CDC WONDER Drug-Overdose Mortality Trends

What this research found

Drug overdose is now a leading cause of injury death in the United States, and tracking it accurately means pinning down a precise case definition and a consistent population standard. K-Dense pulled the national mortality series for 2010 through 2021 from CDC WONDER and the National Vital Statistics System, using the standard sixteen-code drug-poisoning definition and age-adjusting every rate to the 2000 US standard population. The national rate more than doubled over twelve years, from 12.3 to 32.4 deaths per 100,000, and the 2021 state-level spread runs eight-fold from West Virginia to Nebraska.

• The national age-adjusted overdose death rate rose from 12.3 per 100,000 in 2010 (95% confidence interval 12.1 to 12.5) to 32.4 in 2021 (32.2 to 32.6), an increase of 163.4%. Absolute deaths grew from 38,329 to 106,699, up 178.4%.
• The steepest single-year jump in the series coincides with the onset of COVID-19: the rate went from 21.6 per 100,000 in 2019 to 28.3 in 2020, roughly a one-third increase in a single year, before continuing to 32.4 in 2021.
• The trajectory was not monotonic. Growth through 2010-2015 averaged roughly one additional death per 100,000 per year, accelerated sharply in 2016-2017, then briefly plateaued with small declines in 2012 and 2018 — the only two years in the series where the rate fell.
• State burden in 2021 varied eight-fold. West Virginia was highest at 90.9 per 100,000 (86.3 to 95.5, from 1,501 deaths), well clear of second-place District of Columbia at 63.6, while Nebraska was lowest at 11.4 (9.9 to 12.9, from 214 deaths). The median jurisdiction sat at 31.5 and the mean at 34.3.
• Every one of the 51 jurisdictions cleared the reliability threshold by a wide margin — the smallest death count was 105, in South Dakota — so no rate needed suppression for small counts and all jurisdictions were rankable.
• Precision varies systematically with jurisdiction size because the standard error scales with the rate divided by the square root of deaths. The widest interval belongs to the District of Columbia at about plus or minus 5.9, the narrowest to Texas at about plus or minus 0.5, so mid-table rank comparisons are less secure than the extremes, whose intervals do not overlap their neighbours.

How it was done

Cases were defined by the standard NCHS drug-poisoning criterion: an underlying cause of death coded to any of sixteen ICD-10 codes spanning unintentional poisoning, suicide, homicide, and undetermined intent. National counts and rates for 2010 through 2020 came from the CDC WONDER Underlying Cause of Death database, while the 2021 national figure was taken from the final National Vital Statistics System tabulation and 2021 state rates from the NCHS Stats of the States series, because the successor single-race database blocks the group-by queries this analysis required. All rates are directly age-adjusted to the 2000 US standard population so that jurisdictions and years are comparable. Standard errors for 2010-2020 came from CDC WONDER directly; for 2021 they were estimated by a Poisson approximation, and 95% intervals were formed at 1.96 standard errors. The 2021 national death count was cross-validated against the independent sum of the 51 state counts and matched exactly.

Data sources

• CDC WONDER Underlying Cause of Death, Database D76 — national counts and age-adjusted rates, 2010-2020, bridged-race denominators (accessed 10 July 2026)
• NCHS National Vital Statistics System final tabulation, 2021 — single-race denominators
• NCHS Stats of the States, Drug Overdose Mortality — age-adjusted rates for 50 states and the District of Columbia, 2021
• ICD-10 underlying-cause drug-poisoning codes X40-X44, X60-X64, X85, Y10-Y14