Technical Notes

Provisional Death Counts for COVID-19

PAGE 7 of 7

View Table of Contents

The provisional counts for COVID-19 deaths are based on a current flow of 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 is important to note that it can take several weeks for death records to be submitted to National Center for Health Statistics (NCHS), processed, coded, and tabulated. Therefore, the data shown on this page may be incomplete, and will likely not include all deaths that occurred during a given time period, especially for the more recent time periods. Death counts 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. COVID-19 death counts shown here may differ from other published sources, as data currently are lagged by an average of 1–2 weeks.

Comparing data in this report to other sources

Provisional death counts in this report will not match counts in other sources, such as media reports or numbers from county health departments. Death data, once received and processed by National Center for Health Statistics (NCHS), are tabulated by the state or jurisdiction in which the death occurred. Death counts are not tabulated by the decedent’s state of residence. COVID-19 deaths may also be classified or defined differently in various reporting and surveillance systems. Death counts in this report include laboratory confirmed COVID-19 deaths and clinically confirmed COVID-19 deaths. This includes deaths where COVID-19 is listed as a “presumed” or “probable” cause. Some local and state health departments only report laboratory-confirmed COVID-19 deaths. This may partly account for differences between NCHS reported death counts and death counts reported in other sources. Provisional counts reported here track approximately 1–2 weeks behind other published data sources on the number of COVID-19 deaths in the U.S. (1,2,3).

Nature and sources of data

Provisional death counts are based on death records received and processed by NCHS as of a specified cutoff date. National provisional counts include deaths occurring within the 50 states and the District of Columbia. NCHS receives the death records from state vital registration offices through the Vital Statistics Cooperative Program. Provisional data are based on available records that meet certain data quality criteria at the time of analysis and may not include all deaths that occurred during a given time period especially for more recent periods. Estimates of completeness are provided. Therefore, they should not be considered comparable with final data and are subject to change.

Cause-of-death classification and definition of deaths

Mortality statistics are compiled in accordance with World Health Organization (WHO) regulations specifying that WHO member nations classify and code causes of death with the current revision of the International Statistical Classification of Diseases and Related Health Problems (ICD). ICD provides the basic guidance used in virtually all countries to code and classify causes of death. It provides not only disease, injury, and poisoning categories but also the rules used to select the single underlying cause of death for tabulation from the several diagnoses that may be reported on a single death certificate, as well as definitions, tabulation lists, the format of the death certificate, and regulations on use of the classification. Causes of death for data presented in this report were coded according to ICD guidelines described in annual issues of Part 2a of the NCHS Instruction Manual (4).

COVID-19 deaths are identified using the ICD–10 code U07.1. Deaths are coded to U07.1 when COVID-19 is reported as a cause that contributed to death on the death certificate. These can include laboratory confirmed cases, as well as cases without laboratory confirmation. If the certifier suspects COVID-19 or determines it was likely (e.g., the circumstances were compelling within a reasonable degree of certainty), they can report COVID-19 as “probable” or “presumed” on the death certificate (5, 6). COVID-19 is listed as the underlying cause on the death certificate in 87% of deaths, but this varies from year to year (see Tables 1 and 2).

Pneumonia deaths are identified using multiple cause-of-death codes from the 10th Revision of ICD (ICD–10): J12–J18, excluding deaths that involve influenza (J09–J11). Influenza deaths are identified from the ICD–10 codes J09–J11, and include deaths with pneumonia or COVID-19 listed as a contributing cause of death.

Table 1. Percent of COVID-19 deaths with COVID-19 as underlying cause on the death certificate, by year, United States.

Percent of deaths with COVID-19 coded as the underlying cause, by year

Table 2. Percent of COVID-19 deaths with COVID-19 listed as underlying cause on the death certificate, by sex and age, 2020-2022, United States.

Percent of deaths with COVID-19 coded as the underlying cause, by sex and age

Estimated completeness of data

Provisional data are incomplete, and the level of completeness varies by jurisdiction, week, decedent’s age, and cause of death. Until data for a calendar year are finalized, typically in December of the following year, completeness of provisional data cannot be determined. However, completeness can be estimated in a variety of ways. Surveillance systems that rely on weekly monitoring of provisional mortality data, such as CDC’s FluView Interactive mortality surveillance (7), estimate completeness by comparing the count of deaths in a given week of the current year to the average count of deaths in that same week in previous years. These estimates can be generated for specific causes of death, jurisdictions, and age groups, and updated on a weekly or daily basis. For the purposes of COVID-19 surveillance, completeness is approximated by comparing the provisional number of deaths received to the number of expected deaths based on prior years data. Percent of expected deaths provided in this data release are based on the total count of deaths in the most recent weeks of the current year, compared with an average across the same weeks in previous years. These estimates of completeness are calculated by week, month and year, jurisdiction of occurrence, and age group.

It is important to note that the true levels of completeness are unknown, and the estimates provided here are only a proxy. In cases where mortality rates are increasing rapidly, particularly when excess deaths due to a novel cause are occurring, values for completeness for recent weeks may exceed 100% even when NCHS has yet to receive all available data. Conversely, if the number of deaths was elevated in prior years due to a severe flu season, for example, estimated completeness in the most recent weeks may be lower than the true value. To avoid relying too heavily on comparisons to a single week of a single prior year, estimates of completeness included in this release are based on the average counts in a given week across three years from 2017 to 2019.

Percent of expected deaths provided in this release are shown to provide context for interpreting provisional counts of COVID-19 deaths and deaths due to related causes. Where estimated values are high (e.g., greater than 100%), this suggests that mortality is higher in 2020-2023 relative to the same weeks of prior years. Where estimated values of completeness are low, this could indicate that data are incomplete due to delayed reporting, or that mortality is lower in 2020-2023 compared with prior years, or some combination of these factors.

Delays in reporting

Provisional counts of deaths are underestimated relative to final counts. This is due to the many steps involved in reporting death certificate data. When a death occurs, a certifier (e.g. physician, medical examiner or coroner) will complete the death certificate with the underlying cause of death and any contributing causes of death. In some cases, laboratory tests or autopsy results may be required to determine the cause of death. Completed death certificate are sent to the state vital records office and then to NCHS for cause of death coding. At NCHS, about 90% of deaths are automatically processed and coded within seconds, but 10% of deaths need to be manually coded, or coded by a person. Deaths involving certain conditions such as influenza and pneumonia are more likely to require manual coding than other causes of death. Death certificates are typically manually coded within 7 days of receipt, although the coding delay can grow if there is a large increase in the number of deaths. As a result, underestimation of the number of deaths may be greater for certain causes of death than others.

There have been recent improvements in the timeliness of provisional mortality data. Using mortality data from 2021, the data were approximately 65% complete within 2 weeks, 85% complete within 4 weeks, and at least 94% complete within 8 weeks of when the death occurred. Similarly, COVID-19 deaths are 65% complete within 2 weeks, 85% complete within 4 weeks, and 95% complete within 8 weeks of when the death occurred. These completeness estimates reflect improvements in the timeliness of provisional mortality data in recent years. Previous estimates based on 2015 mortality data suggested that completeness was at 27% within 2 weeks, 54% within 4 weeks, and 75% within 8 weeks of when the death occurred.

Estimated distributions of COVID-19 deaths and population size by race and Hispanic origin

The percentages of COVID-19 deaths by race and Hispanic origin were calculated by dividing the number of COVID-19 deaths for each race and Hispanic origin group by the total number of COVID-19 deaths. Percentages may not sum to 100 due to rounding.

In April and May 2020, the majority of COVID-19 deaths in the U.S. occurred in urban areas that have a larger percentage of their populations that are non-Hispanic black, non-Hispanic Asian, or Hispanic, and a smaller percentage that are non-Hispanic white. Weighted population distributions use county-level data to more closely align the population distributions with the specific areas where COVID-19 deaths were occurring. To account for the geographic clustering of COVID-19 deaths, weighted population distributions, which more closely matched the areas initially most affected by COVID-19 deaths, were provided in order to reflect differential risk within the areas most affected by COVID-19 deaths. As the pandemic has become more widespread across the U.S., there is less need to align the population distributions with the specific geographic areas experiencing COVID-19 outbreaks and mortality. While weighted estimates are no longer included in the data visualization, the estimates can still be found in the data file.

The weighted population distributions were calculated as follows. County-level population counts by race and Hispanic origin were multiplied by the corresponding total count of COVID-19 deaths by county (of residence). These weighted counts were then summed to the state (or national) level. The percentage of the population within each race and Hispanic origin group by state (or for the US) was then estimated using these weighted counts. Counties with no COVID-19 deaths received a weight of zero, and thus do not contribute to the weighted population totals. Population counts for counties with large numbers of COVID-19 deaths are upweighted proportional to their numbers of COVID-19 deaths. These weighted population distributions ensure that the population estimates and percentages of COVID-19 deaths represent comparable geographic areas, in order to provide information about whether certain racial and ethnic subgroups are experiencing a disproportionate burden of COVID-19 mortality. For example, assume that 75% of the total number of COVID-19 deaths occurred in a single county, County X, while the other 25% of COVID-19 deaths occurred in County Y, and all other counties reported zero deaths. The weighted population counts for County X would contribute 75% of the total population counts, while the population counts for County Y would contribute 25% of the total, while all other counties with zero COVID-19 deaths would not count toward the total population counts. In other words, County X population counts would be weighted by 0.75, County Y population counts would be weighted by 0.25, and all other county population counts would be weighted by 0. These weighted counts are then summed to a total (either state or US), and then the percent of the population in each race and Hispanic origin group is computed. These weighted distributions ensure that the population distributions are as closely matched to the geographic areas where COVID-19 deaths are occurring, to the extent possible.

Unweighted population distributions by state are provided for context in Table 1 of the Health Disparities: Race & Hispanic Origin page, and are publicly available (see: https://wonder.cdc.gov/single-race-population.html. For example, 60% of the United States population is non-Hispanic white and 13% non-Hispanic black or African American. The majority of COVID-19 deaths have occurred in New York City where the racial distribution is different than the racial distribution of the United States. After weighting the population to reflect the areas experiencing the greatest number of COVID-19 deaths (i.e., up-weighting areas like New York City that have a disproportionate amount of deaths), the weighted percent of the US population that is non-Hispanic white is reduced to less than 50% and the percent that is non-Hispanic black or African American is increased to nearly 25%.

Counties are the smallest geographic unit for which COVID-19 and population data are available. There may be geographic clustering of COVID-19 deaths within counties, and therefore weighting population counts by county may not be sufficient to ensure comparability between the geographic areas where COVID-19 deaths are occurring and the corresponding population estimates.

Quality of Race and Hispanic Origin Data

Race and Hispanic origin data from death certificates are not available for 0.3% of all deaths and 0.7% of deaths involving COVID-19.  Death data for Hispanic, non-Hispanic American Indian or Alaska Native (AIAN), and non-Hispanic Asian or Pacific Islander (API) persons should be interpreted with caution because of inconsistencies in reporting Hispanic origin or race on the death certificate compared with censuses, surveys, and birth certificates. The latest research on race and Hispanic origin misclassification on death certificates shows that deaths are underreported by 33% for non-Hispanic AIAN, 3% for non-Hispanic API, and 3% for Hispanic decedents.  See Technical Notes for additional information (8).

Estimates of disparities and adjusting for age

Disparities are estimated as the difference between the share of COVID-19 deaths experienced by a given race or Hispanic origin group and their share of the population. For example, if 13% of the population is non-Hispanic black, but approximately 25% of COVID-19 decedents are non-Hispanic black, then the difference would be 12%. Positive values indicate that a group represents a greater share of COVID-19 deaths than the population. Negative values for this difference suggest that a given group represents a smaller share of COVID-19 deaths than their share of the population.

Disparities are shown by age group, and age-standardized estimates are also provided. Age-standardized estimates were calculated by multiplying each age-specific distribution by the proportion of the standard 2000 US population in that age group, and then summing these estimates over age groups. Age is therefore controlled for across the different distributions and racial/ethnic groups. Age-standardized distributions show what disparities would look like, assuming that all of the groups had the same age distribution as the 2000 standard population.

Place of Death

Place of death noted on the death certificate is determined by where the death was pronounced and on the physical location where the death occurred (9). Healthcare setting includes hospitals, clinics, medical facilities, or other licensed institutions providing diagnostic and therapeutic services by medical staff. Decedent’s home includes independent living units such as private homes, apartments, bungalows, and cottages. Hospice facility refers to a licensed institution providing hospice care (e.g., palliative and supportive care for the dying), but not to hospice care that might be provided in other settings, such as a patient’s home. Nursing home/long-term care facility refers to a facility that is not a hospital but provides patient care beyond custodial care, such as a nursing home, skilled nursing facility, a long-term care facility, convalescent care facility, intermediate care facility, or residential care facility. Other includes such locations as a licensed ambulatory/surgical center, birthing center, physician’s office, prison ward, public building, worksite, outdoor area, orphanage, or facilities offering housing and custodial care but not patient care (e.g., board and care home, group home, custodial care facility, foster home).

Comparing deaths from different states

Death counts should not be compared across states. Data timeliness varies by state. Some states report deaths on a daily basis, while other states report deaths weekly or monthly. Furthermore, health departments and state vital record offices may be affected by COVID-19 related response activities, which could further delay death certificate reporting. Currently, 61% of US deaths are reported within 10 days of the date of death, but there is variation within states. Twenty states report over 75% of deaths within the first 10 days, while three states report fewer than 1% of deaths within 10 days.

Why are pneumonia and influenza deaths included in this report?

Pneumonia and influenza deaths are included to provide context for understanding the completeness of COVID-19 mortality data and related trends. Deaths due to COVID-19 may be misclassified as pneumonia or influenza deaths in the absence of positive test results, and pneumonia or influenza may appear on death certificates as a comorbid condition. Additionally, COVID-19 symptoms can be similar to influenza-like illness, thus deaths may be misclassified as influenza. Thus, increases in pneumonia and influenza deaths may be an indicator of excess COVID-19-related mortality. Additionally, estimates of completeness for pneumonia and influenza deaths may provide context for understanding the lag in reporting for COVID-19 deaths, as it is anticipated that these causes would have similar delays in reporting, processing, and coding.

Source

NCHS, National Vital Statistics System. Estimates are based on provisional data.

References

  1. Dong E, Du H, Gardner L. An interactive web-based dashboard to track COVID-19 in real time. Lancet Infect Dis. 2020. Available from: https://doi.org/10.1016/S1473-3099(20)30120-1.
  2. Wu J, McCann A, Collins K, Harris R, Huang J, Almukhtar S. Coronavirus in the U.S.: Latest map and case count. New York Times. https://www.nytimes.com/interactive/2020/us/coronavirus-us-cases.html.
  3. National Center for Immunization and Respiratory Diseases (NCIRD), Division of Viral Diseases. Cases in the US. Centers for Disease Control and Prevention. 2020.
  4. National Vital Statistics System. Instructions for classifying the underlying cause of death. In: NCHS instruction manual; Part 2a. Published annually.
  5. National Center for Health Statistics. Guidance for certifying deaths due to COVID–19. Hyattsville, MD. 2020. Available from: https://www.cdc.gov/nchs/data/nvss/vsrg/vsrg03-508.pdf.
  6. National Center for Health Statistics. New ICD code introduced for COVID-19 deaths. Hyattsville, MD. 2020. Available from: https://www.cdc.gov/nchs/data/nvss/coronavirus/Alert-2-New-ICD-code-introduced-for-COVID-19-deaths.pdf
  7. National Center for Immunization and Respiratory Diseases (NCIRD). CDC’s FluView Interactive. Centers for Disease Control and Prevention. Available from: https://www.cdc.gov/flu/weekly/index.htm.
  8. Arias E, Heron M, Hakes JK. The validity of race and Hispanic-origin reporting on death certificates in the United States: An update. National Center for Health Statistics. Vital Health Stat 2(172). 2016.
  9. Centers for Disease Control and Prevention. Medical Examiners’ and Coroners’ handbook on death registration and Fetal death reporting. Hyattsville, Maryland: DHHS Publication; 2003. PHS2003-1110