Prognosticating in patients with advanced cancer--observational study comparing the accuracy of clinicians' and patients' estimates of survival.

BACKGROUND Clinicians' prognoses in patients with advanced cancer are imprecise. The aim of this study was to compare doctors', nurses' and patients' survival predictions and to identify factors which influence accuracy. PATIENTS AND METHODS Some 1018 patients with advanced cancer were recruited. Survival estimates were obtained from the attending doctor, nurse, multidisciplinary team (MDT) and patient (n = 829, 954, 987 and 290 estimates, respectively) and were compared with actual survival. Clinician and patient characteristics were recorded. RESULTS MDTs', doctors' and nurses' predictions were accurate 57.5%, 56.3% and 55.5% of occasions, respectively. Nurses were less accurate than the MDT (P = 0.007) but were no worse than doctors (P = 0.284). Estimates of clinicians and patients were more optimistic (doctors: 31%; nurses: 34%; MDT: 31.1%; patients: 45.1%) than pessimistic (12.7%, 11%, 11.4% and 2.7%). Nurses' accuracy increased if they had reviewed the patient within 24 h. Most patients (61.4%) wanted to know their prognosis. Only 37.1% were willing to offer an estimate regarding their own survival. Patients' prognostic estimates were less accurate than health care professionals' (P < 0.001). CONCLUSIONS MDTs were better at predicting survival than doctors' or nurses' alone. Patients were substantially worse. Among nurses, recency of review was related to improved prognostic accuracy.