BACKGROUND: The diagnostic performance of tools used to screen vulnerability in older cancer patients varies widely. We assessed the diagnostic performance of gait speed (GS) for assessing vulnerability in such patients.
METHODS: All consecutive outpatients 65 years and older were referred for geriatric oncology assessment (GA) before a therapeutic decision between November 2013 and April 2016 in a bicentric observational and prospective cohort study. Vulnerability was defined as impaired score on at least one of the 6 domains of the GA. GS and the G8 index and G8 modified index were assessed at the first geriatric oncology visit during the GA. Sensitivity, specificity, positive and negative predictive value and positive and negative likelihood ratio were estimated. The accuracy of the three tools was analysed by the area under the receiver operating characteristic curve (AUC).
RESULTS: Among 269 included patients (mean [SD] age, 81.3 years [5.9]; 55% women, 94.4% solid tumors; 39.4% with metastasis), 252 (93.7%) had impaired GA. With the GS threshold of 1 m/s, sensitivity was 79.4% (95% CI, 73.8-84.2), specificity 64.7% (38.3-85.8), and AUC 82.0 (74.0-90.0). The corresponding values for the G8 index were 90.1% (85.7-93.5), 35.3% (14.2-61.7), and 79.0 (70.0-88.0) and G8 modified index were 89.3% (84.8-92.8), 64.7% (38.3-85.8), and 84.0 (74.0-92.0).
CONCLUSIONS: GS < 1 m/s with a single measure could be used as a new screening tool for detecting vulnerability in older cancer outpatients. This first external validation of the G8 modified index was very good.