The Impact of Internet Coverage of the March 2011 Japan Earthquake on Sleep and Posttraumatic Stress Symptoms: An International Perspective

To the Editor: On March 11, 2011, a 9.0 magnitude earthquake and tsunami struck Japan, killing 14,000 people and damaging nuclear and petrochemical plants. The immediate and intense media coverage exposed viewers across the world to disturbing images. While television coverage of a nearby disaster has been found to increase the risk for subsequent psychological distress (1), little is known about the effects of Internet coverage of a distant disaster.

We conducted an online survey simultaneously in France, Canada, and the United States within 2 weeks of the event. Adult participants (N=698; 42.7% from Canada, 44.6% from France, 8.2% from the United States, and 4.5% from “other/unspecified” countries) were contacted through online mailing lists and snowballing procedures (i.e., participants were also asked to circulate the link to the survey among their contacts) and asked to report 1) the time they spent viewing television and Internet coverage of the event during the first week afterward; 2) their immediate reactions after learning the news (i.e., peritraumatic distress [2] and dissociation [3]); and 3) any disruptive nocturnal behavior (i.e., trauma-related sleep disturbances such as nightmares [4]) they experienced during the first 10 days after the event.

After the earthquake and tsunami, most participants (64.9%) had increased their media consumption. The amount of television and Internet viewing correlated both with symptoms of peritraumatic distress and dissociation (r>0.22 and p<0.001 in all cases) and with disruptive nocturnal behavior (r>0.17 and p<0.001 in all cases), while 45% of the participants reported at least one disruptive nocturnal behavior. Being female, knowing someone in Japan, and figuring in the amount of time spent on the Internet each predicted at least one disruptive nocturnal behavior in a logistic regression (Table 1). In the second step, peritraumatic dissociation and distress significantly predicted disruptive nocturnal behavior; however, the time spent on the Internet became nonsignificant, suggesting a mediating effect of peritraumatic reactions. This was confirmed by a multiple mediator analysis revealing that the direct effect of Internet viewing on disruptive nocturnal behavior was not significant, while indirect effects through both peritraumatic distress and dissociation were significant (p<0.05), suggesting that peritraumatic reactions might explain the relationship between Internet exposure and disruptive nocturnal behavior. Replicating the analyses separately by country yielded similar results. As a follow-up, in a subsample reassessed 2 months later (109 individuals who provided their e-mail addresses), similar analyses examining predictors of posttraumatic stress symptoms (5) revealed a significant effect of Internet viewing that was mediated by peritraumatic reaction.

Limitations of this study include the convenience sample, the lack of an assessment of disruptive nocturnal behavior and posttraumatic stress symptoms before the event, an expectedly low level of symptoms, and the possible response bias. However, the results suggest that Internet coverage of a distant disaster may induce sleep disturbances and subclinical psychological symptoms of posttraumatic stress in the general population. Such effects, consistent with previous reports (6, 7) of posttraumatic stress disorder symptoms after trauma exposure, were mediated by peritraumatic reactions.