The patient reported outcomes measurement information system—Cancer (PROMIS-Ca): Cancer-specific application of a generic fatigue measure

6537 Background: PROMIS-Ca is a cancer-specific extension of an NIH Roadmap effort to standardize measures of self-reported symptoms (pain; fatigue; anxiety; depression), physical function, and social function (http://www.nihpromis.org). PROMIS provides item banks that support brief-yet-precise computerized adaptive testing and multiple short form options. PROMIS-Ca extends PROMIS to oncology by ensuring the cancer-relevance and performance of its generic item banks. To illustrate, we customized the PROMIS Fatigue item bank to cancer and developed a link (“cross-walk”) between the generic PROMIS-Fatigue and the cancer-specific PROMIS-Ca Fatigue banks. Methods: We first calibrated a 95-item PROMIS Fatigue bank on a sample of 803 people representative of the US general population, with norms set on a T distribution (M=50; SD=10). The 72-item PROMIS-Ca Fatigue bank contains 56 calibrated items from the generic PROMIS Fatigue bank. Its psychometric properties were evaluated on a second sample of 513 cancer patients (72% female, 81% White, mean age=56, mean time since diagnosis=4.1 years). Analyses included item scalability, dimensionality, and IRT model fit. Using general population calibration data, we examined differential item functioning (DIF) between PROMIS-Ca Fatigue and PROMIS Fatigue banks. A cross- walk between these banks was established by equating item parameters using non-DIF items via the Stocking-Lord method. Results: All pre-set analysis criteria were met by 54 items; however, 6 showed DIF between PROMIS and PROMIS-Ca. Stocking-Lord results (using 48 non-DIF items) showed that the cancer sample was slightly more homogeneous and reported more severe fatigue (T-score=53.3) compared to norms. Thus, a fatigue continuum was built by equating item parameters from PROMIS and PROMIS-Ca Fatigue banks. Conclusions: Comparable item parameters between PROMIS and PROMIS-Ca Fatigue banks provide cross-walk information that enables one to measure fatigue with questions relevant to oncology, and compare their fatigue results to the general US population. Further research in this area will enable comparisons across multiple diseases. No significant financial relationships to disclose.