Application Of Risk-adjusted Control Chart In Public Health Monitoring

Public health monitoring is an important part of disease prevention and control.Reasonable and effective public health monitoring can detect the development trend of a certain disease as soon as possible,so as to formulate targeted solutions to protect people's lives and property.During the monitoring process,if the control chart alarms,the relevant personnel will think that the process has deviated from the normal state,and immediately stop the process to find out the specific cause of the deviation.Then find the corresponding solution to restore the entire process to normal.However,if the control chart does not consider the specific situation and send out the wrong alarm signal,it will lead to a lot of waste of resources,including the waste of resources caused by investing time and resources to find out the specific reasons,and the waste of resources caused by stopping the monitoring process and affecting the whole progress.In the previous monitoring process,people usually only monitor the final result,but in reality,there are many unavoidable objective conditions that will affect the final result.If these factors are not considered in the monitoring,the probability of false alarms from the control chart will increase.We cannot use the same standard to monitor all situations.At this time,it is urgent to combine the control chart and risk factors for monitoring.The data in public health are generally count data in the same time interval.cumulative sum control charts and exponentially weighted moving average control charts are more sensitive in monitoring small drift.Based on this background,this thesis has done the following work:(1)This thesis studies the counting data of Poisson distribution in public health,and uses Poisson regression to establish the relationship between risk factors and Poisson results,so that the expected distribution is in dynamic change with the actual situation.(2)Applying the established Poisson regression model and combing the Poisson CUSUM control chart and Poisson EWMA control chart to monitor whether the actual data is in the expected state,and then achieve the real-time monitoring and early warning of the count data in the public health field.(3)The proposed regression-adjusted Poisson CUSUM control chart and the regressionadjusted Poisson EWMA control chart were used to monitor the actual influenza data set,and were compared with the traditional Poisson CUSUM control chart and the traditional Poisson EWMA control chart.It is concluded that the control chart after regression adjustment is more sensitive in monitoring.By connecting the influence factors with the results for online monitoring,the monitoring results can be more reasonable,and the monitoring sensitivity of the control chart can be improved to a certain extent.