Risk-adjusted Bernoulli chart in multi-stage healthcare processes based on state-space model with a latent risk variable and dynamic probability control limits

Abstract

Healthcare-related processes are usually multistage in nature, and the output of each stage is the input of the subsequent stage. However, most existing monitoring schemes consider the individual stages of the healthcare process without considering both inter-stage and intra-stage links. The modeling of multistage healthcare processes based on the state-space model describes the cascade property of the risks and outcomes in subsequent stages. The present paper proposes a Bernoulli state-space model in which unmeasurable risks of processes are considered as a latent risk variable to give a realistic estimate of potential risks. In the proposed monitoring scheme for multistage medical processes, in addition to Parsonnet scores, the other categorical operational covariates are considered. Moreover, dynamic probability control limits are applied to remove the effect of patient risk distributions on in-control average run length performance. Simulation results reveal that the proposed risk-adjusted Bernoulli group exponentially weighted moving average chart for multistage healthcare processes performs satisfactorily under different shifts. Moreover, the proposed monitoring scheme helps to identify the out-of-control stage of the healthcare process to remove the corresponding assignable cause.