Associations of Patient- and Physician-Related Factors With the Optimism of Prognostic Prediction in Patients With Terminal Cancer.

Abstract

BackgroundAccurate prediction of one-month survival is important for end-of-life care planning. However, clinical prediction of survival tend to be optimistic, and the reasons for this, particularly physician-related factors, remain unclear.ObjectiveThis study aims to identify patient- and physician-related factors associated with optimistic prediction of one-month survival.MethodsWe investigated terminally ill cancer patients referred to the palliative care unit at the Medical Center between July 2023 and June 2025 who died within one month after referral. The patients' treating physicians completed a survey regarding the rationale for their prognostic predictions. Patients were classified into an accurate group if predicted survival was within one month and an optimistic group if predicted survival exceeded one month. Physician-reported rationale was categorized into five domains: Tumor Progression, Symptoms, General Condition/Functional Status, Biologic Factors, and Prognostic Prediction Models. We analyzed associations between patient- and physician-related factors and optimistic prognostic predictions.ResultsAmong the 173 patients, 100 (58%) had optimistic prognostic predictions. General Condition/Functional Status and Tumor Progression were the most frequently cited rationale domains, whereas Prognostic Prediction Models were least frequently cited. Oral intake of more than a few mouthfuls, intake of only a few mouthfuls or less, resting dyspnea, and edema were associated with a lower risk of optimistic prognostic prediction (Odds ratio 0.038, 0.002, 0.028, and 0.022, respectively). No physician-related factors were associated with optimism.ConclusionsPrognostic accuracy was associated with patient-related factors but not physician-related factors. Future studies should evaluate strategies to reduce prognostic optimism in routine clinical practice.