Science Brief: Indicators for Monitoring COVID-19 Community Levels and Making Public Health Recommendations

The number of new hospital admissions of patients with COVID-19 per 100,000 population reflects the amount of severe COVID-19 disease within the community. Percent of staffed inpatient beds occupied by patients with COVID-19 is an indicator of local healthcare system usage and remaining capacity. Regardless of the reason for inpatient care, the level of usage of clinical care resources to manage patients with COVID-19 reflects impact on the community and signals when urgent implementation of layered prevention strategies might be necessary to prevent overloading the healthcare system. These indicators are proxies for underlying COVID-19 morbidity and severity of COVID-19 cases, and for the ability of the local healthcare system to support additional people requiring hospital care, including those with COVID-19. In addition, new COVID-19 cases per 100,000 population in the past 7 days serves as a leading indicator - most importantly early in a surge - of anticipated healthcare strain.

The indicators combine to result in three COVID-19 Community Levels: low, medium, and high. The COVID-19 Community Level is determined by the higher of the new COVID-19 hospital admissions and percent inpatient beds occupied by patients with COVID-19 indicators and is adjusted upwards one level if the new COVID-19 cases indicator is ≥ 200 cases per 100,000 population in the last 7 days.

A rapid rise in new cases may forecast increases in new hospitalizations or inpatient beds occupied by COVID-19 patients. Monitoring this leading indicator can allow communities to plan appropriately. When COVID-19 Community Levels are low, it does not indicate that virus is not circulating, or that individuals do not need to take any preventive measures to protect themselves, especially if they are at high-risk for serious disease. As communities see declines in case rates and the burden on the healthcare system eases, this can signal to individuals and communities when to discontinue the use of some layered prevention strategies. Community members can likewise consider these factors in making decisions about individual prevention behaviors based on their level of risk for severe disease or that of the members of their household or contacts.¹⁶

How the indicators of COVID-19 community levels were selected

To identify indicators, criteria were established that reflected priorities for indicator representation.

1. All indicators should have available county-level metrics from either data reported at the county level or allocated to county level from Health Service Areas (HSAs).
2. All indicators should have coverage from all counties in the United States (or with the possibility of allocation to all counties using HSAs).
3. Indicators should directly reflect the intended goals of minimizing medically significant disease or healthcare strain or represent a leading indicator of potential increases in severe disease or healthcare strain.
4. Indicators must represent data reported at least weekly or more often, with sufficient timeliness to permit assessment of COVID-19 Community Levels and use those data to inform decisions about recommended prevention measures in a timely manner.

Based on these criteria, a comprehensive review of historical data and available indicators, including reviews of existing data sources, data systems, and metrics, was conducted. This included an inventory of available indicators and data sources with frequent reporting and displays on CDC COVID Data Tracker. Candidate indicators were compiled that included all available data systems and sources. Historical data and thresholds used in the COVID-19 Community Profile Report and the State Profile Report were reviewed. These two reports provide daily metrics based on established thresholds reflecting trend data at the state and county level. Finally, a review of historical trends in increases and declines in cases, hospital metrics, and other data was conducted. Each candidate indicator on the comprehensive list of indicators was assessed against the pre-established criteria, and those that did not fully meet all criteria were eliminated. For example, death rates provide important information for monitoring the impact of COVID-19 at local, state, and national levels. Nonetheless, they are a lagging indicator and occur in very small numbers when reported frequently, particularly in sparsely populated areas. For this reason, death rates were eliminated from the list of potential indicators but retained as a potential outcome to assess the performance of selected community metrics.

Syndromic surveillance, based on the percent of emergency department visits due to COVID-19 (from the National Syndromic Surveillance Program) serves as an early warning system and holds promise in providing community metrics. However, this data source only reflects 71% coverage of emergency departments in the United States, and thus was not included. Counties and states with access to these data may consider these metrics as an additional, optional indicator when determining COVID-19 community level. When this metric is available, jurisdictions could use thresholds of <4.0%, 4.0-5.9%, and ≥6.0% of emergency department visits with diagnosed COVID-19 (7-day average) for communities with <200 new cases per 100,000 population in the past 7 days, and <4.0% and ≥4.0% for communities with ≥200 new cases per 100,000 population in the past 7 days. If including emergency department visits as an additional metric, communities would use the highest of the three indicators (percentinpatient beds occupied by patients with COVID-19, new COVID-19 hospital admissions, or percent emergency department visits due to COVID-19).

Wastewater surveillance complements traditional surveillance and enables health departments to intervene earlier to focused support communities experiencing increasing concentrations of SARS-CoV-2 in wastewater. Robust, sustainable implementation of wastewater surveillance requires public health capacity for wastewater testing, analysis, and interpretation. Wastewater surveillance is a valuable tool that health departments have used to allocate testing resources and forecast resource needs at the community level. Wastewater surveillance represents innovative data for local jurisdictions to use to inform and interpret COVID-19 Community Levels. Because wastewater surveillance does not provide national coverage, it was not considered a critical indicator for COVID-19 Community Levels.

Candidate indicators were further refined based on additional considerations. Percentof ICU beds occupied with COVID-19 patients was eliminated due to limitations of data from rural hospitals, the fact that it is a lagging indicator of severe disease, and the potential for bias due to small numbers. New COVID-19 hospital admissions with confirmed COVID-19 per 100 staffed beds did not provide added value as the concepts were already represented in other metrics (new COVID-19 hospital admissions per 100,000 population and percent of inpatient beds occupied by COVID-19 patients). Test positivity has limited current utility due to the widespread use of point-of-care and at-home tests, and thus was eliminated. Metrics that reflect percent change (e.g., in new hospital admissions or new cases) were eliminated due to challenges with interpretability of this metric for lay audiences.

Following this thorough review, new hospital admissions with confirmed COVID-19 per 100,000 population in the past 7 days and percent of staffed inpatient beds occupied by patients with confirmed COVID-19 were retained as "best candidate" indicators. New cases per 100,000 population in the past 7 days was also retained to assess performance as a leading indicator.

Data sources

The recommended COVID-19 Community Levels leverage the U.S. Department of Health and Human Services Unified Hospital Data Surveillance System (UHDSS), which monitors national and local trends in healthcare system stress, capacity, and community disease levels for approximately 6,000 hospitals in the United States. Data reported by hospitals to UHDSS represent aggregated counts, and include metrics capturing information specific to hospital capacity, occupancy, hospitalizations, and admissions. Most hospitals are required to report daily to UHDSS (with data backdated if reporting is delayed on weekends or holidays). For full guidance on hospital reporting and a list of data elements and definitions, including those used for the COVID-19 Community Level indicators, please visit: https://www.hhs.gov/sites/default/files/covid-19-faqs-hospitals-hospital-laboratory-acute-care-facility-data-reporting.pdf .

Hospital-based metric calculations for COVID-19 Community Level indicators

1. New admissions of patients with confirmed COVID-19 per 100,000 population (7-day total)
   • Numerator: the total number of patients admitted with laboratory-confirmed COVID-19 to an adult or pediatric inpatient bed each day during the previous 7 days for the specified geographic locality.
   • Denominator: total U.S. Census population for the specified geographic locality (based on 2019 Census population estimates).
   • Missing data: if there are no data reported for a locality for the given 7-day window, the indicator reported is "N/A."
2. Percent of staffed inpatient beds occupied by patients with confirmed COVID-19 (7-day average)
   • Numerator: the average number of adult and pediatric patients hospitalized with laboratory-confirmed COVID-19 each day during the previous 7 days for the specified geographic locality, calculated as the average of valid values within the 7-day period (e.g., if only three valid values, the average of those three is taken).
   • Denominator: the average number of staffed inpatient beds during the previous 7 days for the specified geographic locality, calculated as the average of valid values within the 7-day period (e.g., if only three valid values, the average of those three is taken). Per UHDSS reporting guidance, staffed inpatient beds in a facility are defined as those that are currently set up, staffed, and able to be used for a patient within the reporting period. This includes all overflow, observation, and active surge/expansion beds used for inpatients, ICU beds, surge/hallway/overflow beds that are open for use for a patient, regardless of whether they are occupied or available.
   • Missing data: if there are no data in the locality for the given 7-day window, the indicator reported is "N/A."

These metrics are calculated for all hospital subtypes reporting to UHDSS (including Veterans Administration, Defense Health Agency, and Indian Health Service hospitals), excluding psychiatric, rehabilitation, and religious non-medical hospitals.

The data include new COVID-19 hospital admissions and inpatient beds occupied by patients with confirmed COVID-19. Universal screening for SARS-CoV-2 infection continues to be common practice in clinical care, and incidental infections are common in many healthcare settings. The UHDSS data do not distinguish incidental infections from COVID-19-related hospitalizations. Confirmed SARS-CoV-2 infection could be a contributing factor to a health condition in varying ways that are not immediately clear. Further, infection can create burden on the healthcare system, regardless of whether it is the determining factor in hospitalization. All patients with SARS-CoV-2 infection in the hospital pose a risk to healthcare workers and other patients, even if the patient has mild illness or is asymptomatic and the infection is incidental. In addition, having SARS-CoV-2 infection requires isolation and other precautions that place added burden on the healthcare system. Furthermore, SARS-CoV-2 infection may complicate treatment and clinical course of other health conditions, further contributing to health system burden and severe illness.

The data on inpatient bed utilization include the total number of all staffed inpatient beds in the facility, that are currently set up, staffed and able to be used for a patient within the reporting period. This includes all overflow, observation, and active surge/expansion beds used for inpatients, as well as ICU beds. Because this metric is reported as a percent of staffed beds (including overflow), the denominator may vary over time based on staffing and how many overflow beds are set up. Inclusion of all surge/hallway/overflow beds that are open for use for a patient in the staffed inpatient beds in use by COVID-19 patients may result in an overestimate of remaining capacity.

Geographic unit of analysis for COVID-19 Community Level indicators

Analysis of these data at the local level is complicated by unequal distribution of hospitals within regions, as they are often clustered in large population centers and have service areas that overlap and extend across multiple communities. The unequal distribution of hospitals leads to a mismatch between places where people live and places where they receive care. This issue is particularly relevant in rural communities or those with relatively small populations, many of which have no hospitals or have hospitals with few inpatient beds. More complex hospital care (including critical care) may only be available at the very largest hospitals, often located in metropolitan areas with a very large catchment. Spatial aggregation of small geographic units (e.g., counties) to match the catchment area of local hospitals involving HSAs reflect hospital utilization patterns. HSAs reflect service areas of hospitals and helps reallocate patients to surrounding areas and align with county boundaries.^17-19

An HSA is defined by CDC's National Center for Health Statistics as a geographic area containing at least one county which is self-contained with respect to the population's provision of routine hospital care. Every county in the United States is assigned to an HSA, and each HSA must contain at least one hospital. While "county" has previously been used as a geographic area to display these metrics since it is typically the smallest geographic unit for which national data are available, not all counties in the United States have hospitals. As a result, county-level analyses of hospital data can result in inaccurate local population estimates. (/nchs/data/series/sr_02/sr02_112.pdf ). Use of HSAs in the calculation of COVID-19 Community Level indicators allows for more accurate characterization of the relationship between health care utilization and health status at the local level.

Data sources for case incidence rates

Total cases are based on aggregate counts of COVID-19 cases reported by state and territorial jurisdictions to the Centers for Disease Control and Prevention (CDC). In accordance with the CSTE definition of COVID-19 cases and deaths, counts for many jurisdictions include both confirmed and probable COVID-19 cases. For aggregate state-level data, CDC calculates the number of new cases each day either by using the information provided by states and territorial jurisdictions or by calculating the difference in cumulative counts reported by the state from the day before.

Public link

Please use the above public link if you want to share this noodl on another website.