Objective: The aim of this study was to determine the optimal patient-reported outcome measure (PROM) for assessing symptom benefit in trials of palliative chemotherapy for women with symptomatic ovarian cancer.
Methods: Candidate PROMs were EORTC QLQ-C30 plus ovarian-specific QLQ-OV28, Functional Assessment of Cancer Therapy-Ovarian (FACT-O), FACT Ovarian Symptom Index (FOSI), and gynecologic cancer-specific Symptom Representation Questionnaire. Predefined optimality criteria were inclusion of all symptoms necessary for the specified purpose, recall period covering typical length of palliative chemotherapy, numerical item rating scales, and all necessary symptoms included in a single symptom index. Qualitative and quantitative methods were applied to data from stage 1 of the Gynecologic Cancer Intergroup Symptom Benefit Study to determine the set of necessary symptoms and to objectively assess candidate PROMs against the optimality criteria.
Results: Ten necessary symptoms were identified: pain, fatigue, abdominal bloating/discomfort, sleep disturbance, bowel disturbance, nausea and vomiting, shortness of breath, poor appetite, urinary symptoms, and weight changes. Although QLQ-C30 and QLQ-OV28 together cover all these symptoms, they split them into numerous scales, dissipating potential symptom-benefit signal. Conversely, FACT-O does not cover all necessary symptoms and contains many other HRQoL-related items and treatment side effects, diluting potential symptom-benefit signal when summed into scales. Item response scales and composite scoring of all candidate PROMs were suboptimal to our specific purpose. We therefore developed a new PROM, the Measure of Ovarian Symptoms and Treatment (MOST) concerns, to provide optimal measurement for the specified purpose.
Conclusions: This article documents the development of the MOST, a new PROM designed to assess patient-reported benefits and burden as end points in clinical trials of palliative chemotherapy for women with symptomatic ovarian cancer. The validity, reliability, and statistical efficiency of the MOST, relative to the best candidate scales of existing PROMs, will be assessed in the stage 2 of Gynecologic Cancer Intergroup Symptom Benefit Study.