The Association of the Palliative Performance Scale and Hazard of Death in an Ambulatory Cancer Population

BACKGROUND: Previous studies have reported survival estimates in palliative populations using the Palliative Performance Scale (PPS) (where 100=best status, 0=death). However, little research has examined the association of the PPS with hazard of death in ambulatory populations. OBJECTIVE: We examined the association between performance status and the instantaneous hazard of death in ambulatory cancer patients, using longitudinal PPS scores. METHODS: This retrospective, population-based cohort study included cancer outpatients who had at least one PPS assessment completed between 2007 and 2009. PPS scores were recorded opportunistically by health care providers at clinic or home care visits. We used a Cox proportional hazards model to determine the relative hazard of death based on repeated measures of PPS score, while controlling for other covariates. RESULTS: Among 11,342 qualifying cancer patients, there were 54,207 PPS assessments. The distribution of PPS scores at first assessment were 23%, 56%, 20%, and 1% for PPS scores of 100, 90-70, 60-40, and ≤ 30, respectively. A quarter of the cohort died within 6 months of the first assessment. The relative hazard of death increases by a factor of 1.69 (95% confidence interval [CI]: 1.72-1.67) for each 10-point decrease in PPS score. Thus the hazard of death increases by 8.2 (1.69(4)) times for a person with PPS score of 30 compared with a person with a score of 70. CONCLUSION: The PPS was significantly associated with hazard of death in ambulatory cancer patients; the relative hazard of death increased based on lowered PPS scores. Providers should consider broadening its use to include patients throughout their disease trajectory.