Objectives: Researchers wishing to assess the health-related quality of life (HRQoL) of women with gynecologic cancers have a range of questionnaires to choose from. In general, disease-, treatment-, or symptom-specific questionnaires are assumed to be better able to identify between-group differences (sensitivity) and changes over time (responsiveness) than are cancer-specific or generic questionnaires. However, little work has tested this assumption in oncology. We set out to (a) identify all multidimensional HRQoL questionnaires used in studies with women with gynecologic cancer and (b) evaluate their track records in identifying minimal clinically important differences (MCIDs), with a view to making recommendations.

Methods: We searched MEDLINE using the term quality of life and each gynecologic cancer type, as well as the names of identified questionnaires. We used 10% of the scale range as the threshold for an MCID.

Results: We identified 1 generic (SF-36/SF-12), 3 cancer-specific (European Organisation for the Research and Treatment of Cancer Quality of Life Questionnaire [EORTC QLQ] C30, Functional Assessment of Cancer Therapy-General [FACT-G], and short-form Cancer Rehabilitation Evaluation System [CARES-SF]), and 1 disease-specific (QOL-Ovarian Cancer Patient Version) HRQoL questionnaires and 5 disease-specific (QLQ-OV28, FACT-O for ovarian, QLQ-CX24, FACT-Cx for cervical and FACT-V for vulvar), 1 treatment-specific (FACT and Gynecologic Oncology Group-Ntx for neurotoxicity), and 2 symptom-specific (FACT-Anemia and Functional Assessment of Chronic Illness and Therapy [FACIT]-Fatigue) modules. Twenty-seven articles reported results from 26 studies in which an MCID had been identified. The FACIT's anemia and fatigue subscales were more sensitive, and the neurotoxicity subscale more sensitive and responsive than the FACT-G on at least 1 comparison. However, we found no evidence for superior performance by the FACT-G compared with the SF-36 or EORTC and FACIT disease-specific modules versus the QLQ-C30 and FACT-G. There was also little evidence to favor EORTC versus FACIT questionnaires or vice versa.

Conclusions: The evidence we reviewed offered little support for the hypothesis that disease-, symptom-, or treatment-specific instruments are more sensitive and responsive than cancer-specific or generic questionnaires. However, conclusions were limited by the small number of head-to-head comparisons available. We summarize the clinical contexts in which each instrument identified an MCID to inform choice of questionnaire(s), sample size calculations, and interpretation of results in future studies.