Using decision fusion methods to improve outbreak detection in disease surveillance

A decision support framework for prediction of avian influenza

For years, avian influenza has influenced economies and human health around the world. The emergence and spread of avian influenza virus have been uncertain and sudden. The virus is likely to spread through several pathways such as poultry transportation and wild bird migration. The complicated and global spread of avian influenza calls for surveillance tools for timely and reliable prediction of disease events. These tools can increase situational awareness and lead to faster reaction to events. Here, we aimed to design and evaluate a decision support framework that aids decision makers by answering their questions regarding the future risk of events at various geographical scales. Risk patterns were driven from pre-built components and combined in a knowledge base. Subsequently, questions were answered by direct queries on the knowledge base or through a built-in algorithm. The evaluation of the system in detecting events resulted in average sensitivity and specificity of 69.70% and 85.50%, respectively. The presented framework here can support health care authorities by providing them with an opportunity for early control of emergency situations.

Outbreak minimization v.s. influence maximization: an optimization framework

Background

An effective approach to containing epidemic outbreaks (e.g., COVID-19) is targeted immunization, which involves identifying “super spreaders” who play a key role in spreading disease over human contact networks. The ultimate goal of targeted immunization and other disease control strategies is to minimize the impact of outbreaks. It shares similarity with the famous influence maximization problem studied in the field of social network analysis, whose objective is to identify a group of influential individuals to maximize the influence spread over social networks. This study aims to establish the equivalence of the two problems and develop an effective methodology for targeted immunization through the use of influence maximization.

Methods

We present a concise formulation of the targeted immunization problem and show its equivalence to the influence maximization problem under the framework of the Linear Threshold diffusion model. Thus the influence maximization problem, as well as the targeted immunization problem, can be solved by an optimization approach. A Benders’ decomposition algorithm is developed to solve the optimization problem for effective solutions.

Results

A comprehensive computational study is conducted to evaluate the performance and scalability of the optimization approach on real-world large-scale networks. Computational results show that our proposed approaches achieve more effective solutions compared to existing methods.

Conclusions

We show the equivalence of the outbreak minimization and influence maximization problems and present a concise formulation for the influence maximization problem under the Linear Threshold diffusion model. A tradeoff between computational effectiveness and computational efficiency is illustrated. Our results suggest that the capability of determining the optimal group of individuals for immunization is particularly crucial for the containment of infectious disease outbreaks within a small network. Finally, our proposed methodology not only determines the optimal solutions for target immunization, but can also aid policymakers in determining the right level of immunization coverage.

Performance of Bayesian outbreak detection algorithm in the syndromic surveillance of influenza-like illness in small region

Early warning for Infectious disease outbreak is an important public health policy concern, and finding a reliable method for early warning remains one of the active fields for researchers. The purpose of this study was to evaluate the performance of the Bayesian outbreak detection algorithm in the surveillance of influenza-like illness in small regions. The Bayesian outbreak detection algorithm (BODA) and modified cumulative sum control chart algorithm (CUSUM) were applied to daily counts of influenza-like illness in Tehran, Iran. We used data from September 2016 through August 2017 to provide background counts for the algorithms, and data from September 2017 through August 2018 used for testing the algorithms. The performances of the BODA and modified CUSUM algorithms were compared with the results coming from experts' signal inspections. The data of syndromic surveillance of influenza-like illness in Tehran had a median daily counts of 7 (IQR = 3-14). The data showed significant seasonal trends and holiday and day-of-the-week effects. The utility of the BODA algorithm in real-time detection of the influenza outbreak was better than the modified CUSUM algorithm. Moreover, the best performance was when a trend included in the analysis. The BODA algorithm was able to detect the influenza outbreaks with 4-5 days delay, with the least false-positive alarm. Applying the BODA algorithm as an outbreak detection method in influenza-like syndromic surveillance might be useful in early detection of the outbreaks in small regions.