Abstract

Rotavirus infection is one of the main causes of acute gastroenteritis and has an important impact on hospitalisation. There is no homogeneous surveillance system for rotavirus infections in Europe. The aim of this study is to develop a predictive model in order to estimate the expected rotavirus infections in the population covered by a hospital. A five year study (2000-2004) was developed in a Spanish university hospital. A correlation test between the notifications reported to the Microbiological Information System (SIM) and hospitalisations was carried out, as well as a time series analysis, obtaining the trend and the cyclical components. The predictive model was adjusted using the least squares method. A direct relationship between the microbiological isolations and the hospitalisations was established (=0.925; p<0.001). A significant annual cycle was observed, with the peak of cases occurring in February. The two principal outbreaks that occurred in the study period would have been detected with the predictive model. Expected rotavirus cases and hospitalisations for 2005 and 2006 were obtained. The notifications of rotavirus infections reported to SIM are adequate in order to establish a hospital surveillance system, but a predictive model which provides expected cases is also necessary. Therefore, this tool will be useful to evaluate preventive measures such as rotavirus vaccines, which will soon be available in Europe.