Monitoring of process and risk-adjusted medical outcomes using a multi-stage MEWMA chart

TL;DR

This paper introduces a multi-stage control chart using a multivariate EWMA statistic to monitor healthcare processes at all stages, adjusting for patient risk and dependencies, enabling early detection of small changes.

Contribution

It develops a novel multi-stage control chart based on EWMA for simultaneous, risk-adjusted monitoring of all process stages in healthcare, addressing limitations of traditional final-stage only surveillance.

Findings

Effective detection of small process shifts demonstrated in simulations

Better understanding of stage relationships and impact on outcomes

Potential for improved resource allocation in healthcare quality improvement

Abstract

Most statistical process control programmes in healthcare focus on surveillance of outcomes at the final stage of a procedure, such as mortality or failure rates. Such an approach ignores the multi-stage nature of these procedures, in which a patient progresses through several stages prior to the final stage. In this paper, we develop a multi-stage control chart based on a multivariate exponentially weighted moving average (EWMA) test statistic derived from score equations. This allows simultaneous monitoring of all intermediate and final stage outcomes of a healthcare process, with adjustment for underlying patient risk factors and dependence between outcome variables. Use of the EWMA test statistics allows quick detection of small gradual changes in any part of the process. Three advantages of the approach are: better understanding of how outcomes at different stages relate to each other, explicit monitoring of upstream stage outcomes may help curtail trends that lead to poorer end-stage outcomes and understanding the impact of each stage can help determine the most effective allocation of quality improvement resources. Simulations are performed to test the control charts under various types of hypothesised shifts, and the results are summarised using out-of-control average run lengths.