Four Approaches for Determining Composite Scores for the Measurement of Transition in Cancer Scale

We created the Measurement of Transitions in Cancer Scale to assess patients' perceptions of the extent of change they experience with cancer-related transitions and how well they feel they are managing these transitions. For some transitions, we found that the more change that was reported, the worse management was reported; however, the benchmark by which patients assess how well they have managed may vary with the extent of change.

The aim of the study was to identify approaches to combine reports of extent and management of change.

Among women with breast cancer, we explored relationships of composite measures (arithmetic and geometric means, subtractive and proportional need for improvement) with other indicators of well-being (symptoms, anxiety, depression, uncertainty, self-efficacy, knowledge of care options, medical communication competence). We examined statistical significance using false rate discovery for multiple tests on correlations with clinical outcomes.

Greater extent and less management were significantly associated with higher total symptoms, anxiety, depression, uncertainty, and less self-efficacy in Personal Transitions, but not in Care Transitions. The arithmetic and geometric means had weak correlations with clinical outcomes, whereas the subtractive and proportional need for improvement had significant correlations with most clinical outcomes both in Personal and Care Transitions. The combined proportional need for improvement in Personal Transitions was significantly associated with total symptoms, anxiety, depression, uncertainty, and self-efficacy. The Care Transitions score was also significantly associated with total symptoms, anxiety, uncertainty, and self-efficacy.

These approaches can be applied to other aspects of self-management that require assessment of the extent and management of an experience. The four approaches yield different results. We recommend the need for improvement composites to capture correlations with the clinical outcomes.