Introduction

The introduction of the controlled clinical trial heralded the era of rational treatment. An important landmark was the UK Medical Research Council trial of streptomycin in pulmonary tuberculosis, with random assignment to treatment groups, but some forerunners had already used parallel control groups. Among these were James Lind in 1753 (lemons and oranges to prevent scurvy in sailors), Louis in 1835 (bleeding as a treatment for pneumonia, erysipelas or throat inflammation), Fibinger in 1898 (serum for diphtheria, with alternate assignment), and Ferguson, Davey and Topley in 1927 (vaccines for the common cold, with blinding of patients).

The main principles for design and execution of clinical trials are no different for neurology than for other disciplines. Excellent textbooks have been written to explain these even to the uninitiated(6,7). In this chapter we shall therefore not attempt to summarize and paraphrase the principles of clinical trials. Rather we wish to concentrate on a few special subjects. Our choice has been determined mostly by what we have learned through our own mistakes. We start with some issues that apply to clinical trials in general: pragmatic versus explanatory trials, role of sponsoring industries, methods of randomization, and double blind versus single blind design. The second and greatest part of the chapter is dedicated to measurement of outcome in neurology. This is a notorious minefield, perhaps especially for neurological disorders. It is not our aim to cover every field, but again we have made a selection: stroke, Parkinson's disease, multiple sclerosis, polyneuropathy, migraine, and epilepsy.