Monitoring multivariate progress variability after heart surgery

Patients undergoing cardiac surgery including Coronary Artery bypass grafting or Valve replacements are normally placed in the ICU for routine monitoring of vital signs related to simple and complex aspects of heart function. This is usually achieved by screening several indicators such as Systolic Blood Pressure, Diastolic Blood Pressure, Mean Arterial Pressure and many more. In most clinical monitoring, there are always more than one quality characteristics of interest which are usually correlated and the condition of the patient is assessed by a subgroup of size 1 multivariate data at any given time. Therefore, multivariate quality control charts should be deployed to monitor the progress of the patient. In this paper the optimal Multivariate Exponentially Weighted Mean Squared deviation (MEWMS) control chart is deployed to simultaneously monitor the variability of up to three correlated characteristics that represent patient's progress in the ICU unit using subgroup size one. The results clearly show that disregarding the correlation between the characteristics and monitoring the patient's progress under the current practice that uses individual charts will lead to misdiagnosis. The research is based on real data collected from an ICU unit.