Abstract
The differential prognostic accuracy of the Palliative Prognostic Index (PPI) in hospitalized oncologic patients treated by a palliative care team according to the presence or absence of acute concomitant diseases was analyzed. All patients (n = 322) hospitalized in a palliative unit of a university hospital were included in a 2-year prospective study. PPI was determined at the time of hospital admission and discharge. Patients were grouped into two categories according to the causes of hospitalization (presence and absence of acute concomitant diseases). Metastases, PPI punctuation, refractory symptoms, and the presence of acute concomitant diseases were analyzed as survival-related factors. The absence of acute concomitant diseases and a PPI calculated at admission >4 or >6 were related with survival at 3 and 6 weeks, respectively. After hospital discharge, the accuracy of PPI was lower, decreasing the positive predictive value from 84% (PPI calculated at the time of hospital admission) to 67% (PPI calculated at the time of discharge) for survival <6 weeks. In conclusion, the impact of acute concomitant diseases on survival should be considered in prediction models for patients receiving palliative care.