Modeling of quality of life in older adults with advanced cancer: Data from a URCC NCORP multisite trial.

Abstract

218 Background: While the importance of symptom assessment is well recognized in Palliative Care (PC), limited evidence is available on the growing population of older adults with advanced cancer. Increasingly geriatric assessment (GA), which incorporates validated tools to assess health status, is advocated as a component of the care for older patients with cancer, but is not routinely incorporated into PC evaluations. We tested the hypothesis that variability in QoL for older adults with advanced cancer would be predicted by both symptoms and GA. Methods: Data from an ongoing cluster RCT conducted at 68 oncology practices were analyzed. Inclusion criteria were: > 70 years old, advanced solid tumor diagnosis, impairment in at least 1 GA domain (e.g., function). Multiple reliable and valid objective assessments and self-report measures for each GA domain were completed. Symptoms were assessed by MD Anderson Symptom Inventory (MDASI). Hierarchical regression modeling was conducted. The dependent variable was FACT score at Time 2 (4-6 weeks). The independent (Time 1) variables were entered in 3 steps: 1) gender, race, educational level( < or > high school), cancer type ( GI / Lung or other), receiving chemotherapy; 2) MDASI, 3) GA measures: Polypharmacy, Blessed Orientation and Memory Concentration Test, % weight loss, Short Physical Performance Battery, Instrumental Activities of Daily Living (IADL), Older Adult Resource Survey (OARS) Comorbidity, OARS Medical Social Support, and Generalized Anxiety Disorder-7. Results: N = 342 at time 1 (Mean age 77 years, 43% female, 90% white, 51% high school graduates, 49% GI or Lung cancer, 68% on chemotherapy). N = 303 at time 2(11% attrition due to death or withdrawal). Overall, 46% of variance in QoL was explained (demographic and disease characteristics: 4% [p = .05], symptoms: 32% [p < .0001], GA measures: 10% [p < .0001]). Significant individual GA predictors of worse QoL were lower social support, higher anxiety, impairment in IADLs, and poorer physical performance. Conclusions: Findings suggest that GA explains variance in quality of life scores for older adults with advanced cancer and reinforce the importance of symptom assessment and management in this population. Clinical trial information: NCT02107443.