Performance of Common Clinician- and Self-Report Measures in Assessing the Function of Community-Dwelling People With Metastatic Breast Cancer

Cheville AL, Basford JR, Troxel AB, Kornblith AB. Performance of common clinician- and self-report measures in assessing the function of community-dwelling people with metastatic breast cancer.ObjectiveTo characterize the performance of common clinician- and self-report measures of function in assessing community-dwelling people with metastatic breast cancer.DesignCross-sectional study.SettingA tertiary medical center outpatient cancer clinic.ParticipantsA consecutive sample of community-dwelling patients (N=163) with stage IV breast cancer.InterventionsNot applicable.Main Outcome MeasuresMedical Outcomes Study 36-Item Short-Form Physical Functioning (PF-10) and Role Physical subscales; the Older Americans Resource Study (OARS) activities of daily living (ADL) and instrumental ADL subscales; Karnofsky Performance Scale (KPS); and the FIM Total and FIM Mobility scores.ResultsWith the exception of the PF-10 and Role Physical subscales, which demonstrated floor effects, ceiling effects were detected in all the measures and were particularly persistent in the OARS ADL subscale. Instrument and item score distributions varied markedly across KPS-defined subgroups with FIM Mobility, FIM Total, and OARS subscale score distributions deviating least from the normal in the lowest performing (KPS 40–50) participants. Correlations between self-reported (Role Physical subscales, PF-10, OARS ADL subscales) and the clinician-rated (KPS and FIM scales) scales were moderate to high (r=.55–.82); however, clinician-reported scores were more consistently associated with the presence of physical impairments.ConclusionsIn this population with stage IV breast cancer, ceiling effects limit the discriminatory capacity of the common functional scales assessed in this study. Instruments and items, particularly when ADL based, tend to perform better at lower levels of function (KPS 40–50) and less well at higher levels. Clinician-rated outcomes may have greater capacity to discriminate the presence of physical impairments.