Global COVID-19 Tracker

Published: Sep 9, 2025

Editorial Note: The Policy Actions tracker will no longer be updated as the data source has ceased tracking government responses to COVID-19. For more information, please visit the Oxford Covid-19 Government Response Tracker.

Cases and Deaths

This tracker provides the cumulative number of confirmed COVID-19 cases and deaths, as well as the rate of daily COVID-19 cases and deaths by country, income, region, and globally. It will be updated weekly, as new data are released. As of March 7, 2023, all data on COVID-19 cases and deaths are drawn from the World Health Organization’s (WHO) Coronavirus (COVID-19) Dashboard. Prior to March 7, 2023, this tracker relied on data provided by the Johns Hopkins University (JHU) Coronavirus Resource Center’s COVID-19 Map, which ended on March 10, 2023. Please see the Methods tab for more detailed information on data sources and notes. To prevent slow load times, the tracker only contains data from the last 200 days. However, the full data set can be downloaded from our GitHub page. While the tracker provides the most recent data available, there is a two-week lag in the data reporting.

Note: The data in this tool were corrected on March 18, 2024, to clarify that they represent new cases and deaths over a full week rather than the average per day over a seven-day period.

Policy Actions

Editorial Note

The Policy Actions tracker will no longer be updated as the data source has ceased tracking government responses to COVID-19. For more information, please visit the Oxford Covid-19 Government Response Tracker.

This tracker contains information on policy measures currently in place to address the COVID-19 pandemic. Policy categories currently being tracked include social distancing & closure measures, economic measures, and health systems measures. Policies are tracked at the country-, income-, and region-level. Please see the Methods tab for more detailed information on data sources and notes.

Social Distancing and Closure Measures

As countries continue to implement policies to prevent the transmission of SARS-CoV-2, the virus that causes COVID-19, these tables and charts show which social distancing and closure measures are currently in place by country.

Global COVID-19 Policy Actions

Economic Measures

The COVID-19 pandemic has placed an unprecedented strain on country economies. These tables and charts show which economic-related measures, namely income support and debt relief, are currently in place by country.

Global COVID-19 Policy Actions

Health Systems Measures

The COVID-19 pandemic continues to strain and disrupt global health systems. These tables and charts show which health systems measures are currently in place by country.

Global COVID-19 Policy Actions

Methods

Cases and Deaths

SOURCES

As of March 7, 2023, all data on COVID-19 cases and deaths are drawn from the World Health Organization’s (WHO) Coronavirus (COVID-19) Dashboard. Prior to March 7, 2023, this tracker relied on data provided by the Johns Hopkins University (JHU) Coronavirus Resource Center’s COVID-19 Map, which ends on March 10, 2023. Population data are obtained from the United Nations World Population Prospects using 2021 total population estimates. Income-level classifications are obtained from the latest World Bank Country and Lending Groups. Regional classifications are obtained from the World Health Organization.

Policy Actions

NOTES

Policy actions data include the measure that was in place for each indicator at the country-level as of the end of 2022. Policy actions data will no longer be updated as the data source has ceased tracking government responses to COVID-19. For more information, please visit the Oxford Covid-19 Government Response Tracker.

Social Distancing and Closure Measures

Under ‘Stay At Home Requirements’, exceptions for leaving the house may include anything from being able to leave for daily exercise, grocery shopping, and essential trips, to only being allowed to leave once a week, or one person may leave at a time, etc. Under ‘Workplace Closing’, partial closing includes instances in which a country recommends closing the workplace (or working from home); businesses are open but with significant COVID-19-related operational adjustments; or when workplaces require closing for only some, but not all, sectors or categories of workers. Under ‘School Closing’, partial closing includes instances in which a country has recommended school closures; all schools are open but with significant COVID-19-related operational adjustments; or some schools, but not all, are closed; full closing includes schools that are in session but operating virtually. Under ‘Restrictions On Gatherings’, partial restrictions include restrictions on gatherings of more than 10 people; full restrictions include restrictions on gatherings of 10 people or less. Under ‘International Travel Controls’, partial restrictions include screening and quarantine requirements for those entering the country. Values for ‘Cancel Public Events’ were not recodified.

Economic Measures

Under ‘Income Support’, narrow support includes instances in which a country’s government is replacing less than 50% of lost salary (or if a flat sum, it is less than 50% median salary); broad support includes instances in which a country’s government is replacing 50% or more of lost salary (or if a flat sum, it is greater than 50% median salary). Under ‘Debt/Contract Relief’, narrow support includes instances in which a country’s government is providing narrow relief, such as relief specific to one kind of contract.

Health Systems Measures

Under ‘Vaccine Eligibility’, partial availability includes availability for some or all of the following groups: key workers, non-elderly clinically vulnerable groups, and elderly groups, or for select broad groups/ages. Under ‘Facial Coverings’, recommend/partial requirement includes instances in which a country’s government recommends wearing facial coverings, requires facial coverings in some situations, and requires facial coverings when social distancing is not possible. 

SOURCES

Data on and descriptions of government measures related to COVID-19 provided by the Oxford Covid-19 Government Response Tracker (OxCGRT). For more detailed information on their data collection and methodology, please see their codebook and interpretation guide.

How Do Health Care Prices and Utilization in the United States Compare to Peer Nations?

Authors: Delaney Tevis, Matt McGough, Juliette Cubanski, Matthew Rae, and Cynthia Cox
Published: Sep 4, 2025

This updated chart collection compares indicators of health care utilization and prices in the United States and 11 similarly wealthy countries to investigate whether higher prices or higher utilization of healthcare services drives the high health care expenditures in the U.S. relative to peer nations.

The U.S. spends nearly twice as much on health care per person as peer nations ($13,432 vs. $7,393 per person), meanwhile health care utilization in the U.S. — from doctor visits to surgeries — is generally lower than in other wealthy countries. The evidence continues to support the finding that higher prices – as opposed to higher utilization – explain the United States’ high health spending relative to other high-income countries.

The analysis is part of the Peterson-KFF Health System Tracker, an online information hub dedicated to monitoring and assessing the performance of the U.S. health system.

State and Federal Reproductive Rights and Abortion Litigation Tracker

Last updated on

The Supreme Court’s Dobbs ruling, overturning Roe v. Wade, returned the decision to restrict or protect abortion to states. In many states, abortion providers and advocates are challenging state abortion bans contending that the bans violate the state constitution or another state law. The state litigation tracker presents up-to-date information on the ongoing litigation challenging state abortion policy.

In addition, since the Dobbs decision, new questions have arisen regarding the intersection of federal and state authority when it impacts access to abortion and contraception. Litigation has been brought in federal court to resolve some of these questions. The federal litigation tracker presents up-to-date information on the litigation in federal courts that involves access to contraception and abortion.

Status of Abortion Litigation in State Courts, as of 2/14/2023
Status of Abortion Litigation in State Courts, as of February 15, 2023

State and Federal Reproductive Rights and Abortion Litigation Tracker

Last updated on

The Supreme Court’s Dobbs ruling, overturning Roe v. Wade, returned the decision to restrict or protect abortion to states. In many states, abortion providers and advocates are challenging state abortion bans contending that the bans violate the state constitution or another state law. The state litigation tracker presents up-to-date information on the ongoing litigation challenging state abortion policy.

In addition, since the Dobbs decision, new questions have arisen regarding the intersection of federal and state authority when it impacts access to abortion and contraception. Litigation has been brought in federal court to resolve some of these questions. The federal litigation tracker presents up-to-date information on the litigation in federal courts that involves access to contraception and abortion.

Litigation Involving Reproductive Health and Rights in the Federal Courts, as of February 15, 2023

Medicaid Enrollment and Unwinding Tracker

Published: Sep 4, 2025

Enrollment Data

Note: The data presented below are updated monthly as new Medicaid/CHIP enrollment data become available.

The Medicaid Enrollment and Unwinding Tracker presents the most recent data on monthly Medicaid/CHIP enrollment reported by the Centers for Medicare & Medicaid Services (CMS) as part of the Performance Indicator Project as well as archived data on renewal outcomes reported by states during the unwinding of the Medicaid continuous enrollment provision. The unwinding data were pulled from state websites, where available, and from CMS.

Medicaid/CHIP enrollment trends generally use February 2020 as the baseline month because it was the month prior to the start of the COVID-19 pandemic and implementation of the continuous enrollment provision. During continuous enrollment, which was in place during the three years of the pandemic, states paused Medicaid disenrollments. As a result, when the continuous enrollment provision ended in March 2023, national Medicaid/CHIP enrollment had increased to a record high of 94 million enrollees. Beginning April 1, 2023, states could resume disenrolling people after conducting renewals to verify eligibility for the program, though some states delayed the start of their unwinding periods until May, June, or July 2023. Most states took 12 months to complete unwinding renewals and nearly all states completed renewals by August 2024.

The figures below show Medicaid and CHIP enrollment from February 2020 through the most current month of available data. Some figures also include enrollment for adults and children in Medicaid/CHIP. Key enrollment trends as of May 2025 include:

  • There are 78.1 million people enrolled in Medicaid/CHIP nationally (Figure 1). This represents a 17% decline from total Medicaid/CHIP enrollment in March 2023, but is still 9% higher than Medicaid/CHIP enrollment in February 2020, prior to the pandemic (Figure 2 and Table 1).
  • Several factors likely explain why national Medicaid/CHIP enrollment is higher than pre-pandemic enrollment. The pandemic may have encouraged some people who were previously eligible for Medicaid but not enrolled to newly enroll in the program. During the unwinding, many states took steps to improve their renewal processes, which reduced the number of people who were disenrolled despite remaining eligible. In addition, some states expanded eligibility for certain groups since the start of the pandemic, such as the Affordable Care Act’s (ACA) Medicaid expansion.
  • Medicaid/CHIP enrollment is higher than pre-pandemic levels in all but eleven states (AR, CO, ID, IA, LA, MT, NM, SC, TN, TX, and WV). Enrollment changes from pre-pandemic baseline vary from a 15% decrease in Montana to a 54% increase in North Carolina (Figure 2). Many of the states with the largest increases in enrollment expanded eligibility since the start of the pandemic. For example, five states (NE, OK, MO, SD, and NC) implemented the Medicaid expansion between October 2020 and December 2023 and Maine increased the income limit for children to qualify for Medicaid.
  • In the 49 states and DC with complete enrollment data by age, there are 36.4 million children (48%) and 39.9 million adults (52%) enrolled, a change from pre-pandemic (February 2020) enrollment patterns when children made up a slight majority (51%) of Medicaid/CHIP enrollees (Figure 1).
  • Child enrollment in Medicaid/CHIP is below pre-pandemic enrollment in 15 states, while adult enrollment is below pre-pandemic levels in 8 states (Figure 2).
  • There are 70.8 million people enrolled in Medicaid and 7.3 million people enrolled in CHIP (Figure 1). More states report Medicaid enrollment above their pre-pandemic baselines compared to the number reporting CHIP enrollment above the baseline (Figure 2).
National Enrollment in Medicaid/CHIP, February 2020 to April 2025
Cumulative Percent Changes in Enrollment from February 2020 to April 2025
Total Medicaid/CHIP Enrollment, Selected Time Periods

Unwinding Data – Archived

Note: The data on unwinding renewal outcomes presented below were last updated on September 12, 2024; since most states have now completed the Medicaid unwinding, the information will not be updated again.

As of September 12, 2024 and with nearly complete unwinding data for most states: 

  • Over 25 million people were disenrolled (31% of completed renewals) and over 56 million people had their coverage renewed (69% of completed renewals).  
  • Disenrollment rates varied across states from 57% in Montana to 12% in North Carolina, driven by a variety of factors including differences in renewal policies and procedures as well as eligibility expansions in some states.  
  • Among those who were disenrolled, nearly seven in ten (69%) were disenrolled for paperwork or procedural reasons while three in ten (31%) were determined ineligible.  
  • Among those whose coverage was renewed during the unwinding, 61% were renewed on an ex parte, or automated, basis, meaning the individual did not have to take any action to maintain coverage. 

State Data on Renewal Outcomes

The data on unwinding-related renewal outcomes presented in this section rely primarily on monthly reports that states were required to submit to the Centers for Medicare & Medicaid Services (CMS) during the unwinding period. The data also reflect updates to the monthly reports that states submit three months after the original report submission to account for the resolution of pending cases and any other changes in renewal metrics. For 13 states, data were pulled from dashboards or reports published on state websites that provide more complete information, and for a few additional states, updated monthly reports were pulled from state websites because they were more timely than what is reported on the CMS website. 

To view archived data for specific states, click on the State Data – Archived tab.

 

As of August 1, 2024, States Have Reported Renewal Outcomes for Over Eight in Ten People who were Enrolled in Medicaid/CHIP Prior to the Start of the Unwinding

 

Medicaid Disenrollments

  • As of September 12, 2024, at least 25,198,000 Medicaid enrollees had been disenrolled during the unwinding of the continuous enrollment provision. Overall, 31% of people with a completed renewal were disenrolled in reporting states while 69%, or 56.4 million enrollees, had their coverage renewed.
  • There is wide variation in disenrollment rates across reporting states, ranging from 57% in Montana to 12% in North Carolina. A variety of factors contribute to these differences, including differences in renewal policies and system capacity. Some states adopted policies that promote continued coverage among those who remain eligible and/or have automated eligibility systems that can more easily and accurately process renewals while other states have adopted fewer of these policies and have more manually-driven systems. In addition, North Carolina and South Dakota adopted Medicaid expansion and other states increased eligibility levels for certain populations (e.g., children, parents, etc.) during the unwinding, which may have lowered disenrollment rates in these states.

At least <b>24,838,000</b> Medicaid enrollees have been disenrolled with publicly available unwinding data, as of August 1, 2024

 

  • Across all states with available data, 69% of all people disenrolled had their coverage terminated for procedural reasons. However, these rates vary based on how they are calculated (see note below). Procedural disenrollments are cases where people are disenrolled because they did not complete the renewal process and can occur when the state has outdated contact information or because the enrollee does not understand or otherwise does not complete renewal packets within a specific timeframe. High procedural disenrollment rates are concerning because many people who are disenrolled for these paperwork reasons may still be eligible for Medicaid coverage. 

(Note: The first tab in the figure below calculates procedural disenrollment rates using total disenrollments as the denominator. The second tab shows these rates using total completed renewals, which include people whose coverage was terminated as well as those whose coverage was renewed, as the denominator. And finally, the third tab calculates the rates as a share of all renewals due, which include completed renewals and pending cases.)

Of all people who were disenrolled, 69% were terminated for procedural reasons, as of August 1, 2024

Medicaid Renewals

  • Of the people whose coverage has been renewed as of September 12, 2024, 61% were renewed on an ex parte basis while 39% were renewed through a renewal form, though rates vary across states. Under federal rules, states are required to first try to complete administrative (or “ex parte”) renewals by verifying ongoing eligibility through available data sources, such as state wage databases, before sending a renewal form or requesting documentation from an enrollee. Ex parte renewal rates varied across states from 90% or more in Arizona, North Carolina, and Rhode Island to less than 20% in Pennsylvania and Texas. 

Overall, 61% of people who retained Medicaid coverage were renewed through ex parte processes, as of August 1, 2024

Federal Data on Renewal Outcomes

The data presented here are cumulative unwinding metrics published by CMS. These counts and percentages may differ from the above data, which present renewal metrics reported on state websites when state-reported data are more complete.  

Figure 1 below shows cumulative renewal data reported by CMS during states’ unwinding periods. Renewal data for the months after the end of states’ unwinding period are excluded. The data reflect updated unwinding data reported by states three months after the original monthly reports as they become available.   

Cumulative Medicaid Renewal Outcomes for Reporting States through April 2023

For questions about this tracker, please contact KFFTracker@kff.org

State Data – Archived

Note: The state data presented below were last updated on September 12, 2024; since most states have now completed the Medicaid unwinding, the information will not be updated again. 

The data presented here provide state-level data on enrollment trends and renewal outcomes during the unwinding period. Figure 1 shows total Medicaid enrollment by month starting in January 2023 and, once disenrollments resumed in a state, the cumulative percent change in Medicaid enrollment relative to the month before Medicaid disenrollments started (this baseline month will differ across states). Figure 2 shows renewal metrics for each month of a state’s unwinding period (or cumulative data for the unwinding period for some states). 

For total national Medicaid enrollment, click on the Enrollment Data tab.

Related Resources

Resources on unwinding data

Resources on state policies and preparations for the unwinding

Resources on pre-pandemic enrollment patterns and coverage transitions

KFF’s unwinding explainer

5 Key Facts about Medicaid’s Share of National Health Spending

Published: Sep 3, 2025

Medicaid, as the primary program providing comprehensive coverage of health care and long-term services and supports to about 80 million low-income people in the United States, accounts for one-fifth of all personal health care spending in the United States and a large share of state budgets. During its 60 years since enactment, Medicaid’s share of health insurance coverage and health care spending have incrementally increased; the program has evolved over time through a series of legislative and judicial actions, within the context of broader changes in the health care landscape. Now, landmark changes to Medicaid coverage and enrollment policies are set to roll out over the next several years.

According to the Congressional Budget Office (CBO), the recently enacted reconciliation package is estimated to reduce federal Medicaid spending by $911 billion over 10 years (after accounting for interactions that produce overlapping reductions across different provisions of the law), and to increase the number of uninsured people by 10 million in 2034. The most recent projections for national health spending do not account for the changes in the law; but changes in the law are expected to have big implications for Medicaid coverage and spending that could reverse longstanding incremental trends. Policy changes in the reconciliation package that lead to more uninsured people are likely to increase out-of-pocket spending as a share of national health care spending. Shifts in spending patterns are likely to be more profound over time and beyond the ten year projection period if there are no other changes in federal laws that affect health spending.

To provide historical context for how changes to Medicaid spending may impact national health spending trends, this brief explores how Medicaid spending contributes to national health spending and how different service areas contribute to Medicaid costs. This brief uses National Health Expenditures (NHE) historic data, published annually by the Centers for Medicare & Medicaid Services, which provide estimates of national spending on health care, by payer and by type of service. The analyses in this brief focus on spending for personal health care, which excludes the costs of public health programs and payers’ administrative spending (see Methods).

1. Over time, Medicaid has covered an increasing share of the population and health care costs.

Over the past two decades, the percent of the population enrolled in Medicaid increased by more than 10 percentage points (from 12% in 2000 to 25% in 2023 by NHE enrollment estimates). During the same time, Medicaid’s share of national health spending increased by only 3 percentage points, from 16% to 19%. Medicaid spending is driven by multiple factors, including the number and mix of enrollees, their use of health care and long-term services and supports, the prices of Medicaid services, and state policy choices about benefits, provider payment rates, and other program factors. Some of the faster growth in Medicaid enrollment relative to spending is that enrollment growth over the past two decades was driven by increased enrollment stemming from the Great Recession, implementation of the Affordable Care Act (ACA) Medicaid expansion, and the COVID-19 pandemic. Each of those events spurred increased enrollment of working-age adults and their families, groups that tend to have lower per-enrollee Medicaid costs than older adults and people who come into Medicaid because they need long-term care. CBO estimates predict that changes to Medicaid enacted in the reconciliation package will result in a downward shift in future Medicaid spending and in enrollment. KFF analyses show how these recent changes to Medicaid policy are likely to reduce federal Medicaid spending and impact enrollment, with varying impacts to different states or areas.

Medicaid Covers an Increasing Share of the Population and Health Care Costs

2. Medicaid’s share of spending has grown, but remains lower than that of private insurance and Medicare.

Prior KFF analysis has shown that generally, third party payers cover a greater share of total health spending than in previous decades because more people have gained coverage, especially public coverage, and payers’ spending per enrollee has grown. The total share of national health care spending for each type of coverage reflects the number and mix of enrollees and the amount spent to cover each enrollee. The share of health care spending paid by people out-of-pocket decreased as more spending was paid by Medicare and private insurance. Out-of-pocket spending includes payments for care from people who are not insured, and payments for care from people with health coverage when coverage requires enrollees to pay some of the costs. Common types of out-of-pocket spending among people with coverage include copayments (a flat fee per service), coinsurance (a percentage of the total costs), and deductibles (an amount enrollees pay before coverage kicks in). Historically, one difference between Medicaid and other forms of health coverage was the low out-of-pocket spending. Estimated increases in the uninsured following implementation of the reconciliation package could reverse incremental declines in out-of-pocket costs.

Medicaid’s Share of Spending Has Grown, but Remains Lower Than That of Other Insurance Types

3. Over 70% of Medicaid spending pays for hospital services and long-term care.

Over the last 23 years, the largest share of Medicaid spending paid for hospital services, which accounted for 38% of Medicaid spending on average and varied between 37% and 39% during the 2000-2023 period. Medicaid mirrors broader hospital spending trends; spending on hospitals makes up the largest share of all health care spending, a trend projected to continue. During the same period, long-term care accounted for an average of 37% of Medicaid spending, although it declined somewhat from a high of 40% in 2007 to 36% in 2023. The category with the greatest relative growth was payments to providers such as physicians which increased from 11% in 2000 to 17% in 2023. The percent of spending that paid for prescription drugs decreased from 11% in 2000 to 7% in 2023, which is likely attributable to the enactment of the Medicare prescription drug benefit which took effect in 2006. Prior to that point, Medicaid paid for prescription drugs for low-income Medicare beneficiaries who were also enrolled in Medicaid (e.g., dual-eligible individuals).

Over 70% of Medicaid Spending Pays for Hospital Services and Long-Term Care

4. Medicaid pays for nearly 20% of hospital spending.

Medicaid pays for nearly 20% of hospital spending, a share that has changed little since 2000. During that time, Medicare’s share of national spending on hospital care decreased from 30% in 2000 to 25% in 2023, while the share paid by private health insurance rose from 33% to 37%. People pay a much smaller percentage of hospital spending out-of-pocket compared with other types of health care. Medicaid covered 41% of all U.S. births in 2023; births are the most common reason for a hospital inpatient stay. Medicaid financing for hospitals is complex, but Medicaid studies have shown that Affordable Care Act (ACA) Medicaid expansion is associated with improved hospital financial performance and lower likelihood of hospital closure, particularly in rural areas.

Medicaid Pays for Nearly 20% of Hospital Spending

5. Medicaid is the primary payer of long-term care, most of which is now provided in people’s homes and the community.

Medicaid continues to be the primary payer of long-term care, comprising an increasing share of all spending on long-term care. Medicaid’s share of spending on long-term care rose nine percentage points, growing from 52% in 2000 to 61% in 2023. Since 2000, Medicaid’s spending on care delivered in people’s homes and communities (e.g., home care) increased faster than spending on institutional care such as nursing facilities. In 2000, the 52% of long-care spending that was paid by Medicaid included 29% on home care and 23% on institutional care. In 2023, the 61% that was paid by Medicaid included 47% on home care and only 14% on institutional care. Prior KFF analysis has shown that in 2021, three-quarters of the 5.7 million people who used Medicaid long-term care were receiving home care, although that percentage varies across the states widely. The larger share of people receiving care in the community as opposed to in an institution reflects initiatives to make home care more widely available in recent years and to remove what has been referred to as the “institutional bias” in Medicaid. 

Medicaid is the Primary Payer of Long-Term Care, Most of Which Is Now Provided in People’s Homes and the Community

Methods

This analysis uses National Health Expenditures (NHE) historic data. Unlike other sources of information on health care spending, the NHE data use an accounting structure that captures all expenditures of health care goods and services and investment in the health care sector. Expenditures are classified into high-level service categories and by source of payment. Data sources include federal administrative data, household and individual surveys, surveys of businesses, and economic data from the Bureau of Labor Statistics and the Bureau of Economic Analysis. Medicaid spending estimates are derived from financial reporting through Form CMS-64, except for durable medical equipment estimates which are developed from person-level payment data.

See the NHE Accounts Methodology documentation for additional information including definitions, sources, and methods; CMS publishes both complete documentation and short definitions.

Enrollment: This KFF analysis uses NHE Accounts enrollment data to estimate Medicaid’s share of total health insurance enrollment (Figure 1). NHE Accounts data estimates Medicaid enrollment using the Medicaid Statistical Information System (MSIS) for years 2000-2013, and enrollment projections reported on form CMS-64 for years 2014-2023. The NHE Accounts data estimates for total health insurance enrollment include private health insurance, Medicare, Medicaid, CHIP, and the Departments of Defense and of Veterans Affairs.

Personal Health Care Services: Personal health care services in the NHE Accounts data represent aggregate revenue received by health care providers and retail providers of medical goods and services. Cost estimates for personal health care services expenditures exclude administrative costs, government public health activities, or investments in structures or equipment. KFF analyzes NHE personal health care data using service categories (i.e. “hospitals,” “providers,” or “prescription drugs,”) that align with the classification system used within the NHEA, except for Long-Term Care.

Spending by Payer: NHE Accounts data estimate spending attributable to certain payer categories. KFF uses the payer categories defined by the NHE Accounts data and defines “Other” spending (Figures 2 and 4) below. For personal health care services:

  • Medicaid spending estimates include both state and federal spending on both fee-for-service and managed care enrollees but exclude Children’s Health Insurance Programs (CHIP) spending.
  • Medicare spending estimates include Medicare Parts A, B, and C (Medicare Advantage). Medicare spending estimates also include Medicare Part D and Medicare Advantage Part D. Private supplemental Medicare insurance, i.e. Medigap and employer-sponsored Medicare Part D, is excluded from Medicare spending and included in private insurance spending.
  • Private Insurance spending estimates include premiums and benefits from traditional fully-insured health coverage whether purchased individually or through an employer, self-insured employer health benefit plans, plans purchased through the Affordable Care Act marketplaces, and indemnity plans such as those covering hospital care or long-term care. Private insurance spending estimates also include supplemental Medicare plans (e.g., Medigap).
  • Out-of-Pocket spending estimates include direct consumer spending including coinsurance, deductibles, and any other amounts not covered by insurance. Premiums are included in private insurance spending and excluded from out-of-pocket spending.
  • Other: KFF defines “Other” spendingas personal health care expenditures by the Children’s Health Insurance Program, the Indian Health Services, the Substance Abuse and Mental Health Services Administration, the Veterans Health Administration, federal spending through the Pre-Existing Conditions Insurance Plans (PCIP) or COVID-19 relief funds (e.g. the Provider Relief Fund), direct payments to the needy through general assistance programs (e.g. the State Pharmaceutical Assistance Program), certain state and local programs (e.g. temporary disability insurance or provider subsidies), and property or casualty insurance.

Spending by Service Category: KFF uses the NHE Accounts definitions for hospital care, provider care, and prescription drugs. KFF definitions for “Other” services (Figure 3) and “Long-Term Care (Figures 3 and 5) are included below.

  • Hospital Care includes all services provided by hospitals to patients. These expenditures include the services of resident physicians, inpatient pharmacy, room and board and ancillary costs, hospital-based home health care, and other services billed by hospitals. Services rendered in a hospital by a physician who bills independently are considered Provider Care.
  • Provider Care includes services provided in non-hospital clinics and practices. These include physician-operated practices, outpatient care centers, and certain federally operated clinics and clinics operated by non-physician clinicians (such as private-duty nurses, podiatrists, optometrists, chiropractors, or occupational therapists). Provider care also includes certain medical laboratory services.
  • Prescription Drugs covers “retail” sales of products available only by a prescription, such as drugs, biologics, and diagnostic products.
  • Other: KFF defines Medicaid spending on “Other” services to include personal health care expenditures not attributable to hospitals, providers, prescription drugs, nor long-term care. This includes dental care, durable medical equipment, non-durable medical equipment (e.g. diagnostic tools or wound dressings and other medical supplies), and non-prescription drugs.
  • Long-Term Care: KFF defines long-term care to include spending for nursing care and continuing care retirement communities; home health; and other health, residential, and personal care if it is paid for by Medicaid, individuals who are paying out-of-pocket, the Children’s Health Insurance Program, the Indian Health Services, the Substance Abuse and Mental Health Services Administration, the Veterans Health Administration, general assistance, other federal programs, other state and local programs, school health. See 10 Things about LTSS for more information.

VOLUME 29

Better Prompting May Help Reduce AI Hallucinations, False Vaccine Claims Spread, and Industrial Solvent Promoted as Hidden Cancer Cure


Summary

This volume highlights new research showing that certain prompting techniques can help reduce the risk of AI chatbots amplifying false medical information when users include fabricated terms in their queries. It also examines false claims linking vaccines to sudden infant death syndrome and the promotion of the industrial solvent dimethyl sulfoxide as an allegedly suppressed cancer cure despite a lack of clinical evidence. Lastly, it explores misunderstandings among first responders about the risks of overdose from fentanyl exposure and ongoing myths about sunscreen safety.


The Problem Isn’t Trust in Vaccines, It’s That People Don’t Know Who to Trust

In last week’s “Beyond the Data” column, KFF’s CEO, Dr. Drew Altman, draws on years of KFF polling about vaccines and writes that uncertainty about vaccines stems not from lack of confidence in their safety but from eroding trust in sources of health information, leaving many unsure where to turn for reliable guidance. As Altman notes in the column, just 16% of the public believe mRNA vaccines are unsafe, but only 14% of the public say they have a lot of trust in federal health agencies like the Centers for Disease Control and Prevention (CDC) and Food and Drug Administration (FDA) to ensure the safety and effectiveness of vaccines. The result: the majority of Americans are uncertain and unsure what to believe.


AI & Emerging Technology

Chatbots Prone to Hallucinations When Prompted with Fabricated Terms

KFF / Getty Images

Artificial intelligence (AI) chatbots can convincingly amplify false claims when false medical information is embedded within user questions, according to new research published in Communications Medicine. The study, conducted by researchers at Mount Sinai Medical Center, tested how leading large language models (LLMs) respond when fictional medical terms are included in patient scenarios. The researchers created 300 hypothetical cases containing fabricated diseases, symptoms, or medical tests to evaluate whether chatbots would identify and reject the false information. 

Without additional safety prompts, the AI responses regularly contained hallucinations, a phenomenon that occurs when an LLM fabricates false information instead of relying on evidence. In 65.9% of responses, the chatbot expanded on the fictional medical details, generating confident explanations about treatments for non-existent conditions like “Faulkenstein Syndrome.” Researchers suggested that users may unknowingly include false information in their health queries to AI chatbots, from unreliable sources or misremembered medical terms, and may receive responses that not only repeat but expand upon those inaccuracies.

The study also demonstrated, however, that including additional safeguards in the prompt could reduce this risk. By asking the AI models to use only clinically validated information and acknowledge uncertainty, the rate of hallucinations dropped to 43.1%. The researchers suggested that additional safeguards built into AI models, or included in user prompts, help mitigate the risks of chatbots amplifying false health information. The findings demonstrate the need for further education about AI, including around proper prompting techniques. An August 2024 KFF poll found that most adults (56%) were not confident in their ability to tell the difference between what is true and false when it comes to information from AI chatbots. Users may benefit by developing new prompting strategies, particularly when using AI for health information where falsehoods can cause particular harm. Framing prompts with accurate, verified details, or including safeguards in the prompts, could help prevent chatbots from repeating false information, enhancing their reliability and usefulness.


Recent Developments

False Claims Link Vaccines to Sudden Infant Death Syndrome

Karl Tapales / Getty Images

Widely-shared social media posts in early August repeated unsubstantiated claims that routine childhood immunizations are associated with sudden infant death syndrome (SIDS). The claims circulated online this summer following the publication of a review, authored by a researcher who has previously published studies claiming that vaccines are unsafe, that alleged without evidence that vaccines cause SIDS. One X user, with nearly 800,000 followers, falsely alleged that children who died of SIDS were killed by the vaccine industry. Another account, with more than 100,000 followers, called for a ten-year ban on all vaccines, implying that such a ban would lead to “a society without SIDS” and other illnesses like cancer and autism.

Claims of an association between childhood vaccinations and SIDS have persisted for decades, largely citing the temporal association between the recommended vaccination schedule and incidence of SIDS. Cases of SIDS peak around 2-4 months of age, coinciding with the timing of recommended early childhood vaccines. Multiple large-scale studies, though, have found no causal link between childhood vaccines and SIDS. Research has identified multiple factors that may contribute to SIDS risk, including sleep position, maternal smoking, and premature birth, but vaccination is not among them. Rates of SIDS sharply declined after the American Academy of Pediatrics (AAP) introduced safe sleep guidelines for infants, then stabilized in the 2000s during the same timeframe that the childhood vaccination schedule expanded. Some studies have found that vaccinations may in fact be associated with a lower risk of SIDS. 

Polling Insights: Recent KFF polling has found that notable shares of parents are coming across vaccine-related content on social media. In KFF’s July Tracking Poll on Health Information and Trust four in ten parents (41%) reported seeing information about vaccines on social media in the past 30 days. Previous KFF polling has found that about a third of parents (36%) say they are “not too confident” or “not at all confident” that they can tell what is true versus what is false when it comes to information on social media.

As parents are exposed to vaccine information online, the July Tracking Poll also found that many express distrust in the CDC and in local public health departments when it comes to providing reliable information about vaccines. Overall, four in ten parents say they trust the CDC “not much” or “not at all” to provide reliable information about vaccines and a similar share (38%) express distrust towards their local public health department.

About Four in Ten Parents Report Little to No Trust in the CDC and Their Local Public Health Department To Provide Reliable Vaccine Information

Dimethyl Sulfoxide, An Industrial Solvent, Promoted as Suppressed Cancer Cure

Irfan Setiawan / Getty Images

Some influential X accounts with large followings have promoted dimethyl sulfoxide (DMSO), an industrial solvent and byproduct of paper production, as a “miracle cure” for a range of health issues, often portraying past regulatory restrictions on the substance as evidence of a deliberate suppression of an effective treatment. Social media posts in late July and early August have claimed that the solvent can treat conditions including blindness, tinnitus, skin issues, and cancer. One X account, with more than 1.8 million followers, described a combination of DMSO and the histologic stain hematoxylin as “The 50-Year-Old Cancer Miracle Hiding in Plain Sight” and said that the FDA “chose to bury the evidence.” Other accounts also presented unsupported claims about DMSO as a cancer cure as evidence of a conspiracy, including one with more than 300,000 followers who claimed that the FDA banned the substance despite knowing that it “cured cancer” and “made chemo more likely to work.”

DMSO, which penetrates the skin quickly and is used in some transdermal drug delivery systems, was tested in the 1960s as a pain reliever, but trials were halted after research showed abnormal changes in the eyes of laboratory animals. Research has since continued into DMSO’s potential for pain relief, including for pain from osteoarthritis, but has yielded inconsistent results. The FDA has approved DMSO for treating symptoms of interstitial cystitis, which remains the only FDA-approved use of the substance. While its effectiveness in treating cancer has also been investigated, claims that it is a “cure” overstate the available evidence. Some research in laboratory settings has shown that it may have anti-cancer properties, but other studies have found it may increase the rate at which cancer cells multiply, and no large-scale clinical trials have demonstrated its effectiveness for treating cancer in humans. DMSO has also been shown in some research to adversely interact with effective cancer treatments, including platinum-based chemotherapy drugs like cisplatin, carboplatin, and oxaliplatin.

It has nonetheless been promoted by alternative medicine advocates for decades, and some science communicators have theorized that the FDA’s effective ban of the chemical may have contributed to advocacy for its use. The substance’s promotion as a cancer cure, despite a lack of evidence, may cause patients to delay treatment, which is known to lead to higher mortality rates. Its popularity mirrors unsupported claims about ivermectin’s ability to treat cancer and demonstrates the potential for unproven and potentially harmful treatments to gain attention as trust in health authorities declines.

Police Video Highlights Ongoing Misunderstandings About Fentanyl Overdose Risk

Adobe Stock

Widespread misunderstandings about brief skin contact with fentanyl continue to shape public perception and emergency protocols, despite a decline in opioid deaths that began in mid-2023 and continued through 2024.In early August, a video of a police officer collapsing after handling evidence during an arrest was widely shared on social media, contributing to discussion about the risk of fentanyl exposure among law enforcement. According to the local sheriff’s office, the officer handled a dollar bill containing fentanyl without gloves and was administered naloxone and transported to a local hospital. The sheriff’s description has not confirmed fentanyl as the cause, but it has reinforced misconceptions about the drug, with some commenters falsely suggesting that brief skin contact with fentanyl could cause an overdose. A former narcotics officer repeated the claim on cable television, saying that merely touching fentanyl “could kill someone.” Others, though, including an X user who identified himself as a medical doctor, stated that physical absorption of fentanyl through such brief contact was not possible.

The misunderstanding that brief skin contact with fentanyl can cause an overdose can be traced to a 2016 Drug Enforcement Administration (DEA) video that warned law enforcement officers that such contact could be fatal. Fentanyl can be absorbed through the skin, but a joint statement from the American College of Medical Toxicology (ACMT) and the American Academy of Clinical Toxicology (AACT) clarified that small, unintentional skin exposures to fentanyl powder or tablets are “very unlikely” to cause significant opioid toxicity. According to the statement, it would take 14 minutes with both hands covered with high-absorption fentanyl patches to receive a therapeutic dose. Risk may be greater for workers in certain high-risk situations, like a drug storage or distribution facility where the potential for longer duration exposure is higher, and the National Institute for Occupational Safety and Health (NIOSH) has recommended higher levels of personal protective equipment (PPE) when larger quantities may be encountered. 

Studies have found no confirmed cases of fatal overdoses among first responders due to contact with the drug, and there have been no documented cases of harm from incidental exposure. Still, the perceived risk among first responders remains high, and misunderstandings about the potential of accidental overdose have shaped emergency response protocols. These concerns may result from conflating brief contact during routine emergency calls with prolonged, high-concentration exposures which carry higher risks. A KFF Health News article reported that these misconceptions may have contributed to some first responders being warned to proceed with caution when responding to an overdose. This has prompted concern among some public health advocates that misunderstandings of risk could lead to delayed interventions such as CPR and rescue breaths, techniques that are recommended by the Substance Abuse and Mental Health Services Administration (SAMHSA) as essential steps in responding to opioid overdoses. 

Sunscreen Claims Persist as UV Protection Remains Important

ljubaphoto / Getty Images

Misleading claims that sunscreen is ineffective or toxic have circulated on social media through late summer, with some influencers promoting the idea that chemical sunscreens cause more harm than sun exposure itself. One post, from an account with more than 1.1 million followers, claimed that an increase in melanoma rates was associated with the “toxic chemicals” found in commercially available sunscreen. Others promoted alternative sun protectants, like beef tallow, and claimed that vitamin D from sun exposure helped reduce cancer risk. 

Claims that ingredients in sunscreen are toxic may have originated with a 2020 study that found some chemical ingredients, including oxybenzone and avobenzone, were systematically absorbed by the body in concentrations that exceed the FDA’s limit for requiring further safety testing. The FDA has requested further safety data on these chemicals, but the agency has noted that the fact that an ingredient is absorbed does not mean it is unsafe, nor does its request for additional information. There has been no conclusive evidence that these chemicals are harmful, and dermatologists have noted that they have been safely used for decades. The FDA continues to recommend the use of sunscreen, along with other sun-protective measures, like limiting time in the sun. Some posts also referenced recent recalls of sunscreen products containing benzene, a known carcinogen, incorrectly associating it with oxybenzone and avobenzone. Melanoma diagnosis rates have risen alongside increased use of sunscreen, but this trend also coincided with greater public awareness and improved diagnostic tools. Exposure to ultraviolet (UV) radiation from sun exposure is a known risk factor for developing skin cancers like melanoma, and trials have shown that regular use of sunscreen significantly reduces this risk. There is no evidence that vitamin D reduces the risk of skin cancers, and evidence shows that regular sunscreen use does not cause vitamin D deficiency.

Ongoing myths about supposed harms of sunscreen may lead some to discontinue use, resulting in higher rates of the most common and preventable forms of cancer. UV rays remain a year-round concern, as fall and winter months can still present significant UV exposure. Damage from sun exposure is cumulative, and contributes both to skin cancer risk and accelerated aging. The FDA recommends use of broad spectrum sunscreens with SPF values of 15 or higher. 

About The Health Information and Trust Initiative: the Health Information and Trust Initiative is a KFF program aimed at tracking health misinformation in the U.S., analyzing its impact on the American people, and mobilizing media to address the problem. Our goal is to be of service to everyone working on health misinformation, strengthen efforts to counter misinformation, and build trust. 


View all KFF Monitors

The Monitor is a report from KFF’s Health Information and Trust initiative that focuses on recent developments in health information. It’s free and published twice a month.

Sign up to receive KFF Monitor
email updates


Support for the Health Information and Trust initiative is provided by the Robert Wood Johnson Foundation (RWJF). The views expressed do not necessarily reflect the views of RWJF and KFF maintains full editorial control over all of its policy analysis, polling, and journalism activities. The data shared in the Monitor is sourced through media monitoring research conducted by KFF.

Policy Tracker: Exceptions to State Abortion Bans and Early Gestational Limits 

Last updated on August 26, 2025

states have abortion bans or early gestational limits in effect

states have no health exception

states have no rape or incest exception 

states have no fatal fetal anomaly exception 

Abortion is currently banned in 12 states and 6 states have early gestational limits between 6 weeks and 12 weeks in effect. Nearly all of these bans include exceptions, which generally fall into four categories: to prevent the death of the pregnant person, when there is risk to the health of the pregnant person, when the pregnancy is the result of rape or incest, and when there is a lethal fetal anomaly. Some states have more than one abortion ban or restriction in place. The maps below illustrate the exceptions in each state’s most restrictive gestational limit or total ban. For details hover over each state to read the rollover.  

For more information on the status of state abortion bans, please visit our Abortion in the United States Dashboard

Exceptions to State Abortion Bans and Early Gestational Limits in Effect, as of April 1, 2024
Exceptions in Abortion Bans and Gestational Limits, as of April 15, 2024

Status of State Medicaid Expansion Decisions

Published: Aug 26, 2025

The Affordable Care Act’s (ACA) Medicaid expansion expanded Medicaid coverage to nearly all adults with incomes up to 138% of the Federal Poverty Level ($21,597 for an individual in 2025) and provided states with an enhanced federal matching rate (FMAP) for their expansion populations.

To date, 41 states (including DC)   have adopted   the Medicaid expansion and 10 states   have not adopted   the expansion. Current status for each state is based on KFF tracking and analysis of state expansion activity.

These data are also available in a table format. The map may be downloaded as a Powerpoint.

Status of State Action on the Medicaid Expansion Decision
Key States with Expansion Activity

Medicaid Expansion Resources