Adverse events

Adverse events are harmful or unwanted occurrences observed during or after exposure to a substance, recorded without prejudging whether the substance caused them. The term originates in clinical trials and pharmacovigilance, where a neutral category is needed for everything that goes wrong — so that causation can be evaluated afterward, not assumed at the point of recording.

The deliberate agnosticism is the point. When a trial participant reports a headache, a fall, or a hospitalization, each is logged regardless of whether the study drug is considered responsible. Over many participants, patterns emerge that can implicate or clear a substance.

Outside clinical trials, the same logic applies. Emergency departments, poison control centers, and harm reduction services record adverse events to map what is actually happening across populations using a substance — feeding into real-world safety signals that trials cannot fully anticipate.

How it is done

In clinical trials, adverse events are collected on a defined schedule: participants report symptoms, researchers document clinical findings, and each event is graded by severity and assessed for how plausibly the study drug could be responsible.

Post-market, national pharmacovigilance systems collect adverse event reports from clinicians, pharmacists, and patients through voluntary and mandatory reporting programs. These databases allow regulators to detect harms that appeared only rarely in trials — or appeared only in populations excluded from them.

In harm reduction settings, the concept operates less formally but with real consequences. Emergency department records, drug-checking services, and overdose surveillance networks all generate adverse event data. A harm reduction advisory warning of an unexpectedly potent batch or a novel adulterant is downstream of this kind of reporting.

What it cannot tell you

An adverse event record does not establish that a substance caused the outcome. Causation requires comparison: rates in people who used the substance against rates in people who did not, controlling for what else those groups differ on. A large volume of adverse event reports may reflect that a substance is widely used, not that it is unusually dangerous.

The category also cannot capture what is not reported. Adverse events are systematically undercounted where healthcare access is limited, where use is stigmatized, and wherever people normalize negative consequences as expected. An absence of reports is not evidence that a substance produces no harm — it may indicate a gap in reporting.

A reassuring adverse event profile in trial data may not translate to real-world use. Trial participants are typically screened to exclude people with serious comorbidities or complex polypharmacy — the same characteristics common in the heaviest real-world use. Signals that appear only in those populations can take years to surface through post-market surveillance.