DataTrove-edu

Deaths per week in the US by cause

Tags: Cancer Heart disease

Function family: Trigonometric Topic: Public health

The historical data of deaths per week of the year are sinusoidal for many causes of death. They vary by season with peaks in winter and troughs in summer.

For heart disease, the periodic variation is apparently a combination of stress, overeating/drinking, and extra exertion (like shoveling snow): all apparently lead to more heart attacks.

These are heart-related weekly deaths in the US:

Heart disease deaths per week in the US

And cancer related deaths in the US:

Cancer disease deaths per week in the US

Note: The drop at the far end is expected and comes from reporting delays.

Here are aggregate weekly deaths in the US, from all causes, which show a noticeable increase due to COVID-19 after 2020:

Deaths per week in the US

There is a periodic component and a mild growth trend which may be due to population growth. Thus, modeling these could be purely periodic, or one could add a linear or exponential term.

For example, a possible model for weekly heart-related deaths in the US is shown below

Heart disease deaths per week in the US together with sine + trend model

For a plot with other causes of death, see the Scientific American Article, or the graph below created from the raw data:

Deaths by cause

Cleaned up data

The cause-specific data comes from National Center for Health Statistics. Weekly Counts of Death by Jurisdiction and Select Causes of Death, accessed on February 26, 2026, available from the CDC site https://data.cdc.gov/d/u6jv-9ijr. The aggregate all-cause weekly totals come from National Center for Health Statistics. Excess Deaths Associated with COVID-19, accessed on June 16, 2026, available from https://data.cdc.gov/d/xkkf-xrst.

Sources