BIOL PSYCHIATRY 1992;31:1-3 EDITORIAL

It is an article of faith that he action of a drug at the molecular level somehow determines its effects at the phenomenological * level. Yet, there is a lack of correspondence between what we know about drug actions on receptors, and the drug actions that we see as clinicians. We can often find a drug that will relieve a patient's unpleasant experiences, but finding the right drug for the right patient is more of an art than a science. Among patients with the same diagnosis, some patients respond to one type of drug, others only to another. For example, psychotics may benefit from drugs as diverse as neuroleptics, anticonvulsants, beta-blockers, benzodiazepines and lithium. We construct neuro~rans-mitter models of psychiatric illness, only to find each one has some fatal flaw. The phenomenological responses that define therapeutic and toxic responses are the proper concern of the clinician. There are, however, some relatively new linking variables which relate more directly to events at the neuropharmacological level and, in combination with mediating variables, give a better correspondence b tween the clinical and neurophar-macological levels of discourse The idea of mediating variables is an old story in phenomenological psychopharma