Improved statistics to help monitor disease
Proper public health surveillance which requires effective statistical monitoring is crucial for advancing the health sector in Europe. The EU-funded SMPHS (Improved Statistical Monitoring Procedures for Attributes with Applications in Public Health Surveillance) project worked on improving statistical monitoring techniques to detect changes in disease incidence and mortality rates. It also evaluated current statistical monitoring techniques in implementing and evaluating public health policies. To achieve its aims, the project investigated, developed and tested improved control charts that represent the key statistical tool within process monitoring. As control charts require specific assumptions to yield viable statistics, the project team closely assessed how they function and proposed better techniques to enhance statistical monitoring. In more technical terms, the focus was mainly on developing new cumulative sum and exponentially weighted moving average control charts. The project team also studied the performance of Shewhart-type control charts for zero-inflated processes. This yielded five new cumulative sum control charts that can be used for monitoring non-typical count data. Various types of count data arise very frequently in health-related problems. It also resulted in a new exponentially weighted moving average chart that can be used for monitoring any type of count data. Moreover, the project evaluated the performance of control charts for zero-inflated process with estimated parameters. It outlined an improved set of general inflated discrete probability distributions for many different probability models that are useful for describing several types of count data. The results also included simple control charting techniques that can be exploited for improving statistical monitoring. The outcomes of the project were discussed in several papers published in relevant journals. The team also released the programming codes related to their research, which can assist in designing new statistical monitoring schemes and for reproducing the numerical results discussed in the papers. Overall, the new control schemes and techniques are set to extend more useful solutions to the challenges that undermine current statistical monitoring. They will help in detecting health issues in a wide geographical area more accurately and flag conditions that look out of the ordinary. If these statistical techniques are exploited, important socio-economic benefits and implications may emerge from the results of this project.
Keywords
Public health surveillance, statistical monitoring, SMPHS, control charts, count data
