特定关键词及百度指数与流感病毒活动相关性分析

目的 分析特定关键词的搜索引擎数据与流感病毒活动的相关性,探讨百度指数在流感监测中的应用。方法 根据对流感确诊病例调查选择最优关键词,通过百度搜索得到各关键词的百度指数。并采用Spearman相关分析和数据移动对天津市2014-2015年流感流行季百度指数与哨点监测数据进行相关分析。结果 流感确诊病例中65.04%（173/266）的人在出现症状后通过网络搜索引擎寻求过帮助。“感冒”作为关键词占比最高为67.72%（107/158）。关键词“发烧”的百度指数和阳性率的相关系数最高（r=0.808,P<0.05）,阳性率向前移动1周,与“发烧”百度指数的相关性增加（r=0.827,P<0.05）。结论 关键词“发烧”的百度指数能较好的反映和更早的预测流感病毒的活动。

Correlation between Baidu index of specific keywords and influenaz epidemics

Objective To analyze the correlation between Baidu index of specific keywords and influenza epidemics and to explore the utilization of the index in influenza surveillance.Methods Data on reported influenza like illness (ILI) cases and weekly positive rate of influenza virus isolation among the cases from October 2014 to March 2015 were collected form influenza surveillance sentinel hospitals in Tianjin city and 266 laboratory test confirmed influenza cases were surveyed with a questionnaire.Search queries related to influenza epidemics were gathered from Baidu database for the calculation of Baidu keyword index.Spearman correlation was used to analyze the relationship between influenza epidemics and Baidu keyword index.Results Among the confirmed influenza cases,65.04% performed internet searches about their early symptoms using search engine and "cold" was a keyword with the highest query frequency (67.72%) used by the cases.There was a highly close correlation between Baidu index for "fever" and the weekly positive rate of influenza virus isolation (r=0.808,P<0.05) and the correlation was more close when a one-week lag Baidu index for "fever" was adopted in the analysis (r=0.827,P<0.05).Conclusion Baidu index for "fever" could be considered as a earlier signal for incidence of influenza epidemic.