In a new analysis, KFF researchers find that COVID-19 vaccination rates among nursing home staff increased by 25 percentage points nationally (63% to 88%) from when the Biden administration announced the vaccine mandate for health care workers in August 2021 to after vaccination deadlines passed in March 2022.

Researchers analyzed nursing home-level data from the federal government covering some 14,700 nursing homes, or about 97 percent of all nursing homes in the U.S. They conclude that while a number of factors may have been at work, it appears that the mandate contributed to the increases in staff vaccination rates.

As of March 2022, 12 percent of nursing homes nationally reported that 100 percent of their staff were fully vaccinated, while 39 percent reported staff vaccination rates of over 90 percent but less than 100 percent. The remaining 49 percent of nursing homes reported that fewer than 90 percent of their staff had been fully vaccinated.

While the federal mandate does not explicitly require booster shots, the national booster rate for nursing home staff was 44 percent as of March 2022.

Some nursing home operators warned that requiring health care workers to be vaccinated against COVID-19 could contribute to staffing shortages. Data suggest that the vaccine mandate has not exacerbated such shortages, however.

KFF analysts find that 28 percent of nursing homes in the U.S. reported staffing shortages as of March 2022, down slightly from the peak in January 2022 where nearly one in every three nursing homes reported a shortage.

Staff shortages in nursing homes are a longstanding problem that predates the pandemic. In a separate analysis also released today, KFF analysts summarize federal and state standards related to nursing home staffing prior to COVID-19 and identify changes that states have made to minimum staffing requirements during the pandemic. They also examine state legislative and regulatory actions since the onset of the pandemic that directly affect worker wages and training requirements.

The Biden administration has announced plans to propose new federal minimum staffing adequacy regulations in the next year, and the National Academies of Science, Engineering, and Medicine recently recommended minimum staffing levels as part of its comprehensive report on ways to improve nursing home quality.

Conditions and COVID-19 precautions in nursing homes remain a subject of heightened public interest since over one in five COVID-19 deaths during the pandemic has been in a long-term care facility. While the vaccine mandate remains in effect, litigation challenging the new rule is ongoing. Although cases in Texas and Florida were dismissed after the Supreme Court allowed the rule to take effect, cases in Missouri and Louisiana are still pending.

Over one in five COVID-19 deaths has been in a long-term care facility since the start of the pandemic. Due to the disproportionate impact of COVID-19 on this population early in the pandemic, vaccinations of nursing facility residents and staff were a high priority when the vaccine rollout began in Winter 2020-2021. Since then, the Centers for Medicare and Medicaid Services (CMS) has implemented a health care worker vaccination mandate, which does not include booster shots, for nursing facilities and other providers that participate in Medicare and/or Medicaid. Although some states sued to challenge this rule, the Supreme Court allowed the mandate to take effect while those cases are resolved by the lower courts.

This analysis uses nursing facility-level data reported by the federal government to track the increase in vaccination rates among nursing facility staff nationally and by state between August 2021 (when the vaccine mandate was first announced) and March 27^th, 2022 (after the vaccine deadline for health workers had passed in all states), building on a previous data note. Additionally, this analysis provides state-level information on booster rates among nursing home staff and the prevalence of staffing shortages after all vaccination deadlines had passed. Approximately 14,700 nursing facilities (or about 97% of all nursing facilities) are included in this analysis, and at least 95% of facilities in each state reported reliable data. See methods for more details.

While this analysis does not directly attribute the increase in vaccination rates to the mandate, since many factors could affect these measures, the trends are nonetheless helpful for understanding the impact of different policies on rates of infections, hospitalization, and deaths attributable to COVID over time. Key findings from our analysis include:

• Vaccination Rates: Nursing facility staff vaccination rates increased nationally by 25 percentage points, from 63% to 88%, between August 2021 and March 2022 (Figures 1 and 2). As of March 2022, 12% of nursing facilities nationally reported 100% staff vaccination rates, and 39% of facilities reported rates over 90% but less than 100% (Figure 3). The remaining 49% of facilities reported rates of 90% or below. These data cannot be used to estimate the share of facilities in each state that comply with the mandate since data do not account for vaccine exemptions.
• Booster Rates: As of March 2022, the national booster rate for nursing facility staff was 44%, ranging from 93% in MA to 24% in FL, MO, and MS (Figure 4). While the federal mandate does not explicitly require booster shots, five states (CT, MA, NJ, NY, and CA) that reported a greater than 30-percentage point increase in booster rates from January 2022 to March 2022, as well as some of the highest booster rates in March 2022, had state-level booster mandates during this time period. Nationally, about one-third of facilities reported booster rates among nursing facility staff higher than 50%, while another one-third of facilities reported booster rates under 25% (Figure 5).
• Staff Shortages: More than one in four (28%) of nursing facilities nationally reported staffing shortages as of March 2022 ranging from 63% in AK to 3% in CA (Figure 6). Staffing shortages peaked nationally during the Omicron wave, with nearly one in three facilities reporting a shortage in January 2022. Staffing shortage rates have fallen since then from 33%, and remained relatively stable through March 2022, even after the implementation of the vaccine mandate.

Vaccine Mandate: Requirements and Litigation

In mid-August 2021, the Biden Administration announced that nursing facility staff would be required to be vaccinated against COVID-19 as a condition of participation in the Medicare and Medicaid programs. CMS released the interim final rule establishing the new vaccination requirement in November 2021. Due to litigation, facilities in different states had different deadlines to comply with the new rule. CMS guidance requires 100% of staff in nursing facilities to have completed their vaccinations or have a pending or approved exemption by February 28^th in 25 states plus D.C.; by March 15^th in 24 states; and by March 21^st in the remaining state, Texas.

CMS released a series of guidance with details on enforcement and implementation timelines for facilities to comply with the mandate. After the mandate became effective in all states, including in states that sued to challenge it, the substantive provisions of the guidance were the same for all states, except that the effective dates differed, as explained above. Under this guidance, CMS established three difference enforcement action thresholds:

• Within 30 days after the relevant guidance was issued: 100% of staff must have received at least one dose of COVID-19 vaccine, or have a pending or approved exemption, or a CDC-recommended delay. A facility that is above 80% and has a plan to achieve a 100% staff vaccination rate within 60 days is not be subject to additional enforcement action.
• Within 60 days after the relevant guidance was issued: 100% of staff must have received all necessary doses of the COVID-19 vaccine series, or have a pending or approved exemption, or a CDC-recommended delay.  A facility that is above 90% and has a plan to achieve a 100% staff vaccination rate within 30 days is not be subject to additional enforcement action.
• Within 90 days after the relevant guidance was issued: Facilities that do not comply with the 100% standard may be subject to enforcement action.

Nursing facilities that do not achieve compliance could be subject to enforcement through a number of mechanisms. CMS states that federal and state surveyors will begin surveying facilities for compliance with the rule as part of initial certification, standard recertification or reaccreditation, and complaint surveys 30 days following the issuance of the guidance. Facilities that do not achieve the 100% standard may be subject to enforcement through actions such as civil monetary penalties, denial of payments, and termination of participation from the Medicare and Medicaid programs. CMS in the guidance emphasizes that their “primary goal is to bring health care facilities into compliance” and termination would likely occur “only after providing a facility with an opportunity to make corrections and come into compliance.”

While the vaccine mandate is in effect, litigation challenging the new rule is ongoing. Although cases in Texas and Florida were dismissed after the Supreme Court allowed the rule to take effect, cases in Missouri and Louisiana are still pending. The Missouri case is now back at the district court for a decision about whether to permanently block the rule, while the Louisiana case is currently at the 5^th Circuit Court of Appeals, awaiting an expected return to the district court. Although the Supreme Court decided that the federal government has authority to issue the rule, the states in both active cases are now seeking to amend their complaints to add new claims that were not part of their arguments at the Supreme Court.

Missouri now argues that the rule is arbitrary and capricious since the Omicron variant has replaced the Delta variant, which was prevalent when the rule was issued. Louisiana now argues that requiring state surveyors to enforce the rule violates the “anti-commandeering doctrine” and also challenges subsequent CMS guidance that indicates that state surveyors should be vaccinated. In both the Missouri and Louisiana cases, the district court judges have indicated that they are likely to allow the states to add these new claims, which raises the possibility that the rule could be blocked again – decisions that likely would be appealed.

Nursing Facility Staff Vaccinations

Nursing facility staff vaccination rates increased nationally by 25 percentage points (63% to 88%) between August 2021 (when the vaccine mandate for health workers was initially announced) and March 2022, the deadlines for compliance with the mandate in all states (Figure 1).  While a number of factors may have influenced the rise in vaccination rates, it appears that the vaccine mandate contributed to increases in staff vaccination rates.

By the end of March 2022, after vaccine mandate deadlines had passed in all states, vaccination rates for nursing facility staff varied widely by state, ranging from 76% in Montana to 99% in five states (NY, RI, DC, ME, and MA) (Appendix Table 1). Previous analysis found similar state variation in nursing facility staff vaccination rates prior to the mandate deadlines, which CMS cited as a factor leading to the adoption of the rule. While state variation in vaccination rates has persisted since the deadlines passed, all states reported an increase in staff vaccination rates after the vaccination deadlines passed (Figure 2).

The increase in staff vaccination rates also varied across states between August 2021 and March 2022, ranging from a 37-percentage point increase in Georgia to a nine-percentage point increase in Hawaii (Figure 2). Nursing facilities in ten states reported a 30-percentage point or greater increase between August 2021 and March 2022, while facilities in five states reported an increase of 15 percentage points or less in same time period (Figure 2). This variation may be partially attributed to some states reporting high baseline vaccination rates, such as Hawaii (89%) and California (82%), while other states such as Georgia and Louisiana reported lower rates (48% and 52%, respectively).

After all vaccination deadlines had passed by the end of March 2022, 12% of facilities reported a staff vaccination rate of 100% and 39% of facilities reported staff vaccination rates greater than 90% but less than 100%. The remaining 49% of facilities reported rates of 90% or lower (Figure 3). The share of facilities reporting universal staff vaccination rates ranged from 56% of facilities in Rhode Island to 0% of facilities in Wyoming. The share of facilities reporting near-universal rates of vaccination (91-99%) ranges from 82% of facilities in Connecticut to 16% of facilities in Oklahoma. 24% of nursing facilities reported vaccination rates between 81-90% and roughly the same share (25%) reported a staff vaccination rates of 80% or lower.

It is difficult to assess compliance with the mandate based on these data. The federal data includes the number of staff who report medical contraindications to the COVID-19 vaccine, but does not track other potential exemptions, such as religious exemptions or CDC-recommended delays. Additionally, it is not clear whether all staff with a reported medical contraindication fall under the umbrella of “approved exemptions”. A ProPublica investigation found that nearly 20,000 nursing facility workers, or about one percent of nursing facility staff, claimed a medical exemption for vaccines and an additional 164,000, or about nine percent, declined to get the vaccine for another reason, which could include a religious objection.

Nursing Facility Staff Boosters

The national booster rate for nursing facility staff was 44% as of March 2022 (Figure 4).  Though booster shots have been cited as crucial protection against serious illness and death due to COVID-19, the CMS rule does not require staff to receive booster shots. The CDC began recommending boosters for nursing home residents in September 2021, after data indicated that vaccinations become less effective over time, especially among older adults.  Staff booster rates in nursing facilities increased 17 percentage points (27% to 44%) between mid-January 2022 and March 2022, after vaccination deadlines had passed. January 2022 was chosen as the comparison time period since it is five months after the vaccine mandate was announced, aligning with CDC’s recommendations to receive a booster shot five months after most initial vaccine regimens.  

Booster rates for nursing facility staff varied widely across states, ranging from 93% in Massachusetts to 24% in Florida, Missouri, and Mississippi, as of March 2022 (Figure 4). Eight states reported booster rates above 60%, while 13 states reported booster rates below 30%. This wide variation may reflect differences in attitudes or hesitancy about boosters, as well as state and local policies on boosters. Five states (CT, MA, NJ, NY, and CA) reported a greater than 30 percentage point increase in booster rates from January 2022 to March 2022, as well as some of the highest booster rates in the country in March 2022 (an average of 72%, or nearly 30 percentage points higher than the national average). All five of these states had booster mandates during this time period, which may partially explain why these state-level increases are higher than the national increase of 17 percentage points in this time period.

As of the end of March 2022, after all vaccination deadlines had passed, about one-third of nursing facilities nationally reported staff booster rates of over 50%, another third reported rates between 25-50%, and the third of remaining facilities reported rates of under 25% (Figure 5). There is wide state variation in the share of facilities reporting staff boosters of over 50%, ranging from 99% of facilities in Massachusetts to five percent of facilities in Tennessee. Conversely, about 60% of facilities in Florida, Louisiana, and Missouri reported staff booster rates of under 25%, while zero percent of facilities in Massachusetts fall in this category.

Nursing Facility Staffing Shortages

More than one in four (28%) nursing facilities nationally reported staffing shortages as of March 2022 after all vaccination deadlines had passed, ranging from 63% of facilities in Alaska to 3% of facilities in California (Figure 6). State variation in staffing shortages may be attributable to variation in local economies and labor markets and could also be due to differences in the way in which nursing facilities define and report staff shortages. Previous analysis shows that staffing shortages peaked nationally during the Omicron wave, with nearly one in every three facilities reporting a shortage in January 2022. Rates have fallen slightly since then and remained relatively stable through March 2022, even after the implementation of the vaccine mandate. When issuing the vaccine mandate rule, CMS acknowledged that some staff may leave their jobs because they do not want to receive the vaccine, possibly exacerbating staffing shortages, while also citing examples of vaccine mandates that resulted in high rates of compliance and few employee resignations. Data suggest that the vaccine mandate has not exacerbated staffing shortages to the extent initially hypothesized since shortages have actually fallen nationally since January 2022. There are likely a multitude of factors that impact staffing shortages beyond the vaccine mandate.

Looking Ahead

While it is difficult to isolate the impact of the staff vaccination mandate on vaccination rates, booster rates, and staff shortages, the mandate did likely play a role in increased vaccination rates from August 2021 to March 2022. As new COVID-19 variants emerge and make their way into nursing facilities, maximizing vaccination and booster rates among facility staff will help protect against illness and death of residents and staff, and help maintain sufficient staffing levels. The Biden Administration has announced a number of policies to address staffing issues in nursing facilities, including establishing minimum staffing levels for Medicare and Medicaid-certified nursing facilities, supporting state efforts to tie Medicaid reimbursement rates to increases in staff compensation, and launching a national nursing career pathways campaign. The National Academies of Sciences, Engineering, and Medicine also released a comprehensive report on quality of care in nursing facilities and recommended enhanced and updated federal staffing standards as one component of a wide-ranging package of nursing facility reforms. Since the onset of the COVID-19 pandemic, some states have adopted changes to increase state-level minimum staffing requirements. It will be important to watch whether and how lessons from the pandemic are integrated into new federal and state policies to improve nursing facility care quality and protect residents and staff.

Appendix Table

Despite occasional anecdotal reports of people having trouble finding a doctor who takes their insurance, KFF researchers find in a new analysis that the vast majority of non-pediatric office-based physicians accept new Medicare patients, as well as new private insurance patients.

Eighty-nine percent of such physicians accepted new Medicare patients in 2019, and 91 percent accepted new private insurance patients, according to the analysis, which uses data from the federal 2019 National Electronic Health Records Survey (NEHRS). Among primary care physicians the rates are lower, but still comparable for new patients with Medicare (83%) and private insurance (86%).

In addition, the analysis found that roughly the same percentage of non-pediatric office-based physicians accepted new Medicare patients in 2011 (88%) as in 2019. And the share of physicians who accept new private insurance patients has risen considerably over the period — up 10 percentage points from 81 percent in 2011 to 91 percent in 2019.

The new analysis also contains state-level data. KFF researchers find that in 23 states, at least 90 percent of all non-pediatric office-based physicians accepted new Medicare patients in 2015-2017. Across all states the share ranged from 76 percent in Washington D.C. to 95 percent in three states (Iowa, Minnesota, and Pennsylvania), similar to the range for privately-insured patients.

The findings are notable because some proponents of doing more to contain health care costs have proposed adopting Medicare rates, or a multiple of Medicare rates, for those who now have private insurance, since Medicare rates tend to be lower. Skeptics of that approach – which could involve directly regulating prices, or creating a Medicare-like plan as a public option or as part of a single payer system — have raised concerns that moving towards Medicare rates would cause access problems.

Across a range of specialties, KFF analysts find that the share of physicians accepting new Medicare patients is similar to the share accepting new patients with private insurance. The share of physicians accepting new patients is lowest among psychiatrists, with 60 percent accepting new Medicare patients and 59 percent accepting new privately-insured patients.

KFF analysts also find that a very small share (1%) of non-pediatric physicians formally opted-out of the Medicare program in 2022, with psychiatrists being the most likely to opt-out (7.5% ). In all states except for Alaska and Colorado, fewer than 2 percent of physicians have opted-out of Medicare.

The fact that the vast majority of physicians are accepting new patients does not mean that all patients will have access to all clinicians, since providers may only be in the networks of some private insurance or Medicare Advantage plans.

For more data and analyses about Medicare visit kff.org.

Over the past decade, health care prices paid by private insurers have risen sharply. These higher prices contribute to higher health care costs per person, which contribute to higher premiums and out-of-pocket spending among the privately insured. In contrast to Medicare, which pays physicians based on a fee schedule, prices for those with private insurance are negotiated by payers and providers, and on average, private insurers pay 143% more for physician services than Medicare. The difference in Medicare and private prices has served as the basis for proposals to adopt Medicare rates, or a multiple of Medicare rates, for those with private insurance, or as part of a Medicare-like plan through a public option or single payer plan.

Over the years, some have questioned whether the level of Medicare payments for physicians are too low to cover the costs of providing care and have raised concerns that moving towards Medicare rates would cause access problems. Previous KFF analysis has found that even though Medicare payments to physicians are lower, the vast majority of Medicare beneficiaries report having a usual source of care and high levels of satisfaction with the quality of their medical care. In addition, two separate KFF analyses, using data from 2012-2013 and 2020, found that the nearly all physicians accept Medicare, and only a small fraction “opt out” of Medicare altogether.

This brief builds on and updates the previous KFF analyses by drawing on the 2019 National Electronic Health Records Survey (NEHRS) to examine: 1) the share of non-pediatric office-based physicians accepting new patients with Medicare or private insurance and how these rates have changed over time; and 2) the share of physicians accepting new patients by physician specialties, geographic areas, and physician and practice characteristics across Medicare and private insurance. We are only able to assess whether doctors report accepting any new patients and cannot evaluate whether all new patients are accepted or how timely or convenient it is to schedule an appointment. This analysis further examines the extent to which non-pediatric physicians are opting out of Medicare, by specialty and state, drawing on 2022 opt-out affidavit data published by the Centers for Medicare & Medicaid services and data on the total number of active state-licensed physicians from Redi-Data, Inc (See Methods box for details).

Key findings include:

• Across almost every dimension we examine, the share of physicians accepting new Medicare patients is similar to the share accepting new patients with private insurance, with the only exception being the share of obstetricians and gynecologists accepting new Medicare patients (93%) is somewhat lower than the share accepting new privately-insured patients (99%). Physicians accepting new patients with private insurance will not be participating in every plan’s network, and similarly will not be in every Medicare Advantage plan network.
• Most non-pediatric office-based physicians accept new Medicare patients (89%), as well as new private insurance patients (91%).
• In 2011 and 2019, 88% and 89% of non-pediatric office-based physicians accepted new Medicare patients, respectively, with modest fluctuations in the years in between. The share accepting new patients with private insurance increased by 10 percentage points over the period, rising from 81% in 2011 to 91% in 2019, a share similar to Medicare.
• Across most specialties, the majority of physicians (at least 80%) accept new patients with Medicare and private insurance. Almost all surgical specialists accept new Medicare and privately-insured patients (96% for both). Among primary care physicians the rates are lower, but comparable for new patients with Medicare (83%) and private insurance (86%), similar to the share of physicians in other medical specialties accepting new Medicare (87%) and privately-insured (88%) patients. The share of physicians accepting new patients is lowest among psychiatrists with 60% accepting new Medicare patients and 59% accepting new privately-insured patients.
• In 23 states, at least 90% of all non-pediatric office-based physicians accepted new Medicare patients in 2015-2017. Across states, the share of physicians accepting new Medicare patients ranged from a high of 95% in three states (Iowa, Minnesota, and Pennsylvania) to a low of 76% in the District of Columbia, similar to the range across states for privately-insured patients.
• A very small (1%) share of non-pediatric physicians formally opted-out of the Medicare program in 2022, with psychiatrists accounting for a disproportionate share of physicians opting-out. In all states except for Alaska and Colorado, less than 2% of physicians have opted-out of Medicare.

Findings

Acceptance of new patients by office-based physicians by type of insurance coverage

Differences in payment rates may affect whether people with different types of insurance coverage have comparable access to physician services. We use data from the National Electronic Health Records Survey (NEHRS) for 2011 to 2019 (the most recent year of data) to assess whether there are differences in the share of non-pediatric office-based physicians who are accepting new patients across Medicare and private insurance.

The vast majority of non-pediatric office-based physicians accept new patients, with similar shares accepting new Medicare and privately-insured patients. Overall, 95% of office-based physicians accepted new patients in 2019. The share of physicians accepting new patients was similar for Medicare (89%) and private insurance (91%), with no statistically significant differences between the payers (Figure 1). The share of physicians accepting any new patients is higher than the share accepting Medicare or private insurance because some physicians only accept patients with one type of insurance but not others, including types of insurance not broken out here (i.e., self-pay patients).

Between 2011 and 2019 the overall share of non-pediatric office-based physicians accepting new patients was stable and a consistent share accepted new Medicare patients. In 2011 and 2019, 88% and 89% of office-based physicians accepted new Medicare patients, respectively, with modest fluctuations in the years in between. During this same period, the share of physicians accepting new privately-insured patients increased from 81% in 2011 to 91% in 2019, becoming more similar to the share that accepted new Medicare patients (Figure 1).

Share of physicians accepting new Medicare and privately-insured patients, by specialty

To understand whether access varies more for some types of physicians, we examine whether the share of physicians accepting new patients is different for primary care, medical and surgery specialty groups, and among 12 common specialties.

A smaller share of primary care physicians accept new Medicare or privately-insured patients than physicians in other medical or surgical specialties. To look at differences by specialty type, we pooled 2017 and 2019 NHERS survey data. Overall, a smaller share of primary care physicians (89%) reported having an open practice—meaning they accept any new patients—compared to other medical specialists (95%) and surgical specialists (99%). This pattern was consistent across Medicare and private insurance. A smaller share of primary care physicians accepted new Medicare (83%) or privately insured patients (86%), while the vast majority of surgical specialists accepted new patients with either type of insurance (96% for both). Among other medical specialists, 87% were accepting new Medicare patients and 88% were accepting new privately-insured patients (Table 1).

In most specialties, similar shares of physicians accept new Medicare and privately-insured patients. Except for obstetrics and gynecology, there were no statistically significant differences in the share of physicians accepting new patients with Medicare and private insurance. A larger share of obstetricians and gynecologists accepted new privately-insured patients (99%) than new Medicare patients (93%) (Figure 2, Table 1). Notably, almost 1 in 5 of general/family practice and internal medicine physicians were not taking new Medicare patients, similar to the share not taking new privately-insured patients.

Among all non-pediatric, office-based physicians, psychiatrists were least likely to accept new patients regardless of insurance type. Among psychiatrists, similar shares were accepting new Medicare (60%) and privately-insured (59%) patients (Figure 2, Table 1). For both Medicare and private insurance, the share of psychiatrists accepting new patients was at least 20 percentage points lower than the share of physicians in general/family practice accepting new patients.

Variation in acceptance of new patients across states, in counties that have provider shortages, and by metropolitan status

To understand whether access to physician services varies geographically, we examined the share of office-based physicians accepting new patients by state, metropolitan status, as well as for counties that are designated as primary care or mental health professional shortage areas (HPSAs). Examining whether the share of physicians accepting new Medicare or privately-insured patients varies in HPSA designated counties highlights how differences in payment rates interacts with the supply of providers to affect access to physician services.

state-level shares of physicians accepting new patients

In every state, a large majority of non-pediatric office-based physicians accept new Medicare patients, with some variation among states. In 23 states, at least 90% of physicians accepted new Medicare patients (Figure 3). The five states with the highest share of physicians accepting new Medicare patients were Iowa (95%), Minnesota (95%), Pennsylvania (95%), South Carolina (95%), and North Dakota (94%). The states with the smallest share of physicians accepting new Medicare patients were the District of Columbia (76%), Utah (79%), Oregon (79%), Hawaii (79%), and Maryland (80%). This pattern was generally similar for privately-insured patients, where the share of physicians accepting new patients ranged from 96% (Arkansas) to 72% (the District of Columbia).

county and rural status variation

The Health Resources and Services Administration (HRSA) designates counties as primary care and/or mental health professional shortage areas (HPSAs) if the availability of providers or services is lower than necessary to support the population’s needs. When a county is designated as a HPSA, there are additional incentives for health care professionals to provide services, including a 10 percent bonus payment for providing care to Medicare beneficiaries and eligibility for special federal and state student loan repayment programs for providers that accept Medicare, Medicaid and CHIP patients in these areas. We examine whether the lower supply of providers in HPSA-designated areas is correlated with fewer physicians accepting new patients.

The vast majority of physicians in counties with primary care or mental health shortages accept new Medicare patients, with similar rates for privately-insured patients. In counties where the whole county is designated as a primary care HPSA, 89% of physicians accepted new Medicare patients and 91% of physicians accepted new privately-insured patients. Similarly, in mental health HPSAs, most physicians accepted both new Medicare (89%) and privately-insured (90%) patients. While the share of physicians in these areas accepting new patients is generally high, access can still be problematic since there are not enough physicians available in these areas to provide care.

Most (89%) physicians practicing in rural areas accept new Medicare patients, similar to the share of physicians accepting new privately-insured patients (91%). The share of physicians accepting new patients were similar across Medicare and private insurance in both rural and urban areas (Table 1).

Variation in acceptance of new patients by physician and practice characteristics

The composition of physicians is changing as more women enter medicine, the share of physicians who hold a Doctor of Osteopathic Medicine (DO) rather than a Doctor of Medicine (MD) is increasing, and doctors practice medicine longer, raising the average age of physicians in the United States. Additionally, physicians are increasingly practicing as part of large or multi-specialty groups, rather than in solo or small practice settings. As the composition of the physician workforce and how physicians practice changes, understanding the relationship between physician and practice characteristics and the acceptance of new patients, and whether this varies by payer, is helpful in assessing how access to physician services may evolve over time. We examine physician acceptance rates by physician characteristics (age, sex, degree type) and practice characteristics (group size, group setting), using 2017 and 2019 pooled data from the NEHRS.

The share of physicians accepting new Medicare patients was generally similar across physician demographic groups, but younger physicians were more likely to take new patients. There were no statistically significant differences in the share of physicians accepting new patients between Medicare and private insurance by physician age, sex or degree type (Table 1). Across both Medicare and private insurance, physicians who were younger than age 65 were more likely than physicians ages 65 and older to take new patients, with 89% and 91% of physicians under age 65 accepting new Medicare and privately-insured patients, respectively, compared to 82% and 84% of physicians age 65 and over. A similar trend was evident for physicians younger than age 50 compared to physicians ages 50 and older. There were no statistically significant differences within insurance groups by physician sex or degree type.

Solo practicing physicians were least likely to report that they were accepting new patients with both Medicare and private insurance. The share of physicians accepting new Medicare and privately-insured patients was similar when comparing by the size and practice type. By group size, a smaller share of physicians in solo practices accepted new Medicare (81%) or privately-insured (83%) patients than in practices with 2 physicians (89% for Medicare, 92% for privately-insured), 3-5 physicians (91% for Medicare, 94% for privately insured), and 6-10 physicians (92% for Medicare, 93% for privately-insured). Additionally, a smaller share of physicians in solo practices accepted new privately-insured patients (83%) than physicians in practices with or 11 or more physicians (87% for Medicare, 89% for privately-insured); however, differences between physicians in solo practices and physicians in practices with 11 or more physicians were not statistically significant for new Medicare patients.

A larger share of physicians in multispecialty group practices accepted new Medicare (90%) patients compared to physicians in single-specialty practices (86%) (Table 1, Figure 4). Differences between Medicare and privately-insured patients and between single- and multi-specialty practices for private insurance were not statistically significant.

Physician Opt-Out Rates in the Medicare Program

Physicians and other health providers may participate in Medicare by either 1) agreeing to accept Medicare’s fee schedule amounts as payment-in-full (i.e., participating providers) or 2) accepting Medicare patients but choosing whether to accept Medicare’s approved amount on a claim-by-claim basis (i.e., non-participating providers). Physicians can also choose to “opt-out” of the Medicare program entirely by submitting an opt-out affidavit to Medicare. Physicians who have opted-out of Medicare enter into private contracts with their Medicare patients, allowing them to bill their Medicare patients any amount they deem appropriate and Medicare patients are responsible for the entire cost of these services. Prior KFF analyses found that a historically small share of physicians have chosen to opt-out of Medicare entirely. Our analysis examines the extent to which non-pediatric physicians are opting-out of Medicare, by specialty, and by state. We use opt-out affidavit data from the Centers for Medicare & Medicaid services as of March 2022. To calculate the total number of active state-licensed physicians, we draw January 2022 data from Redi-Data, Inc.

Only 1% of all non-pediatric physicians have formally opted-out of the Medicare program in 2022, with opt-rates varying by specialty, and highest for psychiatrists (7.5%). As of March 2022, 10,105 non-pediatric physicians have opted-out of Medicare, accounting for a very small share (1.0%) of the total number of active physicians, consistent with the shares reported in 2013 and 2020. Opt-out rates vary by specialty and are highest for psychiatrists (7.5%), physicians in plastic and reconstructive surgery (3.8%), and neurology (2.9%) (Figure 5). Among the 10,105 physicians opting out of Medicare, psychiatrists account for 42% of these physicians.

In all but two states (Alaska, Colorado), less than 2 percent of non-pediatric physicians have opted-out of Medicare. As of March 2022, Alaska (3.3%), Colorado (2.2%), District of Columbia (1.9%), Wyoming (1.9%), and Idaho (1.9%) have the highest rates of non-pediatric physicians who have opted out of Medicare. States with the lowest physician opt-out rates include North Dakota (0.1%), West Virginia (0.1%), South Dakota (0.2%), Nebraska (0.3%), and Minnesota (0.4%).

Discussion

Our analysis finds that the vast majority of non-pediatric office-based physicians accept new patients. We also find that the share accepting new Medicare patients largely mirrors the share accepting new privately-insured patients, despite higher payments from private insurers relative to Medicare. This suggests that Medicare’s relatively lower payment rates for physician services do not generally lead to fewer physicians accepting new Medicare patients. We note, however, that due to data limitations we were unable to differentiate the extent to which physicians accepting new patients are accepting all, most, or just some new patients, or the timeliness of getting appointments and needed care. We also cannot assess how many physicians participate in any or all private insurance or Medicare Advantage networks, or the impact of concierge medicine on patients’ access to physician services.

The high share of non-pediatric, office-based physicians accepting new patients – both Medicare and privately insured – was consistent across most specialties, geographic areas, and physician and practice characteristics. Increasingly, hospitals are acquiring medical practices and providing more services in a hospital setting. Due to data limitations, we are unable to assess how these trends may affect access to physician services, either overall or for patients with different types of insurance.

Notably, psychiatrists are significantly less likely than other physicians to accept new patients with either Medicare or private insurance. This finding does not suggest that Medicare rates specifically contribute to access problems for beneficiaries, but it does raise a concern about access to psychiatrists and potentially other mental health practitioners, an issue that was likely exacerbated by the COVID-19 pandemic with the increase in the share of adults reporting symptoms of anxiety disorder and/or depressive disorder.

Despite concerns that lower physician fees paid by Medicare relative to fees paid by private insurance may disincentivize physicians from taking Medicare patients, our analysis suggests that Medicare beneficiaries continue to have good access to physicians, similar to people with private insurance.

Appendix

The Enrollment Growth is Expected to Cost States Over $47 Billion Through FY 2022, But States Will Have Received $100 Billion to Cover New Medicaid Costs and Provide Additional General Fiscal Relief

Between 5.3 million and 14.2 million low-income people could lose Medicaid coverage following the end of the public health emergency and federal continuous enrollment requirement, KFF researchers find in a new analysis. We arrive at those estimates after examining two potential scenarios about how states might manage the resumption of eligibility redeterminations and disenrollments.

KFF analysts also project that, from fiscal year 2020 through fiscal year 2022, states will have collectively spent $47.2 billion to cover additional people enrolled in Medicaid because of the continuous enrollment requirement and will have received $100.4 billion in new federal funds to cover those costs and provide general fiscal relief through enhanced Medicaid matching funds.

Based on the analysis, we estimate that enrollment will have risen by 25 percent, to a projected 110.3 million, by the end of fiscal year 2022. That would be 22.2 million more people than were enrolled in fiscal 2019, before the pandemic. We attribute most of the increase to the continuous enrollment requirement that helped to ensure coverage for enrollees during the pandemic and halted churn or temporary lapses in coverage due to enrollees cycling on and off the program in short periods of time.

In a recent KFF survey of state Medicaid directors in the field in February 2022, state officials projected that median Medicaid enrollment would decline by 5 percent next year following the end of the public health emergency. Another KFF survey from January 2022 found states expect on average about 13 percent of Medicaid enrollees to be disenrolled following the end of the continuous enrollment requirement (with estimates ranging from 8% to over 30%).These recent survey results illustrate the great deal of uncertainty surrounding how states will implement the end of the continuous enrollment requirement and how many individuals will lose coverage.

In the new analysis, KFF researchers examine two potential scenarios based on these survey results, estimating that a 5 percent decline in Medicaid enrollment would translate into a decline of 5.3 million enrollees and a 13 percent decline in Medicaid enrollment would translate into a decline of 14.2 million enrollees. There was a wide range in state estimates, so in any given state the share of enrollees losing coverage could be lower or higher than these illustrative estimates.

The biggest losses are expected to be among adults made eligible for coverage by the Affordable Care Act’s Medicaid expansion, other adults eligible for Medicaid other than due to a disability, and children, as these were the groups that experienced the largest increase during the pandemic. Researchers did not examine how many of the people who lose Medicaid might be able to find coverage from other sources such as the ACA Marketplace, how many might become uninsured and how many might still be eligible for Medicaid but face barriers in retaining coverage.

In exchange for abiding by the continuous enrollment requirement, states became eligible for a 6.2 percentage point bump in the federal Medicaid match rate (“FMAP”) to help them manage rising enrollment and rapidly provide broader fiscal relief by allowing states to reduce state Medicaid spending at a time when states experienced significant drops in revenue early in the pandemic.

KFF analysts find that over time states will have received pandemic-related federal Medicaid money amounting to more than double the higher state Medicaid costs they have faced due to increased enrollment. We did not examine how states used the fiscal relief. State spending on Medicaid is likely to increase following the expiration of the public health emergency and the loss of enhanced federal funding, even after factoring in the expected declines in Medicaid enrollment.

States that have not expanded Medicaid under the ACA have tended to draw more federal fiscal relief than expansion states during the pandemic, chiefly because the enhanced federal matching money does not apply to the ACA adult expansion population.

The new analysis uses data from a companion KFF analysis also released today. In the second analysis, KFF researchers more closely examine typical Medicaid enrollment patterns before the pandemic.

The full analyses are:

• Fiscal and Enrollment Implications of Medicaid Continuous Coverage Requirement During and PHE and After the PHE Ends
• Unwinding the PHE: What We Can Learn From Pre-Pandemic Enrollment Patterns

For more data and analyses related to Medicaid and the pandemic, visit kff.org

Executive Summary

The Families First Coronavirus Response Act (FFCRA), passed during the coronavirus pandemic, includes a Medicaid continuous enrollment requirement that prohibits states from disenrolling people from Medicaid in exchange for additional federal funds. The continuous enrollment requirement has prevented coverage loss and churn (moving off and then back on to coverage) among enrollees during pandemic and, along with other factors, has contributed to the substantial enrollment growth in Medicaid. From February 2020 to January 2022, total Medicaid/CHIP enrollment has increased by 15.7 million enrollees (22.1%). The continuous enrollment requirement will last until the end of the month in which with national public health emergency (PHE) ends. Currently, the PHE is in effect through mid-July 2022 and, at this time, it is expected that the continuous enrollment requirement will end August 1, 2022. The Biden administration has said it will give states 60 days’ notice before the PHE ends, and so if state are not notified by mid-May, the PHE will likely be extended further.

To help inform the current policy discussion and better understand the factors that contribute to Medicaid enrollment changes, this brief examines typical enrollment patterns for Medicaid and CHIP¹  and uses 2018 Medicaid claims data to gain insight into the effects of the continuous enrollment requirements by eligibility group. Overall, we find that:

• Roughly 2% of Medicaid enrollees come on or leave the program in an average month, although there is variation across eligibility groups.
• A policy to require continuous enrollment would result in sharp reductions in monthly disenrollment rates and would also reduce monthly enrollment rates due to reductions in churn.
• Based on data from 2018, the effect of continuous coverage would have greater impact on enrollment for children and non-elderly adults as they are more likely to experience higher rates of churn or income fluctuations affecting eligibility.

These findings from the simulation may help explain how the current continuous enrollment requirement may have contributed to Medicaid enrollment increases and what may happen with the unwinding of the PHE. A description of the data and methods underlying this analysis is in the Methods section at the end of the brief.

What factors contribute to enrollment changes over time?

Medicaid enrollment is driven by several factors, including economic conditions and policy choices. Medicaid is a countercyclical program, meaning that typically during economic downturns, more people qualify and enroll in Medicaid. Economic downturns were the primary driver of Medicaid enrollment growth in the 2001 recession and again during the Great Recession beginning in 2007. Medicaid enrollment also reflects policy changes. For example, Medicaid enrollment growth peaked in FY 2015 due to implementation of the Affordable Care Act (ACA) coverage expansions. Also, prior to the pandemic, enrollment declined in FY 2018 and FY 2019 and was relatively flat in FY 2020 before increasing sharply in 2021.

Enrollment trends reflect the net effect of people coming onto the program as well as people leaving the program. People typically enroll or disenroll from coverage for three main reasons:

• Change in income (for example, a drop in income may make someone newly eligible, while an increase in income may make someone ineligible for Medicaid);
• Change in circumstance other than income (for example, children may age out of coverage, people may move to another state or die); and
• Barriers to renewing coverage that are not based on ongoing eligibility but may result in disenrollment (for example, forms to renew coverage may be confusing or someone may miss a deadline resulting in disenrollment).

It is often difficult to understand the relative magnitude of each of these factors in contributing to net enrollment changes over time.

What do data show about enrollment patterns?

In general, data show that about as many people enroll in Medicaid in a given month as disenroll, with small differences leading to changes in net monthly enrollment. In 2018, the average monthly disenrollment rate was 2.1% while the monthly enrollment rate (people beginning an enrollment span) was 1.9%, which resulted in a net monthly enrollment decline of 0.2% (Figure 1). Earlier research attributed this enrollment decline to a robust economy and to policies and processes in place that may have boosted disenrollment, such as more frequent checks on income or barriers to renewing coverage. At the same time, our analysis of data for other recent years showed similar patterns (approximately 2% of enrollees coming on or off the program each month), with the difference in monthly new enrollment and disenrollment rates leading to an overall increase or decrease in enrollment. As noted above, we expect patterns would have been different (with much higher enrollment rates relative to disenrollment rates) during the Great Recession and after the implementation of the ACA.

Average monthly enrollment and disenrollment rates vary by eligibility group (Figure 1). In 2018, both enrollment and disenrollment rates for elderly enrollees and enrollees with a disability were lower than rates for children and adults. In other words, in a given month, fewer elderly and disabled enrollees were entering and leaving the program relative to total enrollment in those groups compared to children and non-elderly adults. The elevated enrollment rates and disenrollment rates among children and adults reflect that they are more likely to experience income fluctuations affecting eligibility and also leading to higher rates of churn.

Not all enrollment and disenrollment reflects people newly entering or permanently leaving Medicaid. Another analysis that examined churn within Medicaid shows that, among people enrolled in 2018, approximately 10% experienced a gap in enrollment of less than a year. This suggests that the enrollment in each month reflects individuals who had been enrolled, then disenrolled, and then were re-enrolling after a gap in Medicaid coverage, as well as those who were newly eligible. Similarly, disenrollments in each month reflect some people who may re-enroll in Medicaid after a short gap, as well as those who transitioned from Medicaid to other coverage or become uninsured. As noted above, children and non-elderly adults experience the highest rates of churn compared to people with disabilities or enrollees ages 65 and over.

What happens to enrollment patterns under a continuous enrollment requirement?

To help understand the effects of the current Medicaid continuous enrollment policy, we examined what would have happened if there were a similar policy prohibiting disenrollment in place in 2018. Halting disenrollment (except for enrollees who were disenrolled due to a move out of state or death) lowered the average monthly disenrollment rate compared to actual 2018 (0.2% vs. 2.1%, respectively). Halting disenrollments in the simulation also lowered the average enrollment rate compared to actual 2018 (1.4% vs. 1.9%, respectively) due to eliminating churn. Over the course of the year, the simulated continuous enrollment policy resulted in a net increase in enrollment (12.9%, Figure 2) from January to December 2018 compared to the slight decrease in enrollment observed in actual 2018 (-1.9%). When looking at the monthly enrollment changes, the first few months of the simulation experienced slightly faster growth than the later months.

The simulated continuous enrollment policy led to much higher enrollment growth for children and non-elderly adults compared to people with disabilities and those ages 65 and older (Figure 3). This result was also expected, as these groups typically have higher turn-over (i.e., greater disenrollment and enrollment rates) and so they would experience bigger effects from a policy to stop disenrollment.

Children and expansion adults accounted for 76% of the enrollment change due to the simulated continuous enrollment policy for 2018 (Appendix Table 1). These findings are as of December 2018, or after 12 months of the simulated continuous enrollment policy. While the total number of enrollees are not complete because we exclude 10 states from the analysis, we estimate that actual enrollment declined by 1.2 million or 1.9% in 2018 for the 41 states included in our analysis. Under the simulated continuous enrollment policy, enrollment would have increased by 8.0 million or 12.9%. The net change would be an increase of 9.2 million in enrollment. We estimate that, of this net growth, children account for 43% of this net change and expansion adults make up 33% of the change. Other non-elderly adults not enrolled through the Medicaid expansion made up 18% of the net change. These three eligibility groups represent the largest share of total Medicaid/CHIP enrollment²  and also experienced higher average monthly enrollment and disenrollment rates in actual 2018. People with disabilities made up 3% of enrollment due to this net enrollment growth and people ages 65 and older made up 2%.

What to watch going forward?

Understanding historic enrollment patterns and the effects of a simulated policy to halt disenrollment in 2018 can help inform the current continuous enrollment requirement in place. Key findings suggest that, first, halting disenrollment has a significant effect on overall enrollment growth. Second, growth may be steeper when the policy is first put into place, but growth rates may slow as a result of lower churn. Third, the policy would have a significantly disproportionate effect on children and non-elderly adults relative to people with disabilities and those ages 65 and older. Finally, the simulated growth patterns reflect recent experience with Medicaid growth. We observed steep growth in Medicaid enrollment in calendar year 2020 when the continuous coverage policy was put into effect but, similar to the simulation, observed growth rates have slowed over time.

Overall growth rates may be harder to compare given the sharp differences in economic conditions in 2018 compared to recent experience. The fact that observed growth rates from February 2020 through January 2021 showed 13.5% growth, slightly higher than the one-year results of the simulation (12.9%), suggest that much of current growth is attributable to the continuous coverage rather than new enrollment. The economic effects of the COVID-19 recession are complex. Unlike prior recessions that resulted in large shifts in income and losses of employer coverage, this recession largely affected low-wage workers who may have already been eligible for Medicaid.

Finally, as the PHE expires and states are able to resume more routine renewals and disenrollments, our analysis (in addition to current enrollment reports) shows that a large share of non-elderly adults and children could be at risk of losing coverage, even if many continue to be eligible. Efforts to conduct outreach, education and provide enrollment assistance can help ensure that those who remain eligible are able to retain coverage and those who are no longer eligible can transition to other sources of coverage. The Centers for Medicare and Medicaid Services (CMS) released new guidance in March 2022 emphasizing continuity of coverage for enrollees after the end of the continuous enrollment requirement, and the guidance also extended the timeline for states to complete all redeterminations after the continuous enrollment requirement ends from 12 to 14 months (states will still need to initiate all redeterminations within 12 months). However, a recent survey of state officials shows that, as of January 2022, many states had not made key decisions about how to handle the resumption of disenrollments after the end of the PHE. State policy choices about how to prioritize pending eligibility actions, engage with stakeholders, conduct outreach to enrollees, and provide enrollment assistance will have important implications for the continuity of coverage for millions of Medicaid enrollees, especially non-elderly adults and children.

Methods

Data Source and Linkage

Our analysis is based on the Transformed Medicaid Statistical Information System (T-MSIS) Analytical Files (TAF) Research Identifiable Files (RIF). We include beneficiaries who were enrolled at any point in 2018. We use the 2018 Demographic Eligibility (DE) Base file to determine eligibility pathway based on last-best eligibility data. We draw enrollment dates from the 2017-2019 DE Dates files. Our analysis uses enrollment dates from 2017 and 2019 so that we are able to identify people who were enrolled from December 2017 into January 2018, as well as December 2018 into January 2019. Data for 2019 are from the preliminary version of the T-MSIS RIF. All other DE Base and DE Dates files are final versions (Release 1).

State Exclusion Criteria

We use 41 states in our analysis. We excluded 10 states (FL, IN, KY, ME, MS, NE, OK, OR, UT, and WY) due to missing or inconsistent data based on state-level information available at the DQ Atlas as well as our own analysis We relied on data quality assessments from DQ Atlas to exclude states that have a “medium concern”, “high concern”, or “unusable” data quality assessment for: (1) the average monthly Medicaid/CHIP enrollment compared to an external enrollment benchmark (Medicaid/CHIP Performance Indicator Data) (excluded five states (IN, KY, ME, MS, NE)); (2) the average length of enrollment gaps, and (3) the percent of beneficiaries with overlapping Medicaid and S-CHIP enrollment spans (no states were excluded based on measures 2 or 3). We further excluded states based on: (4) the percent of beneficiaries missing an eligibility group code (threshold of >=10%, excluded OR); (5) the percent of beneficiaries with only one enrollment span (threshold of >=99.5, excluded FL, KY, and WY); and (6) the percent of beneficiaries with three or more enrollment spans in a year (threshold of >=5%, excluded OK). Notably, our exclusion criteria related to the number of enrollment spans in a year (measures 5 and 6) are less restrictive than that in the DQ Atlas because we did not want to make assumptions about the number of enrollment spans in our analysis, but did want to remove extreme cases that are the most likely to represent inaccurate enrollment dates.

Beneficiary Linking, Eligibility Classification, and Exclusion

We linked individuals across years using BENE_ID, which are unique enrollee identifiers created by the Chronic Conditions Warehouse and are recommended for use when combining data for multiple years. We also use BENE_ID to link the DE Base and DE Dates files. We assigned restricted benefit status and eligibility group code using a last-best approach for 2018, which assigns eligibility based on the most recent eligibility code in 2018. We classified eligibility groups using a hierarchy that first checked if the eligibility group code was missing, then checked for medically needy eligibility, disability (under age 65), and expansion adult. Any enrollees in the DE Base file that had a non-missing eligibility code and had not been assigned an eligibility group through this hierarchy were then were assigned by age to children (ages 20 and under), adults (ages 21 to 64), and aged (ages 65 and over).

The 2018 DE Base files for 41 states in our analysis contained 74.0 million observations coded as full-benefit after removing a small number of “dummy” records that represent enrollees who have claims data but no eligibility data provided by the states. We removed observations missing eligibility codes (151,000 observations), people qualifying through a medically needy pathway (1.5 million observations), people missing a BENE_ID for linking files (953,000 observations), and people with duplicated BENE_ID (102,000 observations). After merging with the DE Dates files, there were less than 3,000 enrollees who did not have a matching BENE_ID in the DE Dates and DE Base files and were dropped from the final sample. Our final sample included 71.3 million unduplicated, full-benefit enrollees.

Calculating Enrollment Trends

The DE Dates files provide a start and end date for every enrollment span in our time period. However, the earliest start date in any DE Date file in January 1^st of that year, and the latest end date is December 31^st of that year. Using the 2017 and 2019 DE Dates files, linked by BENE_ID, we were able to identify people who started enrollment spans in January 2018 (versus people who were continuously enrolled from December 2017) and people who disenrolled in December 2018 (versus people who were continuously enrolled into January 2019). Before calculating monthly enrollment and disenrollment rates, we merged all overlapping and contiguous enrollment spans for enrollees, which we defined as enrollment spans that are separated by one day or less. For example, if a person has two enrollment spans with an end date and a start date that are one day apart (i.e., the person disenrolled and re-enrolled the next day), we considered these spans to be contiguous and merged them into one enrollment span. After merging overlapping and contiguous enrollment spans, we counted the number of enrollment starts and disenrollments in a month (the enrollment and disenrollment rates). Enrollment counts are de-duplicated counts of people enrolled at any point in the month. Average monthly net change in enrollment is the average of monthly enrollment change from January 2018 to December 2018.

Simulating A Continuous Enrollment Policy

To simulate a continuous enrollment policy in 2018, we assume that no one is disenrolled from Medicaid and CHIP except in cases of death or moving to a different state. We used the 2018 DE Base file to identify death dates. We used the 2018 and 2019 DE Bases files to identify people who moved states and then re-enrolled in Medicaid and CHIP (and used BENE_ID for linking files). Notably, we do not have information about why a person was disenrolled or length of time before they enrolled in another state (we only can see that they were enrolled in another state after disenrollment). In total, there were 1.5 million people disenrolled in the simulation due to death or moving out of state.

After accounting for death and moving, we de-duplicated enrollees with multiple spans in 2018. We then used each enrollee’s earliest enrollment date in 2018 (or, if continuously enrolled from 2017, kept the date from 2017), and extended coverage through the year unless the enrollee had a death date or “move out” date. Finally, using this simulated sample, we calculated monthly enrollment rates, disenrollment rates, and total enrollment as described above.

Our simulation has important differences from the current continuous enrollment requirement under the FFCRA. First, our simulation applies to Medicaid and CHIP; however, while the current continuous enrollment requirement under the FFCRA does not apply to separate CHIP programs. In total, 3.6% of enrollees in our sample were only enrolled in separate CHIP programs throughout 2018 (and never enrolled in Medicaid through the year). Additionally, our simulation does not match the economic conditions of the pandemic, and the simulation does not account for other related factors such as elevated death rates or state or federal policy actions facilitating enrollment in Medicaid, CHIP, or Marketplace coverage.

Appendix

1. Our simulation of a continuous enrollment policy in 2018 applies to both Medicaid and the Children’s Health Insurance Program (CHIP). Although the continuous coverage policy under the FFCRA does not apply to separate CHIP programs, it does apply to CHIP Medicaid expansion programs. See the Methods section for more details on the simulation. ↩︎
2. January 2018 enrollment was the same in the simulation analysis and in actual 2018. The share of total full-benefit Medicaid/CHIP enrollment for each eligibility group in January 2018 were as follows: children (48%), expansion adults (22%), other adults (13%), people with disabilities (11%), and individuals ages 65 and older (7%). See Appendix Table 1 for enrollment counts. ↩︎

Introduction

Under longstanding immigration policy, federal officials can deny entry to the U.S. or adjustment to lawful permanent resident (LPR) status (i.e., a “green card”) to someone they determine to be a public charge. On September 9, 2022, the Biden Administration published new public charge inadmissibility regulations.¹  that have largely codified 1999 field guidance governing public charge determinations and these regulations went into effect on December 23, 2022.²  The Biden Administration returned to use of this 1999 field guidance in March 2021, when it stopped applying public charge regulations implemented by the Trump Administration in 2019 that had newly considered the use of noncash assistance programs, including Medicaid, in public charge determinations.^3 ,4  Under the final rule and the 1999 field guidance, the government will not consider use of noncash benefit programs, including Medicaid coverage, except for long-term institutionalization, when making public charge determinations. A primary stated aim of the final public charge rule is to address chilling effects of the 2019 rule that led many immigrant families, including citizen children in these families, to not seek assistance, including health coverage and care, for which they were eligible.^5 ,6 ,7 

This brief provides an overview of the public charge rule and discusses its potential impacts for health coverage and care of immigrant families.

History of Public Charge Policies

The concept of denying entry to immigrants on the basis of becoming a “public charge” was first introduced into federal legislation as part of the Immigration Act of 1882, which was the first act to regulate immigration at the federal level.⁸  This Act permitted the government to prevent any person “unable to take care of himself or herself without becoming a public charge” from entering the country. A subsequent revised version of the Act allowed the government to deny entry to “paupers or persons likely to become a public charge,” and allowed the deportation of any who did become a public charge within a year. Certain immigrants, including refugees and asylees and other humanitarian immigrants, are exempt from public charge determinations under law.⁹  Immigration officials must consider, at a minimum, an individual’s age, health, family status, assets, resources, and financial status, and education and skills when making a public charge inadmissibility determination. However, federal legislation does not define who should be considered a public charge, leaving substantial discretion to immigration officials to make those determinations.

In 1999, the Department of Justice issued Field Guidance on Deportability and Inadmissibility on Public Charge Grounds. This field guidance defined a public charge as “an alien who has become or is likely to become primarily dependent on the federal government for subsistence as demonstrated by either (i) the receipt of public cash assistance for income maintenance or (ii) institutionalization for long-term care at government expense.“ In response to confusion about whether the use of Medicaid, the Children’s Health Insurance Program (CHIP), and other non-cash programs applied in public charge determinations, the guidance further clarified that immigration officers should not place any weight on the receipt of non-cash public benefits (other than institutionalization) with respect to determinations of admissibility or eligibility for adjustment to LPR status.

In 2019 the Trump Administration issued new regulations that broadened the scope of programs that the federal government would consider in public charge determinations to include previously excluded health, nutrition, and housing programs. The regulations also redefined a public charge as an “alien who receives one or more public benefits for more than 12 months in the aggregate within any 36-month period (such that, for instance, receipt of two benefits in one month counts as two months),” and defined public benefits to include federal, state, or local cash benefit programs for income maintenance and certain health, nutrition, and housing programs, including non-emergency Medicaid for non-pregnant adults, the Supplemental Nutrition Assistance Program (SNAP), and several housing programs. It also identified characteristics that would be considered as negative factors and heavily weighted negative factors that increase the likelihood of someone becoming a public charge, including prior receipt of public benefits (for more than 12 months within the prior 36 months) and having income below 125% of the federal poverty level (FPL) ($28,787 for a family of three as of 2022), as well as positive factors and heavily weighted positive factors that decrease the likelihood of someone becoming a public charge.

Impacts of the 2019 Trump Administration Changes

Research suggests that the 2019 Trump Administration changes to public charge policy and other immigration policy changes increased fears among immigrant families about participating in programs and seeking services, including health coverage and care.¹⁰  Although very few people subject to public charge determinations are eligible for the public programs listed in the 2019 rule due to the programs’ restrictions on eligibility for immigrants, the rule had much broader chilling effects on program participation across immigrant families.¹¹  Fears and confusion related to the rule led to individuals forgoing enrollment in or disenrolling themselves and their children from programs. Prior KFF analysis estimated that the rule could lead to disenrollment of between 2.0 to 4.7 million Medicaid and CHIP enrollees who are noncitizens or citizens living in a family with a noncitizen if the rule led to disenrollment rates ranging from 15% to 35%. A 2021 KFF survey of Hispanic adults found that 1 in 4 potentially undocumented Hispanic adults and over 1 in 10 lawful permanent resident Hispanic adults reported that they or a family member did not participate in a government assistance program in the past three years  due to immigration-related fears (Figure 1).¹²  Similarly, a KFF survey of Asian community health center patients found that over 50% reported not having enough information about how recent immigration policy changes, including the public charge rule, impact them and their family, and 1 in 4 reported that they or a family member in their household avoided participating in a publicly-funded health, nutrition, or housing program in the past 12 months due to immigration-related fears.¹³ 

Research also suggests that immigration-related fears have affected willingness to access COVID-19 vaccines. For example, KFF COVID-19 Vaccine Monitor survey data from May 2021 found that almost 4 in 10 unvaccinated Hispanic adults (rising to 58% of those who are potentially undocumented) said they were concerned that they may be required to provide a government-issued identification or Social Security number to get vaccinated and about one third (rising to 63% of the potentially undocumented) were concerned that getting the vaccine would negatively affect their own or a family member’s immigration status.

Overview of the 2022 Final Rule on Public Charge

The Biden Administration reversed the public charge policy changes implemented by the Trump Administration. Following an array of court actions challenging implementation of the 2019 Trump Administration changes to public charge policies, in March 2021, the Biden Administration filed motions asking the Supreme Court to dismiss the federal government’s appeal to maintain the rule and reinstated use of the 1999 field guidance to govern public charge determinations. It subsequently proposed a new public charge rule on February 24, 2022, which would largely codify the 1999 field guidance. The final rule was published on September 9, 2022 and went into effect on December 23, 2022.

A primary purpose of the 2022 public charge rule is to address the chilling effects of the 2019 rule on immigrant families’ participation in public programs, including Medicaid and CHIP. As mentioned in the preamble, “DHS believes that, in contrast to the 2019 Final Rule, this proposed rule would effectuate a more faithful interpretation of the statutory concept of ‘‘likely at any time to become a public charge’’; avoid unnecessary burdens on applicants, adjudicators, and benefits-granting agencies; and mitigate the possibility of widespread ‘‘chilling effects’’ with respect to individuals disenrolling or declining to enroll themselves or family members in public benefits programs for which they are eligible, especially by individuals who are not subject to the public charge ground of inadmissibility.” The preamble also discusses the 2019 rule’s chilling effects on immigrant families during the COVID-19 pandemic, contributing to ‘‘reduce[d] access to vaccines and other medical care, resulting in an increased risk of an outbreak of infectious disease among the general public,’’ which could perpetuate inequities in morbidity and mortality due to the pandemic.

The 2022 public charge rule returned to the 1999 public charge definition of someone who is likely to become primarily dependent on the federal government as demonstrated by the use of cash assistance programs for income maintenance or government-funded institutionalized long-term care, such as mental health or nursing home care (Table 1). Under the final rule, the federal government will only consider cash assistance programs, including Supplemental Security Income (SSI), Temporary Assistance for Needy Families, and state, local, and Tribal cash assistance to pay for basic needs such as rent, food, and utilities. It also will consider “long-term government assistance for institutionalization,” including that provided in a nursing home or mental health institution. Long-term institutionalization at government expense would be the only category of Medicaid-funded services to be considered in public charge determinations. The 2022 rule clarifies that “long-term institutionalization” excludes institutionalization for short periods for rehabilitation and institutionalization that violates federal discrimination laws (such as the Americans with Disabilities Act).¹⁴  The rule also specifies that it excludes receipt of home and community-based services, including those provided through Medicaid. Moreover, to mitigate chilling effects, the 2022 rule clarifies that applying for a public benefit, being approved for benefits in the future, assisting someone else to apply for benefits, or being in a household or family with someone who receives benefits would not constitute receipt of public benefits for consideration of public charge. This represents a change from the 1999 Field Guidance, which allowed for consideration of a noncitizen’s family’s reliance on public cash benefits as the sole means of support for the family.

The rule applies a forward-looking test to public charge determinations, where adjudicators predict whether a person could become a public charge in the future based on present factors including age, family status, income and resources, education, and health status. Under the 2022 rule, the aforementioned statutory minimum factors must be analyzed in their totality, and no single factor, other than the lack of a sufficient Affidavit of Support, when required, would control the decision. Under the 2019 rule, public charge determinations also were based on a totality of circumstances, but the rule further specified that certain factors should we weighted as negative or positive or heavily weighted negative or positive factors in determinations. For example, officials were directed to consider receiving public benefits for more than 12 months in a 36-month period as a heavily weighted negative factor. The 2022 rule eliminates this practice of separately defining each factor and assigning its weight, noting that, “each inadmissibility determination must be individualized and based on the evidence presented in the specific case, and the relative weight of each factor and associated evidence is necessarily determined by the presence or absence of specific facts. Consequently, the designation of some factors as always ‘‘heavily weighted’’ suggested a level of mathematical precision that would be unfounded and inconsistent with the long-standing standard of considering the totality of the individual’s circumstances.” In addition, under the 2022 public charge rule, U.S. Citizenship and Immigration Services officers are required to articulate a reason for every public charge determination.

Implications

The Biden Administration's changes to public charge policies are intended to reduce fears of accessing programs, but it will likely require sustained community-level efforts to rebuild trust and reduce fears among families. Prior experience suggests that outreach and education from trusted members of the community will be important for helping to alleviate fears and that direct one-on-one assistance will be key for facilitating enrollment of eligible people into programs for which they and their children are eligible, including Medicaid and CHIP.¹⁵  The Biden Administration recently increased funding for Navigators to help eligible individuals enroll in Marketplace, Medicaid, and CHIP coverage with a focus on individuals with limited English proficiency (LEP), communities of color, and rural areas, which may facilitate increased enrollment of eligible individuals in immigrant families.¹⁶  However, broader multilingual outreach efforts focused on helping families understand the changes to the rule will also likely be important.

Even with community level outreach and enrollment efforts, some families may remain fearful and confused about the policy. For example, in earlier research, some families have expressed concerns that rules may change in the future. Moreover, some families may remain confused about the policy. In particular, the continued inclusion of long-term institutional care covered by Medicaid in public charge determinations prevents clear and simple messaging that all Medicaid coverage is excluded from public charge determinations. This consideration of long-term institutional care also disproportionately affects people with disabilities and older adults, who are more likely to use long-term institutionalization services.

Increased enrollment of eligible uninsured immigrants may help to close gaps in health coverage for immigrant families, but many uninsured immigrants remain ineligible for coverage. Overall, noncitizen immigrants are significantly more likely than citizens to be uninsured.¹⁷  In 2020, among the nonelderly population, 26% of lawfully present immigrants and about four in ten (42%) undocumented immigrants were uninsured compared to less than one in ten (8%) citizens. Moreover, among citizen children, those with at least one noncitizen parent were significantly more likely to be uninsured as those with citizen parents (10% vs. 4%). Most (81%) uninsured lawfully present immigrants are eligible for a coverage option under the ACA but are not enrolled, so facilitating enrollment of eligible individuals may help to narrow these coverage gaps. However, other immigrants, including undocumented immigrants, remain ineligible for federal health coverage options, and will continue to face increased challenges accessing health care.

1. U.S. Citizenship and Immigration Services Department of Homeland Security (2022), “Public Charge Ground of Inadmissibility,” https://www.federalregister.gov/documents/2022/09/09/2022-18867/public-charge-ground-of-inadmissibility, accessed September 13, 2022. ↩︎
2. Department of Justice (1999), “Field Guidance on Deportability and Inadmissibility on Public Charge Grounds, https://www.govinfo.gov/content/pkg/FR-1999-05-26/pdf/99-13202.pdf, accessed April 15, 2022. ↩︎
3. U.S. Citizenship and Immigration Services Department of Homeland Security (2019), “Inadmissibility on Public Charge Grounds,” https://www.govinfo.gov/content/pkg/FR-2019-08-14/pdf/2019-17142.pdf, accessed April 15, 2022. ↩︎
4. Kaiser Family Foundation (KFF) (2019), “Changes to “Public Charge” Inadmissibility Rule: Implications for Health and Health Coverage,” https://modern.kff.org/racial-equity-and-health-policy/fact-sheet/public-charge-policies-for-immigrants-implications-for-health-coverage/#:~:text=In%20August%202019%2C%20the%20Trump,entry%20into%20the%20U.S.%20or, accessed April 22, 2022. ↩︎
5. Artiga, S., Garfield, R., & Damico, A. (2019), “Estimated Impacts of Final Public Charge Inadmissibility Rule on Immigrants and Medicaid Coverage,” https://modern.kff.org/report-section/estimated-impacts-of-final-public-charge-inadmissibility-rule-on-immigrants-and-medicaid-coverage-key-findings/, accessed April 15, 2022. ↩︎
6. Artiga, S., Damico, A., & Garfield, R. (2019), “Potential Effects of Public Charge Changes on Health Coverage for Citizen Children,” https://modern.kff.org/racial-equity-and-health-policy/issue-brief/potential-effects-of-public-charge-changes-on-health-coverage-for-citizen-children/, accessed April 15, 2022. ↩︎
7. Tolbert, J., Artiga, S., & Pham, O. (2019), “Impact of Shifting Immigration Policy on Medicaid Enrollment and Utilization of Care among Health Center Patients,” https://modern.kff.org/medicaid/issue-brief/impact-of-shifting-immigration-policy-on-medicaid-enrollment-and-utilization-of-care-among-health-center-patients/, accessed April 22, 2022. ↩︎
8. University of California, Berkeley (2020), “The History of the Public Charge and Public Health,” https://pha.berkeley.edu/2020/12/29/the-history-of-the-public-charge-and-public-health/, accessed April 21, 2022. ↩︎
9. U.S. Citizenship and Immigration Services (2021), “Public Charge Fact Sheet,” https://www.uscis.gov/archive/public-charge-fact-sheet, accessed April 22, 2022. ↩︎
10. Artiga, S., & Ubri, P. (2017), “Living in an Immigrant Family in America: How Fear and Toxic Stress are Affecting Daily Life, Well-Being, & Health,” https://modern.kff.org/racial-equity-and-health-policy/issue-brief/living-in-an-immigrant-family-in-america-how-fear-and-toxic-stress-are-affecting-daily-life-well-being-health/, accessed April 22, 2022. ↩︎
11. Migration Policy Institute (2020), “The Public-Charge Rule: Broad Impacts, But Few Will Be Denied Green Cards Based on Actual Benefits Use,” https://www.migrationpolicy.org/news/public-charge-denial-green-cards-benefits-use, accessed April 22, 2022. ↩︎
12. Hamel, L., Artiga, S., Safarpour, A., Stokes, M., & Brodie, M. (2021), “KFF COVID-19 Vaccine Monitor: COVID-19 Vaccine Access, Information, and Experiences Among Hispanic Adults in the U.S.,” https://modern.kff.org/coronavirus-covid-19/poll-finding/kff-covid-19-vaccine-monitor-access-information-experiences-hispanic-adults/, accessed April 15, 2022. ↩︎
13. Artiga, S., Hill, L., Corallo, B., & Tolbert, J. (2021), “Asian Immigrant Experiences with Racism, Immigration-Related Fears, and the COVID-19 Pandemic,” https://modern.kff.org/report-section/asian-immigrant-experiences-with-racism-immigration-related-fears-and-the-covid-19-pandemic-findings/, accessed April 15, 2022. ↩︎
14. U.S. Citizenship and Immigration Services Department of Homeland Security (2022), “Public Charge Ground of Inadmissibility,” https://www.dhs.gov/sites/default/files/2022-02/22_0217_nprm-public-charge_0.pdf, accessed April 26, 2022. ↩︎
15. Kaiser Family Foundation (KFF) (2011), “Connecting Eligible Immigrant Families to Health Coverage and Care: Key Lessons from Outreach and Enrollment Workers,” https://modern.kff.org/racial-equity-and-health-policy/issue-brief/connecting-eligible-immigrant-families-to-health-coverage/, accessed April 22, 2022. ↩︎
16. Centers for Medicare and Medicaid Services (2021), “Biden-Harris Administration Quadruples the Number of Health Care Navigators Ahead of Healthcare.Gov Open Enrollment Period,” https://www.cms.gov/newsroom/press-releases/biden-harris-administration-quadruples-number-health-care-navigators-ahead-healthcaregov-open, accessed April 15, 2022. ↩︎
17. Kaiser Family Foundation (KFF) (2022), “Health Coverage of Immigrants,” https://modern.kff.org/racial-equity-and-health-policy/fact-sheet/health-coverage-of-immigrants/#:~:text=In%202020%2C%20among%20the%20nonelderly,in%20ten%20(8%25)%20citizens., accessed April 15, 2022. ↩︎

Asian, Native Hawaiian, and Other Pacific Islander (NHOPI) people are a diverse and growing population in the United States. Asian people are the fastest-growing racial or ethnic group in the United States, with their population increasing 81% from 10.5 million to 18.9 million between 2000 and 2019.

Given the diversity of the population, broad data on Asian and NHOPI people often mask underlying disparities among subgroups of the population. For example, there is a nearly 8-fold difference in uninsured rates among Asian and NHOPI subgroups, ranging from 4% for Japanese people to 32% for Mongolian people. Uninsured rates also vary by citizenship status among Asian and NHOPI people, with noncitizens more likely to lack coverage across groups.

Data are also often missing to identify and address disparities. Enhancing data to better understand the experiences of Asian and NHOPI people is of particular importance at this time, given growing levels of racism and discrimination amid the COVID-19 pandemic, including a significant uptick in hate incidents against Asian and NHOPI people. A KFF survey of Asian community health center patients found that 1 in 3 respondents experienced more discrimination in 2021 than when the COVID-19 pandemic began and that many reported negative experiences due to their race or ethnicity, ranging from receiving poor service to being verbally or physically attacked.

Source

KFF analysis of 2019 American Community Survey 1-yr Public Use Microdata Sample, 2021 KFF survey of Asian community health center patients