Validation of a Four Item Fatigue Screen (FIFS) in Cancer-Related Fatigue (CRF) (708)

Abstract

1.Assess the Four Item Fatigue Scale relative to the Brief Fatigue Inventory.2.Be able to see the relationship of the Four Item Fatigue Scale and the ICD-10 criterion for fatigue. Questionnaires for CRF are long & burdensome. A FIFS questionnaire was tested for content validity and sensitivity with the Brief Fatigue Inventory (BFI) and ICD10 criteria. Cancer patients completed BFI & FIFS in random order. FIFS included 4 questions: 1. Do you have fatigue:never, sometimes, usually, or always 2. On a scale of 0–10 (0: no fatigue 10: severe fatigue) how would you rate fatigue now 3. What is the level of your fatigue now: none, mild, moderate, severe 4. Is your fatigue over the last day: worse, the same, better. ICD10 criteria were completed by investigators. 1 week later, both tools were completed in opposite order. 65 patients were needed for 80% power for Pearson's correlation of 0.7. Bonferroni corrections were made for multiple variables; P < 0.0035 was significant. 65 completed ICD-10, BFI & FIFS days 1 & 7. Spearman correlation was 0.7 (0.52–0.88) day 1, 0.73 (0.55 0.91) day 7;0.59 (0.38–0.80) for change over time. BFI score correlated with ICD10-1, 3, 7, 10B.BFI regression analysis on day 7 correlated with ICD10-1, ICD10-7 & ICD10-10 (P = 0.024, 0.026, 0.045 respectively), these disappear when adjusted for BFI on day 1 (P = 0.10, 0.99, 0.12 respectively). For changes in BFI from day 1 to 7, only ICD10-5 criteria was significant (P = 0.019) & drops out of the model when adjusted for day 1. For FIFS, ICD10-7 (P = 0.003) & ICD10-10 (P = 0.04) were significant & persisted after adjustment for baseline FIFS scores. Several ICD10 factors were heavily represented (A, 1, 3, 8, B, C). FIFS & BFI correlated well in time & modestly over time. FIFS is valid for CRF. Certain ICD-10 items better predict fatigue severity.