Independent validation of the modified prognosis palliative care study predictor models in three palliative care settings.

Context: Accurate prognostic information in palliative care settings is needed for patients to make decisions and set goals and priorities. The Prognosis Palliative Care Study (PiPS) predictor models were presented in 2011, but have not yet been fully validated by other research teams.

Objectives: The primary aim of this study is to examine the accuracy and to validate the modified PiPS (using physician-proxy ratings of mental status instead of patient interviews) in three palliative care settings, namely palliative care units, hospital-based palliative care teams, and home-based palliative care services.

Methods: This multicenter prospective cohort study was conducted in 58 palliative care services including 16 palliative care units, 19 hospital-based palliative care teams, and 23 home-based palliative care services in Japan from September 2012 through April 2014.

Results: A total of 2426 subjects were recruited. For reasons including lack of followup and missing variables (primarily blood examination data), we obtained analyzable data from 2212 and 1257 patients for the modified PiPS-A and PiPS-B, respectively. In all palliative care settings, both the modified PiPS-A and PiPS-B identified three risk groups with different survival rates (P<0.001). The absolute agreement ranged from 56% to 60% in the PiPS-A model and 60% to 62% in the PiPS-B model.

Conclusion: The modified PiPS was successfully validated and can be useful in palliative care units, hospital-based palliative care teams, and home-based palliative care services.