Methodological approaches to estimate physical resilience in older adults: a comparison across two clinical settings

Various methodological approaches exist to estimate physical resilience in older adults (phenotype, expected recovery, adapted ageing), but their comparative performance remains unclear. This study evaluated their agreement and predictive performance across two clinical contexts.

We applied three methodological approaches in two cohorts: older adults receiving chemotherapy (TENT) and those presenting with acute illness at the emergency department (APOP). Physical functioning was assessed using Katz Activities of Daily Living and Lawton Instrumental Activities of Daily Living scales. The phenotype approach measured functional change after the health stressor; the expected recovery approach compared actual to predicted recovery based on clinical characteristics; and the adapted ageing approach assessed baseline function relative to expected levels given clinical risk profile. Outcomes at 12 months included mortality and a composite endpoint of functional decline, quality of life decline or mortality. Agreement was assessed using Cohen's kappa and predictive performance using area under the curve (AUC).

The TENT cohort included 330 patients (median age 75 years; 43% female) and APOP included 2111 patients (median age 78 years; 54% female). Agreement between approaches was poor to moderate (κ = -0.10 to 0.64). All approaches showed only moderate discriminative ability for 12-month mortality (AUC 0.55-0.69) and composite outcomes (AUC 0.52-0.66). The physical resilience approaches provided limited added discrimination beyond simple predictors: functional score at follow-up (AUC 0.65-0.69) and age (AUC 0.48-0.67).

Various physical resilience approaches identified different patient groups as resilient and demonstrated limited prognostic value. Current approaches may inadequately capture the dynamic construct of physical resilience; higher-frequency longitudinal measures may better quantify physical resilience for clinical practice.