A Comprehensive Study on Monitoring Treatment Outcomes Using Risk- adjusted CUSUM Control Charts

Rapidly determining difficulties in the quality of care is critical to patient’s well-being. Without proper inspection procedures, such issues can go unnoticed for years. Although cumulative sum (CUSUM) charts are effective for quality control, there is limited methodology for predicting survival results. They are powerful tools for detecting deviations from expected performance over time and are particularly suited for monitoring both continuous and binary outcomes in clinical settings. This study aims to an advanced statistical technique for evaluating treatment outcomes and monitoring quality in healthcare using CUSUM control charts. A Cox proportional hazards model was employed to estimate covariates and cumulative baseline hazard function, enabling risk-adjusted evaluations. Both CGR-CUSUM and BK-CUSUM charts were constructed. These charts were calibrated to detect changes in survival and treatment success during a predefined follow-up period. In parallel, a Bernoulli CUSUM chart was developed for binary outcomes, such as treatment success or failure, using logistic regression to adjust for covariates. By combining survival analysis and risk-adjusted quality control, the method ensures high sensitivity to performance changes while accounting for variability in patient characteristics. This integrative approach made by using success (survival control charts estimation software) package in R programming. The use of CUSUM charts for risk-adjusted monitoring gives a reliable solution to enhancing clinical outcomes and maintaining standards in healthcare environments.