1013 - Concordance of Perceptions of Cancer Prognosis in Patients and Radiation Oncologists
Presenter(s)
S. Joshi1,2, A. Tam1, P. Tang1, R. Garg1, J. A. Novak3, M. Loscalzo4, Y. R. Li1, A. Amini1, and C. J. Ladbury1; 1Department of Radiation Oncology, City of Hope National Medical Center, Duarte, CA, 2California University of Science and Medicine, Colton, CA, 3Department of Radiation Oncology, Rogel Cancer Center University of Michigan Health, Ann Arbor, MI, 4Department of Supportive Care Medicine, City of Hope National Medical Center, Duarte, CA
Purpose/Objective(s): Prognosis, which describes the expected course of disease including the severity and likelihood of curative treatment, is a key component of oncologic care. Previous studies have shown that discussions surrounding prognosis influence patients’ treatment preferences and expectations from treatment. Limited research has been performed on how the patient’s understanding of their prognosis aligns with those of radiation oncologists (RO). In this study, we aimed to describe concordance between patients and ROs, and among different ROs, hypothesizing that there is a low concordance rate (with a kappa coefficient of <0.4) in the perception of prognosis among patients and radiation oncologists assessing based on case summaries.
Materials/Methods: Between September 2020 and June 2022, we conducted a survey study on patient expectations from cancer treatments. Patients who were undergoing radiation treatment at our department were asked to assess their disease prognosis using a Likert scale, ranging from “cure is very likely, in the range of 76-100%” to “cure is not at all likely, in the range of 0-25%”. After survey collection, we reviewed each patient’s medical chart and generated case summaries for them. Three ROs reviewed these case summaries and gave their rating of the patients’ prognosis using the same Likert scale from the survey study, while blinded to the patient’s rating. Concordance rates and kappa coefficients (?) were summarized. Statistical analyses were performed using STATA.
Results: A total of 285 patients were surveyed (51.2% female and mean age of 57.9 years). The most common cancer diagnoses were breast, hematologic, and lung. The overall concordance (patient vs. 3 ROs) was 22.8% (? =0.34). The concordance rate improved to 42.8% when patient and two of three ROs were compared (patient vs. 2 of 3 ROs). The concordance rates of patients with each RO ranged from 32.3% to 45.6% (? ranged 0.13 to 0.23). The concordance rate among the ROs alone was 48.4% (? =0.50) and 93.7% (? ranged 0.42 to 0.68) among all three ROs and two of three ROs, respectively. When the Likert scores were averaged, patients overestimated their prognosis by a median of 0.33 points (IQR 0-1.33) when compared to RO.
Conclusion: In this study, we found low concordance of prognosis among patients and ROs. These findings may suggest there are currently gaps in our communication, and more importantly, expectations from treatment. The results also suggest that part of the discordance may stem from patients considering their prognosis was better than estimated by the ROs. Future studies are needed to better characterize these perceptions of both patients and ROs to facilitate shared decision making and improve patient satisfaction.