Introduction Prediction of life expectancy in terminally ill patients is an important end-of-life care issue. Amongst the most used validated tools for prognostication are the Palliative Performance Scale (PPS) and the Palliative Prognostic Index (PPI). There is no reliable tool to evaluate the prognosis of patients when life expectancy is less than 3 weeks. The objective is to improve the accuracy of the survival prediction of terminally ill patients (with or without cancer) in a palliative care residence setting: short and intermediate-term life expectancy.

Methods This was a prospective open study on all patients admitted at the Teresa Dellar Palliative Care Residence (TDPCR) during a period of 1 year. PPI and PPS scores were assessed at admission (Day 0), Week 1, 2, 3, and monthly thereafter or until patient death. We also systematically documented daily, 7 physical signs of impending death (Designated Short Term Prognosis Signs: DSTPS). A questionnaire was used to determine if families/patients wanted to know the prognosis and reasons for it. Primary endpoints included the prognosis determination using PPI and PPS scores at every time point vs the actual survival time; and the number of days between the first date of occurrence of each of the DSTPS item and the actual date of death.

Results/Application 285 patients were included in this study (217 with cancer): 44% males; 56% females. Median age: 82 years old, median survival: 8 days. Most families (83%) found usefulness of prognosis. At admission, most patients had a PPS of 20 or 30%. Almost all patients exhibited the totally bedbound DSTPS item (200 patients) which was associated with the longest survival time. The other DSTPS items were also present in most patients associated with various survival times.

IMPACT A prognostic model is being prepared based on the results of this study.