Medicare COVID-19 Data Release Blog

By Seema Verma, Administrator, Centers for Medicare & Medicaid Services
Twitter: @CMSgov

The Centers for Medicare & Medicaid Services (CMS) released preliminary data on COVID-19 derived from Medicare claims. The data provides a highly instructive picture of the impact of COVID-19 on the Medicare population, further confirming a number of long understood patterns in the disease such as the elevated risk for seniors with underlying health conditions. Additionally, in presenting stark socio-economic and racial/ethnic disparities, the data makes the transition from a fee-for-service system to a value-based one that fosters accountability for outcomes more urgent than ever.

Between January 1 and May 16, 2020, over 325,000 Medicare beneficiaries were diagnosed with COVID-19, and nearly 110,000 of those were hospitalized. These new findings come from the snapshot of COVID-19 data that CMS released today.

The snapshot presents preliminary information gathered from Medicare data, and is part of the Trump Administration’s consistent commitment to data transparency throughout the coronavirus pandemic in an effort to inform public health efforts. It also builds on our efforts over the past three years to make data in our programs more available to researchers.

The snapshot breaks down COVID-19 cases and hospitalizations for Medicare beneficiaries by state; race/ethnicity; dual eligibility for Medicare and Medicaid; age; gender; and urban/rural areas. Findings from the snapshot reinforce previous findings by the Centers for Disease Control and Prevention (CDC) that older Americans and those with chronic health conditions are at the highest risk for COVID-19 and should continue to take extra precautions to protect their health and safety. The data also show that COVID-19 has disproportionately impacted racial and ethnic minority groups and lower income adults, further confirming longstanding healthcare disparities in these populations.

CMS usually releases this type of claims and encounter information on an annual basis when the data are more complete. However, as part of our efforts to ensure the public has this vital information as soon as it’s available during this public health emergency, we’re releasing preliminary COVID-19 case and hospitalization data today. The data will be updated on a monthly basis as more claims and encounter records are received. We also anticipate releasing similar information on Medicaid beneficiaries, once our Transformed Medicaid Statistical Information System data is complete enough for public reporting.

The Medicare claims data on which this snapshot is based is different from the separate Centers for Disease Control and Prevention (CDC) COVID-19 case reporting system in which data are voluntarily reported to CDC by state and local health departments.

Monitoring for COVID-19 a
Work on the snapshot began earlier this year when we started checking for COVID-19 by analyzing the claims and encounter data for International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) diagnosis codes that could indicate the coronavirus. Once an official ICD-10-CM diagnosis code for COVID-19, (U07.1), was created on April 1, we began monitoring the use of that specific code. Prior to that, in addition to looking for the ICD-10-CM diagnosis code for “other coronavirus as the cause of diseases classified elsewhere,” we monitored the use of ICD-10-CM diagnosis codes such as “acute respiratory stress syndrome,” and “other viral pneumonia.” It is important to note that weekly counts for beneficiaries with these diagnosis codes in early 2020 remained steady, and tracked trends from 2019 data through mid-March. Once an official diagnosis code for COVID-19, (U07.1), was created on April 1, we began monitoring the use of that specific code.

Key Findings on Coronavirus among Medicare Beneficiaries
Over 325,000 Medicare beneficiaries have been diagnosed with COVID-19 this year through May 16, translating to 518 cases per 100,000 beneficiaries.

The 110,000 Medicare beneficiaries hospitalized with a COVID-19 diagnosis this year through May 16 translates to 175 hospitalizations per 100,000 beneficiaries. Among those hospitalized with COVID-19, the five most prevalent chronic conditions for Medicare fee-for-service beneficiaries were: hypertension (79%); hyperlipidemia (60%); chronic kidney disease (50%); anemia (50%); and diabetes (50%).

The snapshot data also show that 28% of hospitalized beneficiaries died in the hospital, and 27% were discharged to their homes. The remainder were discharged to skilled nursing facilities (21%) or other healthcare settings. Half of hospitalizations (50%) were less than 8 days, while 9% were 21 days or longer. Medicare payments for fee-for-service hospitalizations totaled $1.9 billion, with an average of $23,094 per hospitalization.

Not surprisingly, the COVID-19 hospitalization rate is highest in New York and New Jersey. Additionally, the rates are also higher for Black beneficiaries (465 hospitalizations per 100,000 beneficiaries), Hispanic beneficiaries (258/100K), and among beneficiaries who are age 85 or older (379/100k). But beneficiaries eligible for Medicare because they have end stage renal disease (ESRD) have the highest COVID-19 hospitalization rate, with 1,341 cases per 100,000 beneficiaries. This finding for beneficiaries with ESRD confirms previous CDC findings that “people with chronic kidney disease undergoing dialysis” (i.e., people with ESRD) might be at higher risk for severe illness from COVID-19. Patients with ESRD are also more likely to have chronic comorbidities associated with increased of COVID-19 complications and hospitalization, such as diabetes and heart failure.

Beneficiaries enrolled in both Medicare and Medicaid (dually eligible individuals or duals) also have a higher rate of COVID-19 hospitalizations, with 473 hospitalizations per 100,000 beneficiaries. For comparison, the rate for beneficiaries enrolled only in Medicare is 112 hospitalizations per 100,000. The rate of COVID-19 hospitalizations for dually eligible individuals is higher across all age, sex, and race/ethnicity groups. This is not unexpected as previous research and analysis have shown that dually eligible individuals experience high rates of chronic illness, with many having long-term care needs and social risk factors which can lead to poor health outcomes. In particular, duals also experience higher rates of poverty.

Data Limitations
The data source for the snapshot is Original Medicare fee-for-service claims, Medicare Advantage encounter data, and Medicare enrollment information. We collect this data to pay providers for services delivered to beneficiaries in Original Medicare and to support program activities, such as risk adjustment, in Medicare Advantage. It’s not collected for public health surveillance. As a result, caution must be used when interpreting the data, and findings may not match other publicly available data sources. These include CDC data collected for public health surveillance purposes, and which use different data sources, different time periods for reporting, and different methodologies for identifying a COVID-19 case/hospitalization.

There are also a number of other reasons for caution when using this data. For example:

  • Medicare claims and encounter data contain information on healthcare services provided to Medicare beneficiaries, and CMS can use ICD-10-CM diagnosis codes on these data to identify COVID-19. However, the data do not include other clinical information such as the results of a laboratory test for COVID-19.
  • The International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) diagnosis code for the “COVID-19,” (U07.1), became effective on April 1, 2020. Consequently, for the first three months of 2020, CMS could only identify potential COVID-19 cases and hospitalizations using the ICD-10-CM diagnosis code, “Other coronavirus as the cause of diseases classified elsewhere” (B97.29). CMS decided to include these potential COVID-19 cases and hospitalizations in our snapshot, but other researchers may choose to limit their analysis to cases and hospitalizations with the ICD-10-CM diagnosis code U07.1.
  • Diagnosis code accuracy depends on how clinicians’ document, e.g., omitting information or using synonyms or abbreviations to describe a patient’s condition, and on medical coder experience and training. As a result, it’s possible that diagnosis information from claims and encounters is less reliable than clinical information collected in other ways, such as chart reviews.
  • There will always be a delay or “claims lag” between when a service is provided and when the claim or encounter for that service appears in our database. Providers have up to a year to submit fee-for-service claims to CMS, and Medicare Advantage Organizations must collect encounters before submitting them, so across both programs the data can take a year or more to be complete. There may also be a longer lag in claims due to the pandemic, but we don’t yet know the impact. Due to claims lag, all data presented in our snapshot is preliminary and will continue to change as CMS processes additional claims and encounters for the reporting period.

A Call to Action
CMS is making our snapshot of COVID-19 data available because it’s essential to understanding how the pandemic is affecting Medicare beneficiaries and to respond accordingly. While we’ve long known that seniors are disproportionately at risk for COVID-19, this snapshot suggests the critical importance of social risk factors, particularly income as shown by the higher rate of COVID-19 cases and hospitalizations among dually eligible individuals.

But our current healthcare payment system needs to address the social determinants of health in order to address disparities and improve outcomes. We must change payment structures to create incentives for doctors to focus on the health of the whole person rather than simply the delivery of care. We have to hold providers accountable for the outcomes they achieve, and poor health outcomes for minorities or those of a lower socio-economic status is not acceptable. Until we move to a system that incentivizes value over volume and starts paying doctors for better health outcomes, we’ll never be able to adequately address the social determinants of health.

Since our current fee-for-service system still often rewards the volume of care provided, it should not come as a surprise that spending grows inexorably – with little benefit on health outcomes. But under a value-based system, providers are rewarded for actually keeping patients healthy and improving quality and outcomes. In addition, a value-based system allows information about outcomes to be transparent, so they can make healthcare choices on the basis of this information.

Government processes are slow-moving. The result is that many of the rules and requirements that CMS has promulgated over the years have become out-of-date, and they are standing in the way of innovation and progress. Designed for a fee-for-service system, they often incentivize the perpetuation of that system. We’re doing everything we can to change that.

CMS has been doubling down across-the-board to encourage the move to value-based care. But our efforts are not just about new value-based payment models; we’re adopting a broader strategy to enable this transformation. That’s why the CMS Innovation Center is testing new opportunities for providers to accept higher levels of risk, and also new financial arrangements that ease providers into value-based agreements. As part of this broader effort, we moved late last year to remove unnecessary government obstacles that impede improved care and lower costs for American patients by proposing the first ever major update to the regulations that interpret the Stark Law.

One area where we’re developing innovative payment models is for kidney disease patients, including those with End Stage Renal Disease. An Executive order from President Trump made way for these models, which add financial incentives for providers to manage care for Medicare beneficiaries in order to delay the onset of dialysis and to incentivize kidney transplants. Over the next few months, we’ll be finalizing rules on these models as well as rules that will update the Organ Procurement Organization Conditions for Coverage.

Ensuring better care for dually eligible individuals is another strategic initiative for CMS, and we’ve created opportunities for Medicaid directors to partner with us to better serve duals through approaches that work best for the unique needs of their states. We also encourage governors and Medicaid directors to especially focus on minimizing the risks of coronavirus transmission among dually eligible individuals. And we call on governors to partner with CMS on new approaches to improving outcomes for dually eligible individuals.

To meet the unprecedented challenge of the coronavirus, we need to be equipped with insights and information. The insights we’ve gained from our snapshot demonstrate that the transition to a value-based system—one designed to address the social determinants of health—has never been more urgent.

This article was originally published on The CMS Blog and is republished here with permission.