Diagnostic criteria in clinical trials

It is evident that diagnostic criteria used to select patients in a clinical trial must be carefully chosen. No matter how carefully the trial is conducted or how exact the outcome measures, unless one is certain that the patient population is homogeneous for the disease characteristic under study, the results will be ambiguous. Most commonly, the patient population is chosen on the basis of the clinical diagnosis. Trials are most common in diseases that have been easy for the clinician to recognize. Duchenne dystrophy, Guillain-Barre, and chronic inflammatory demyelinating neuropathy have been the subjects of well-known trials. Within the past few years, there has been a flurry of trials in ALS. This brief essay focuses on the way in which our use of diagnostic criteria has grown and examines the purpose of diagnostic criteria rather than repeating what is already extensively documented in the literature. ^[1,2]

The application of the scientific method to clinical trials is relatively recent. Only two decades ago, numerous reports were accepted by journals documenting the apparent reversal of serious illness in isolated patients. These clinical descriptions were seldom malicious fabrications but rather honest attempts to describe phenomena that had deeply impressed the patient and the physician. If the lucky investigator happened to have such a patient on the nostrum of the week, a publication was guaranteed. Such reports were often accepted for publication on the premise that they held a flicker of hope in diseases that were otherwise fatal. Reviewers adopted the "who knows, there just might be something in it" approach. Uniformly, nothing came of these reports except a growing realization that there was more to the science of testing drugs than could be left to a neurologist's intuition. In the 1970s, there arose a groundswell that moved neurology into the clinical trials arena …