Methodological aspects of clinical trials in non-ulcer dyspepsia with special reference to selectional factors.

The sampling process preceding a clinical trial is associated with numerous fallacies. This is especially true in non-ulcer dyspepsia, which is a prevalent but poorly defined condition, probably with heterogenic etiologies. Therefore, in order to make subsequent inferences possible, it is crucial that rigorous criteria for adherence to the target population--and the studied sample--are established in advance. Several factors, however, preclude a strictly random sample from the target population. Firstly, only those who find their way to medical services are included in trials, and different circumstances may affect the patient's decision to seek medical advice. Secondly, the doctor's decision to refer a patient to a study is influenced by a number of factors that might introduce bias. As an example, an on-going study of non-ulcer dyspepsia is presented. With the application of strict symptomatic criteria, and the exclusion of circumscribed organic lesions of the upper gastrointestinal tract by means of endoscopy, it was possible to define a population in which very few additional pathological findings were made at thorough diagnostic work-up.