Sunday, May 29, 2016

Does Google know about a suicide before it happens?

According to the CDC, suicide is in the top 10 leading causes of death for people in the United States between the ages of 10 and 64. Among individuals in age groups 10-14, 15-24, and 25-34, it is the 2nd leading cause of death. In the year 2014 alone, we lost 42,772 Americans to suicide. Truthfully, this is probably somewhat of an underestimate because suicide is often miscategorized by as "unintentional injury," leading to false records.

If you are a clinical psychologist, you spend a lot of time thinking about ways to detect when a patient is entertaining thoughts that life is no longer worth living. Unfortunately, many completed suicides are unpredictable, and occur in moments where extreme hopelessness intersects with impulse and access to means. As a field there are many brilliant scientists and clinicians working on ways to identify and help individuals at this extreme of human suffering.

Among them is Dr. Christine Ma-Kellams, University of La Verne, and colleagues who were interested in understanding whether Google search trends can be used to predict suicide, and whether these trends are more effective in predicting suicide rates than our existing measures.

To answer this question, they pulled together data from several different sources. First, they found data from the CDC National Vital Statistics System on the number of completed suicides in the United States. From the U.S. Census Bureau, she collected demographic data that included information like income, population, home-ownership rates, unemployment, and percent of the population under the poverty line,  age, and racial categories. From the National Survey on Drug Use and Health, they collected nationally representative data on suicide vulnerability as reported on the existing gold-standard, clinical measures for suicide risk. Finally, from Google trends, they recorded the relative frequency of google searches for the terms "suicide," "how to suicide," "how to kill yourself," and "painless suicide" compared to the search term "weather." All of the data used in the study were from the years 2008-2009.

They found that the frequency of these Google search terms was significantly associated with the rate of completed suicides recorded by the CDC. They also found that frequency of these search terms was more predictive of suicide rates than the existing self-report measures we use to estimate suicide risk.

The Google search terms weren't perfect, though. They were less effective at accurately predicting suicide rates in states with lower incomes, higher crime rates, and a larger minority population. Also, it's important to acknowledge the limitations of this study. Even though this data was pulled from many different sources, is nationally representative, and cover two years, there is no way for us to know which direction the effect is going. We think these data mean that people are searching for "how to commit suicide" and then those same people are completing suicide, but it is just as plausible that individuals completed suicide, and then people in their community went online and searched for these terms. It is true that a single suicide in a community can inspire increases in discussions of suicide among the members of that community, but either way the problem to be solved is the same. Find a way to help people who feel like life isn't worth living, and prevent suicide. Google can help us find those people.

So, what does this mean? Google knows where you've been, where you're going, what you want, and how you want it. As it turns out, Google also knows who is thinking of committing suicide. Knowledge is power, and here power is life. Google is already implementing the use of sponsored ads for suicide hotlines that target individuals searching for terms just like the ones in this research study. But we are only at the beginning of understanding how to leverage this type of data in ways that can save lives. For example, can we target specific communities in the wake of a tragedy or disaster when suicide rates increase? Can we create sophisticated programs for online chatting for people going through a moment of hopelessness? Can we use the data to identify communities for whom more mental health resources would prevent these feelings of hopelessness? What ideas do you have about how to harness the power of the internet to reduce suicide rates?

Need Help? Know someone who does? Contact the National Suicide Prevention Lifeline
at 1-800-273-TALK (1-800-273-8255) or use the online Lifeline Crisis Chat.  You’ll be connected to a skilled, trained counselor in your area. Both are free and confidential. For more information, visit National Suicide Prevention Lifeline

Ma-Kellams, C., Or, F., Baek, J. H., & Kawachi, I. (2015). Rethinking Suicide Surveillance Google Search Data and Self-Reported Suicidality Differentially Estimate Completed Suicide Risk. Clinical Psychological Science, 2167702615593475.

Photo credit: Garrett Sears via

No comments:

Post a Comment

Get the next post via email:

Believe in our mission too?