OBJECTIVES: This study evaluated the psychometric properties of the FACIT-Fatigue patient-reported outcome (PRO) instrument (Yellen et al., 1997) in patients with iron deficiency anemia (IDA). METHODS: As part of a randomized, controlled clinical trial IDA patients (n=808) completed the FACIT-Fatigue, Linear Analogue Scale Assessments (LASAs), and Medical Outcomes Study Short-Form 36 (SF-36) PRO instruments at five weekly visits. Baseline and Week 3 data, including hemoglobin, were used to validate the FACIT-Fatigue scale in this IDA patient population. RESULTS: The FACIT-Fatigue showed little to no ceiling or floor effects and produced reasonable response variability across items. Further, test-retest reliability (ICC=.87) and internal consistency reliability (Baseline α=.93) were both acceptable. Hypothesized convergent validity associations between other PROs and the fatigue scale were largely supported, although the hemoglobin-fatigue association was not as strong as expected (Baseline r =0.04; ns, Week 3 r=0.15; p<0.001). As predicted, known-groups validity showed that patients receiving intravenous iron treatment with ferumoxytol reported less fatigue than their placebo-treated counterparts (p<0.001). Further, a series of t-tests showed that patients with severely low hemoglobin levels reported significantly greater fatigue than those with mild, moderate, and severe values (ps<0.05).  Finally, the FACIT-Fatigue demonstrated ability to detect change using the SF-36 Vitality domain as a PRO anchor. Specifically, patients with worsened vitality reported worsened fatigue (ns), whereas patients grouped according to much, moderate, and minimal vitality improvements reported fatigue improvements (all ps<0.001; effect sizes increasing across groups accordingly). CONCLUSIONS: This study demonstrated that the FACIT-Fatigue provides a psychometrically valid assessment of fatigue in patients with IDA. Specifically, the scale was reproducible and internally consistent in an IDA population and exhibited construct validity in terms of predicted convergent/ discriminant correlations and across known groups. Finally, the FACIT-Fatigue was able to detect change when change occurred in a PRO anchor.