Risk-adjusted frailty-based CUSUM control chart for phase I monitoring of patients’ lifetime

Abstract

Monitoring the mortality associated with a surgical procedure leads to the proper decision making in a healthcare system. However, the surgical outcomes depend not only on the risk factors of each patient but also on other categorical influential covariates which cannot be easily measured. Ignoring the unmeasured covariates leads to the poor performance of monitoring procedures. To deal with this significant issue, a general Phase-I risk-adjusted cumulative sum control chart is proposed using a combination of accelerated failure time and frailty models to monitor surgical outcomes. Extensive simulation studies are conducted which reveal that the proposed frailty-based CUSUM chart outperforms the simple CUSUM chart. Moreover, the proposed approach performs better than the existing dummy-based CUSUM chart in terms of detection power. Subsequently, the results of a real cardiac surgery dataset indicate that the inclusion of frailty variables in the risk-adjustment model can effectively model the heterogeneity of the surgical data.

KEYWORDS:

• Discrete frailty model
• cumulative sum (CUSUM) control chart
• accelerated failure time (AFT) model
• dummy variable
• test power

Disclosure statement

No potential conflict of interest was reported by the author(s).