Adding items that assess changes in activities of daily living does not improve the predictive accuracy of the Palliative Prognostic Index

Background: Changes in activities of daily living in cancer patients may predict their survival. The Palliative Prognostic Index is a useful tool to evaluate cancer patients, and adding an item about activities of daily living changes might improve its predictive value.

Aim: To clarify whether adding an item about activities of daily living changes improves the accuracy of Palliative Prognostic Index.

Design: Multicenter prospective cohort study.

Setting: A total of 58 palliative care services in Japan.

Participants: Patients aged >20 years diagnosed with locally extensive or metastatic cancer (including hematological neoplasms) who had been admitted to palliative care units, were receiving care by hospital-based palliative care teams, or were receiving home-based palliative care. Palliative care physicians recorded clinical variables at the first assessment and followed up patients 6 months later.

Results: A total of 2425 subjects were recruited and 2343 of these had analyzable data. The C-statistic of the original Palliative Prognostic Index was 0.801, and those of modified Palliative Prognostic Indices ranged from 0.793 to 0.805 at 3 weeks. For 6-week survival predictions, the C-statistic of the original Palliative Prognostic Index was 0.802, and those of modified Palliative Prognostic Indices ranged from 0.791 to 0.799. The weighted kappa of the original Palliative Prognostic Index was 0.510, and those of modified Palliative Prognostic Indices ranged from 0.484 to 0.508.

Conclusion: Adding items about activities of daily living changes to the Palliative Prognostic Index did not improve prognostic value in advanced cancer patients.