Perceptions of medical status and treatment goal among older adults with advanced cancer

Background: Patient-reported medical status and treatment goal are measures of prognostic understanding with demonstrated relationships to important clinical and patient-reported outcomes in the general cancer population. Among older adults, relationships between these measures and other patient-reported (quality of life [QOL], symptoms, functional impairment) and clinical (hospitalization risk, survival) outcomes remains unclear.

Methods: We enrolled patients ≥70 with advanced gastrointestinal cancers, collecting patient-reported medical status (terminally ill vs not), treatment goal (curative vs non-curative), QOL (EORTC-Elderly Cancer Patients), symptoms (Edmonton Symptom Assessment System [ESAS]), and functional impairment (activities of daily living [ADLs]). We also obtained information about hospitalizations and survival. To explore relationships between patient-reported medical status, treatment goal, QOL, symptoms, functional impairment, hospitalizations, and survival, we used regression models adjusted for age, sex, and education.

Results: Of 103 patients, 49.5% reported terminally ill status and 64.0% a non-curative treatment goal. Terminally ill status was associated with worse QOL (EORTC illness burden: 53.59 vs 35.26, p = 0.001), higher symptom burden (ESAS: 28.15 vs 16.79, p = 0.002), more functional impairment (ADLs: 3.63 vs 5.24, p = 0.006), greater hospitalization risk (HR = 2.41, p = 0.020), and worse survival (HR = 1.93, p = 0.010). We did not find associations between patient-reported treatment goal and these outcomes.

Conclusions: In older adults with advanced cancer, report of terminally ill status was associated with other important patient-reported and clinical outcomes, suggesting disease severity may inform illness perceptions. We did not find similar associations for patient-reported treatment goal, indicating that questions related to medical status and treatment goal measure different constructs and more nuanced measures are needed.