Yoshida K, Mochizuki Y, Fukuyama Y
Hokkaido Furano Health Center.
Nihon Koshu Eisei Zasshi. 1991 May;38(5):324-32.
The existence of temporal clustering of suicides was analyzed by comparing the distribution of weekly numbers of suicides under age 20 in Japan (1982-1987) with a Poisson distribution. The linkage between newspaper reports about teenage suicides and these clusters of suicides was examined in an ecologic study. The following results were obtained: 1) The distribution of weekly numbers of suicides under age 20 was not consistent with a Poisson distribution. The observed numbers for weeks with clustering was higher than the expected numbers. Therefore it is suspected that teenage suicides was not an independent occurrence but clustered in specific weeks with clustering observed in April and between the 15th and 25th weeks in 1986. 2) The age and sex characteristics of those who committed suicide during the observed clusters every April did not differ from that of suicide victims for the rest of the year. The incidence ratio for suicide between 15th and 25th weeks in 1986 compared to the same period for other years was higher in girls than in boys, and higher in younger teenagers than older teenagers. Analysis by prefecture of suicides which occurred between the 15th and 25th weeks in 1986 showed high incidences for Hokkaido, Tohoku, Chugoku, and Shikoku blocks, and low incidences for the Kanto and Kinki blocks. 3) Twenty news articles reporting teenage suicides with headings much larger than four columns appeared in the Asahi Newspaper in 1982-87. Significant linkage between newspaper reports about teenage suicides and trend of weekly numbers of suicides in 1985 and 1986 was not observed. 4) Weekly numbers of suicides significantly increased following newspaper reports on January 23rd 1985 and on April 8th 1986, describing a suicide caused by battering, and a jumping suicide of a young singer who was popular among teenagers, respectively. However no increase in the number of suicides was observed following other newspaper reports.
通过将日本（1982 - 1987年）20岁以下青少年每周自杀人数的分布与泊松分布进行比较，分析自杀事件的时间聚集性。在一项生态学研究中，考察了报纸关于青少年自杀的报道与这些自杀聚集事件之间的联系。得到以下结果：1）20岁以下青少年每周自杀人数的分布与泊松分布不一致。出现聚集现象的周的观察值高于预期值。因此怀疑青少年自杀并非独立发生，而是在特定周聚集，1986年4月以及第15至25周出现了聚集现象。2）每年4月观察到的聚集期间自杀者的年龄和性别特征与当年其他时间自杀受害者的特征没有差异。1986年第15至25周与其他年份同期相比，女孩的自杀发病率高于男孩，青少年中年龄较小者高于年龄较大者。按县分析1986年第15至25周发生的自杀事件，北海道、东北、中国和四国地区发病率高，关东和近畿地区发病率低。3）1982 - 1987年《朝日新闻》上出现了20篇标题大于四栏的报道青少年自杀的新闻文章。未观察到报纸关于青少年自杀的报道与1985年和1986年每周自杀人数趋势之间存在显著联系。4）1985年1月23日和1986年4月8日报纸分别报道了一起殴打致死自杀事件和一名受青少年欢迎的年轻歌手跳楼自杀事件后，每周自杀人数显著增加。然而，其他报纸报道后未观察到自杀人数增加。
