Purpose and Methods

For surveillance of COVID-19 and its cause, SARS-CoV-2, in the United States, CDC is using multiple surveillance systems run in collaboration with state, local and territorial health departments, public health, commercial and clinical laboratories, vital statistics offices, health care providers, emergency departments and academic partners to monitor COVID-19 disease in the United States. COVIDView provides a weekly summary and interpretation of a variety of surveillance systems that will be used to track the progression and severity of COVID-19 disease throughout the course of the pandemic. The data summarized in COVIDView draws from a combination of existing influenza and viral respiratory disease surveillance systems, syndromic surveillance systems, and reporting of laboratory results. These systems, when evaluated together, create an ongoing picture of the spread of SARS-CoV-2 and its effects in the United States and provide data to inform the U.S. national public health response to COVID-19.  The data presented in COVIDView each week are preliminary and may change as more data are received.

The U.S. COVID-19 Surveillance goals are to:

• Monitor spread and intensity of COVID-19 disease in the United States
• Understand disease severity and the spectrum of illness
• Understand risk factors for severe disease and transmission
• Monitor for changes in the virus that causes COVID-19
• Estimate disease burden
• Produce data for forecasting COVID-19 spread and impact

Surveillance System Components

1. Virologic Surveillance

Public health laboratories, commercial laboratories and clinical laboratories located throughout all 50 states, Puerto Rico, and the District of Columbia report SARS-CoV-2 testing results to CDC. At this point in the outbreak, all laboratories are performing primary diagnostic functions; therefore, the percentage of specimens testing positive across laboratory types can be used to monitor trends in COVID-19 activity. As the outbreak progresses, it is possible that different types of laboratories will take on different roles and the data interpretation may need to be modified.

All laboratories report each week the total number of respiratory specimens tested for SARS-CoV-2 and the number positive; public health laboratories also report the age of the person tested, if available. The weekly percentage of specimens positive for SARS-CoV-2 is presented for all laboratory types and the weekly percentage of specimens positive by age group (0-4 years, 5-17 years, 18-49 years, 50-64 years, and ≥65 years) is reported for specimens tested at public health laboratories.  These data are presented at a national level and for each of the 10 HHS regionsexternal icon.

2. Outpatient and Emergency Department Illness Surveillance

Two syndromic surveillance systems are being used to monitor trends in outpatient and emergency department visits that may be related to COVID-19. Each system monitors a slightly different syndrome, and together these systems provide a more comprehensive picture of mild to moderate COVID-19 illness than either would individually. Both systems are currently being affected by recent changes in health care seeking behavior, including increasing use of telemedicine and recommendations to limit emergency department (ED) visits to severe illness, as well as increased social distancing. These changes affect the numbers of people and their reasons for seeking care in the outpatient and ED settings.

The U.S. Outpatient Influenza-like Illness Surveillance Network (ILINet) provides data on visits for influenza-like illness (ILI) (temperature of 100°F [37.8°C] or greater) and a cough and/or a sore throat without a known cause other than influenza) to approximately 2,600 primary care providers, emergency departments, and urgent care centers in all 50 states, Puerto Rico, the District of Columbia, and the U.S. Virgin Islands.  During the 2018-19 influenza season, approximately 60 million patient visits were captured in ILINet. Sites with electronic medical records use an equivalent definition as determined by public health authorities. Mild COVID-19 illness presents with symptoms similar to ILI, so ILINet is being used to track trends of mild COVID-19 illness and allows for comparison with prior influenza seasons.

Each week, health care providers around the country report to CDC the total number of patients seen for any reason and the number of those patients with influenza-like illness (ILI) by age group (0-4 years, 5-24 years, 25-49 years, 50-64 years, and ≥65 years). A subset of health care providers also report the number of patients seen for any reason by age group. For the 2019-2020 influenza season, these providers account for more than half the patient visits captured in ILINet.

The national overall percentage of patient visits to health care providers for ILI reported each week is calculated by combining state-specific data weighted by state population. This percentage is compared each week with the national baseline of 2.4% for the 2019-2020 influenza season. The baseline is developed by calculating the mean percentage of patient visits for ILI during non-influenza weeks for the previous three seasons and adding two standard deviations. A non-influenza week is defined as periods of two or more consecutive weeks in which each week accounted for less than 2% of the season’s total number of specimens that tested positive for influenza in public health laboratories. Due to wide variability in regional level data, it is not appropriate to apply the national baseline to regional data; therefore, region-specific baselines are calculated using the same methodology.

Regional baselines established for the 2019-2020 influenza season are:

Region 1 — 1.9% (Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island, and Vermont)

Region 2 — 3.2% (New Jersey, New York, Puerto Rico, and the U.S. Virgin Islands)

Region 3 — 1.9% (Delaware, District of Columbia, Maryland, Pennsylvania, Virginia, and West Virginia)

Region 4 — 2.4% (Alabama, Florida, Georgia, Kentucky, Mississippi, North Carolina, South Carolina, and Tennessee)

Region 5 — 1.9% (Illinois, Indiana, Michigan, Minnesota, Ohio, and Wisconsin)

Region 6 — 3.8% (Arkansas, Louisiana, New Mexico, Oklahoma, and Texas)

Region 7 — 1.7% (Iowa, Kansas, Missouri, and Nebraska)

Region 8 — 2.7% (Colorado, Montana, North Dakota, South Dakota, Utah, and Wyoming)

Region 9 — 2.4% (Arizona, California, Hawaii, and Nevada)

Region 10— 1.5% (Alaska, Idaho, Oregon, and Washington)

The national percentage of patient visits to health care providers for ILI by age group is calculated for the subset of data from providers that report total patient visits by age group. Due to wide variability in the percentage of visits for ILI in different age groups, it is not appropriate to apply the national or regional baselines to age group specific data.

ILI Activity Indicator Map: — Data collected in ILINet are also used to produce a measure of ILI activity for all 50 states, the District of Columbia, New York City, Puerto Rico and the U.S. Virgin Islands. Activity levels are based on the percent of outpatient visits due to ILI in a jurisdiction compared with the average percent of ILI visits that occur during weeks with little or no influenza virus circulation (non-influenza weeks) in that jurisdiction. The number of sites reporting each week is variable; therefore, baselines are adjusted each week based on which sites within each jurisdiction provide data. To perform this adjustment, provider level baseline ratios are calculated for those that have a sufficient reporting history. Providers that do not have the required reporting history are assigned the baseline ratio for their practice type. The jurisdiction level baseline is then calculated using a weighted sum of the baseline ratios for each contributing provider.

The activity levels compare the mean reported percent of visits with ILI for the current week to the mean reported percent of visits due to ILI for non-influenza weeks.  There are 13 activity levels that correspond to the number of standard deviations below, at or above the mean for the current week compared with the mean of the non-influenza weeks.  The levels are classified as minimal (levels 1-3), low (levels 4-5), moderate (levels 6-7), high (levels 8-10) and very high (11-13).  More specifically:

• Level 1 – below the mean
• Level 2 – less than 1 standard deviation above the mean
• Level 3 – more than 1, but less than 2 standard deviations above the mean
• Level 4 – more than 2, but less than 3 standard deviations above the mean
• Level 5 – more than 3, but less than 4 standard deviations above the mean
• Level 6 – more than 4, but less than 5 standard deviations above the mean
• Level 7 – more than 5, but less than 6 standard deviations above the mean
• Level 8 – more than 6, but less than 7 standard deviations above the mean
• Level 9 – more than 7, but less than 8 standard deviations above the mean
• Level 10 – 8 to 11.9 standard deviations above the mean
• Level 11 –12 to 15.9 standard deviations above the mean
• Level 12 – 16 to 19.9 standard deviations above the mean
• Level 13 – at least 20 standard deviations above the mean

The ILI Activity Indicator map reflects the level of ILI activity, not the extent of geographic spread of influenza, within a jurisdiction. Therefore, outbreaks occurring in a single city could cause the state to display high activity levels. In addition, data collected in ILINet may disproportionally represent certain populations within a state, and therefore, may not accurately depict the full picture of influenza activity for the whole state. Differences in the data presented here by CDC and independently by some state health departments likely represent differing levels of data completeness with data presented by the state likely being the more complete.

Emergency Department (ED) visits captured through the National Syndromic Surveillance Program (NSSP) are also being used to monitor COVID-19-like illness. NSSP is a collaboration among CDC, federal partners, local and state health departments, and academic and private sector partners to collect, analyze, and share electronic patient encounter data received from multiple health care settings. To track trends of potential COVID-19 visits, visits for COVID-19-like illness (CLI) (fever and cough or shortness of breath or difficulty breathing or the presence of coronavirus diagnosis code) and ILI to a subset of emergency departments in 47 states are being monitored. Visits meeting the ILI or CLI definition that also have mention of flu or influenza are excluded. The percentage of ED visits for each syndrome nationally and for each of 10 HHS regions is calculated weekly.

3. Hospitalization Surveillance

Laboratory-confirmed COVID-19-associated hospitalization rates are monitored through the COVID-19-Associated Hospitalization Surveillance Network (COVID-NET). COVID-NET conducts all-age, population-based surveillance for laboratory-confirmed COVID-19-associated hospitalizations in more than 250 acute care hospitals in 99 counties in the 10 Emerging Infections Program (EIP) states (CA, CO, CT, GA, MD, MN, NM, NY, OR and TN) and four Influenza Hospitalization Surveillance Project (IHSP) states (IA, MI, OH and UT). In total, ~10% of the U.S. population is covered by this surveillance system.

Cases must be a resident of a designated catchment area and hospitalized within 14 days of a positive SARS-CoV-2 test. Testing is performed at the discretion of health care providers. Cases are identified through active review of notifiable disease and laboratory databases and hospital admission and infection control practitioner logs. Data gathered are used to estimate age-specific hospitalization rates on a weekly basis and describe characteristics of persons hospitalized with COVID-19 illness.

Patient charts are reviewed to determine if any of the following categories of underlying medical conditions are recorded in the chart at the time of hospitalization:

• Asthma/reactive airway disease;
• Blood disorder/hemoglobinopathy;
• Cardiovascular disease;
• Chronic lung disease;
• Chronic metabolic disease;
• Gastrointestinal/liver disease;
• Immunocompromised condition;
• Neurologic disorder;
• Obesity;
• Pregnancy status;
• Prematurity (pediatric cases only);
• Renal disease; and
• Rheumatologic/autoimmune/inflammatory conditions.

COVID-19-associated hospitalization rates by race/ethnicity are calculated using hospitalized COVID-NET cases with known race and ethnicity for the numerator and NCHS bridged-race population estimates for the denominator. Rates are adjusted to account for differences in age distributions within race/ethnicity strata in the COVID-NET catchment area; the age strata used for the adjustment include 0–17, 18–49, 50–64, 65–74, 75–84, and 85+ years.

A minimum set of variables (age, sex, race and ethnicity, hospital admission and discharge dates, in-hospital death, date of in-hospital death, and SARS-CoV-2 testing data) are collected for all patients to generate age-stratified rates. Detailed medical chart reviews are performed on all patients younger than 18 years of age or pregnant at time of admission. However, because of the large number of hospitalized COVID-19 cases identified through COVID-NET surveillance, a sampling scheme has been implemented for detailed medical record abstraction (including signs/symptoms at admission, underlying medical conditions, intensive care unit admission, mechanical ventilation and discharge diagnoses) for patients aged 18 years or older. Random numbers are autogenerated and assigned to each case as soon as a case identification number is entered into the surveillance database; random samples of cases, stratified by age group and surveillance site, are drawn using these random numbers. Due to the sampling methodology, counts are not shown, and weighted percentages are presented for the sampled variables in adults. The sampling strategy differs by surveillance site and months and is described in the below table:

* Sampling strategies for September 1, 2020 and beyond will be updated accordingly.

COVID-NET hospitalization data, including rates for different age groups and by surveillance site, and clinical characteristics, are available on the COVID-NET interactive data page.

4. Mortality Surveillance

 National Center for Health Statistics (NCHS) mortality surveillance data – NCHS collects death certificate data from state vital statistics offices for all deaths occurring in the United States. The provisional counts for coronavirus disease (COVID-19) deaths are based mortality data in the National Vital Statistics System. National provisional counts include deaths occurring within the 50 states and the District of Columbia that have been received and coded as of the date specified. It can take several weeks for death records to be submitted to NCHS, processed, coded, and tabulated. Death counts for earlier weeks are continually revised and may increase or decrease as new and updated death certificate data are received. COVID-19 death counts shown here may differ from other published sources, as data currently are lagged by an average of 1–2 weeks.

For COVIDView, the percentage of total deaths occurring in a given week that had pneumonia, influenza and/or COVID-19 (PIC) listed as a cause of death is calculated. PIC deaths are identified based on ICD-10 multiple cause of death codes J09-J18.9 or U07.1. PIC is being monitored in order to provide a more accurate representation of COVID-19 related mortality than would monitoring COVID-19 alone. Deaths due to COVID-19 may be classified as pneumonia deaths or influenza deaths (deaths due to “flu” or “flu-like illness”) in the absence of positive SARS-CoV-2 test results. The combined PIC categorization also prevents double counting of deaths with more than one PIC cause listed on the death certificate.  NCHS surveillance data are aggregated by the week of death occurrence.  The PIC percentage for earlier weeks are continually revised and may increase or decrease as new and updated death certificate data are received from the states by NCHS.

The PIC percentage is compared to a seasonal baseline of P&I deaths that is calculated using a periodic regression model incorporating a robust regression procedure applied to data from the previous five years.  An increase of 1.645 standard deviations above the seasonal baseline of P&I deaths is considered the “epidemic threshold,” i.e., the point at which the observed proportion of deaths is significantly higher than would be expected at that time of the year in the absence of substantial influenza, and now COVID-related mortality.

Additional provisional death counts from NCHS are also available on NCHS’ National Vital Statistics System.