[Statistical analysis of trend of human Salmonella infections in Poland in 1995-2007].

: The epidemiological situation of human salmonellosis in Poland in the successive years (between 1995 and 2007), presented in this study, was illustrated by using the notification rates expressed as the number of confirmed cases (of total human Salmonella infections and infections associated with chosen serovars) per 100,000 inhabitants. The notification rates were calculated basing on the published data of human salmonellosis in Poland (1995-2007) and the total population in each year. Their distribution over time was presented in a form of the specific notification rate graphs. The aim of the study was the statistical analysis of observed changes in level of human Salmonella infections in Poland in 1995-2007. Particularly, we tried to investigate whether the general decrease of Salmonella infection notification rates observed over investigated period can be recognized as significant in the statistical meaning, and whether the changes in infections caused by the most frequently isolated serovars show the clear tendency, i.e. whether it is possible to say about the permanent decrease or increase of these infections. The trends in notification rates were analysed using simultaneously three different statistical tests including chi-square for trend, linear regression, and Poisson regression. The results of all tests taken together were used for trend estimation. The decreasing/increasing tendency of evaluated rates was found to be characteristic if it showed statistical significance (p < 0.05) in all of these tests simultaneously. The results revealed statistically significant and decreasing trends in notification rates only for salmonellosis in total and for infections caused by S. Enteritidis. In the case of seven other statistically analysed serovars (i.e., S. Typhimurium, S. Hadar, S. Infatis, S. Virchow, S. Newport, S. Mbandaka, S. Agona) it is not possible to say explicitly that their notification rates do not show a certain tendency over time, however changes revealed by them in the investigated period are not essential enough or their distribution is not regular enough that there are no grounds to ascribe a definite trend to them. They are not found to be significantly increasing or decreasing over time, and their distribution shows rather features of random walk.