Risk-oriented model for predicting epidemiological situation with crimean-congo hemorrhagic fever (on the example of Stavropol region)

Our research object was a multi-factor prediction of risks related to even a single case of Crimean-Congo hemorrhagic fever (CCHF) on a territory of a particular administrative district in a RF subject (on the example of Stavropol region). Risk-oriented model aimed at yearly prediction of CCHF occurrence was created with heterogeneous sequential statistics clarification procedure. We considered monthly climatic parameters (air temperature, relative air humidity, precipitations quantity, snow mantle size, and air pressure) and epidemiologic data (number of CCHF cases last year and number of settlements where CCHF cases were registered) as predictors for new CCHF cases occurrence. To check our prediction model precision, we took data on risk factors from 2011 to 2015 for each administrative district in Stavropol region. Threshold level of a positive solution probability was set at 99 % (error probability was equal to 1 %). We tested our prediction model as per retrospective data collected in 2013–2016. It allowed us to predict even a single pa-tient with CCHF occurrence for each administrative district in Stavropol region in 2017. In the course of data analysis we detected high precision in potential prediction results. Totally we revealed six false-positive and two false-negative (actually erratic) results but they can result from objective factors, for example insufficient diagnostics of the disease, as well as imported cases. The ob-tained data can be applied in practical activities of Rospotrebnadzor offices aimed at planning and organizing CCHF prevention. The next stage in the prediction model development will be creation of a technique for calculating an expected number of CCHF cases for each administrative district where at least one case of the disease is predicted in the forthcoming year.