Quality of Data Affects Detection of Turning Points in Flu Pandemic

Management of emerging infectious diseases such as the 2009 influenza pandemic A (H1N1) poses great challenges for real-time mathematical modeling of disease transmission due to limited information on disease natural history and epidemiology, stochastic variation in the course of epidemics, and changing case definitions and surveillance practices.

Researchers Ying-Hen Hsieh, David N Fisman and Jianhong Wu report in BMC Research Notes that application of the Richards model to Canadian H1N1 data shows that detection of turning points is affected by the quality of data available at the time of data usage. Using the Richards model, robust estimates of R0 were obtained approximately one month after the initial outbreak in the case of 2009 A (H1N1) in Canada.

The Richards model and its variants are used to fit the cumulative epidemic curve for laboratory-confirmed pandemic H1N1 (pH1N1) infections in Canada, made available by the Public Health Agency of Canada (PHAC). The model is used to obtain estimates for turning points in the initial outbreak, the basic reproductive number (R0), and for expected final outbreak size in the absence of interventions. Confirmed case data were used to construct a best-fit 2-phase model with three turning points. According to the researchers, R0 was estimated to be 1.30 (95% CI 1.12-1.47) for the first phase (April 1to May 4) and 1.35 (95% CI 1.16-1.54) for the second phase (May 4 to June 19). Hospitalization data were also used to fit a one-phase model with R0=1.35 (1.20-1.49) and a single turning point of June 11.

To access the study from BMC Research Notes, CLICK HERE.

Reference: Hsieh YH, Fisman DN and Wu J. On epidemic modeling in real time: An application to the 2009 Novel A (H1N1) influenza outbreak in Canada. BMC Research Notes 2010, 3:283