External validation and diagnostic value of the Elderly Functional Index version 2.0 for assessing functional status and frailty in older Danish patients with gastrointestinal cancer receiving chemotherapy: A prospective, clinical study.

ABSTRACT

INTRODUCTION: Patient perspectives on functioning are often overlooked in oncology practice. This study externally validates the ELderly Functional Index (ELFI), a patient-reported measure for assessing multidimensional functioning, in older patients with gastrointestinal cancer receiving chemotherapy. The study compares ELFI scoring methods, evaluates its diagnostic value with geriatric oncology tools, and proposes a cut-off point for clinical use. MATERIALS AND METHODS: Danish patients aged ≥70 years with gastrointestinal cancer undergoing chemotherapy from a prospective, observational study were included. Two ELFI scoring methods, item-based and domain-based, were compared. Internal consistency reliability, validity, and correlations between ELFI, its component scales, and measures of functioning/frailty (including Eastern Cooperative Oncology Group Performance Status [ECOG-PS], Geriatric-8 [G8], Vulnerable Elders Survey-13 [VES-13], Timed-Up-and-Go [TUG], and 30-s chair stand test [30CST]) were investigated. Sensitivity and specificity analyses evaluated the ability of ELFI to predict frailty outcomes and identified frailty thresholds. Receiver operating characteristic analyses assessed the diagnostic ability of ELFI, alongside other measures, for oncological outcomes and frailty differentiation. Equipercentile equating methods enabled ECOG-PS, ELFI, and G8 mapping. RESULTS: One hundred fifty-four patients (median age 73.5 years, range 70-85) undergoing curative- or palliative-intent chemotherapy (49%) were included. ELFI demonstrated good internal consistency (Cronbach's alpha = 0.82) and acceptable convergent, structural, and discriminant validity. ELFI showed moderate to very strong correlations with its component scales (r = 0.40-0.93), and weaker correlations with frailty measures (r = 0.02-0.60). ELFI score < 80 indicated frailty risk, with almost fivefold risk of ECOG-PS 2 at follow-up (odds ratio[OR] = 4.8, 95% confidence interval [CI] 1.4-15.9), and predicted G8, VES-13, TUG, and 30CST frailty at follow-up, not completing planned chemotherapy (OR = 3.1; 95%CI 1.5-6.2), mono-therapy (OR = 3.5; 95%CI 1.5-8.1), initial dose reduction (OR = 4.9; 95%CI 2.0-12.1), and shorter overall survival (hazard ratio = 2.0, 95%CI 1.4-3.0). A preliminary crosswalk between ECOG-PS, ELFI, and G8 was established. DISCUSSION: ELFI was validated as a concise patient-reported measure of functional status in older patients with cancer and its relationship to frailty. ELFI demonstrated comparable predictive ability to other tools for oncological outcomes. Both scoring methods yielded similar results, with the domain-based method (ELFI v2.0) endorsed for consistency. ELFI v2.0 score of 80 was suggested as the frailty threshold in this population, supporting its clinical utility.