Why this exists

Read this first, because it is the point of the page. The subject of this analysis is a drug supply. It is not a patient, a caseload or a population of individuals. Nothing in this component identifies a person, and it is not designed to.

The reason it exists is that the substances killing people change faster than any official reporting cycle. CDC surveillance found carfentanil-involved deaths rising roughly sevenfold in a year, from 29 in the first half of 2023 to 238 in the first half of 2024. An adulterant can enter a local supply, kill people for months and appear in a published dataset a year later. By then the pattern has moved.

The vocabulary problem is real too. What a substance is called locally changes constantly, and it changes precisely so that it is not recognized. A harm-reduction worker, an emergency clinician and a public health analyst in the same county can be describing the same thing with three different words and not know it.

How it works

• Public, aggregate sources. Openly posted content, in aggregate, analyzed for vocabulary and geography. No account-level analysis, no identification of posters, no correlation of posts with patients.
• Vocabulary tracking. New street names for substances, and shifts in what existing names refer to — which is often the first observable sign that a supply has changed composition.
• Geographic aggregation. Output is at the level of a region, not a person. A heat map of a county is a public-health instrument. A map of an individual would be something else entirely, and this program does not produce one.
• Corroborated before it is used. A signal from open sources is a hypothesis. It is checked against toxicology and overdose data — see emerging threat analysis — before anybody acts on it.
• Output goes to warning, not enforcement. The product is “this is in the supply here, this is what it is being called, warn people” — routed to public health communication, harm-reduction services and clinicians.

Questions this raises

Does IntellaRx monitor patients’ social media accounts?

No. This component analyzes publicly posted material in aggregate for drug-market vocabulary and geography. Individuals in the program are not searched for, not matched to accounts and not analyzed. Building that capability is outside the scope of this program. See what supply signal analysis actually looks at.

Who receives the output?

Public health communication, harm-reduction services and clinicians — so that people who use drugs can be warned about what is circulating. The product is a warning, not an intelligence lead. See what we do and do not do.