Fatigue is a well-known, distressing side-effect of cancer treatment that persists for some cancer survivors. Existing fatigue measures may be problematic in that they may (1) not characterize psychological and physical aspects of fatigue and (2) suffer from ceiling or floor effects, particularly when used among cancer survivors. The purpose of this study was to analyze the psychometric properties of various fatigue measures in breast cancer survivors (BCS) including content validity, internal consistency reliability, and construct validity. Psychometric theory and the Common Sense Model guided the analysis.
The sample included 70 BCS (Mean age=51; SD=9) participating in a randomized clinical trial testing a medication to treat hot flashes. Participants were married (75%), employed (74%), Caucasian (88%), taking tamoxifen (47%) and 35 (SD=41; median 20) months post-treatment.
Fatigue measures included: 1) F_POMS-sf; 2) Piper Fatigue Scale); 3) Circumplex Octant 6 (low activation, negative affect), and 4) the MOS-SF36 Vitality subscale. Construct validation measures included: 1) CES-D; 2) STAI; 3) PANAS; 4) Ham-D; 5) MOS-SF36 subscales; 6) Marlowe-Crowne Social Desirability Scale (MCSDS) and 7) the PSQI.
Content validity: All fatigue measures address physical qualities and intensity of fatigue, however, only 2 addressed psychological qualities and only 1 addressed fatigue duration and distress. Reliability: Cronbach’s alphas were acceptable (.77 to .96). Construct validity: Fatigue measures were moderately to highly correlated (.52 to .80). Relationships among fatigue measures and with construct validation variables (discriminant, known groups) were as expected. However, measures suffered from ceiling or floor effects.
Findings suggest new measures of fatigue, such as the Cancer Treatment-Related Fatigue Representation Scale (CTRFRep) should (1) address all dimensions of fatigue and (2) be evaluated for ceiling/floor effects. Findings can guide the selection of fatigue measures in future studies of breast cancer survivors.
Supported by NINR/NIH grants #R01 NR05261; F31NR008834-01A1 and PHS grant #5T32NR07066.
Session #1207 - Oncology: & Quality of Life
The 29th Annual MNRS Research Conference (April 1-4, 2005)