In March, when Netflix began streaming its original teen suicide mystery series 13 Reasons Why, it took a few days for people to start freaking out. But soon, schools started sending home notes warning parents about the show’s graphic depictions of suicide and rape. Psychologists wrote op-eds denouncing its disregard for the World Health Organization’s suicide portrayal guidelines. News outlets published more than 600,000 stories about it. And then there was Twitter.
People tweeted about the show more than 11 million times in the three weeks following its premiere (more than any other show so far in 2017). They wanted to know why the show never touched on the subject of mental illness and why it chose to turn teenage turmoil into a glorified scavenger hunt. And they worried that the show would spread a suicide contagion, inspiring copycats and nudging those close to the edge.
Into this fray entered John W. Ayers, a computational epidemiologist who studies public health at San Diego State University. He saw a chance to inject some empiricism into the controversy.
“There was a tremendous amount of debate going on all based on deeply personal experiences that wasn’t going anywhere,” he said. “We saw a need for real data.”
So on Monday, Ayers and his collaborators at Johns Hopkins, the University of Washington, and the University of Southern California published the first data on whether 13 Reasons Why is bad for public health. And according to them, the answer is yes. But that’s not necessarily cause to cancel your Netflix subscription.
Ayers leads a multidisciplinary team of scientists who take big data not traditionally used in public health and turn them into rapid insights. They’ve done things like use Twitter data to call attention to a spike in distracted driving incidents thanks to Pokémon Go players behind the wheel. They realized they could help settle the debate over 13 Reasons Why by looking at how the show influenced what people were searching for on Google. For one, it was easier data to get than death certificates (the CDC maintains a free database of deaths by cause, but it only has records through 2015). And previous studies have found correlations between suicide search trends and suicide rates.
So Ayers and his colleagues grabbed search queries from the US between March 31, 2017, the series’ release date, and April 18, a date the team selected as a cutoff because news of former NFL player Aaron Hernandez’s prison suicide might have contaminated the results otherwise.
They looked at all searches containing the word “suicide,” except for those accompanied by the word “squad,” for obvious reasons. And they also analyzed common queries related to but not mentioning suicide. Then they compared the results to a hypothetical scenario over the same time frame had 13 Reasons Why never been released, based on forecasts using historical search trends.
What they found was kind of shocking. Suicide queries went up nearly 20 percent in the 19 days following the show’s release—between 900,000 and 1.5 million more suicide related searches than expected. At first glance, it looked like exactly what the show’s creators wanted—a massive increase in awareness and people engaging with the subject. But, in this case, the details seem to matter as much as the overall impression: of the 16 queries with the biggest spikes, about a third of them were for specific methods for ending one’s life.
“If it was truly raising awareness, we’d see a very different outcome,” Ayers said. While they observed an uptick in queries related to getting help (hotlines, prevention services), it was accompanied by search terms like “quick suicide,” “painless suicide,” and “how to kill yourself.” “That’s troubling,” Ayers said. “Even with the best of intentions, it’s clear this show has real world consequences.”
Not everyone is so sure. The obvious flaw with search data is you can’t attribute intention to every query, said Sean D. Young, who heads up both the UCLA Center for Digital Behavior and the University of California’s Institute for Prediction Technology. Because Google starts automatically populating search boxes, anyone who takes to the search engine out of pure curiosity might wind up spiraling toward suicide voyeurism. Is that fascination morbid? Sure. But it doesn’t necessarily equal suicidal ideation. “It helps you pick up what people have learned about suicide from the show,” Young said. “But it’s important to not get freaked out that the show is causing people to go kill themselves. I don’t see the evidence for that.”
Especially because you could make the same case for anything that draws attention to 13 Reasons Why, like, say, this article, or Ayers’ study. Will suicide-related search terms get a little bump along with the study’s publication? Probably. But it’s pretty hard to go from that to saying it will have caused the loss of human life.
Now, that’s not to say data from search engines, social media platforms, or even wearables can’t be useful for public health surveillance. Both Young and Ayers are building tools that accurately monitor and predict health issues. Both researchers are already collaborating with federal agencies like the Centers for Disease Control and the Bureau of Alcohol, Tobacco and Firearms to validate these models. In a few years, you could imagine the CDC spotting a blip in the matrix—a movie or a show or a video game triggering some kind of public health scare—and stepping in before the damage is done.
For now though, the show will go on. It’s scheduled for a second season starting in 2018, and began shooting in California this summer despite calls from people like Ayers to put production on hold. There’s no word from Netflix on whether or not it plans to implement any changes, either to the series’ trigger warnings, helpline callouts, or the on-camera content itself. (The company declined to comment for this story). In the meantime, 13 Reasons Why is still available on demand to anyone with a Netflix login. And that final scene, yeah, you can rewind it and watch it as many times as you want, even if you probably shouldn’t.