OBJECTIVE: This study evaluated the comparability of two 5-point symptom self-report rating scales: Intensity (from "not at all" to "very much") and Frequency (from "none of the time" to "all of the time"). Questions from the Functional Assessment of Chronic Illness Therapy (FACIT)-Fatigue 13-item scale was examined. METHODS: Data from 161 patients (60 cancer, 51 stroke, 50 HIV) were calibrated separately to fit an item response theory-based rating scale model (RSM). The RSM specifies intersection parameters (step thresholds) between two adjacent response categories and the item location parameter that reflects the probability that a problem will be endorsed. Along with patient fatigue scores ("measures"), the spread of the step thresholds and between-threshold ranges were examined. The item locations were also examined for differential item functioning. RESULTS: There was no mean raw score difference between intensity and frequency rating scales (37.2 vs. 36.4, p = n.s.). The high correlation (r = .86, p < .001) between the intensity versus frequency scores indicated their essential equivalence. However, frequency step thresholds covered more of the fatigue measurement continuum and were more equidistant, and therefore reduced floor and ceiling effects. SIGNIFICANCE OF RESULTS: These two scaling methods produce essentially equivalent fatigue estimates; it is difficult to justify assessing both. The frequency response scaling may be preferable in that it provides fuller coverage of the fatigue continuum, including slightly better differentiation of people with relatively little fatigue, and a small group of the most fatigued patients. Intensity response scaling offers slightly more precision among the patients with significant fatigue.
OBJECTIVE: This study evaluated the comparability of two 5-point symptom self-report rating scales: Intensity (from "not at all" to "very much") and Frequency (from "none of the time" to "all of the time"). Questions from the Functional Assessment of Chronic Illness Therapy (FACIT)-Fatigue 13-item scale was examined. METHODS: Data from 161 patients (60 cancer, 51 stroke, 50 HIV) were calibrated separately to fit an item response theory-based rating scale model (RSM). The RSM specifies intersection parameters (step thresholds) between two adjacent response categories and the item location parameter that reflects the probability that a problem will be endorsed. Along with patient fatigue scores ("measures"), the spread of the step thresholds and between-threshold ranges were examined. The item locations were also examined for differential item functioning. RESULTS: There was no mean raw score difference between intensity and frequency rating scales (37.2 vs. 36.4, p = n.s.). The high correlation (r = .86, p < .001) between the intensity versus frequency scores indicated their essential equivalence. However, frequency step thresholds covered more of the fatigue measurement continuum and were more equidistant, and therefore reduced floor and ceiling effects. SIGNIFICANCE OF RESULTS: These two scaling methods produce essentially equivalent fatigue estimates; it is difficult to justify assessing both. The frequency response scaling may be preferable in that it provides fuller coverage of the fatigue continuum, including slightly better differentiation of people with relatively little fatigue, and a small group of the most fatigued patients. Intensity response scaling offers slightly more precision among the patients with significant fatigue.
Authors: Neil B Alexander; George E Taffet; Frances McFarland Horne; Basil A Eldadah; Luigi Ferrucci; Susan Nayfield; Stephanie Studenski Journal: J Am Geriatr Soc Date: 2010-05 Impact factor: 5.562
Authors: Roy G Elbers; Marc B Rietberg; Erwin E H van Wegen; John Verhoef; Sharon F Kramer; Caroline B Terwee; Gert Kwakkel Journal: Qual Life Res Date: 2011-10-20 Impact factor: 4.147
Authors: Gale Harding; David Cella; Don Robinson; Parthiv J Mahadevia; Jason Clark; Dennis A Revicki Journal: Health Qual Life Outcomes Date: 2007-06-14 Impact factor: 3.186