Liability

Liability, as used in pharmacology and harm reduction, names a substance's propensity to produce a particular outcome — identified by the qualifier in front of it. Abuse liability describes how readily a substance tends to be taken in ways that go beyond intended use. Dependence liability describes how frequently sustained use produces a state in which the body or mind requires the substance to function normally. Tolerance liability describes how quickly repeated exposure diminishes a substance's effects.

Each type is a population-level characterisation, not a prediction about any individual. The term exists because comparing risks across compounds requires a shared vocabulary — researchers and regulators needed a way to rank substances before clinical approval or scheduling decisions, and this framework provided it.

How it is done

Liability is established by converging several types of evidence: animal self-administration studies, controlled trials in humans, and epidemiological data from populations with long-term access to the substance.

For abuse liability, the key signals are how readily subjects seek out and re-administer a compound, whether use tends to escalate over time, and whether that pattern persists even when conditions make it costly. Dependence liability is assessed by observing withdrawal syndromes when use stops after a sustained period — tracking their prevalence, clinical character, and duration. Tolerance liability is mapped by following how effective doses shift across repeated exposures, and by measuring changes in receptor sensitivity.

The result is a comparative characterisation. A substance is typically described as carrying low, moderate, or high liability relative to others in its class or against a known reference compound.

What it cannot tell you

A liability rating is a population tendency, not an individual forecast. The most common wrong conclusion from a reassuring figure is that low population risk translates directly to low personal risk.

Individual factors — genetic vulnerability, prior trauma, co-occurring mental health conditions, route of use, and social environment — can shift where a person actually lands relative to the population figure. Substances carrying a "low" dependence liability have still produced severe dependence in individual users; the rating describes lower frequency across a group, not zero occurrence within it.

Liability ratings also do not transfer cleanly across routes of administration. Figures derived from one route may not reflect outcomes when the pharmacokinetics change substantially — as they often do.

Combined use falls entirely outside the scope of a single substance's liability rating. When two or more substances are used together, the emergent risk cannot be read off their individual figures. That analysis belongs in the interaction layer, where it can carry the evidence it requires.