Original ResearchGoogle and suicides: what can we learn about the use of internet to prevent suicides?
Introduction
In 2014, suicide was the ninth leading cause of death for the American population and the second most important for youth (aged 20–24 years). Overall, 42,773 people committed suicide in 2014 of whom 3245 were aged between 20 and 24 years.1
With people spending on average more than 20 h online per week,2 the internet represents a potentially important outreach tool to prevent suicides. It is a particularly well-suited platform to discuss such taboo topics, because it provides total anonymity.3 In spite of the growing importance of this media and its potential role to prevent suicides, little is known about the relationship between online searches and the number of suicides.
First, the number of suicides could be uncorrelated to search propensity, because people searching for suicide terms have no intention to commit suicidea or because suicidal people take action without consulting the internet. In both cases, resources devoted to an online presence should be channelled towards other outreach strategies.
Second, the number of suicides could be negatively correlated to search propensity. Suicidal people or their relatives could use the internet to look for help. If websites like those appearing when googling ‘I want to kill myself’ do indeed prevent suicides, the propensity to Google suicide-related terms would be negatively correlated with the number of suicides. Such a negative coefficient would be consistent with the fact that suicide prevention organizations are successfully preventing suicides.
Third, if these websites are unable to prevent at-risk individuals looking for help from committing suicide, the relationship between search propensity and the number of suicides would be positive. In such a case, websites designed to prevent suicides may need to reconsider their outreach strategy. Moreover, such a positive correlation could be explained by the fact that at-risk people tend to consult pro-suicide websites following their internet search4 or could be due to friends or relatives searching for suicide-related terms just following a suicide as a coping mechanism.b
Results from previous studies provide no consensus on this relationship partially due to their small samples. Some studies find no significant relationship,c whereas others do find a positive relationship.d
This article contributes to this literature by considering a large sample consisting of monthly data for 50 states during the period stretching from January 2006 to November 2014 (5388 observationse). This sample is more than 50 times larger than the samples used in previous studies, and therefore enables a more thorough analysis. Moreover, this article studies a later period which has not been considered in previous work. With the increasing use of the internet, more up-to-date studies are necessary. Finally, this article is the first to consider the suicides of different groups to make a first attempt at understanding who is performing such searches.
Overall, there is a positive relationship between the contemporaneous search intensity and the number of suicides: a one percentage point increase in search propensity is related to a yearly increase of 54 suicides in the United States. The magnitude of this relationship is particularly strong after 2010 possibly explaining the non-results of previous studies conducted before 2010. Moreover, overall search propensity is a statistically significant predictor for the number of youth (aged 16–35 years) suicides and for the number of male suicides.
The rest of the article is divided in the following way. I first present the data. I then describe some the methodological challenges. Finally, I show the results and discuss briefly their implications.
Section snippets
Data
Data from Google searches is provided by Google Trends. This website supplies the propensity to search ratios by state for any keyword or combination thereof. These ratios are relative to the period with the highest search ratio for a given state. For example, assume that the word ‘suicide’ was googled 1000 times in period 1 in state 1. Moreover, suppose it was entered 2000 times in the search engine in period 2 in state 1. In such a case, Google Trends would report a ratio of 0.5 for period 1
Results
Table 1 provides evidence that search propensity in a given month is significantly positively correlated with the total number of suicides in this month and state. Both strategies (omitting low population states and tobit method) provide very similar estimates.
An increase of one percentage point in the propensity to search for a suicide-related term is related to a 0.0982 percentage points (Table 1, column 6) increase in the suicide propensity relative to the month with the highest propensity.
Discussion
Overall, these results provide some guidance for practitioners. There is clear evidence that online search behaviour is positively related with suicides suggesting that suicide prevention organizations should develop a stronger online presence. This positive relationship is particularly strong for males and for youth suggesting that suicide prevention websites should be designed with the needs of both groups in mind. Since these data are unable to identify the attributes of the person
Acknowledgements
The author wishes to thank Jennifer McMullin for excellent research assistance. All errors are mine.
Ethical approval
The data used is publicly available and can be obtained upon request. There was therefore no need for an approval from the ethics board of my institution.
Funding
I did not receive any money to specifically conduct this research.
Competing interests
To my knowledge, I have no conflict of interest.
References (14)
Internet monitoring of suicide risk in the population
J Affect Disord
(2010)- et al.
Association of internet search trends with suicide death in Taipei City, Taiwan, 2004–2009
J Affect Disord
(2011) National vital statistics reports
(2016)Adults media use and attitudes
(2016)- et al.
Social desirability bias in CATI, IVR, and Web surveys the effects of mode and question sensitivity
Public Opin Q
(2008) - et al.
Googling suicide: surfing for suicide information on the Internet
J Clin Psychiatry
(2008) Does the volume of Internet searches using suicide related search terms influence the suicide death rate: data from 2004 to 2009 in Japan
Psychiatry Clin Neurosci
(2011)