Nearly all states adopted some form of social distancing measures during the last couple of months in response to the coronavirus pandemic.

These measures are the primary means of controlling the spread of a highly communicable virus and avoiding health care system overload by limiting close contact between people in the absence of a vaccine or curative treatment. While there has been broad public support of strict stay at home orders and other social distancing measures to protect public health, there also have been protests alleging these measures impermissibly restrict individual rights in Wisconsin, Michigan, California, and elsewhere, and lawsuits challenging them filed in a number of states.

Map accurate as of June 1 – For updated information, see tracker here.

This issue brief explains the legal basis for stay at home orders and provides an overview of current legal challenges filed in state and federal courts. Historically, the government’s authority to impose measures that temporarily restrict individual rights during emergencies has been recognized as necessary to protect public health, even when those measures limit rights protected by the Constitution that Americans hold important, such as the rights to free speech, peaceful assembly, interstate travel, and free exercise of religion. Although the Wisconsin Supreme Court’s decision striking down nearly all of that state’s stay at home order, based on its interpretation of state law, received recent media attention, most courts to date generally have allowed stay at home orders to remain in place. While some of the specific issues in current lawsuits could become moot as states start to re-open, the general issue about the scope of government authority to limit individual rights in a public health crisis is likely to return to the courts as states re-institute stay at home orders during future COVID-19 outbreaks in the coming months.

What is the legal basis for stay at home orders?

Mandatory social distancing measures during public health emergencies, such as stay at home orders, are based on states’ general authority to protect the general health, safety, morals, and welfare, known as the police power. The police power is a very broad power through which governments regulate individual rights to protect the interests of society as a whole. Common examples of the police power are safety regulations to reduce fire hazard, zoning laws that regulate land use, and laws prohibiting gambling or prostitution. Examples of social distancing measures adopted under the police power in the current coronavirus pandemic include mandatory stay at home orders, mandatory traveler quarantines, closures of non-essential businesses, bans on large gatherings, school closures, and limits on bars and restaurants and other public places.

Because social distancing measures are adopted under states’ police power, there is state variation in the onset of social distancing measures, the specific measures adopted, and their duration. The police power is the main power possessed by state governments. The federal government has limited powers that are specially enumerated in the Constitution. All remaining government powers, including the general police power, are vested in the states under the 10th Amendment.

The U.S. Supreme Court’s recognition of state use of the police power to regulate individual rights in the interest of protecting public health dates back to 1905. In Jacobson v. Massachusetts, the Court upheld a government requirement for smallpox vaccination when that virus was spreading. The Court set out the legal test, still applied today, which provides that a state’s exercise of the police power to promote public safety during a public health emergency will be upheld unless the order has no real or substantial relation to public health or the measure beyond all question is a plain palpable invasion of fundamental rights.

In upholding the smallpox vaccination requirement, the Court emphasized that a free society requires some restrictions on individual rights to protect public health and safety and promote general social order. The Court noted that, “Upon the principle of self-defense, of paramount necessity, a community has the right to protect itself against an epidemic of disease which threatens the safety of its members.” Rather than being absolute in all circumstances, individual constitutional rights are subject to “manifold restraints to which every person is necessarily subject for the common good.” This balance is necessary because: “Society based on the rule that each one is a law unto himself would soon be confronted with disorder and anarchy. Real liberty for all could not exist under the operation of a principle which recognizes the right of each individual person to use his own, whether in respect of his person or his property, regardless of the injury that may be done to others.”

How have courts decided lawsuits challenging stay at home orders issued in response to the COVID-19 pandemic to date?

Most courts to date generally have allowed stay at home orders issued during the current crisis to remain in place to protect public health, despite restrictions on individual rights such as free speech, peaceful assembly, travel, and free exercise of religion. Some courts also have been asked to decide disputes between state legislatures and Governors about the exercise of government power to issue these orders. Selected cases are described below and summarized in Table 1.

Free speech, peaceful assembly, and economic harm

In a case that reached a final decision on the merits, Friends of Danny Devito v. Wolf, the Pennsylvania State Supreme Court upheld the Governor’s closure of physical operations of all non-life-sustaining businesses to reduce coronavirus spread. The Governor’s order was challenged by a realtor, a golf course owner, and a candidate for the state legislature (no relation to the famous actor), who alleged that the order violated their rights to free speech, free assembly, and due process. The Pennsylvania state court found the order to be reasonably necessary to meet the current public health emergency, which involves a virus that spreads “exponentially” and where enforcement of social distancing is currently the only mitigating tool. In this context, the court noted that temporary business closure is not unduly oppressive when faced with protecting the lives and health of all state citizens. The court also observed that the power to enact laws to promote the general welfare is one of the most essential government powers and its “least limitable.” In early May 2020, the U.S. Supreme Court denied the business owners’ application for a stay, allowing the Governor’s order to remain in effect while the business owners ask the Supreme Court to review the case.

Other courts have allowed stay at home measures to remain in place while lawsuits are pending. A California federal district court denied a temporary restraining order in a case brought by a gun shop owner and political candidate who sought to hold public protests and rallies at the state capital, deferring to state officials’ “informed efforts to protect all citizens especially the most vulnerable against a deadly pandemic.” Another federal district court denied a request to block Maryland’s stay at home order in a lawsuit brought by a number of plaintiffs, including members of the state legislature who wanted to attend political rallies and businesses that had lost substantial money because they were deemed non-essential and forced to close. The Maryland court found that, “The executive orders at issue have a ‘real or substantial relation’ to the public health crisis.” The Maryland plaintiffs have appealed that decision to the 4th Circuit Court of Appeals. Other federal lawsuits have been brought by abortion protestors, after they were arrested for violating stay at home orders in North Carolina and Michigan; neither case resulted in a decision on the merits because the orders expired or were amended, making them moot.

Travel Restrictions

A Michigan state court refused to issue a preliminary injunction in a case challenging the scope of the Governor’s stay at home order. The plaintiffs, five Michigan residents, claimed that the intrastate travel restrictions violated their right to procedural due process. The court found that while the order severely restricted individual constitutional rights, the restrictions were reasonable in light of the current public health crisis. The court also emphasized the temporary nature of the restrictions, compared to the “all too permanent” impact on “those who contract the virus and cannot recover and their family and friends.” In contrast, a Kentucky federal court struck down a provision in the Governor’s order requiring individuals to self-quarantine for 14 days after entering the state in a case brought by state residents who wanted to travel freely in and out of state. The Kentucky court found that the requirement was not narrowly tailored to achieve the government’s purpose because it applied to an individual visiting a friend eight miles away in Ohio, but not if the visit occurred eight miles away within Kentucky.

Free exercise of religion

One of the most politically charged provisions of the stay at home orders has been restrictions on public gatherings as they apply to religious worship. There have been multiple examples from around the world of religious services acting as super-spreader events for COVID-19. Federal appeals courts have split on the issue of whether social distancing orders should be lifted or remain in place while lawsuits alleging violations of the free exercise of religion are pending. In the two religious cases seeking U.S. Supreme Court review to date, the Court has allowed social distancing orders to remain in place while the appeals are pending. In the litigation challenging restrictions on in-person religious services, the churches and other houses of worship contend that religious services are being singled out and not being treated equally with other business operations. As states begin to re-open, some lawsuits are alleging that churches should be included in earlier rather than later re-opening phases.

On May 22nd, President Trump declared places of worship “essential” operations that should hold services in person regardless of state orders. While the President lacks the authority to override state orders, many states are starting to open up and allow some in person religious services. The CDC issued specific guidelines for faith based worship services, which it later amended to emphasize religious rights (Box 1). The amended CDC guidelines now provide that faith based organizations should be treated the same as similarly situated entities. The factual question is what kind of business is the correct comparison for religious services: movie theaters, live concerts, grocery stores or office buildings. In a concurring opinion, denying the request from a California church to allow churches to be treated like other businesses allowed to open, Chief Justice Roberts noted that, “Similar or more severe restrictions apply to comparable secular gatherings, including lectures, concerts, movie showings, spectator sports, and theatrical performances, where large groups of people gather in close proximity for extended periods of time.” In his dissent Justice Kavanaugh asserted, “California’s latest safety guidelines discriminate against places of worship and in favor of comparable secular businesses. Such discrimination violates the First Amendment.”

In two cases appealed from the 9th and 7th Circuit Court of Appeals, the Supreme Court denied requests brought by churches in California and Illinois to block state stay at home orders while the appeals are pending. The California church contended the Governor’s decision to place churches in Stage 3 instead of Stage 2 of the state reopening plan violates their right to free exercise of religion. The 9th Circuit upheld a California federal district court’s denial of a temporary restraining order. The court stated:

“Where state action does not ‘infringe upon or restrict practices because of their religious motivation’ and does not ‘in a selective manner impose burdens only on conduct motivated by religious belief,’ it does not violate the First Amendment. We’re dealing here with a highly contagious and often fatal disease for which there presently is no known cure. In the words of Justice Robert Jackson, if a ‘[c]ourt does not temper its doctrinaire logic with a little practical wisdom, it will convert the constitutional Bill of Rights into a suicide pact.’ “

In a 5-4 vote the Supreme Court denied the California church’s request for an emergency stay of the state’s restriction on worship services. In a concurring opinion, Chief Justice Roberts emphasized that the Constitution delegated the decisions to keep states safe to local politicians: “The precise question of when restrictions on particular social activities should be lifted during the pandemic is a dynamic and fact-intensive matter subject to reasonable disagreement” but entrusted by the Constitution to politically accountable  elected officials of the state.

Like the California church, the Illinois church contended that the state’s reopening plan imposed unique limits on religious worship services that are not imposed on other businesses. After the Illinois church appealed the 7th Circuit decision allowing the social distancing measures to stay in place, the Illinois Department of Public Health issued new guidelines allowing in-person worship services, with recommendations for maximum capacity and social distancing. The Supreme Court denied the request for a stay from the IL church because the new guidelines made the request for a stay moot.

Similarly, in a case brought in state court by 10 churches alleging violation of their religious rights, the Oregon State Supreme Court stayed a temporary restraining order issued by a lower state court. The Oregon State Supreme Court is expected to rule on the merits of the case soon.

Other federal district courts in California and New Mexico also have denied requests for temporary restraining orders in cases challenging state social distancing orders brought by churches. These courts found that the orders did not violate the right to free exercise of religion because the orders are neutral, generally applicable, and related to addressing the pandemic. For example, the California federal district court noted that while the order temporarily burdens individual rights to a greater degree than normally permissible, the right to free exercise of religion does not include the liberty to expose the community to a communicable disease. The New Mexico district court observed that when the state faces a major public health threat, its police powers are at a maximum.

In contrast to the above decisions, the 5th and 6th Circuit Courts of Appeals have lifted bans on public gatherings in cases alleging infringement on the right to free exercise of religion. The 5th Circuit Court of Appeals enjoined enforcement  of the City of Holy Spring, Mississippi’s stay at home order, while the district court decides the merits of the case. The case was brought by a church alleging that the city’s designation of churches as “nonessential” violates the rights to free exercise of religion and free speech. The church did provide assurances that it will satisfy the “requirements entitling similarly situated businesses and operation to reopen.”  In two related cases in Kentucky, the 6th Circuit Court of Appeals held that the public gathering limitations violated the plaintiffs’ right to free exercise of religion under KY’s Religious Freedom and Restoration Act and the First Amendment. In the case brought by Maryville Baptist Church, while the 6th Circuit allowed the plaintiffs to hold drive-in services, the district court subsequently allowed in-person services that abide by social distancing requirements. Similarly, a North Carolina federal district court issued a temporary restraining order in a challenge brought by a church, stating, “There is no pandemic exception to the Constitution of the United States or the Free Exercise of the First Amendment.” The plaintiffs later voluntarily dismissed their case after the Governor issued a new executive order exempting all religious and spiritual gatherings from the public gathering limitations.

Conflicts Between State Legislatures and Governors

In cases disputing the scope of authority of the executive branch vs. the legislature under state law to impose social distancing measures, the Wisconsin State Supreme Court lifted nearly all of the state public health secretary’s stay at home order, while the Kansas State Supreme Court upheld the Governor’s order. While these cases each turn on specific provisions of state law, they also speak to the larger policy and political debate about the extent of executive authority to limit individual rights during a public health crisis. The Wisconsin decision is the only case thus far to lift nearly all provisions of a stay at home order. The case was brought by the state legislature, contesting the state public health secretary’s authority to issue an extended stay at home order without going through an administrative rule making process. While most courts have emphasized that stay at home orders serve to protect people during a public health emergency, the Wisconsin state court in a four to three decision instead highlighted the importance of the administrative rulemaking process to “safeguard all people” (Box 2). The Kansas Supreme Court allowed the Governor’s executive order limiting religious gatherings to 10 people to remain in place, finding that a seven member Legislative Coordinating Council does not have the power to overrule the Governor. The Governor brought suit after the Council voted to revoke her order. Both of these cases involved opposing political parties in control of different branches of state government.

Following the Wisconsin decision, the Republican controlled Michigan state legislature sued the Democratic Governor challenging her stay at home order. The Michigan state district court found that the Governor does have the authority to issue stay at home orders under state law. The Legislature has appealed this decision and is seeking emergency-bypass review by the state Supreme Court. In another case, Republican members of the Illinois House of Representatives are challenging the Democratic Governor’s stay at home orders alleging the Governor has exceeded the his authority under the state Emergency Management Agency Act. The plaintiffs filed the case in state court, but the Governor filed the paperwork to move the case to federal court. The plaintiffs are contesting the move to federal court. The Trump Administration issued a statement of interest in the case, supporting the plaintiffs’ position to keep the case in state court.

Looking Ahead

Adoption of social distancing measures continues to be a subject of policy as well as political debate. Courts generally have upheld these measures in the interest of protecting public health during emergencies, despite acknowledging the restriction on individual rights, in cases brought by individuals and groups seeking to fully exercise those rights. Three cases to date have asked for Supreme Court review. The Supreme Court has denied stays in all three cases in which plaintiffs have appealed to block state stay at home orders.  Other cases are pending review at district courts and courts of appeals. While state and local actions to open up might make some of the current litigation moot, states may decide to re-instate some of these orders if there are new flares in COVID-19 over the coming year, and we can expect similar legal challenges.

As states recently have begun to roll back some social distancing measures, public opinion on these changes also differs somewhat by political party affiliation. According to the May 2020 KFF Health Tracking Poll, about half of adults – across political party identification – say they think their state is moving at the right speed in easing restrictions. However, more than three times as many Republicans (32%) as Democrats (9%) say their state is opening “too slowly,” and one-fifth (22%) of independents say the same. Four in ten Democrats (39%) say their state is opening “too quickly” compared to 28% of independents and 13% of Republicans. States’ scattershot approach to adopting stay-at-home orders raised concerns about ongoing community transmission in the U.S., while recent movements by some states to ease restrictions similarly raise concerns about new outbreaks. In the coming months and years, many policymakers and other stakeholders will be studying the economic and public health consequences that resulted from varying state approaches to balancing public health vs. individual rights and weighing the effects on the economy.

In May 2020, the Trump Administration announced the number of Medicare Part D plans that will be participating in 2021 in a Medicare Innovation Center model to address out-of-pocket costs for insulin products for Part D enrollees. Under the voluntary model, participating Part D plans can offer coverage of insulin for a flat monthly copayment of no more than $35, in contrast to varying cost-sharing amounts during different phases of the Part D benefit under the current design. The model comes in response to rising prices for insulin, which have attracted increasing scrutiny from policymakers, leading to congressional investigations and overall concerns about affordability and access for people with diabetes who need insulin to control blood glucose levels.

Among people with Medicare (including people 65 and older and younger adults with long-term disabilities), one third (33%) had diabetes in 2017, up from 18% in 2000. The rate of diabetes in the overall population has been trending up, and is highest among people 65 and over. Being overweight, having high blood pressure or high cholesterol, older age, and physical inactivity are among the risk factors for type 2 diabetes (by far the most common type). Though not all people with diabetes need insulin, for those who do, it is a life-saving medication and essential to maintaining good health.

In light of heightened attention to insulin and the Trump Administration’s new Part D model to address out-of-pocket costs for insulin for Medicare beneficiaries, we analyzed out-of-pocket spending on insulin by beneficiaries enrolled in Part D drug plans, variation in Part D plan formulary coverage and tier placement of insulin products, and trends in prices for insulin (not accounting for rebates), based on data from the Centers for Medicare & Medicaid Services (CMS) (see Data and Methods for details). We also discuss the Administration’s Part D demonstration, which will take effect in 2021, and policy proposals introduced in the 116^th Congress that address cost concerns related to insulin.

Key Takeaways

• Aggregate out-of-pocket spending by people with Medicare Part D for insulin products quadrupled between 2007 to 2017 from $236 million to $984 million, reflecting both an increase in the number of users and price increases for insulin.
• Among insulin users without Part D low-income subsidies (LIS), average annual per capita out-of-pocket spending on insulin increased by 79% over these years, from $324 in 2007 to $580 in 2017. Average annual growth in costs was 6%, which exceeded the 1.6% average annual rate of growth in inflation over this time period.
• If Part D enrollees had paid 12 months of $35 copays for insulin in 2017, as under the Administration’s new Part D model, annual costs for one insulin product would have been $420, or $160 (28%) lower than average annual costs paid by non-LIS Part D insulin users in 2017.
• Formulary coverage and tier placement of insulin products vary across Part D plans, but in 2019, a large number of Part D plans placed insulin products on Tier 3, the preferred drug tier, which typically had a $47 copayment per prescription during the initial coverage phase. However, once enrollees reach the coverage gap phase, they face a 25% coinsurance rate, which equates to $100 or more per prescription in out-of-pocket costs for many insulin therapies, unless they qualify for LIS. Paying a flat $35 copayment rather than 25% coinsurance during the coverage gap phase could translate to meaningful savings on many insulin products.
• Among the 22 insulin therapies that have been on the market since at least 2013, 16 products had average annual price increases of more than 10% between 2013 and 2018, as measured by the change in total Part D spending per dosage unit between these years, which far exceeded the 1.5% rate of inflation over the same time period. Although CMS’s Part D pricing data does not account for rebates, it is relevant in considering out-of-pocket costs for people who use insulin, because in some cases what people pay out of pocket is based on the list price.
• The Administration’s new model that allows participating Part D plans to cap monthly out-of-pocket insulin costs could help some beneficiaries with their insulin costs, if they are enrolled in a participating Part D plan and use an insulin product covered by their plan with a $35 copayment. But the new model applies to a subset of plans and enrollees (those without low-income subsidies in enhanced plans, which cover 80% of non-LIS Part D enrollees in non-employer Part D plans in 2020), and not all insulin products have to be covered by all participating plans. The new model also does not address underlying list price increases for insulin or affordability concerns for people who are uninsured or covered by other sources of coverage.

In addition to the Administration’s new Part D model, several bills have been introduced in the 116^th Congress that aim to lower insulin out-of-pocket costs, and prices, including proposals to cap or eliminate Medicare beneficiaries’ monthly out-of-pocket costs for insulin and insulin supplies, allow Medicare to negotiate the price of insulin (and other) products, let the government manufacture insulin products directly, allow importation of insulin, use international reference pricing for insulin, and limit drug price increases to the rate of inflation.

Out-of-Pocket Spending on Insulin by Medicare Part D Enrollees

Aggregate Out-of-Pocket Spending for Insulin in Part D

Aggregate out-of-pocket spending by people with Medicare Part D for insulin products quadrupled between 2007 to 2017, from $236 million to $984 million (Figure 1). The number of Medicare Part D enrollees using insulin nearly doubled over this time period (from 1.6 million to 3.1 million beneficiaries), which indicates that higher aggregate out-of-pocket spending was not solely a function of more Medicare beneficiaries using insulin.

Average Annual Per Capita Out-of-Pocket Costs for Insulin

Annual out-of-pocket spending on insulin products among Medicare Part D insulin users who do not receive low income subsidies was $580 in 2017, a 79% increase since 2007 ($324). Average annual growth in insulin out-of-pocket costs was 6% over these years, which far exceeded the 1.6% average annual rate of growth in inflation over this time period.

Some enrollees spent considerably more than the average (Figure 2). For example, 10% of non-LIS insulin users spent nearly $1,500 on insulin in 2017 and 1% spent over $2,600. Higher-than-average out-of-pocket spending is due to a greater number of prescription fills for insulin products and higher out-of-pocket costs per insulin prescription. In other words, taking more than one insulin product and taking newer, more expensive formulations leads to higher out-of-pocket costs.

Out-of-Pocket Spending Per Insulin Prescription

Among Medicare Part D insulin users who do not receive low-income subsidies, out-of-pocket costs per prescription increased by 50% across insulin products available in both 2007 and 2017, from $39 to $57 per prescription, on average (Figure 3). For all insulin products available in 2017, out-of-pocket spending per prescription by non-LIS Part D enrollees ranged from $20 for Humulin N, an intermediate-acting insulin, to $128 for Humulin R U-500, a short-acting concentrated insulin (Table 1).

For some products, the percentage increase in out-of-pocket costs per prescription between 2007 and 2017 was higher than average, including an 87% increase for Novolin N and a 63% increase for Novolin R, both short-acting insulin products. In terms of actual out-of-pocket costs per prescription, these were among the lower-cost insulin products in both 2007 and 2017, and less commonly used than several other insulin products (Table 1).

Part D Formulary Coverage and Tier Placement of Insulin

Formulary Coverage and Tier Placement of Insulin Therapies

Formulary coverage and tier placement are key factors determining how much Medicare Part D enrollees without low-income subsidies pay out of pocket for their prescriptions, including insulin. In 2019, most insulin products were covered by at least 50% of all Part D plans, including both stand-alone prescription drug plans and Medicare Advantage drug plans (Table 2). Although coverage of insulin products was not universal, as for drugs in the so-called protected classes, it exceeds the minimum requirement that plans cover at least two chemically distinct drugs in each class.

Even when insulin products are covered by Part D plans, there is variation in tier placement and cost-sharing requirements when enrollees are in the initial coverage phase. A large number of plans placed insulin products on Tier 3, the preferred drug tier, with a $47 copayment per prescription during the initial coverage phase; more plans used this combination of tier placement and cost-sharing requirement for coverage of insulin than any other combination in 2019 (Table 3, Table 4).

This means that for Part D enrollees, the typical out-of-pocket cost for insulin when they are in the initial coverage phase is not pegged to the list price, as it would be with coinsurance. But some plans charged coinsurance for Tier 3, with a median coinsurance rate 25%, and a handful of plans placed some insulin products on Tier 4 (the non-preferred drug tier), with a median coinsurance rate of 48%. These coinsurance rates translate into varying out-of-pocket costs for insulin users depending on the underlying list price, as described below.

Variation in Out-of-Pocket Spending for Insulin Products by Tier Placement and Cost-Sharing Requirements

Because some insulin products are considerably more expensive than others, differences in tier placement and cost-sharing requirements can translate to wide variation in out-of-pocket spending for any given insulin therapy and from one insulin therapy to another (Figure 4).

For example, Xultophy has the highest median total cost for a 30-day fill among all pen injector insulin products, at more than $1,000, while Basaglar has the lowest, at $334. A beneficiary enrolled in a Part D plan that placed these drugs on Tier 4 with the most common coinsurance requirement (48%) would pay more than $500 per prescription for Xultophy in the initial coverage phase of the Part D benefit, which is three times what they would pay for Basaglar ($160). Both amounts are more than they would pay for these products if their plan charged a $100 copayment for Tier 4 drugs, and substantially higher than if the drugs were covered on Tier 3 with a $47 copayment.

Variation in Out-of-Pocket Spending for Insulin Products by Part D Benefit Phase

A distinctive feature of the Part D benefit is that out-of-pocket costs paid by non-LIS enrollees may not be uniform for each prescription fill from one month to another, depending on which phase of the benefit they are in. The Part D defined standard benefit has several phases where cost sharing for enrollees varies, including a deductible, an initial coverage phase, a coverage gap phase, and catastrophic coverage. Between the initial coverage phase and the coverage gap phase, cost sharing paid by non-LIS enrollees for any given drug switches from however much a plan charges for that drug based on its specific formulary design to a uniform 25% coinsurance rate, followed by a uniform 5% coinsurance rate that non-LIS enrollees pay in the catastrophic coverage phase. Based on the underlying list price of a medication, enrollees’ costs can increase substantially when they move from the initial coverage phase to the coverage gap phase.

For insulin products, this variation in cost sharing translates to varying out-of-pocket costs from one benefit phase to another. Figure 5 shows out-of-pocket costs for a selection of insulin products where in the initial coverage phase, the most common tier placement in 2019 was Tier 3 with a $47 copayment. For all of these products, out-of-pocket costs per prescription increase to $100 or more during the coverage gap phase, when cost sharing is based on a 25% coinsurance rate. For example, for Levemir, an enrollee would pay $47 in the initial coverage phase, but during the coverage gap phase (after total drug costs exceeded $3,820 in 2019), their out-of-pocket costs would rise to $226 per prescription, which is 25% of the $906 list price of Levemir. Out-of-pocket costs would decrease to $45, or 5% of the list price, during the catastrophic phase (after annual out-of-pocket spending exceeded $5,100 in 2019).

Trends in Insulin Prices

Among the 22 insulin therapies that have been on the market since at least 2013, 16 products had average annual price increases of more than 10% between 2013 and 2018, as measured by the change total Part D spending per dosage unit between these years, which far exceeded the 1.5% average annual growth in inflation over the same time period (Figure 6).

It is important to note that this spending measure, and total spending as reported in Part D claims, do not account for rebates, which are not publicly available at the individual product level. The Government Accountability Office has estimated rebates of 41% in 2016 for drugs in the therapeutic class that includes insulin. This is higher than the estimated 20% rebate for Part D covered drugs overall in 2016. While gross total Part D spending on insulin increased from $1.4 billion in 2007 to $13.1 billion in 2017, net total spending would be significantly lower after accounting for rebates.

The level of rebates offered by manufacturers for specific drugs influences coverage and tier placement decisions by Part D plans, which in turn affects out-of-pocket costs paid by enrollees. It has been suggested that the magnitude of list price increases for insulin products over time can be attributed to manufacturers offering increasingly larger rebates for these drugs in exchange for preferred formulary coverage.

Although CMS’s Part D pricing data does not account for rebates, it is relevant in considering out-of-pocket costs for people who use insulin, because in some cases what people pay out of pocket is based on the list price. As a result, percentage increases in prices for insulin products can translate to higher out-of-pocket spending over time.

Proposals to Address Insulin Costs for Medicare Beneficiaries

Innovation Center Model

In March 2020, the Administration announced a voluntary Innovation Center model, the “Part D Senior Savings Model,” that allows participating Part D plans to offer coverage of insulin for a flat monthly copayment of no more than $35 for enrollees not receiving low-income subsidies (LIS), starting in 2021. (Most LIS enrollees already pay a much lower copayment of no more than $8.95 for brands in 2020, so the model does not apply to them.) Participating Part D plans – enhanced plans only, not basic plans – can offer insulin products from participating manufacturers at a flat $35 copayment during all phases of the benefit (deductible, initial coverage, coverage gap, and catastrophic). Compared to basic Part D plans, enhanced plans can offer a lower deductible (the standard is $435 in 2020), reduced cost sharing (under the basic standard benefit, cost sharing is a uniform 25% coinsurance or actuarially equivalent amounts), and/or a higher initial coverage limit (the standard limit is $4,020 in 2020).

CMS recently announced that more than 1,750 enhanced Part D plans will participate in the model in 2021; for reference, this equals roughly 50% of all enhanced plans in 2020, including both Medicare Advantage and stand-alone drug plans and roughly 40% of all Part D plans this year. It is not yet known how many Part D plans of each type will be offered in 2021, or how many plans in total.

Each of the three major insulin manufacturers – Eli Lilly, Sanofi, and Novo Nordisk – has agreed to participate in the model for 2021. Participating manufacturers will continue to provide a 70% price discount in the coverage gap, but the discount will be calculated before the application of the supplemental benefit (i.e., the $35 monthly copay for insulin). Under current law, the manufacturer discount is calculated after the application of the supplemental benefit if a plan lowers cost sharing in the coverage gap. This creates a disincentive for plans to offer supplemental benefits in the coverage gap because plan liability increases and manufacturer liability decreases.

For example, for a drug that costs $500, under current law in a plan with no supplemental benefit, the manufacturer discount is $350 (70% of $500), beneficiaries pay $125 (25% of $500), and plans pay $25 (5% of $500) (see table below). If, instead, the plan charged a flat $35 copayment in the coverage gap, the plan’s total liability increases to $465 ($500-$35), and the manufacturer 70% discount is calculated off of the $35 remainder (i.e., after the application of the supplemental benefit). The Part D model modifies this calculation, thereby reducing the amount plans pay and increasing liability for manufacturers.

Under the model, CMS expects that there will be more insulin claims filled in the coverage gap phase and fewer enrollees passing into the catastrophic coverage phase. This change, plus the changes to plan and manufacturer liability for insulin claims in the coverage gap, translates into an estimated $250 million in additional coverage gap discounts over five years, and similar savings for Medicare.

As an additional financial incentive for plans to participate in the model, participating plans can also receive financial protection through a narrower risk corridor, which would reduce potential losses if a plan’s actual costs are higher than expected. Currently, plans bear full liability for actual costs up to 5% higher than expected costs, beyond which higher costs are paid by both plans and the government. The model reduces the initial risk corridor threshold to 2.5% for participating plans that have a significantly greater number of insulin-dependent diabetics enrolled relative to other similar plan types. Higher costs can also be passed on to enrollees in the form of higher premiums, and CMS has estimated that the model will in fact increase enhanced plan premiums by approximately $1 per month due to higher plan liability.

Implications for Part D Enrollees

Based on our analysis of actual spending in 2017, we found that average annual out-of-pocket costs by Part D insulin users not receiving low-income subsidies was $580. If instead they had paid 12 months of $35 copays, as under the Administration’s model, annual costs would have been $420 for one insulin product, or $160 (28%) lower than average annual costs paid by non-LIS Part D insulin users in 2017. Paying a flat $35 copayment during the coverage gap phase could translate to meaningful savings on many insulin products, with typical list prices in the hundreds of dollars, which equates to out-of-pocket costs of $100 or more under the current benefit design where beneficiaries pay 25% coinsurance in the coverage gap phase.

Two key features of the model will limit its scope, however. The model relies on voluntary plan participation (though participation seems relatively robust for 2021), and only enhanced Part D plans can participate (though most enrollees – 80% of non-LIS Part D enrollees in non-employer plans – are in these enhanced plans in 2020). If Part D enrollment in 2021 is similar to 2020, one-fifth of non-LIS Part D enrollees (those who are in basic plans) would not stand to benefit from the model unless they switch from a basic to an enhanced plan, and some additional share of current enhanced plan enrollees who are in plans that have chosen not to participate would need to switch into a participating plan. Prior work shows that most enrollees do not switch during the open enrollment period, though it is possible that insulin users would be more willing switch plans if joining a participating plan would lower their out-of-pocket costs.

Another potential concern is that enhanced Part D plans have higher monthly premiums than basic Part D plans, and as mentioned earlier, CMS expects that participating enhanced plans will have even higher premiums based on higher plan costs. On average, the monthly premium for enhanced stand-alone drug plans in 2020 is $47, compared to $30 for basic stand-alone drug plans – or more than 50% higher. This translates to more than $200 in higher annual premiums, on average. Based on these average premiums amounts and our estimate of possible savings based on 2017 out-of-pocket costs, the higher premium for an enhanced plan could fully offset savings from lower out-of-pocket costs for one insulin product (though beneficiaries could see other savings in these enhanced plans, such as a lower deductible).

A final concern is that participating Part D plans do not have to cover all insulin products at the $35 monthly copayment amount, just one of each dosage form (vial, pen) and insulin type (rapid-acting, short-acting, intermediate-acting, and long-acting). For insulin products not covered by a participating plan, users of these products would not benefit from the model unless they switch to a product from a competing manufacturer. Making this switch could be difficult for insulin users who have been stabilized on a particular insulin medication (or medications) and for whom consistency in treatment is important in managing diabetes.

Legislative Proposals

Several bills have been introduced in the 116^th Congress aiming to curb insulin costs for consumers. Some of these proposals target insulin directly, while others address drug prices as a whole and would therefore impact insulin prices. Proposals that target out-of-pocket costs specifically for insulin under Medicare Part D include the Lower Costs, More Cures Act (Rep. Walden, R-OR), which includes a provision to impose a $50 cap on monthly out-of-pocket costs incurred by Medicare beneficiaries for insulin and insulin supplies, and the Insulin Access for All Act (Rep. Rush, D-IL), which would eliminate all cost sharing for insulin and related supplies for Medicaid and Medicare beneficiaries.

Some current proposals rely on strategies that could help control prescription drug costs more generally to target insulin prices specifically. These strategies include importation, international reference pricing, and manufacturer rebates.

• The Affordable Insulin Act (Rep. Welch, D-VT) allows individuals, pharmacies, and wholesalers to import insulin from Canada and other qualifying countries.
• The End Price Gouging for Insulin Act (Sen. Merkley, D-OR) authorizes the government to establish reference prices for insulin products sold under any federal health program, including Medicare and Medicaid, based on existing prices in certain economically comparable countries, and to impose civil penalties on manufacturers that are non-compliant to these prices.
• The Insulin Price Reduction Act (Sen. Shaheen, D-NH) prohibits manufacturer rebates for insulin products unless they are passed on at the point of sale.

Several proposals target drug costs more broadly and would have effects on insulin prices and spending.

• The Prescription Drug Pricing Reduction Act (Sen. Grassley, R-IA, and Sen. Wyden, D-OR), limits Part D drug price increases to the rate of inflation, requiring manufacturers to pay rebates equal to any increase exceeding inflation rates.
• The Affordable Drug Manufacturing Act (Sen. Warren, D-MA) establishes a federal office to authorize and oversee the manufacture of generic drugs in case of shortages and lack of competition and to set fair market prices for such drugs. While the bill is not exclusively aimed at insulin prices, it targets insulin for manufacture due to lack of competition and high prices.
• The Elijah E. Cummings Lower Drug Costs Now Act, passed by the House of Representatives in December 2019, authorizes the government to negotiate prices for a limited number of high-cost, single source Part D drugs, and explicitly lists insulin among those drugs to be negotiated. This bill also includes an inflation rebate provision.

Conclusion

Rising prices for insulin have attracted increasing scrutiny from policymakers. The Administration’s new model that allows participating Part D plans to cap monthly out-of-pocket insulin costs for their enrollees will help some beneficiaries with their insulin costs, if they are enrolled in a participating Part D plan and use an insulin product covered by their plan with a $35 copayment. But the new model applies to only a subset of plans and enrollees, and not all insulin products have to be covered by all participating plans. The new model also does not address underlying insulin list price increases or affordability concerns for people who are uninsured or covered by other sources of coverage besides Medicare Part D.

Our analysis of Part D claims shows that higher insulin prices have led to an increase in out-of-pocket spending by Medicare beneficiaries for insulin, and higher out-of-pocket costs for some insulin products relative to others. At the same time, the typical formulary placement for insulin products in Part D plans insulates many people on Medicare from the high cost of this drug since many plans typically charge a flat dollar copayment for covered insulin products during the initial coverage phase. However, when insulin users reach the coverage gap and face 25% coinsurance, or when an insulin product is covered on a higher tier or not covered at all, out-of-pocket costs are higher.

It is important to note that this analysis does not include the cost of insulin supplies, nor does it include the cost of other drugs that people with diabetes might take for other conditions, so it is not the full picture of the out-of-pocket prescription drug cost burden facing insulin users in Medicare. These findings demonstrate why the cost of insulin and other prescription drugs is an ongoing concern for patients and public and private payers, and a pressing issue for policymakers.

The federal government has deemed workers in the food and agricultural sector part of the essential critical infrastructure workforce. Moreover, under recent a Presidential Executive Order, meat and poultry processing plants must continue operations to prevent disruption in the food supply chain. Workers in these industries face risk for coronavirus exposure due to their continued work outside the home, with many facing increased risks due to close quarters in their working environment. The Centers for Disease Control and Prevention (CDC) and other analysis have identified coronavirus outbreaks in meatpacking facilities. This data note analyzes key characteristics of food production workers to provide insight into who these risks affect and the health and financial implications of the COVID-19 outbreak for these workers and their communities. It finds:

• Roughly 3.4 million individuals work in food production industries.
• While the majority of food production workers are White (51%), Hispanic workers (34%) make up twice as many food production workers compared to their share of all workers (17%). A higher proportion of food production workers are noncitizens compared to all workers (22% vs 8%).
• Reflecting the higher shares of noncitizen and Hispanic workers, 37% of food production workers speak a language other than English at home, and a quarter (25%) are less than fluent in English.
• Despite being more likely than all workers to work full-time (82% vs. 76%), food production workers are more likely to have household income below 200% of the federal poverty level (FPL), which is $12,760 for an individual and $26,200 for a family of four as of 2020 (29% vs. 19%). Food production workers, in particular Black and Hispanic workers, also are more likely to be uninsured compared to workers overall (17% vs. 10%).
• Food production workers are more likely to be living in non-metro areas than workers overall (29% vs. 13%).

These findings show that many food production workers have limited ability to absorb income decreases, creating disincentives for them to miss work even if they feel ill and increasing the risk of them experiencing financial challenges if they do miss work. Restrictions on immigrant eligibility for public programs and federal financial relief may compound these potential financial challenges given the relatively high share of noncitizen workers. Moreover, food production workers may face increased barriers to testing and treatment due to their relatively higher uninsured rate. Workers who live in immigrant families may also face barriers to care due to immigration-related fears. At a broader level, outbreaks among food production workers may exacerbate the disproportionate impact COVID-19 is having on people of color and widen underlying health, social, and economic disparities.

Background

The federal government has deemed workers in the food and agricultural sector part of the essential critical infrastructure workforce. Moreover, under recent a Presidential Executive Order, meat and poultry processing plants must continue operations to prevent disruption in the food supply chain. Workers in these industries face risk for coronavirus exposure due to their work outside the home, with many facing increased risks due to close quarters in their working environment. While essential workers in many industries face ongoing risk, the CDC has identified barriers to effective prevention and control of COVID-19 within meat processing plants and noted that socioeconomic challenges might contribute to workers in this industry working while feeling ill. Workers in other food production and processing industries may also be at increased risk for exposure due to their working environment, which likely requires close contact with colleagues.

Few states break out data on coronavirus cases and deaths by workplace. However, as of April 2020, the CDC had identified nearly 5,000 confirmed cases and 20 COVID-19 related deaths among workers within 115 meat or poultry processing plants in 19 states reporting data to the CDC. It further found that, among 17 states reporting the number of workers in their affected facilities, 3% of workers tested positive, but this rate was higher in several states, including Washington (7.1%), Wisconsin (8.8%), South Dakota (17.3%), and Iowa (18.2%). Other analysis finds that, as of May 28, there had been at least 18,500 cases tied to facilities in 33 states and at least 69 reported worker deaths in 21 states. Ongoing investigations and news reports highlight the risk for food processing plants to become COVID-19 “hot spots.” Emerging reports also are identifying spread of the virus among farmworkers, with one farm in Tennessee recently identifying all 200 of its employees as infected with the virus.

This data note analyzes key characteristics of food production workers to provide insight into who these risks affect and the health and financial implications of the COVID-19 outbreak for these workers and their communities. It is based on Kaiser Family Foundation analysis of 2018 American Community Survey data. We identify food production workers as those employed in animal processing, farming, or food manufacturing such as commercial bakeries or beverage plants. See methods for more details.

Who are Food Production Workers?

Roughly 3.4 million individuals work in food production industries. The largest shares of these workers are in animal production and processing (38%) and crop production (30%); 7% work in seafood, and the remaining 25% work in other food manufacturing, including bakeries, beverage manufacturers, and fruit and vegetable preserving and other specialty food processing (Figure 1).

Demographic Characteristics

Food production workers are more likely to be male compared to workers overall, and the majority are less than age 50. Over two-thirds of all food production workers are male (69%), compared to 52% of all workers. Workers in crop production have the highest share of male workers (73%), while a lower share of seafood production workers is male (62%) (Appendix, Table 1). The majority of food production workers are less than age 50 (67%), but 33% are age 50 or over, including 6% who are 65 or older.

Just over half (51%) of food production workers are White, and the remaining 49% are people of color, compared to 38% of all workers who are people of color (Figure 2). This pattern reflects that Hispanic workers make up twice as many food production workers compared to their share of workers overall (34% vs. 17%). The share of food production workers who is Black is less than the share of all workers (10% vs. 12%). Among crop production workers, over half (55%) are people of color, including nearly half who are Hispanic (49%), and 3% who are Black. The total share of seafood production workers who are people of color is similar to that of all food production workers (50%), but they include a higher share of Black workers (14%) and a lower share of Hispanic workers (26%) compared to all food production workers (Appendix, Table 1).

The majority of food production workers are citizens, but more than 1 in 5 (22%) are noncitizens, compared to 8% of all workers who are noncitizens (Figure 3). The share of workers who are noncitizen is higher among crop production workers (33%) and lower among seafood production workers (16%).

Reflecting the higher shares of noncitizen and Hispanic workers, 37% of food production workers speak another language than English at home, and a quarter (25%) are less than fluent in English. In contrast, 22% of all workers speak another language than English at home and 8% are less than fluent in English. Among workers in crop production, nearly half (49%) report speaking another language other than English at home and over a third (36%) are less than fluent in English, compared to 37% and 25% of all food production workers, respectively. The shares of seafood production workers who speak another language than English at home and who are less than fluent in English are lower at 32% and 19%, respectively.

Income and Health Insurance

Over eight in ten food production workers are employed full time (82%), higher than the share of full-time employees among workers overall (76%) (Figure 4). Workers in crop production have the lowest share of full time full-time workers (74%), while seafood production workers are more likely to work full-time (87%).

Despite being more likely than all workers to be employed full-time, nearly three in ten (29%) food production workers have household incomes below 200% FPL compared to just less than one in five workers (19%) overall. The share of workers who have household incomes below 200% of FPL is higher among crop production workers (39%), while lower among seafood workers (20%). The FPL is $12,760 for an individual and $26,200 for a family of four as of 2020.

Compared to workers overall, food production workers are less likely to have employer sponsored coverage (59% vs. 68%), more likely to be covered by Medicaid (12% vs. 9%), and more likely to be uninsured (17% vs. 10%). Over one in four (26%) crop production workers lack health insurance coverage, reflecting a substantially lower rate of employer coverage (38%) compared to all food production workers. In contrast, seafood production workers have a lower uninsured rate (10%) due to a higher rate of employer coverage (74%) relative to all food production workers.

There are meaningful differences in household income and insurance coverage among food production workers by race/ethnicity. The share of food production workers who work full-time is similar across race/ethnicity groups. However, compared to White food production workers, Black and Hispanic food production workers are more likely to live in households with income below 200% FPL (35% and 44% compared to 18%). Similarly, Black and Hispanic food production workers are more likely to be uninsured than White workers (16% and 31% compared to 9%).

Geographic Distribution

Nearly three in ten (29%) food production workers live outside of metro areas, compared to 13% of all workers. The share living outside a metro area is significantly higher for workers in animal production and processing (40%), and lower for seafood production workers (18%).

Over one in four (28%) food production workers are in just three states California (17%), Texas (6%), and Illinois (4%) (Appendix, Table 2). Food production workers constitute more than 5% of the workforce in Nebraska, Iowa, Arkansas, Idaho and South Dakota. Animal production and processing composes a large share of food production workers in Mississippi, Arkansas and Wyoming, where crop production is a larger share in Washington, California and North Dakota.

What are the implications for risks posed by COVID-19?

Together, these findings show that, while the majority of food production workers are White and citizens, they include disproportionately larger shares of Hispanic and noncitizen workers compared to workers overall. Moreover, despite higher rates of full-time employment, food production workers are more likely to have household income below 200% FPL and to be uninsured. Further, food production workers are more likely to live outside of metro areas compared to workers overall.

Food production workers have limited ability to absorb income reductions, creating disincentives for them to miss work and increasing the risk of financial challenges for those who do miss work. As noted by the CDC, socioeconomic challenges may contribute to individuals working while feeling ill, particularly if there are management practices such as bonuses that incentivize attendance. Those who do miss work due to illness may experience financial challenges and problems paying for basic needs if they do not have sick leave and lose pay. Restrictions on immigrant eligibility for public programs and federal financial relief may compound these challenges given the higher share of noncitizens among food production workers. The CDC suggests that programs that actively encourage symptomatic workers to stay home can support infection control, including paid sick leave and the avoidance of incentives that might encourage individuals to work while symptomatic.

Food production workers may face increased challenges accessing COVID-19-related testing and treatment services due to a relatively higher uninsured rate and larger share of noncitizens. The higher uninsured rate among food production workers may increase barriers to testing and treatment. Even though federal legislation provides resources to cover testing costs for uninsured individuals, individuals may not know where to obtain a test and/or may forego testing out of fear of having to pay out-of-pocket. Aside from barriers to testing, individuals also may face barriers to obtaining treatment. Although federal legislation provides funding to reimburse providers for treating uninsured patients with COVID-19, there are questions about sufficiency of the funds to cover costs. Workers who live in immigrant families may also face barriers to testing and treatment due to restrictions on immigrant eligibility for health coverage and fears about potential negative consequences on their or a family member’s immigration status. The United Food and Commercial Workers International Union has called on the Administration to prioritize testing for essential workers, including those in meatpacking and food processing.

Outbreaks among food production workers may exacerbate the disproportionate impact COVID-19 is having on people of color. Data illustrate the stark impacts COVID-19 is having on people of color, who account for a higher share of confirmed cases and deaths relative to their share of the population in many states. Given that people of color make up a disproportionate share of food production workers, outbreaks among these workers could amplify these disparities and widen preexisting underlying health, economic, and social disparities. The diversity among food production workers also points to the importance of providing outreach and education to workers in linguistically and culturally appropriate ways. The CDC recommends periodic infection control and occupational safety and health training for all workers and supervisors tailored to literacy levels and preferred languages as part of its prevention and mitigation strategies among meat and poultry facilities. In addition, because food production workers are more likely to live outside of metro areas, outbreaks among these workers may affect rural communities that could experience particular challenges given their relative lack of hospital capacity, older average age, and higher shares of their populations with underlying health conditions.

In sum, food production workers may face particular health care access and financial challenges associated with the COVID-19 outbreak. Addressing their needs may help protect the health and financial stability of workers themselves, but also plays a role in preventing widening of underlying racial and ethnic disparities in health, social, and economic factors given the diversity of the food production workforce.

This analysis is based on KFF analysis of the 2018 American Community Survey (ACS), 1-year file. The ACS includes a 1% sample of the US population; the subset of food production workers examined in this data note includes over 33,000 observations. We define the food production workforce as all individuals who earned at least $1,000 during the year and indicated that their job was in one of the food processing industry codes listed below. We include only adults age 18 and up in the analysis.

We roll up food production industries related to animal production and processing into one category based on the following categories: animal production and aquaculture; animal slaughtering and processing; animal food, grain, and oilseed milling, and dairy product manufacturing. Similarly, we include beverage manufacturing; bakeries and tortilla manufacturing, except retail bakers; sugar and confectionary productions; and not specified food industries in the food manufacturing category.

The ACS asks respondents about their health insurance coverage at the time of the survey. Respondents may report having more than one type of coverage; however, we sort individuals into only one category of insurance coverage based on a hierarchy of coverage (detailed here).

The ACS divides the United States into Public Use Microdata Areas (PUMAs), or geographic areas representing at least 100,000 people. PUMA’s vary in size based on population density. The USDA Economic Research Service defines metro and non-metro areas. Using the USDA method, we identify 80% of the PUMAs as solely in non-metro areas, 9% containing metro areas, and 12% containing both metro and non-metro areas. We categorize the PUMAs containing both metro and non-metro areas based on where the majority of the population lives. Using this approach, we categorize a total of 14% of the population living in a PUMA was classified as a non-metro area.