Responsiveness of a Single-Item Indicator versus a Multi-Item Scale: Assessment of Emotional Well-Being in an International Adjuvant Breast Cancer Trial

• Published in: Medical care 1996-03, Vol.34 (3), p.234-248
• Main Authors: Hürny, Christoph, Bernhard, Jürg, Coates, Alan, Peterson, Harriet F., Castiglione-Gertsch, Monica, Gelber, Richard D., Rudenstam, Carl-Magnus, Collins, John, Lindtner, Jurij, Goldhirsch, Aron, Senn, Hans-Jörg
• Format: Article
• Language: English
• Subjects: Adjectives; Adult; Breast cancer; Breast Neoplasms - drug therapy; Breast Neoplasms - psychology; Chemotherapy; Chemotherapy, Adjuvant - psychology; Clinical trials; Data Interpretation, Statistical; Emotional expression; Emotional states; Female; Humans; Middle Aged; Neoplasm Recurrence, Local - psychology; Postmenopause; Psychological Tests; Quality of Life; Self-Assessment; T tests; Treatment Outcome; Wellbeing
• ISSN: 0025-7079; 1537-1948
• DOI: 10.1097/00005650-199603000-00004

A single-item linear analogue self-assessment scale for mood was compared with a 28-item adjective checklist for emotional well-being. To confirm its concurrent validity and responsiveness to treatment and recurrence in patients with breast cancer, emotional well-being was assessed every 3 months for 2 years and at 1 and 6 months after recurrence in 1,169 patients who were premenopausal and 960 patients who were postmenopausal. These patients were enrolled in two International Breast Cancer Study Group randomized clinical trials in operable breast cancer conducted from 1986 to 1993. To assess concurrent validity, Pearson's correlations between the linear analogue self-assessment scale and the adjective checklist were calculated for each time-point within each treatment group and for the two assessments after recurrence. Responsiveness to treatment and recurrence were analyzed using paired t tests and the squared ratio of these t tests, an estimate of relative efficiency. Concurrent validity of the mood linear analogue self-assessment was consistently confirmed across four language groups. Both measures were responsive; out of 24 changes over time, 19 were in the expected direction for the linear analogue self-assessment scale (P ≤ 0.05 for 9 of 19) and 17 for the adjective checklist (P ≤ 0.05 for 10 of 17). The linear analogue self-assessment scale was less but sufficiently efficient for detection of treatment effects, with relative efficiency estimates ranging from 0.16 to 2.45 and a median of 0.66 among the comparisons with relatively stable estimates (|t| ≥ 1.0) and more efficient for recurrence than the adjective checklist. The mood linear analogue self-assessment scale is a valid indicator of emotional well-being in patients with breast cancer in large multicenter, multicultural trials in which comprehensive scales are less feasible. This investigation supports the clinical relevance of linear analogue self-assessment scales as indicators of components of quality of life in cancer clinical trials.