Four screening instruments for frailty in older patients with and without cancer: a diagnostic study

Background: Frailty in older patients might influence treatment decisions. Frailty can be determined using a Comprehensive Geriatric Assessment (CGA), but this is time-consuming and expensive. Therefore we assessed the diagnostic value of four shorter screening instruments. Methods: We tested the abbreviated CGA (aCGA), the Vulnerable Elders Survey-13 (VES-13), the Groningen Frailty Indicator (GFI) and the Geriatric 8 (G8). A full CGA including functional status, cognitive status, depression, nutrition and comorbidity was used as reference. A minimum of 85% for both sensitivity and specificity was predefined as acceptable. Data were collected through personal interviews by trained interviewers. We assessed people aged >= 70 years: 108 patients with recently diagnosed cancer recruited in hospitals and 290 without cancer recruited by general practitioners in the Netherlands and Belgium. Frailty was defined as having impairment in at least two domains of the full CGA. We used original cut-offs for the screening instruments and calculated sensitivity, specificity, positive and negative diagnostic values and the percentage classified as frail. Results: Sensitivity of aCGA was 79% and 87% for patients with and without cancer; specificity was 59% and 64%. Sensitivity of VES-13 was 67% and 82% for patients with and without cancer; specificity was 70% and 79%. Sensitivity for GFI was 76% (in both groups) and specificity 73% (in both groups). Sensitivity for G8 was 87% and 75% for patients with and without cancer; specificity was 68% (in both groups). Conclusions: No screening instrument was acceptable according to our predefined minimum of 85% for both sensitivity and specificity. The diagnostic value of the investigated instruments is rather poor and one could wonder about their additional value to clinical judgment