Little has been published about optimizing medical monitoring protocols, although monitoring accounts for over half of all medical laboratory tests. Monitoring is a form of surveillance consisting of repeated testing intended to detect a specified change in a patient indicating a change in his prognosis, need for treatment or need for a change in treatment. The concept of monitoring overlaps with those of screening and diagnosis. One may monitor a physiologic or pathologic process, a therapeutic or noxious agent, comparing results with the patient's previous results or with a group reference range. Pertinent questions include: what are the indications, what is to be monitored, which tests are to be used, when to start, how frequently to test, when to stop. Preliminary improvement of monitoring protocols may be achieved by applying common sense guidelines for medical decision making; using best estimates of test characteristics, of possible risks and benefits of testing, not testing, treating and not treating. Sensitivity analysis may help determine when further clinical research is needed.