Prevalence characteristics of cerebral infarction among residents in Hebei province, 2015 – 2018: a time series analysis

  Objective   To examine prevalence characteristics of cerebral infarction (CI) among residents in Hebei province from 2015 to 2018, and to explore the application of time series model in the prediction of CI incidence for the prevention and control of the disease.

  Methods   From National Chronic Non-Communicable Disease Surveillance System, we collected the data on registered cerebral infarction incidents in 18 surveillance points in Hebei province from January 1, 2015 through December 31, 2018. The prevalence characteristics of CI and its changing trend were analyzed and a time series model was established to predict annual and monthly incidence rate of CI in 2019.

  Results  During the 4-year period in the province, the average annual CI incidence rate were 311.73/100 000, with the yearly incidence rate of 283.22, 296.36, 322.93, and 342.55 per 100 000 for the year from 2015 to 2018 and a significantly increasing trend (χ² = 624.353, P < 0.001). The CI incidence rate in the male residents were significantly higher than that in the female residents (348.81/100 000 vs. 273.60/100 000, χ² = 1 680.960; P < 0.05). The age-specific CI incidence rate was significantly different among various age groups, with the rate of 2.58, 48.96, 263.21, 650.61, 1 641.77, 2 083.59, and 3 000.88 per 100 000 for the residents aged < 35, 35 – 44, 45 – 54, 55 – 64, 65 – 74, 75 – 84, and ≥ 85 years, respectively (χ² = 381 051.465, P < 0.001). The established best fitted model was ARIMA (autoregressive integrated moving average) (0, 1, 1)(2, 1, 0)₁₂; the model's residual sequence was white noise sequence (P > 0.05); and the parameters for the established model were as following: Akaike information criterion (AIC) = 214.480, Bayesian information criterion (BIC) = 220.698, root mean square error (RMSE) = 2.97, mean absolute error (MAE) =1.62, and mean absolute percentage error (MAPE) = 7.55%. Based on the model, the predicted yearly CI incidence rate was 362.46/100 000 for the residents in 2019, which was higher than that in previous years; the predicted monthly CI incidence rates ranged between 23.62/100 000 and 33.33/100 000 and the variation of the rates for the year of 2019 was similar to that in previous years.

  Conclusion   The incidence rate of cerebral infarction was at a high level and increased yearly during 2015 – 2018 among residents in Hebei province; the incidence rate was relatively higher in the male and the elderly residents. Time series model can be used to predicate the incidence rate for the prevention and control of cerebral infarction.