PROMIS pediatric measures validated in a longitudinal study design in pediatric oncology.

Abstract

This study assessed the responsiveness to change over time and theorized associations of Patient-Reported Outcomes Measurement Information System (PROMIS) pediatric measures in children and adolescents in treatment for cancer to determine measure readiness for use in cancer clinical trials. We administered eight PROMIS (three symptom, two psychological, and three performance) pediatric short-form measures and the Symptom Distress Scale (SDS) to 96 pediatric oncology patients at three time points during a course of chemotherapy. We assessed responsiveness using paired t tests and generalized estimating equation (GEE) models, calculated standardized response mean (SRM) values for PROMIS measures, and examined scores over three data points (T1-T3). Guided by the theory of unpleasant symptoms (TOUS), we examined associations among the PROMIS measures, the SDS, and other variables using GEE. The paired t tests showed statistically significant changes in two psychological measures and one performance measure from T1 to T2; three symptom, two psychological and two performance measures from T2 to T3; and three symptom and two psychological measures from T1 to T3. Findings from GEE models indicate PROMIS pediatric measures had statistically significant short-term and long-term changes, controlling for demographic and clinical variables. One performance measure did not achieve significant change at any time point. We found positive support for theorized relationships in the TOUS. Most of the PROMIS pediatric measures demonstrated changes over time and had significant relationships as theorized, thus supporting concurrent and construct validity of these measures when administered to pediatric oncology patients during a course of chemotherapy. This evidence supports the measures' readiness for use in clinical trials.