value-of-certified-community-behavioral-health-clinics).

Sources of funding for CCBHCs have been reviewed extensively by the Center for Health Care Strategies (https://www.chcs.org/media/Planning-for-CCBHC-Program-Sustainability-Lessons-from-State-Medicaid-Leaders.pdf) and the National Center for State Courts (https://www.ncsc.org/__data/assets/pdf_file/0019/71380/CCBHCs.pdf). Their impact has been further reviewed by the National Council for Mental Wellbeing (https://www.thenationalcouncil.org/resources/2022-ccbhc-impact-report/) and the National Council for State Legislatures (https://www.ncsl.org/state-legislatures-news/details/the-value-of-certified-community-behavioral-health-clinics).

Access to behavioral health among people with Medicaid

According to a 2022 report from the Medicaid and CHIP Payment and Access Commission (MACPAC), many Medicaid beneficiaries with mental health conditions face challenges accessing needed treatment. MACPAC estimates that in 2018, 50 percent of Medicaid beneficiaries with serious mental illness did not receive necessary care despite reporting need. Barriers include limited state coverage of mental health services and low provider acceptance of Medicaid patients. The lack of adequate access has serious impacts, with this population demonstrating elevated rates of inpatient psychiatric hospitalization and criminal justice system involvement compared to privately insured peers.

As reviewed by MACPAC, data from the State Health Access Data Assistance Center indicates that in 2018, approximately one in five non-institutionalized adults aged 18-64 had a mental illness, while about half of all Americans will experience mental illness during their lifetime. While some live with mild to moderate conditions, others have serious mental illness. Regardless of insurance status, many report difficulty accessing care, especially those with serious conditions—estimates indicate 50 percent of adults with serious mental illness indicated needing but not receiving treatment in 2018, compared to 20 percent of those with mild or moderate illness.

The Americans with Disabilities Act and Supreme Court Olmstead decision affirms that individuals with serious mental illness, including those covered by Medicaid, are entitled to treatment in the most integrated, community-based setting appropriate to their needs and desired by the patient. However, beneficiaries still face substantial barriers accessing services outside institutional placements. Compared to privately insured populations, Medicaid beneficiaries with mental illness are less likely to receive outpatient therapy but more likely to undergo inpatient psychiatric hospitalization (https://www.macpac.gov/wp-content/uploads/2022/06/Chapter-2-Access-to-Mental-Health-Services-for-Adults-Covered-by-Medicaid.pdf).

Collaborative care model

The CoCM has demonstrated significant benefits in improving health outcomes and reducing costs across various medical settings. Evidence indicates that CoCM enhances access to mental health care and is more effective and cost-efficient than standard care for treating common mental illnesses (https://pubmed.ncbi.nlm.nih.gov/36595989). Specifically, in primary care settings, CoCM has been shown to reduce mortality among elderly patients without increasing hospital use (https://pubmed.ncbi.nlm.nih.gov/11213273). Furthermore, systematic reviews and meta-analyses confirm that CoCM improves mental and physical outcomes for individuals with mental disorders across a wide variety of care settings (https://pubmed.ncbi.nlm.nih.gov/22772364). In terms of cost-effectiveness, studies have shown that CoCM implementation in the Veterans Affairs health system was associated with reductions in mortality and healthcare costs (https://pubmed.ncbi.nlm.nih.gov/31517791; https://pubmed.ncbi.nlm.nih.gov/37016822).Collaborative care models have also reduced emergency and hospital utilization, lowering disease-specific costs (https://pubmed.ncbi.nlm.nih.gov/33877778). When implemented in outpatient general mental health clinics, the implementation of CoCM was associated with significant reductions in inpatient costs (https://pubmed.ncbi.nlm.nih.gov/32732780). Similarly, in a study involving Medicaid beneficiaries with depression, the CoCM led to significantly lower odds of emergency department visits and inpatient medical admissions, contributing to more efficient healthcare utilization (https://pubmed.ncbi.nlm.nih.gov/37221885). The Collaborative Care Model has demonstrated cost-effectiveness across over 80 randomized controlled trials (https://www.psychiatry.org/psychiatrists/practice/profession-al-interests/integrated-care/get-paid/medicaid-payment-and-collab-orative-care-model). An economic analysis found collaborative care could save $3,363 per patient over 4 years compared to usual care (https://blog.proemhealth.com/the-collaborative-care-model-and-integrated-care-7-things-to-know).

Medicaid funding of the Collaborative Care Model

1. Utilizing Medicaid Health Homes

The Affordable Care Act’s Medicaid Health Home State Plan Option under Section 2703 provides a mechanism for states to implement the Collaborative Care Model. Health Homes coordinate primary, acute, behavioral, and long-term care for Medicaid beneficiaries with chronic conditions, including serious mental illness (https://www.chcs.org/media/HH_IRC_Collaborative_Care_Model__052113_2.pdf; https://integrationacademy.ahrq.gov/expert-insight/tips/collaborative-care-model-approach-integrat-ing-physical-and-mental-health-care).

• The Collaborative Care Model aligns well with the required Health Home services like comprehensive care management, care coordination, transitional care, and use of health IT (https://www.chcs.org/media/HH_IRC_Collaborative_Care_Model__052113_2.pdf).
• States have flexibility in defining Health Home services, allowing reimbursement for both in-person and virtual Collaborative Care activities (https://www.psychiatry.org/psy-chiatrists/practice/professional-interests/integrated-care/get-paid/medicaid-payment-and-collaborative-care-model).

2. Leveraging New Payment Models

New payment models are facilitating adoption of the Collaborative Care Model:

• Some Medicaid plans now provide coverage for the Collaborative Care Management (CoCM) billing codes (99492, 99493, 99494) (https://www.psychiatry.org/psychiatrists/practice/professional-interests/integrated-care/get-paid/medicaid-payment-and-collaborative-care-model).
• Federally Qualified Health Centers and Rural Health Clinics can bill for CoCM services using HCPCS code G0512 (https://www.psychiatry.org/psychiatrists/prac-tice/professional-interests/integrated-care/get-paid/medicaid-payment-and-collaborative-care-model).

3. State-Specific Initiatives

Some states are implementing the Collaborative Care Model through legislation, Medicaid waivers, and state-funded programs:

• Washington State’s Mental Health Integration Program is a large-scale collaborative care program in safety-net clinics (https://integrationacademy.ahrq.gov/expert-insight/tips/collaborative-care-model-approach-integrating-physical-and-mental-health-care).
• Minnesota’s DIAMOND program used collaborative care principles to develop patient-centered medical homes (https://integrationacademy.ahrq.gov/expert-insight/tips/collaborative-care-model-approach-integrating-physical-and-mental-health-care).
• Oregon, Arkansas, and Nevada have pursued Medicaid Section 1115 waivers to support integrated care models like the Collaborative Care Model (https://www.psychiatry.org/psychiatrists/practice/professional-interests/integrated-care/get-paid/medicaid-payment-and-collaborative-care-model).

Chapter 8 | Beyond Survival: Supporting Mothers and Children to Thrive

Maternal mortality statistics

The United States has the highest maternal mortality ratio among developed countries, which worsened from 2000 to 2017 as ratios rose from 16.9 to 19.7 maternal deaths per 100,000 live births while globally rates declined (https://www.semanticscholar.org/paper/5ec5d313fe1388c1e341100ff26f1fb2c3e04ebb). Stark racial disparities underlie this crisis, with Black women facing maternal mortality ratios 3 to 4 times higher than white women (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9914526/). In some states, Black women’s mortality is 6 times higher (https://pubmed.ncbi.nlm.nih.gov/32769421/). For American Indian/Alaska Native populations, 2018-2019 mortality ratios reached 30.4 per 100,000, 2.5 times higher than for white women (https://www.semanticscholar.org/paper/5ec5d313fe1388c1e341100ff26f1fb2c3e04ebb). These racial disparities persist even adjusting for socioeconomic factors, reflecting systemic racism’s key role, further exacerbated as Black Americans faced 3 times higher COVID-19 mortality (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9914526/; https://pubmed.ncbi.nlm.nih.gov/32769421/). Further estimates of racial disparities in maternal mortality have been summarized by the Kaiser Family Foundation (https://www.kff.org/racial-equity-and-health-policy/issue-brief/racial-disparities-in-maternal-and-infant-health-current-status-and-efforts-to-address-them/).

Relationships of structural racism and implicit bias to racial disparities in maternal mortality

Multiple studies demonstrate how structural racism and implicit bias drive racial disparities in maternal mortality, especially among Black women. A multi state analysis found Black women had substantially higher adjusted odds of in-hospital mortality or end-organ injury compared to white women (odds ratio 2.5), reflecting systemic factors like structural racism (https://pubmed.ncbi.nlm.nih.gov/37819719). Qualitative research reveals how biases and barriers directly impact Black women’s experiences, including racially insensitive and marginalizing treatment (https://pubmed.ncbi.nlm.nih.gov/33601991).

A systematic review identified six studies linking structural racism to adverse maternal outcomes, underscoring the complexity measuring this impact which varies by factors like population and covariates analyzed (https://pubmed.ncbi.nlm.nih.gov/36401939).

Causes of maternal mortality in the United States

According to recent data, the causes of maternal deaths in the U.S. vary considerably depending on timing. During pregnancy, hemorrhage (11% of deaths) and cardiovascular conditions like high blood pressure (8%), stroke (7%), and other cardiac conditions (15%) are the leading causes. At birth and shortly after, infection accounts for 13% of deaths, the leading single cause during delivery. However, over half of maternal deaths occur in the extended postpartum period. In the 43 weeks following delivery, weakened heart muscle (cardiomyopathy) makes up 11% of deaths, while mental health conditions including substance use and suicide represent a combined 9% of fatalities during this period. The diversity and phases of maternal mortality highlight gaps in care after childbirth, when new mothers have often already been discharged (https://www.common-wealthfund.org/publications/issue-brief-report/2020/dec/mater-nal-mortality-united-states-primer). The CDC has further detailed strategies for prevention of pregnancy-related deaths nationally (https://www.cdc.gov/reproductivehealth/maternal-mortality/pre-venting-pregnancy-related-deaths/state-strategies.html).

The CenteringPregnancy model

The CenteringPregnancy model has demonstrated significant quantitative benefits in terms of health outcomes and cost savings among Medicaid populations in the United States. A study found that participation in CenteringPregnancy was associated with a 36% reduction in the risk of premature birth among Medicaid beneficiaries, leading to substantial cost savings of an average of $22,667 in health expenditures for every premature birth prevented (https://link.springer.com/article/10.1007/s10995-016-1935-y; https://pubmed.ncbi.nlm.nih.gov/26979611/). Additionally, there was a 44% reduction in the incidence of delivering a low birth weight (LBW) infant among participants of the CenteringPregnancy program, with each case averting approximately $29,627 in health expenditures (https://link.springer.com/article/10.1007/s10995-016-1935-y). Infants of CenteringPregnancy participants also had a reduced risk of requiring a neonatal intensive care unit (NICU) stay by 28%, further contributing to the cost-effectiveness of the program (https://link.springer.com/article/10.1007/s10995-016-1935-y). A study conducted inSouth Carolina showed that after considering the state investment of $1.7 million in the CenteringPregnancy program, there was an estimated return on investment of nearly $2.3 million, underscoring the cost-effectiveness of the model in reducing healthcare expenditures for Medicaid (https://link.springer.com/article/10.1007/s10995-016-1935-y). Beyond individual health benefits and cost savings, the CenteringPregnancy model has been shown to improve overall health system outcomes, including higher rates of prenatal care utilization, increased patient satisfaction, and better maternal knowledge of pregnancy and childbirth (https://scholarcommons.sc.edu/cgi/viewcontent.cgi?article=7491&context=etd; https://pn3pol-icy.org/pn-3-state-policy-roadmap-2023/us/group-prenatal-care/).Further details on funding the model at FQHCs have been detailed online (https://www.centeringhealthcare.org/uploads/files/Potential-Payment-Models-for-FQHC-Centering-Programs_pdf.pdf).

Nurse-Family Partnership

The Nurse-Family Partnership (NFP) model, developed by David Olds, Ph.D., is a well-established, evidence-based community health program that has been rigorously evaluated through randomized controlled trials (RCTs) over the past 45 years. These trials have consistently demonstrated significant benefits for first-time mothers and their children, particularly those facing social and economic challenges. The NFP model was tested in three separate RCTs conducted in different locations and with diverse populations: Elmira, NY(1977) with 400 low-income white participants in a semi-rural area; Memphis, TN (1987) with 742 low-income Black participants in an urban setting; and Denver, CO (1994) with 735 low-income Hispanic participants in a different urban environment. The results from these trials have shown substantial improvements across various dimensions, including a 48% reduction in child abuse and neglect, a 56% reduction in emergency room visits for accidents and poisonings, a 50% reduction in language delays by the age of 21 months, 67% fewer behavioral/intellectual problems at age 6, 32% fewer subsequent pregnancies, a 61% reduction in maternal arrests, an 82% increase in the months mothers were employed, and a 59% reduction in child arrests at age 15. The NFP has not only demonstrated immediate benefits but also long-term positive impacts on the lives of participants, with ongoing research by Olds and his team at the Prevention Research Center for Family and Child Health at the University of Colorado, including 14 follow-up studies that track the short- and long-term outcomes of the program participants. To ensure fidelity to the original model, the Nurse-Family Partnership uses a web-based performance management system that helps collect and report data on family characteristics, needs, services provided, and progress toward program goals, ensuring that the program’s implementation aligns with the outcomes observed in the RCTs. The Nurse-Family Partnership stands as a prime example of how evidence-based interventions, validated through the gold standard of randomized controlled trials, can effectively address significant public health challenges, making it a model for preventive programs aimed at improving the health and well-being of vulnerable populations. See further details online (https://www.nursefamilypartnership.org/about/proven-results/).

Impact of Medicaid expansion on maternal mortality and perinatal/postnatal child outcomes

The impact of Medicaid expansion on maternal mortality and prenatal birth outcomes has been a significant area of study, particularly following the implementation of the Affordable Care Act (ACA). The ACA Medicaid expansion has been associated with reduced racial and ethnic disparities in maternal mortality rates, with some studies reporting that expansion states saw a decrease in maternal mortality rates compared to non-expansion states, particularly among minority groups (https://pubmed.ncbi.nlm.nih.gov/34982621/). The extension of postpartum Medicaid coverage has also been linked to a decrease in racial disparities in maternal mortality, with expansion states seeing a decrease in the uninsured rate among women who gave birth in the past year and reduced racial disparities in preterm birth and low birth weight, contributing to a decrease in infant mortality (https://pubmed.ncbi.nlm.nih.gov/33439701/).

However, the evidence on the impact of Medicaid expansion on general birth outcomes such as low birth weight and preterm births is mixed, with some studies finding improvements in these outcomes and others reporting minimal or no significant changes (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10012219/; https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11001670/; https://pubmed.ncbi.nlm.nih.gov/35132151/). Studies focusing on specific conditions, such as gestational diabetes, found that Medicaid expansion led to better maternal and neonatal outcomes in states with ACA implementation compared to those without, including lower rates of adverse outcomes like cesarean sections, low Apgar scores, NICU admissions, and assisted ventilation (https://pubmed.ncbi.nlm.nih.gov/38445784/).

Research analyzing the impact of Medicaid expansion on infant mortality also presented mixed results. A study using CDC’s linked birth/death files from 2011 to 2017 found no statistically significant change in infant mortality in the full sample post-PPACA Medicaid expansion, but when focusing on states that adopted the 2003 birth certificate form and excluding states with a Medicaid waiver, reductions in infant mortality were observed, particularly for babies born to white mothers, while this decrease was not evident for babies born to Black mothers (https://pubmed.ncbi.nlm.nih.gov/32913131/).

A quasi-experimental study on Medicaid expansion and perinatal health outcomes found significant improvements in insurance coverage and access to care for pregnant women, with the percentage of Medicaid-covered deliveries increasing by 2.3 percentage points in expansion states compared with non-expansion states. However, there were no significant changes in the rate of women initiating prenatal care in the first trimester, indicating that while coverage improved, access to early prenatal care did not significantly change (https://pubmed.ncbi.nlm.nih.gov/31599841/).

Safe Start CHW program

The study “Safe Start Community Health Worker Program: A Multisector Partnership to Improve Perinatal Outcomes Among Low-Income Pregnant Women With Chronic Health Conditions” by Cunningham et al. evaluated the impact of the Safe Start program, a community health worker (CHW) initiative aimed at improving peri-natal outcomes for low-income pregnant women with chronic health conditions in Philadelphia, Pennsylvania. The program represents a partnership between a high-volume, inner-city, hospital-based prenatal clinic, a community-based organization (Maternity Care Coalition), a large Medicaid insurer, and a community behavioral health organization. As of June 2019, the program had enrolled 291 women, with CHWs providing services such as accompanying women to prenatal visits, providing health education, facilitating social service referrals, and offering emotional support. The prospective cohort analysis compared Safe Start participants to a control group of 300 eligible women who declined to participate or were not approached due to CHW capacity. The study found that Safe Start participants were significantly more likely to be Black and have hypertension and less likely to report substance use than the comparison group. Controlling for these differences, Safe Start participants had lower odds of inadequate prenatal care (adjusted odds ratio [AOR] = 0.37; 95% confidence interval [CI] = 0.27, 0.53) and antenatal inpatient admissions (AOR = 0.58; 95% CI = 0.35, 0.96)and higher odds of postpartum visit attendance (AOR = 1.47; 95% CI= 1.05, 2.06) and contraception use (AOR = 1.57; 95% CI = 1.06, 2.34) than the comparison group. While there were no differences in rates of neonatal intensive care unit admissions, the length of stay among babies admitted to the neonatal intensive care unit was significantly shorter among babies born to Safe Start participants (adjusted incidence rate ratio = −0.14; 95% CI = −0.23, −0.05; https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7204480/).

International policies related to improved maternal mortality and early childhood outcomes

Several social policy interventions have been shown to improve maternal mortality and early childhood outcomes. Paid parental leave is one such intervention, with evidence showing that it is associated with lower rates of infant mortality, low birth weight, and maternal depression (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7204480/). A study in California found that each additional week of paid leave was associated with a 3.6% decrease in the share of underweight births and a 7.3% decrease in late prenatal care utilization (https://jamanetwork.com/journals/jamapediatrics/ful-larticle/382597). In OECD countries, an additional month of paid maternity leave was associated with a 13% lower infant mortality rate (https://publichealth.jhu.edu/2023/solving-the-black-mater-nal-health-crisis). The estimated cost-benefit ratio for paid parental leave policies ranges from 1.5 to 2.8, indicating substantial returns on investment (https://www.semanticscholar.org/paper/cbcd3b911c90e0365b61dcd64b5cc1c49c8a9819).

High-quality early childhood education programs have also been linked to improved cognitive and social-emotional development, better educational attainment, and higher earnings in adulthood.The Perry Preschool Program, a high-quality preschool program for low-income children, showed a return on investment of $7.16for every dollar spent due to increased earnings, reduced crime, and improved health outcomes. The Abecedarian Project, an intensive early childhood education program, demonstrated significant improvements in educational attainment, employment, and reduced likelihood of criminal behavior, with a cost-benefit ratio of 2.5 to 4.1. Minimum wage increases have been linked to improved birth outcomes, including lower rates of low birth weight and preterm birth.A study estimated that a $1 increase in the minimum wage could lead to a 1.6% decrease in low birth weight and a 0.5% decrease in preterm births, resulting in substantial cost savings for healthcare and long-term economic benefits (https://pubmed.ncbi.nlm.nih.gov/32913131/).

Chapter 9 | Pursuing Universal Coverage: Cautionary Lessons from Medicare-Medicaid Integration and Medicare Advantage

Initiatives to integrate Medicare and Medicaid

Challenges to the experience of dually-eligible patients have been cataloged by Rachel Werner of the University of Pennsylvania (https://thehill.com/opinion/healthcare/4652356-its-time-to-end-the-medicare-medicaid-merry-go-round/amp/).

As of the time of this writing in 2024, a number of states in the United States are actively pursuing initiatives to integrate Medicare and Medicaid services for dual-eligible individuals, with the primary aim of improving care coordination and health outcomes for those who qualify for both Medicare and Medicaid. According to a 2023 issue brief by the Kaiser Family Foundation (KFF), nearly all states are leveraging strategies to coordinate care for dual-eligible individuals, with many states employing multiple strategies to achieve this integration (https://www.kff.org/medicaid/issue-brief/medicaid-arrangements-to-coordinate-medicare-and-medicaid-for-dual-eligible-individuals/). The brief highlights that in 2022, 28 states used Medicaid managed care to cover some or all benefits for dual-eligible individuals, allowing enrollees to receive services through a health plan that coordinates with Medicare, potentially including provisions for paying Medicare cost-sharing or providing additional services (https://www.kff.org/medicaid/issue-brief/medicaid-arrangements-to-coordinate-medicare-and-medicaid-for-dual-eligible-individuals/).

The Center for Health Care Strategies (CHCS) supports states in advancing models that integrate the financing and delivery of services for dual-eligible individuals, providing technical assistance and resources to help states build their capacity for integration, focusing on improving the member and family experience of care, increasing care quality, and reducing costs (https://www.chcs.org/topics/medicare-medicaid-integration/). The integration of Medicare and Medicaid is also facilitated through Dual Eligible Special Needs Plans (D-SNPs), which are Medicare Advantage plans specifically designed to provide targeted care and manage benefits for dual-eligible individuals. The Bipartisan Budget Act of 2018 strengthened Medicare-Medicaid integration requirements for D-SNPs, which must have contracts with state Medicaid agencies that meet specific integration criteria (https://www.cms.gov/medicaid-chip/medicare-coordination/qualified-beneficiary-program/d-snps-integration-unified-appeals-grievance-requirements).

Impact of Medicaid Advantage on cost and quality of care

Payments of taxpayer funds to Medicare Advantage (MA) plans have increased every year since 2015, reaching $12.8 billion in excess payments in 2023 (https://www.kff.org/medicare/issue-brief/medi-care-advantage-in-2023-premiums-out-of-pocket-limits-cost-shar-ing-supplemental-benefits-prior-authorization-and-star-ratings/).

Recent analyses by the Medicare Payment Advisory Commission (MedPAC) have shed light on the growing disparity between Medicare Advantage (MA) payments and the spending that would have occurred in traditional Medicare (TM). In their 2024 report, the independent academic researchers in MedPAC found that coding intensity and favorable selection are causing MA payments to exceed TM spending by a significant margin. The commission projects that in 2024, this disparity will reach its largest yet, amounting to $88 billion in excess spending on MA compared to FFS (https://www.medpac.gov/wp-con-tent/uploads/2023/10/MedPAC-MA-status-report-Jan-2024.pdf).

MedPAC’s analysis also revealed that 52%, or 31.6 million, of Medicare beneficiaries are currently enrolled in MA plans. The report highlighted that MA enrollment is concentrated among a few large firms, potentially limiting competition in the market. This concentration raises concerns about the extent to which beneficiaries can benefit from robust competition among insurers.

Furthermore, MedPAC reiterated its long-standing argument that the quality bonus program in MA is costly and ineffective in judging quality. The commission noted that the program accounts for at least $15 billion in annual MA payments but has serious flaws, such as relying on too many measures, some of which are based on small samples, and not promoting the use of high-value care or providing meaningful information about local plan quality.

MedPAC’s analysis found that MA plans have a stronger financial incentive to code more diagnoses than their FFS counterparts, lead-ing to higher payments and greater inequity across MA plans. From 2020 to 2022, chart reviews and health risk assessments accounted for nearly half of the coding intensity in MA.

The impact of Medicare Advantage (MA) on cost and quality of care compared to traditional Medicare (TM) has been a topic of extensive research, yielding mixed and nuanced results despite the high cost to taxpayers of MA.

A systematic review of peer-reviewed literature published between 2010 and 2020, which included 69 studies, found that 65% of the 273 analyses showed a statistically significant relationship between MA and quality or cost outcomes, with 52% of these analyses favoring MA, with the remainder being neutral or favoring TM (https://pubmed.ncbi.nlm.nih.gov/34533909/).

Several studies have observed no clear advantage of MA over TM. For example, a 2021 study examining the quality of care and outcomes among Medicare patients hospitalized with heart failure found no significant differences between MA and TM in terms of in-hospital mortality (3.3% vs. 3.5%, P = 0.57), 30-day mortality(7.6% vs. 8.1%, P = 0.18), or 30-day readmission rates (21.4% vs.21.8%, P = 0.40) (https://pubmed.ncbi.nlm.nih.gov/32529714/). Similarly, a 2020 study investigating Medicare Advantage and post-discharge quality found no significant differences in 30-day readmission rates between MA and TM (17.2% vs. 17.4%, P = 0.64) (https://pubmed.ncbi.nlm.nih.gov/33877772/).

Furthermore, a 2018 study revealed that MA enrollees were more likely to enter lower-quality nursing homes compared to TM enrollees. The study found that the mean nursing home star rating was3.17 for MA enrollees and 3.30 for TM enrollees (P < 0.001), and MA enrollees had a 2.3 percentage point higher probability of entering a 1-star nursing home (P < 0.001) (https://pubmed.ncbi.nlm.nih.gov/29181535/).

A cross-sectional study of 6,965 low-income adults aged 65 and older found no significant differences between MA and TM in most measures of healthcare access, preventive care use, and affordability. For instance, the study reported no significant differences in the proportion of beneficiaries who delayed care due to cost (8.1% in MAvs. 7.4% in TM, P = 0.51), had difficulty paying medical bills (11.9%in MA vs. 10.4% in TM, P = 0.28), or received an influenza vaccination(59.4% in MA vs. 58.3% in TM, P = 0.61) (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9175080/).

Medicare Advantage risk score gaming

Two recent Health Affairs articles expose how Medicare Advantage (MA) plans have been gaming the risk adjustment system to inflate payments from the Centers for Medicare and Medicaid Services (CMS). MA plans are paid using a bidding process based on risk scores that estimate members’ expected healthcare costs, with higher risk scores yielding higher payments from CMS (https://www.healthaffairs.org/content/forefront/medicare-advantage-direct-con-tracting-and-medicare-money-machine-part-1-risk-score-game; https://www.healthaffairs.org/content/forefront/medicare-advan-tage-direct-contracting-and-medicare-money-machine-part-2-building-aco).

MA plans have been observed to artificially increase risk scores by adding more diagnosis codes during routine office visits, a practice known as “coding intensity” or “risk score gaming.” A 1% rise in risk scores can trigger a $9 billion increase in federal spending on MA plans. From 2004-2013, coding intensity grew at 6.8% peryear, outpacing the rate of medical inflation. A recent study found that 35.5-46.5% of incremental risk scores from health risk assessments and chart reviews were not accompanied by actual increased resource use, suggesting intensive coding not justified by patient needs (https://pubmed.ncbi.nlm.nih.gov/37542373/).

To maximize profits from coding intensity, MA plans acquired physician practices and built vertically integrated Accountable Care Organizations (ACOs). Owning ACOs allows MA plans to directly employ and control physicians, pressuring them to upcode patient diagnoses or face termination. From 2016-2020, the number of physicians vertically integrated into MA plans grew by 49%.

The consequences of unchecked coding intensity are overpayments to MA plans. The Government Accountability Office estimates MA plans were overpaid $12 billion in 2020 due to coding practices.From 2007-2017, MA risk scores increased 19.2% while fee-for-service Medicare scores only rose 5.1%. Despite CMS attempting risk adjustment to account for coding intensity, MA plans continue finding loopholes to game the system. From 2018-2020, MA coding intensity grew 3.2%.

Despite evidence of gaming by MA plans, CMS has failed to properly audit them or recoup overpayments. CMS relies on MA plans to self-report their own coding errors, creating a conflict of interest. While the Department of Justice has intervened in whistleblower lawsuits alleging over $1 billion in overpayments to specific MA plans, systemic reforms to MA payment models and oversight remain lacking.

Recent studies highlight the need for legislative and regulatory action. One analysis concludes regulatory approaches to improve risk adjustment and recoup overpayments have the highest potential impact (https://pubmed.ncbi.nlm.nih.gov/37497876/).