HERC Seminar: Estimating survival: does patient reported quality of life matter?

Abstract

Standard approaches to modelling outcomes commonly include a range of predictors such as age, sex and co-morbidities. However, there is growing evidence that quality-of-life (QoL) is also an important predictor of
mortality.

Whilst the effects of standard risk factors have been adequately captured to describe mortality in most models, the omission of QoL indicates the assumption of full independence between QoL and life expectancy (LE). This omission could give rise to systematic bias when measuring quality-adjusted life years (QALYs).

Using data from a registry cohort of patients undergoing total knee replacement (TKR), potential correlations between baseline utility, change in utility and survival are explored using parametric survival models. Outcomes such as LE, QALYs, and incremental cost-effectiveness ratios (ICERs) were calculated and compared.