FROM NOISE TO CHARACTERIZATION TOOL: ASSESSING BIASES AND PERFORMANCE OF INFLUENZA SURVEILLANCE METHODS THROUGH QUALITATIVE AND QUANTITATIVE STUDIES
In response to the threat of pandemic flu outbreaks, emerging novel influenza surveillance systems, such as syndromic surveillance, now utilize pre-diagnostic data to monitor flu activities. Biases due to the changes in health related behavior, and the background noise caused by seasonal flu epidemics, may limit its utility for early detection. Situational awareness, the ability of monitoring and characterizing flu transmission over time, is considered to be a more reasonable goal for conducting influenza surveillance. Indeed, disease surveillance is a process, the product of which reflects both real illness and public awareness of the disease. Biases are therefore embedded in each surveillance systems, and need to be assessed to better provide situational awareness for decision making purpose.Through qualitative studies including systematic review of the scientific literature, official documents and news reports, as well as interviews with selected health officials from health departments, objectives, utility and performance of influenza surveillance systems, especially syndromic surveillance systems, were assessed using pH1N1 outbreak as an critical event. The results show, syndromic surveillance provided situational awareness that allowed early recognition of the connection between outbreaks in the United States and Mexico, and led to early response. The utility of syndromic surveillance at health departments is limited, due to the lack of targeted analytical tools and clarified objectives based on systemic assessment of the biases.Influenza surveillance data from Georgetown University and Hong Kong government were used to explore the possibility of using biases as a characterization tool. Bayesian Hierarchical model was applied to estimate the statistical relationships between influenza surveillance data and the informational environment, such as the alerts from HealthMap and web queries from Google. The model identified the types of characteristics in surveillance systems that are less likely to be influenced by the information environment, and systems that might monitor the same population or are subject to the same biases in its process.More than just the noise in the system, informational environment associated biases can be used as a characterization tool, which can facilitate practitioners to make informed decisions on choosing surveillance systems to serve for specific situation awareness purposes.
MetadataShow full item record
Showing items related by title, author, creator and subject.
Employing Quantitative and Qualitative Analytic Methods for Internet Biosurveillance of Pandemic Influenza Iarocci, Emily Ann (Georgetown University, 2013)Public health surveillance has grown rapidly alongside the availability of big data and an increasingly interconnected global community. Internet biosurveillance, one facet of these surveillance activities, uses unstructured, ...
What Can We Learn From Chinese Randomized Controlled Trials? A Systematic Review and Meta-Analysis of Chinese Venlafaxine Studies Koesters, Markus; Zhang, Ying; Ma, Yong Chun; Weinmann, Stefan; Becker, Thomas; Jin, Wei Dong (2011-04)This systematic review evaluated Chinese trials examining the efficacy of venlafaxine in the treatment of depression. Chinese databases CNKI and VIP and western databases were searched for blinded randomized controlled ...
Effect of Ethnicity on Performance in a Final Objective Structured Clinical Examination: Qualitative and Quantitative Study Wass, Val; Roberts, Celia; Hoogenboom, Ron; Jones, Roger; Van der Vleuten, Cees (2003-04-12)