Provisional US overdose death data by state

NarcoScope maps the CDC National Center for Health Statistics VSRR dataset at its published grain: jurisdiction, 12-month-ending period and substance class. The checked-in extract spans 2015 through the February 2026 reporting window across 54 national, state and jurisdiction records.

Open the overdose mapCompare independent evidence lanesOpen the CDC dataset

What the mortality record measures

PeriodProvisional deaths in rolling 12-month-ending windows, not one-month counts.
GrainJurisdiction + period + substance class; some classes can overlap.
Revision stateReported counts, predicted counts where supplied, completeness and suppression remain distinct.

The public extract covers all-drug deaths plus cocaine, heroin, all opioids, psychostimulants and synthetic opioids. “Psychostimulants” is a cause-of-death code class, not a synonym for methamphetamine. Suppressed cells are omitted, never filled with zero.

Questions the data can—and cannot—answer

Where can I find provisional overdose death data by state?

CDC’s VSRR dataset is the authoritative public source. NarcoScope provides an exploratory map and keeps the publisher link, reporting window, jurisdiction, substance class and provisional status alongside the chart.

Are the newest counts final?

No. Provisional records are revised as death certificates and cause coding arrive. A predicted count is an adjustment supplied in the source, not an observed final total, and it is kept separate from the reported count.

Can mortality be attributed to a shipment, exporter or country?

Not from this aggregate dataset. It has no shared incident, shipment or exporter identifier with precursor or seizure records. Showing two trends together does not create the missing causal join.

Public-health boundary: these are aggregate mortality statistics, not medical advice or a real-time surveillance feed. For an emergency in the United States, call emergency services; for treatment information, use an appropriate official health service.

Use the record responsibly

Related research