| Description |
It is widely recognized that health- related quality of life (HRQL) is often a secondary research endpoint. Most HRQL instruments are norm referenced with prescribed indicators. The Patient Generated Index is an individualized patient reported outcome measure of HRQL. In contrast to many other measures, the PGI does not have standardized items, allowing for individuals to define their own quality of life. Evidence supports the PGI as a valid measure of HRQL; however, the PGI has seen little use in oncology populations and none in the radiation treatment setting. The purpose of this prospective study was to evaluate the psychometric properties of the PGI in an outpatient oncology patient population (N= 98) receiving radiation treatment using cognitive interview and repeated measure methods. Eighty-six people participated in the study. Feasibility of using the PGI was supported by a high study retention rate (95%), a low completion time (mean = 4 min) and low missing score rates (10%-17%). Test-retest reliability was also supported (r = .75, p < .001; paired ?=.14,/>=.89). Content validity was supported; however response errors due to high cognitive burden and survey formatting were of concern. PGI scores inversely correlated with National Community Cancer Network Distress levels (r= -.49, - 55, -.44; p<.001) and with role (r=.31, .40,,38; p< .01), emotional (r=.33, .41, .33;p< .01) and social functioning (r=.27, .49, .42; p< .05), pain (r=-.29, -.39, -.39; p<M) and fatigue subscales (r=-.35, -.25, -.47; p< .05) of the European Organization of Research and Treatment of Cancer Quality of Life Questionnaire, supporting convergent validity. PGI scores were sensitive to differences between participants reporting high or low levels of distress at each measurement time (t = 4.42, 3.32 and 2.9,/?<.05). The PGI detected a small to moderate change in HRQL over time (ES=.51 and .38 respectively) as measured by the PGI. Study data support the reliability and validity of the PGI as a measure of HRQL. Further studies using random sampling methods and more diverse study populations are needed. Studies using computerized or interview survey administration methods or further refinement of the PGI are warranted to reduce participant response errors. |
