Serum drug concentration monitoring can be an invaluable aid to patient management, particularly in certain pathological conditions when individualisation of dosage is particularly critical. To be clinically useful, however, drug levels must be interpreted in the context of all factors that could influence the correlation between the concentration of the drug in plasma and the intensity of action. Several such factors may be operating in acute and chronic disease states. For example, a number of pathological conditions are associated with marked changes in the fraction of free, pharmacologically active drug in plasma and this will result in disruption of the normal relationship between total serum drug level and effect, as seen for phenytoin in uraemia. An altered response to a given serum drug level in disease states may also be caused by changes in tissue distribution, by abnormal accumulation of pharmacologically active metabolites in plasma or by changes in end-organ responsiveness. The latter are best illustrated by the altered sensitivity to digoxin in patients with various conditions, including hypokalaemia and thyroid disease. In addition to the factors listed above, consideration should also be given to potential interactions with concomitantly used drugs and to the possibility of analytical errors, especially in view of the evidence that the performance of otherwise reliable drug assays may be grossly impaired in certain diseases (e.g. uraemia), due to abnormal plasma composition and/or accumulation of interfering metabolites. In view of these complexities, a correct interpretation of serum drug levels requires a good knowledge of clinical pharmacology and a close collaboration between physician and laboratory. In any case, serum drug concentrations, like other laboratory tests, are not a substitute for careful patient observation, and any decision about drug treatment should be primarily based upon evaluation of the clinical state and, whenever possible, direct measurement of drug effects.