Anderson and Rosenberg 1998 — Age Standardization of Death Rates Implementation of the Year 2000 Standard

Summary

Anderson and Rosenberg document the U.S. switch from the 1940 to the 2000 standard population for computing age-adjusted death rates (AADRs), effective with data year 1999. The paper shows that the 2000 standard — substantially older, with a $65+$ weight of $12.4\%$ vs. $6.8\%$ — raises the all-cause AADR by approximately $82\%$ in absolute level (from $503.9$ to $918.5$ per $100{,}000$ in 1995), compresses the black/white AADR ratio from $1.60$ to $1.41$ as a statistical artifact, and amplifies cause-specific rates for causes concentrated at older ages (stroke AADR increases $2.4\times$). Period life expectancy is unaffected.

Key Claims

• The 2000 standard population has a substantially larger weight for ages $65+$ ($0.124$ vs. $0.068$ under the 1940 standard), reflecting the older age structure of the contemporary U.S. population.
• All-cause AADR in 1995: $503.9$ per $100{,}000$ (1940 standard) $\to$ $918.5$ per $100{,}000$ (2000 standard), an $82\%$ level increase that does not represent a real change in mortality.
• Causes of death concentrated at older ages show the largest amplification; causes primarily affecting younger ages change less.
• Stroke AADR: $26.7 \to 63.9$ per $100{,}000$ in 1995 ($2.4\times$ increase from standard switch alone).
• Black/white AADR ratio: $1.60$ (1940 standard) $\to$ $1.41$ (2000 standard) — a compression artifact because racial disparities are highest at young ages ($\approx 2.0$) and lowest at old ages ($\approx 1.1$), and the older-weighted 2000 standard down-weights the age groups with the largest disparity.
• Period life expectancy $e_0$ is derived from the life table's own internal stationary population and is unaffected by the choice of standard population.
• Healthy People 2010 goals required recalibration to the 2000 standard; Healthy People 2000 goals were set under the 1940 standard.

Concepts Introduced or Extended

• Age Standardization

Entities Mentioned

• Robert N. Anderson
• Harry M. Rosenberg
• National Center for Health Statistics

Quotes

"Beginning with 1999 data, all age-adjusted death rates from the National Center for Health Statistics will be computed using the year 2000 standard population."

My Take

Essential reference for any researcher using U.S. AADR data spanning the 1999 break. The $82\%$ level shift and the racial ratio compression are the most practically important artifacts. The paper is a methodological note, not an epidemiological study — its value is entirely as documentation of the standard switch and its computational consequences.