National Institute of Infectious Diseases.
Teikyo University Graduate School of Public Health.
Biosci Trends. 2018;12(6):636-640. doi: 10.5582/bst.2018.01279.
In Japan, national official surveillance for influenza has been performed at about 5,000 sentinel hospitals/clinics by the National Official Sentinel Surveillance of Infectious Diseases (NOSSID). Meanwhile, all electronic medical claims nationwide in the National Database of Electronic Medical Claims (NDBEMC) were recently disclosed by the Ministry of Health, Labour and Welfare of Japan. We compared the regional variation of influenza incidence among prefectures between the NOSSID and NDBEMC. The data were extracted from NOSSID and the NDBEMC for the 2010/2011 through 2013/2014 seasons. We compared the data of both datasets season by season by using Spearman's rank correlation in each season. Spearman's rank correlation values for the four seasons were 0.7823, 0.3907, 0.4961 and 0.4543, and their p-values were less than 0.00005, 0.0066, 0.0004 and 0.0013, respectively. Statistically, regional variation of influenza incidence in NOSSID is not imprecise, but its correlation with the NDBEMC dataset is not so high. It is important to note this fact when interpreting regional variation in NOSSID.
在日本，国家官方流感监测由全国传染病官方监测哨点监测系统（NOSSID）在大约 5000 家监测医院/诊所进行。同时，日本厚生劳动省最近公布了全国电子病历数据库（NDBEMC）中的所有电子医疗索赔。我们比较了 NOSSID 和 NDBEMC 中流感发病率在县之间的地区差异。从 NOSSID 和 NDBEMC 中提取了 2010/2011 年至 2013/2014 年四个季节的数据。我们使用每个季节的斯皮尔曼等级相关系数来比较两个数据集的季节性数据。四个季节的斯皮尔曼等级相关系数分别为 0.7823、0.3907、0.4961 和 0.4543，其 p 值分别小于 0.00005、0.0066、0.0004 和 0.0013。从统计学上讲，NOSSID 中流感发病率的地区差异并不不准确，但与 NDBEMC 数据集的相关性并不高。在解释 NOSSID 中的地区差异时，注意这一事实很重要。
