The US cancer death rate falls below 115 per 100,000

Where the rate stands now

• Latest final figure. NCHS Data Brief 548 (published 29 Jan 2026) gives the age-adjusted cancer death rate as 141.8 in 2023 and 139.4 in 2024, a 1.7% fall. https://www.cdc.gov/nchs/products/databriefs/db548.htm
• Latest dashboard figure. The Mortality Dashboard's downloadable table (data.cdc.gov dataset 489q-934x, rows last updated 23 Jun 2026) shows 137.6 for the 12 months ending Q3 2025. The technical notes say 2025 rates on the dashboard still use Vintage 2024 population estimates. https://www.cdc.gov/nchs/nvss/vsrr/mortality-technical-notes.htm
• A one-time drop in the level. NCHS's provisional 2025 report (VSRR No. 44, 2 Jul 2026) switched to the Census Vintage 2025 population base, which is built from the enumerated 2020 census. That base revised the 2024 population up by 1.7% at ages 65–74, 3.4% at 75–84 and 5.8% at 85+. https://www.cdc.gov/nchs/data/vsrr/vsrr044.pdf
• My estimate of the new level. I used 2000-standard direct age adjustment over 11 age groups, Census files nc-est2024- and nc-est2025-agesex-res.csv, and cancer deaths by age from CDC WONDER provisional data (dataset D176, as of 4 Oct 2026).
  • The method exactly reproduces NCHS's all-cause rates: 722.1 for 2024 and 689.2 for 2025.
  • It gives cancer rates of 139.4 for 2024 (matching the published figure) and about 133.9 for 2025 on the new base.
  • On a consistent Vintage 2025 basis, 2020–2025 runs 144.3 → 142.9 → 140.6 → 138.3 → 136.4 → 133.9. That is −1.5% a year on average and −1.9% in 2025.
  • January–August 2026 deaths imply roughly −2% to −2.3% for 2026.
• Publication timing.
  • The NCHS 2026 release calendar lists both the Q4 2025 dashboard update and "Mortality in the United States, 2025" for December 2026. Neither counts, because their periods end before 16 Sep 2026.
  • The dashboard lag has grown from about 5–7 months in 2024–25 to 9–12 months now.
  • The first eligible period is the 12 months ending 30 Sep 2026, likely published mid-to-late 2027 at about 131–132.
  • By the end of 2031, the latest published period will probably be calendar 2030, or at most the 12 months ending Q1/Q2 2031.

Reference class and what YES would take

Using NCHS's long series (data.cdc.gov 6rkc-nb2q through 2018, then the Data Briefs), the fastest 5-year average decline since 1991 was 1.93% a year (2014–2019). The average since 1996 is about 1.4% a year. The 2026 Annual Report to the Nation gives −1.6% a year for men and −1.1% for women over 2019–2023.

Getting from about 133.9 (2025) to below 115 by the 2030 data needs at least 3.0% a year; by mid-2031 data it needs about 2.7% a year. That is roughly 1.5–2 times the best five years on record. Lung cancer, which drives most of the progress, is now only about 21% of cancer deaths, so its pull on the overall trend is shrinking. At 1.8–2% a year, the 2030 rate would be about 121–123.

Model

I ran a Monte Carlo with these inputs:

• Trend: drawn from a normal distribution with mean about 1.75–1.85% a year and standard deviation 0.35–0.4%, plus year-to-year noise.
• Population-estimate revisions: fat-tailed, partly self-correcting level shifts with a standard deviation of about 1.2% each. Past revisions moved the cancer rate by +1.2%, −0.3% and −2.2%.
• Publication lag: 7–12 months.
• A 15% chance the dashboard is discontinued, leaving only annual data.

Cumulative probability of YES comes out at about 0.03% by end-2029, about 0.1% by end-2030, and 0.3–1.6% by end-2031 depending on parameters. I added a tiny hazard for a typo, a data error or a methodology change. The hazards stay near zero until the 2030 data start to appear, giving about 1% cumulative by end-2031.

Paths and counterarguments

• Path to YES. It would take several years of 3%-plus annual declines, for example from fast uptake of immunotherapy, antibody-drug conjugates or other breakthroughs. Alternatively, a 2.5% a year trend combined with another large upward revision of the elderly population could do it. Even 2.2% a year plus a second 2.5% downward population shift leaves 2030 at about 117.
• Strongest case for higher. Declines have recently sped up (about 1.9% in 2025 and about 2.2% in early 2026), and the 2025 population switch alone cut the published level by about 2%. A repeat of that switch would matter.
• Why NO is more likely.
  • Population revisions can go either way.
  • Policy headwinds include Medicaid coverage cuts and research funding cuts.
  • Longer NCHS lags shrink the effective window.
  • A switch to ICD-11 would, if anything, move some blood cancers into the malignant category and raise the rate.

Reading adopted

Only NCHS-published rates on the 2000 US standard population count, for periods ending on or after 30 Sep 2026. Partial-year CDC WONDER query outputs (for example, the 2026 "provisional and partial" rate of 101.8, which uses the full-year 2024 population) are user-run calculations for incomplete years, so I excluded them.

Question. By the last day of each quarter, has CDC's National Center for Health Statistics reported an age-adjusted US cancer death rate below 115 per 100,000 for any calendar year or any twelve months ending in a quarter?

Definitions.

• Cancer death rate: the national underlying-cause mortality rate for all malignant neoplasms combined (ICD-10 C00–C97), not deaths with cancer listed only as a contributing cause.
• US, age-adjusted: the total population of the 50 states and District of Columbia, all sexes and racial and ethnic groups combined, with rates directly standardized to the 2000 US standard population and expressed per 100,000. Rates for territories, individual states, demographic subgroups, or individual cancer sites do not qualify.
• Eligible period: a complete January–December calendar year or a twelve-month period ending March 31, June 30, September 30, or December 31. The Observation period convention applies.
• Below 115: the published point estimate is strictly less than 115, using the precision published by NCHS. A displayed value of 115.0 does not qualify; a confidence interval extending below 115 is insufficient.

Resolves YES if NCHS publishes a qualifying rate for an eligible period in final or provisional mortality statistics. An initial estimate or a subsequent revision can qualify. No attribution to AI or particular treatments is required. Later revisions do not undo a YES under the Resolution window convention.

Resolves NO if no qualifying rate has been published by the horizon. Exclude crude rates, age-specific rates, cancer incidence or survival statistics, multi-year averages, forecasts, and annualized rates for a single quarter. Exclude independent calculations from death counts or microdata and figures published only by organizations other than NCHS. Publications that only repeat or revise figures for observation periods ending on or before the opening date do not qualify.

Resolving source. NCHS publications and official NCHS statistical tables only, a narrower source requirement than Public confirmation (A2). Qualifying sources include the annual Mortality in the United States Data Brief (the cancer row in the leading-causes age-adjusted death-rate table, e.g. Table 4 of Data Brief 548); Deaths: Final Data in National Vital Statistics Reports; and Vital Statistics Rapid Release's Mortality Dashboard and its downloadable tables, selecting Cancer, Age-adjusted, Total US population, and 12 months ending with quarter. Other NCHS reports or successor tables reporting the same measure also qualify. Apply the Resolution window convention and Resolution date (A3); for updated tables, use the dated release that first publishes the qualifying value, not the end of the observation period.