Effect of comprehensive geriatric assessment on hospitalizations in older adults with frailty initiating curatively intended oncologic treatment:The PROGNOSIS-RCT study

Introduction: Frailty constitutes a risk for unplanned hospitalizations in older adults with cancer. This study examines whether comprehensive geriatric assessment (CGA) as an add-on to standard oncologic care can prevent unplanned hospitalizations in older adults with frailty and cancer who initiate curative oncological treatment. Materials and Methods: This randomized controlled trial included older adults aged ≥70 with frailty (Geriatric 8 [G8] ≤14), and solid cancers who initiated curative oncological treatment. Participants were randomized 1:1 to either standard oncologic care (control) or standard oncologic care supplemented with CGA-guided interventions (intervention). Baseline characteristics were retrieved prior to randomization. The primary endpoint, the between-group rate ratio of unplanned hospitalizations within six months of treatment initiation, was analyzed using negative binominal regression. Analyses were performed using an intention-to-treat approach, followed by per-protocol analysis, including participants receiving CGA within 30 days of randomization, and preplanned subgroup analyses based on treatment modality and Geriatric 8 screening. Secondary endpoints included acute hospital contacts, treatment adherence, and toxicity. Results: From November 1, 2020 to May 31, 2023, 173 participants were enrolled. Median age was 75 (interquartile range 72–79), 51.5% were female, 58% had a G8 score > 12, and 84% had Eastern Cooperative Oncology Group performance status 0–1. The most common cancer sites were lung (23%), upper gastrointestinal (15%), and breast (13%). The rate (per person-years) of unplanned hospitalization was 1.32 in the intervention group and 1.81 in the control group, with a between-group rate ratio of 0.74 (95% confidence interval [CI] 0.45–1.23, P = 0.25) favoring the intervention. The between-group rate ratio increased in the per-protocol analysis (0.64 [95% CI 0.37–1.10, P = 0.10]). Similarly, no significant between group differences were found in treatment adherence, rate of ...