Prevalence and Predictors of Physician-Patient Discordance in Prognostic Perceptions in Advanced Cancer

BACKGROUND: Discordance between physicians’ and patients’ prognostic perceptions in advanced cancer care threatens informed medical decision-making and end-of-life preparation, yet this phenomenon is poorly understood. We sought to: (1) describe the extent and direction of prognostic discordance, patients’ prognostic information preferences in cases of prognostic discordance, and physicians’ awareness of prognostic discordance; and (2) examine which patient, physician, and caregiver factors predict prognostic discordance. MATERIALS AND METHODS: Oncologists and advanced cancer patients (median survival ≤12 months; n = 515) from 7 Dutch hospitals completed structured surveys in a cross-sectional study. Prognostic discordance was operationalized by comparing physicians’ and patients’ perceptions of the likelihood of cure, 2-year mortality risk, and 1-year mortality risk. RESULTS: Prognostic discordance occurred in 20% (likelihood of cure), 24%, and 35% (2-year and 1-year mortality risk) of physician-patient dyads, most often involving patients with more optimistic perceptions than their physician. Among patients demonstrating prognostic discordance, the proportion who preferred not knowing prognosis varied from 7% (likelihood of cure) to 37% (1-year mortality risk), and 45% (2-year mortality risk). Agreement between physician-perceived and observed prognostic discordance or concordance was poor (kappa = 0.186). Prognostic discordance was associated with several patient factors (stronger fighting spirit, self-reported absence of prognostic discussions, an information source other than the healthcare provider), and greater physician-reported uncertainty about prognosis. CONCLUSION: Up to one-third of the patients perceive prognosis discordantly from their physician, among whom a substantial proportion prefers not knowing prognosis. Most physicians lack awareness of prognostic discordance, raising the need to explore patients’ prognostic information preferences and perceptions, and to tailor prognostic communication.