Many patients do not take their medication as prescribed by their physician, either intentionally or unintentionally. This phenomenon, most referred as nonadherence, leads to suboptimal treatment response and increased healthcare costs. To detect nonadherence, in addition to indirect methods, we also use direct methods that detect the presence of the drug in patient's body. However, by simple assessment based on the presence or the absence of a measurable concentration of the drug or its metabolite in a blood sample, it is difficult to distinguish complete adherence from partial adherence, i.e., a situation where the patient takes the drug irregularly, or from masked adherence, where the patient takes his medicine only once before a medical appointment. A possible refinement may be allowed using pharmacokinetic simulations, which estimate whether the measured value corresponds to the expected drug level after a single or short-term administration, or whether it represents steady state confirming long-term adherence. Not only an analysis of a parent drug, but also its metabolite with more favourable pharmacokinetic properties (especially with longer half-life) can be used. Since, in many cases, nonadherence can be caused by several different reasons, multiple approaches in its detection and management are required and a multidisciplinary approach should be involved.