The York Research Database

Quality of life is increasingly being used as a measure of outcome in comparative cancer clinical trials. Even when quality of life studies are well funded vis-a-vis data collection and processing, it is unlikely that data sets will be 100% complete, leading to problems at the data analysis stage. Missing data can arise as a consequence of whole assessments being missed due to death or other reasons, or as a result of single items being omitted from particular questionnaires. This paper addresses the latter category of missing data and examines the effects on subscale scores of imputing values for missing items from the Hospital Anxiety and Depression Scale and the psychological complaints subscale of the Rotterdam Symptom Checklist.