National Center for Health Statistics death records

What to take away

What NCHS actually collects

National Center for Health Statistics death records are statistical, not biographical. The National Center for Health Statistics (NCHS) is the statistical agency inside the Centers for Disease Control and Prevention. It does not investigate deaths or keep biographies.

Through the National Vital Statistics System, NCHS compiles mortality data supplied by vital-statistics offices in the fifty states, the District of Columbia and New York City. The Deaths and Mortality page lists the full set of products.

Key NCHS death-data resources include:

Nothing about funeral arrangements, obituaries or burial locations enters the file. The output is statistical, which is why it can confirm a death without confirming a story.

One rule trips up more comparisons than any other. NCHS publishes by place of residence, not place of death. A person who lived in Georgia and died in a Manhattan hospital counts as a Georgia death.

A local report names the hospital. A national table names the home address. Both can be correct, and neither is a contradiction to report.

How a death reaches the national file

The National Vital Statistics System is not a database you can query by name. It is a cooperative agreement among jurisdictions, and state law sets its terms.

The NVSS - Mortality Data page describes that collection and publication process.

The Social Security Administration death file is a separate system, and it does not carry cause.

Transmission timing varies by state. Some send within days, others batch monthly, so a death filed in one state can appear in national provisional files weeks before a death from the same week in a neighboring state.

Provisional death counts: what they confirm

Provisional death counts are the fastest product NCHS releases. They rest on records received so far, update on a rolling schedule, and are labeled incomplete on purpose.

They show aggregate deaths, dates, states of residence and broad cause groups for the published period. They cannot verify that a particular person died, establish an individual's cause or manner of death, or reveal anything from the medical narrative.

A death under investigation may sit in a residual category for months. A sudden spike or drop in a provisional series is frequently a reporting artifact rather than an event, which is the trap our note on why no cause column appears in some monthly tables sets out.

Provisional figures are also revised. A number quoted today can differ next month, so cite the revision date alongside the figure.

National Vital Statistics Reports: the citable version

National Vital Statistics Reports are the agency's formal publications, with methods sections and tables built to be cited. They are the closest thing to an authoritative document in this field.

A typical report covers one full data year and arrives well after that year closes. The delay buys accuracy: the file must close, late records must be added, and codes must be reviewed.

These reports carry rates, not just counts. Crude death rates, age-adjusted rates and life expectancy tables all live here. Age adjustment matters, because a state with an older population shows a higher crude rate with no difference in risk.

A final annual file typically records around 3.3 million deaths, a figure that shifts by a few percent each year. Table 1 of a typical NVSR report lists leading causes with counts and age-adjusted rates.

Use them for context, never for names. A report will say how many people died of a cause in a year, a state and an age band. It will not say who.

The Products - National Vital Statistics Reports - Homepage lists current and archived issues. For a story that runs for years, a published table is the stable reference point while coverage changes around it.

Cause-of-death data and where it fails

NCHS publishes cause-of-death data in the Multiple Cause of Death (MCD) and Underlying Cause of Death (UCD) files. The data come from medical certification on the certificate and are coded to the International Classification of Diseases. The underlying cause is the condition or injury that began the chain leading to death.

Five limits recur:

A certificate and a news report can therefore disagree without either being false. Our guide to facts, records and changes walks through how cause wording shifts over time.

Treat any single cause figure as open until the final file closes. Treat any figure quoted without a source year as unusable.

Checking a reported celebrity death against NCHS data

This is a short exercise, not a research project. The aim is to separate what official data supports from what a headline asserts.

For a named example, an obituary or memorial for Elvis Presley is an individual account. NCHS mortality tables can provide population-level context, but they cannot identify him in public mortality data or confirm the wording of an obituary.

A worked example. A report says an actor died in Florida and sources say the cause was an overdose. Florida provisional counts will show total deaths and a drug-related category, not the actor or one death.

You can publish that Florida recorded deaths in that category for that period. You cannot publish that NCHS confirms the overdose. The same discipline keeps the monthly view honest.

When a death is prominent, the certificate may be amended after an investigation closes, and that can take a year or more. Until it does, the official record is incomplete, and saying so is the accurate report.

Where NCHS stops and state records begin

NCHS publishes statistics. It does not issue certificates and will not mail a copy. The underlying record stays with the state or local vital records office where the death was filed, which is also where a certified copy is requested, usually for a fee and often under eligibility restrictions.

The Where to Write for Vital Records - Homepage gives state-by-state instructions, including addresses, fees and waiting periods. Rules diverge sharply. Some states release certificates to anyone. Others limit them to immediate family, legal representatives or people with a documented interest.

Other archives sit outside this system. The National Archives and Records Administration holds federal records, not state death certificates. The Library of Congress Chronicling America project holds old newspapers, which is where obituary text lives. The Federal Trade Commission Funeral Rule governs funeral pricing, not death records.

The division is simple. Use NCHS for counts, rates and cause groups. Use the state office for the certificate. Use newspapers and archives for the biography. Our overview of year-by-year totals shows how the two halves fit together.

State files are not identical. A field collected in California may be optional elsewhere, and national totals are built from uneven parts. That unevenness is a known feature of the system, not a scandal. When causes of death appear in the yearly files, they reflect what each jurisdiction actually submitted.

Common questions

Can I search NCHS data by a celebrity's name?

No. Public NCHS mortality tables are not searchable by a celebrity's name. The National Death Index is a separate index used for authorized record matching, not a public obituary search. For an individual obituary or memorial, consult a newspaper or archive; certificate access belongs to the relevant state or local office.

How long until a death appears in provisional counts?

It depends on the state. Some transmit within days and others batch monthly. A death under investigation can sit in a pending category for months before the cause is coded.

Do provisional counts get corrected?

Yes. Provisional files are revised as more records arrive and as causes are amended. Note the revision date before quoting a figure, and check whether a final file has replaced it.

What separates a death certificate from an NCHS statistic?

The certificate is the legal record, filed with a state or local office, and it names the person. The statistic is a coded, de-identified entry in a national file used for counts and rates.