Background: Many studies have shown that media reporting of suicide incidents can trigger suicidal behaviours in viewers and readers. Yet little is known about the exact timing and duration of the imitative effects.
Methods: We estimated the Poisson regression model using original data on 109 celebrity suicides and daily suicide counts (n = 8035) in Japan from 1989 through 2010. Various fixed effects were included in the model to control for the effects of seasonal variations and time-specific shocks.
Results: The media reports on celebrity suicides were associated with an immediate increase in total suicides. The total number of suicides increased by 4.6% (95% confidence interval (CI): 2.4-6.7) on the day that media reports on celebrity suicides were published. The increase during the post-report period lasted for about 10 days after the publication of news reports. The average effect of celebrity suicides on total suicides over the 10-day post-reporting period was estimated to be highest when the suicide by nationally recognized politicians was reported (14.8%; CI: 10.9-18.7), whereas reports on the deaths of entertainment celebrities were followed by a 4.7% increase (CI: 2.9-6.5) in suicide counts.
Conclusions: This study presents evidence that media reports on celebrity suicides have an immediate impact on the number of suicides in the general population. Our findings also highlight the importance of responsible and cautious media reporting on suicide.
Keywords: Japan; Suicide; imitation; media; mental health.