Psychometric Evaluation of Two Scales Assessing Functional Status and Peripheral Neuropathy Associated With Chemotherapy for Ovarian Cancer: A Gynecologic Oncology Group Study

Abstract

To evaluate the psychometric properties of two adapted scales, one for functional status and one for peripheral neuropathy secondary to neurotoxic chemotherapy. Repeated measures methodologic design conducted within a Gynecologic Oncology Group (GOG) phase III clinical trial that randomly assigned patients with advanced epithelial ovarian cancer to cisplatin and cyclophosphamide or cisplatin and paclitaxel. 8 GOG institutions participating in the GOG clinical trial. 88 evaluable outpatients enrolled in the GOG clinical trial. Sample size at time 1 (T1) was 88 patients and at time 2 (T2) was 67 patients. All scales were administered at T1 (prior to initiation of chemotherapy) and T2 (after six cycles of chemotherapy but prior to second-look laparotomy). Internal consistency reliability, criterion validity, and construct validity were evaluated, and clinical application was explored. Self-reported peripheral neuropathy and functional status (comprised of physical function and role function subscales), the GOG performance status scale, and the GOG toxicity criteria. Reliability coefficients at T1 were physical function = 0.83, role function = 0.96, and peripheral neuropathy = 0.91; at T2, they were physical function = 0.83, role function = 0.92, and peripheral neuropathy = 0.89. At T1, physical function and role function correlated positively with performance status. Peripheral neuropathy correlated positively with GOG toxicity criteria used at T2. Principal component factor analysis suggested that the functional status scale had a two-factor structure with factors representing general and specific mobility and that the peripheral neuropathy scale also had a two-factor structure with factors representing foot and hand neuropathy. The physical function, role function, and peripheral neuropathy scales have internal consistency, reliability, criterion validity, and construct validity. However, revision of the scales should address modification of specific questions and consider increasing the Likert scale from a four-point to a five- or seven-point scale to enhance clinical sensitivity and application. With minor modifications, these scales should be useful in assessing physical function, role function, and peripheral neuropathy in patients who receive agents that may cause peripheral neuropathy.