M. El Moheb1, C. Shen1, E. Rabinovich1, S. Kim1, C. Tsung1, K. Cummins1, R.G. Witt1, D.R. Brenin1, S. Showalter1, A. Tsung1 1University Of Virginia, Department Of Surgery, Charlottesville, VA, USA
Introduction: Health-related quality of life (HRQOL) is the perceived physical and mental health of an individual and is an important factor in the care of cancer patients. Evidence comparing individual patients' HRQOL pre- and post-diagnosis is limited, and few studies have identified predictors of HRQOL changes. Our study aims to identify predictors of change in HRQOL across two key transitions: (1) pre- to post-diagnosis, and (2) at diagnosis versus two years later in elderly patients with breast cancer, with the goal of identifying at-risk patients for a significant decline in HRQOL.
Methods: This study analyzed SEER-MHOS data, which links cancer data with HRQOL data collected as part of the Medicare Health Outcomes Survey (MHOS) program. Female breast cancer patients diagnosed between 1998 and 2019, >65 years old, with no prior cancer history, and at least two HRQOL surveys (SF-36/VR-12) were included. Two cohorts were analyzed: patients with surveys pre- and post- diagnosis (cohort 1), and those with surveys at diagnosis and two years post-diagnosis (cohort 2). Bayesian linear regression identified predictors of physical (PCS) and mental (MCS) score changes, incorporating demographic, socioeconomic, comorbidity, and cancer-related variables.
Results: Cohort 1 included 1,552 patients (mean [SD] age: 73 [7.1]), and cohort 2 included 891 patients (age: 72 [7.5]). In cohort 1, mean PCS and MCS scores decreased by 2.5 and 1.1 points at diagnosis from baseline. In cohort 2, PCS decreased by 0.2 points, while MCS increased by 1.2 points at two years post-diagnosis. In cohort 1, advanced stage (?PCS: -2.1 to -6.5, ?MCS: -2.0 to -5.1) and baseline limitations in activities of daily living (ADL) (?PCS: -2.0 to -2.6, ?MCS: -2.0) were most predictive of HRQOL decline, while high education (?MCS: +2.6 to +3.2) and positive baseline health perception were protective (?PCS: +1.4 to +3.4) (probability of direction (PD) of all estimates >90%) (Figure). In cohort 2, baseline ADL limitations (?PCS: -1.6 to -5.5, ? MCS: -2.5) and poor health perception (?PCS: -3.2) predicted HRQOL decline, while excellent health perception (?MCS: +4.1), and high income (?MCS: +2.9) were protective (all PD >90%). Surgery, reconstruction, and radiation did not impact HRQOL in either cohort.
Conclusion: In elderly breast cancer patients, advanced stage and functional limitations predict decline in HRQOL, while positive health perception and higher education are protective. These factors remain consistent from pre- to post-diagnosis, and through the two-year follow-up period, regardless of treatment type. This study identifies patients at-risk for HRQOL decline allowing for strategic implementation of supportive care plans to improve cancer outcomes.