The determination of whether a new molecular entity has abuse potential involves a variety of assessments. These can include safety studies, chemistry studies, receptor binding studies, pharmacokinetic studies, animal behavioral studies, human abuse potential studies, an evaluation of adverse events that occurring during clinical safety and efficacy studies, and evidence of abuse from epidemiological studies (if available).

The human abuse potential studies play an important role in the assessment of abuse potential of new drugs. These studies are typically randomized, double-blind, placebo- and positive-controlled crossover investigations in which there are usually at least four treatments. These treatments consist of at least two doses of a test drug, at least one dose of a positive control drug (a drug with known abuse potential that is scheduled under the Controlled Substances Act), and placebo. Study subjects are healthy recreational drug users who have experience with drug class associated with the test drug. Such a study has multiple abuse potential measures, multiple primary endpoints, and multiple comparisons. In recent years, besides general human abuse potential studies, a lot of human abuse potential studies were conducted for assessing abuse deterrent opioids.

There are many statistical issues in human abuse potential studies. For example, misuse multiplicity adjustments for controlling type I error rate, underestimating the sample size of the study, unfair comparison between an intact extended release test drug and the same dose of a crushed immediate release positive control, etc.

In this presentation, I will discuss some of these issues and give my recommendations.