Abstract
Background: Instruments commonly used in quality of life research may yield valid and reliable reflections of a person's state at a particular point in time. Yet, discrete items within those scales may be insensitive indicators for a specific population sample or intervention, thus creating what often appears to be chaotic data.
Objectives: To evaluate the relative sensitivity of each item in three commonly used instruments for measuring quality of life in breast cancer survivors.
Methods: Using data from 150 breast cancer survivors, absolute change scores were calculated for each of 98 items comprising the three scales. A Live Item Index Technique score was calculated by multiplying Magnitude and Breadth factors for each of the 98 items. The proportion of variance that each item contributed was calculated, and the items were ranked in terms of sensitivity.
Results: Forty-eight "live" items accounted for 62% of total change score variance. The most sensitive item was over 12 times more sensitive than the least sensitive item. On average, live items were 1.7 times more sensitive than less sensitive items.
Discussion: If appropriate to meet study objectives, in designing similar studies, one may consider removing the "dead" items that never change. With fewer items as end point measures, the study would become more efficient and a smaller sample size would be needed, thereby reducing research cost. Also, Live Item Index Technique can be used as an exploratory and developmental tool for constructing more effective interventions.