An Effective Method for Online Disease Risk Monitoring

ABSTRACT

Many diseases can be prevented or treated if they can be detected early or signaled before their occurrence. Disease early detection and prevention (DEDAP) is thus important for health improvement of our society. Traditionally, people are encouraged to check their health conditions regularly so that readings of relevant medical indices can be compared with certain threshold values and any irregular readings can trigger further medical tests to find root causes or diseases. One limitation of such traditional DEDAP methods is that they focus mainly on the data collected at the current time point and historical data are not fully used. Consequently, irregular longitudinal pattern of the medical indices could be neglected and certain diseases could be left undetected. In this article, we suggest a novel and effective new method for DEDAP. To detect a disease by this method, a patient’s risk to the disease is first quantified at each time point, and then the longitudinal pattern of the risk is monitored sequentially over time. A signal will be triggered by a large cumulative difference between the longitudinal risk pattern of the patient under monitoring and the longitudinal risk pattern of a typical person without the disease in concern. Both theoretical arguments and numerical studies show that it works well in practice. Supplementary materials for this article are available online.

KEYWORDS:

• Disease early detection
• Longitudinal data
• Online monitoring
• Risk factors
• Sequential test
• Statistical process control

Supplementary Materials

ComputerCodesAndData.zip: This zip file contains computer codes of our proposed method and the stroke data used in the article.

Acknowledgments

The authors thank the editor, the associate editor, and two referees for their constructive suggestions and comments, which improved the quality of the article substantially.