A method has been developed for simultaneously comparing the usefulness of many treatments of established value for symptomatic medical conditions. Medical assessment of outcome is not employed. Instead patients are required to assess treatments prescribed during the course of ordinary general practice rather than under the strictly controlled settings of most clinical trials. Outcome incorporates patient compliance and treatment acceptability and is based on patients' subjective judgements of the usefulness of randomly allocated treatments as recorded in self-completed diaries, which are mailed directly to a trial centre. Thus large and more representative samples are achieved through minimizing the efforts required, both of participating doctors and of patients. Although the approach was originally developed and tested for the comparison of hay fever treatment regimens, we believe that it can be adapted to compare many other treatments where patient-reported symptoms validly describe the outcome of interest. The feasibility of the approach was tested in two pilot studies, and it has been employed successfully in a two-year trial comparing seven hay fever treatments. Aspects of analysing such trials are discussed.