Statistical Ramifications of Linear Analogue Scales in Assessing the Quality of Life of Cancer Patients

Abstract

In evaluating cancer treatments, it is important to know both about the quality and duration of the patient’s survival. This study was initiated to investigate the properties of scales that could be used to assess quality of life. 115 breast cancer patients from the Princess Margaret Hospital in Toronto were asked to evaluate their quality of life using linear analogue scales. Thirty-one variables covered the areas of general health and major problems associated with breast cancer derived from clinical experience and the opinions of patients with the disease. Patients found the linear analogues simple to use, and the broad areas considered by each scale made the mental effort much less than what would be involved with extensive evaluative inventories.

Our general health items were compared with the Sickness Impact Profile from which they were derived to assess the validity of the method. Our patient population produced reliable test-retest results that were stable for both factor analysis and multivariate regressions. Also, we can report the results of some sample size investigations for both reliability and stability of data collected in this and other studies.