Trend equation prediction in medical and pharmaceutical studies (the example of respiratory diseases development in children in the region)

Abstract

The article discusses the incidence forecast on the example of respiratory diseases in children in Nizhny Novgorod. The study used the official statistics of the respiratory diseases prevalence in children of 0-14 age group in 2001-2015. Prediction analysis was based on the trend lines, which displayed the functions from the mean values of the analyzed indicators. The function selection was determined by the temporal data changes. The trend calculation involves three stages: data entry into an Excel spreadsheet, plotting a graph, the selection of a trend line parameters. Then a function is selected that approximates the correlation The choice of the function is determined by the highest determination coefficient (not less than 0,5000). The correlation in question was best approximated by a second-de gree polynomial interpolation. The trend line determination coefficient in this case was 0,9292. The predictive accuracy, that was determined as the mean absolute percentage error (MAPE), was 3%. The predictive accuracy based on a second-degree polynomial interpolation was 97%, which is very high. The found correlation was used in 3-year forecast. Mathematical techniques and informational technologies can significantly improve the quality of care. The information technology becomes one of the leading factors in healthcare development. However, mathematical models and computational algorithms for monitoring, predicting incidence, spread and prevention of various nosologies are still to be developed.