THE PROBLEM OF SAMPLE ATTRITION IN CONTROLLED TREATMENT TRIALS

Abstract

It is virtually impossible to prevent sample attrition during the course of a controlled treatment evaluation study conducted in a hospital. Possible consequences of treatment-related and non-treatment-related dropouts on results have been briefly presented and six approaches or partial solutions to the sample attrition problem have been discussed. Sample attrition can be kept to a minimum through the use of well designed experiments based on adequate information regarding the major components of the experiment; subjects, treatments, setting and criterion measures.