Diagnostic accuracy of the palliative prognostic score in hospitalized patients with advanced cancer.

PURPOSE: To evaluate the predictive accuracy of the Palliative Prognostic (PaP) score in patients with advanced cancer under the care of an oncologist. PATIENTS AND METHODS: The PaP score was calculated in 100 consecutive patients with advanced cancer hospitalized under the care of a medical or radiation oncologist at a university teaching hospital in Australia. The attending oncologist predicted the survival duration for the purpose of the scoring. The positive predictive value of the PaP score was evaluated. Survival analysis was performed to compare the survival of the three prognostic groups. RESULTS: Assessable survival data were available for 98 patients. The overall median survival was 12 weeks (interquartile range, 7 to 25 weeks). The PaP score divided the heterogeneous patient sample into three isoprognostic groups related to the chance of surviving 1 month, with 64 patients in group A (> 70% chance), 32 patients in group B (30% to 70% chance), and four patients in group C (< 30% chance). The estimated median survival of the three groups was 17 weeks (95% CI, 12 to 26 weeks), 7 weeks (95% CI, 4 to 12 weeks), and less than 1 week (95% CI, < 1 to 3 weeks), respectively. These survival differences were highly significant (log-rank test of trend, chi1(2) = 25.65; P < .0001). The 1-month survival of the groups was 97%, 59%, and 25%, respectively. CONCLUSION: When oncologists' survival estimates are used, the PaP score is able to identify accurately three isoprognostic patient groups, irrespective of the cancer type. The PaP score may help reduce the uncertainty of formulating a prognosis in patients with advanced cancer.