A risk-adjusted multi-attribute cumulative sum control scheme in health-care systems

Hospitals increasingly use control charts to monitor clinical processes and their outcomes. In medical context, control charts should have stable performance when different patients with different levels of risk enter the hospital. In order to monitor multi-attribute medical processes, we propose a new control chart with entities having different levels of risk. First, risk-adjusted multivariate cumulative sum control chart (RA-MCUSUM) is developed. Then, simulation experiences are performed to demonstrate the application and to evaluate its performance in terms of in-control average run length (ARL₀) stability with the one of a standard MCUSUM chart. The results show that while the standard MCUSUM shows a sensitive performance, the RA-MCUSM has a robust performance when different entities with different levels of risk enter the system.