Journal of Statistical and Econometric Methods

On The Use of Truncated Zero Inflated Binomial (ZIB) Control Chart for Monitoring Tuberculosis Disease

• [ Download ]
• Times downloaded: 9821

• Abstract

  The design of various statistical methods for monitoring rare health events shows significance of the issue in health sectors. Rare health events, as attribute quality characteristics cannot be monitored by ordinary Shewhart np charts since overdispersion occurs. A good approach to this problem is the use of control charts based on zero inflation in a binomial (ZIB) distribution. In this distribution, it is assumed that random shocks occur with some probability, and upon the occurrence of such random shocks, health event failures can be found, such that the number of failures in each sampling subgroup follows a binomial distribution. This study develops a truncated ZIB control chart applying probability limits in Shewhart based control limits for monitoring ZIB distributed observations. As the most widespread criteria, Average Run Length approach is used to evaluate the performance of this chart. The use of truncated zero inflation in a binomial (TZIB) control chart is also investigated by a real case study, using the number of patients who undergone Tuberculosis treatment and later resulted to Drug Resistant Tuberculosis (DRTB) in General Hospital, Igarra, Akoko-Edo Local Government of Edo State. Results are compared with the traditional number of proportion defective (np) chart.