This article proposes a new class of control charts that may be used for monitoring and improving the quality of care. Unlike conventional control charts that rely on observed performance data, these charts use risk-adjusted data in addition to the observed data. The resulting time-ordered charts are capable of reducing time-to-time variation that may stem from uncontrollable changes in patient mix over time. Depending on how observed and risk-adjusted data are combined, proposed charts are categorized under the framework of either additive or multiplicative models. Risk-adjusted rates are obtained using multivariate logistic regression models. It was found that the risk-adjusted control charts could be effective in reducing biases that arise from variation in patient mix. These charts can potentially achieve higher sensitivity and specificity compared with ordinary control charts.
Marilyn K. Hart, PhD, is Professor, Quality Improvement, College of Business, University of Wisconsin Oshkosh
Kwan Y. Lee, PhD, is Project Director, Division of Research, Joint Commission on Accreditation of Healthcare Organizations, Oakbrook Terrace, Illinois
Robert F. Hart, PhD, is Adjunct Professor, College of Business, University of Wisconsin Oshkosh
James W. Robertson, Associate Project Director, Division of Research, Joint Commission on Accreditation of Healthcare Organizations, Oakbrook Terrace, Illinois.
The authors wish to thank Haixiao Chen, Christine McGreevey, Jodi Neikirk, and Brent James for their help and encouragement.