Development and psychometric properties of the PROMIS® pediatric fatigue item banks

Purpose: This paper reports on the development and psychometric properties of self-reported pediatric fatigue item banks as part of the Patient-Reported Outcomes Measurement Information System (PROMIS). Methods: Candidate items were developed by using PROMIS qualitative methodology. The resulting 39 items (25 tiredness related and 14 energy related) were field tested in a sample that included 3,048 participants aged 8-17 years. We used confirmatory factor analysis (CFA) to evaluate dimensionality and differential item functioning (DIF) analysis to evaluate parameter stability between genders and by age; we examined residual correlations to evaluate local dependence (LD) among items and estimated the parameters of item response theory (IRT) models. Results: Of 3,048 participants, 48 % were males, 60 % were white, and 23 % had at least one chronic condition. CFA results suggest two moderately correlated factors. Two items were removed due to high LD, and three due to gender-based DIF. Two item banks were calibrated separately using IRT: Tired and (Lack of) Energy, which consisted of 23 and 11 items, respectively; 10- and 8-item short-forms were created. Conclusion: The PROMIS assessment of self-reported fatigue in pediatrics includes two item banks: Tired and (Lack of) Energy. Both demonstrated satisfactory psychometric properties and can be used for research settings.