Detection of Multiple Diseases by Sequential Risk Monitoring Using a Dynamic Screening System

Early detection and prevention of diseases, particularly chronic conditions such as cardiovascular diseases (CVD), are essential for improving patient outcomes and reducing the burden on healthcare systems. Since disease risk factors are typically observed sequentially over time, early detection inherently involves sequential decision-making. This process is complicated by challenges such as time-varying data distributions and serial correlations in the observed data. While existing statistical process control (SPC) methods can effectively detect a single disease, there is an increasing demand for approaches capable of addressing multiple diseases simultaneously. In practice, patients often present with multimorbidities, such as myocardial infarction (MI) and stroke, which share common risk factors like high blood pressure, elevated cholesterol, and abnormal blood glucose levels. Developing a unified framework to jointly address these conditions can enhance the efficiency and cost-effectiveness of disease detection and prevention strategies. This paper introduces a novel dynamic screening system for the early detection of multiple diseases. The proposed method quantifies a patient’s risk for various diseases at each observation point using a flexible longitudinal data modeling approach. It then compares the patient’s risk profile with the baseline risk pattern of non-diseased individuals. By sequentially monitoring cumulative risk deviations, the system triggers a signal when a patient exhibits abnormally high risk for any disease of interest. Numerical studies demonstrate the robustness and effectiveness of the proposed approach under various scenarios. The method is applied to a dataset from the Framingham Heart Study. The results showcase its ability to identify early risks for MI and stroke, two major manifestations of CVD, highlighting its potential as a valuable tool in clinical practice.