上海市静安区居民意外跌落死亡率时间序列分析与预测

目的 分析静安区居民意外跌落死亡率变化趋势,为制定预防性干预措施提供参考依据。方法 应用时间序列模型拟合静安区1975-2009年意外跌落死亡率资料,并预测该地区2010-2014年意外伤害死亡率。结果 1975-2009年期间该地区人群意外跌落死亡2814人,占伤害死亡首位(41.99%);所建立的AR(1)模型较好地拟合了既往时间序列的意外跌落死亡率,AIC=207.42,SBC=210.53,参数估计与模型检验均有统计学意义,2010-2014年预测意外跌落死亡率分别为17.07/10万、17.72/10万、18.07/10万、18.25/10万、18.35/10万;模型预测值的动态趋势与实际情况基本一致,各年的实际值均在预测值95% 的可信区间范围内。结论 静安区意外跌落死亡率适合应用AR模型拟合,并可用于该地区意外跌落死亡率的短期预测分析。

Change trend and prediction of unintentional fall mortality in Jing'an district of Shanghai:time series analysis

Objective To analyze the trend for unintentional fall mortality in residents of Jing'an district,Shanghai,and to provide basic evidences for injury interventions.Methods Data on unintentional falls among residents of Jing'an district between 1975-2009 were collected and analyzed with time series analysis and time sequence model was adopted to predict unintentional fall mortality for the years of 2010-2014 in the district.Results During the period in the district,the total number of deaths due to unintentional fall was 2814,accounting for 41.99% of total injury deaths and being the top cause of accidental injury death.An autocorrelation model(AR[1])was established for the fitting of unintentional fall mortality variation,with an Akaike's information criterion(AIC)value of 207.42 and a Schwartz Bayesian criterion(SBC)value of 210.53,and both the parameter estimation and model test of the model presented significant result.Based on the model,the predicted unintentional fall mortality in the district for the years of 2010-2014 are 17.07/100000,17.72/100000,18.07/100000,18.25/100000,and 18.35/100000,respectively;the predicted trend in the mortalities is consistent to the observed status and all the predicted mortality rates are within the range of 95% confidence interval of the actual rates.Conclusion The data on unintentional fall mortality in Jing'an district is appropriate to be fitted to AR(1)model and the established model could be used for the prediction of short-term unintentional fall mortality in the district.