Epidemiological analysis and prediction model research of influenza-like cases in children in Hohhot City

Objective To understand the etiological monitoring results and epidemiological characteristics of pediatric influenza-like illness (ILI) in Hohhot, and their correlation with meteorological factors and air quality factors, and to establish the prediction model of pediatric ILI, so as to provide scientific basis for prevention and control of pediatric influenza and early warning prediction. Methods The monitoring data of ILI cases among children in Hohhot from 2014 to 2017, as well as the meteorological data and air quality factor data in the same period were collected. The correlation of meteorological factors and air quality factors with pediatric ILI was analyzed. The seasonal autoregressive moving average (SARIMA) model of pediatric ILI in this region was established, and the goodness of fit and prediction accuracy of the models before and after the introduction of external factors were analyzed and compared. Results From 2014 to 2017, 58 864 ILI cases among children were reported, with an average ILI report percentage of 9.05%. A total of 1 502 specimens were isolated with a positive isolation rate of 9.85%. In the past four years, pediatric influenza in Hohhot had been mixed with different subtypes of influenza A H1, H3 and influenza B, but the dominant strains were different each year. There were significant differences in viral isolation rates among different age groups and different genders (χ2=43.54 and 7.12, P<0.05) . The number of pediatric ILI weekly cases in Hohhot had obvious seasonal and periodic characteristics. The number of pediatric ILI weekly cases was negative correlated with the weekly mean temperature and weekly mean humidity (rs=-0.290, -0.210, P<0.05) . The weekly mean temperature was incorporated into the SARIMA (1, 0, 1) (0, 1, 1) 52 prediction model for the number of pediatric ILI cases, and the goodness of fit and prediction accuracy were improved. Conclusions From 2014 to 2017, pediatric influenza epidemic in Hohhot mixed influenza A H1, H3 and influenza B, and showed an alternate occurrence. Weekly mean temperature is related to the prevalence of ILI in children. The SARIMA model including weekly mean temperature can be used as a technical method for early prediction of influenza epidemic. Key words: Child; Pediatric influenza-like cases; Meteorological factors; Air quality; Seasonal auto-regression moving average model