• Yusra

Mining Social Media Data Can Show Mental Health Trends and Help Prevent Self-Harm

In the winter of 2014, a 24-year-old student became so overwhelmed with anxiety and stress that she began to have suicidal thoughts. Her life was full of upheaval – she had just finished graduate school and was working as a teaching assistant at a nearby university. But it wasn't until a friend found her passed out in her apartment after overdosing on prescription drugs that she was finally diagnosed with bipolar II disorder.


Upon release from the hospital, the young woman was referred to a therapist who helped her better understand her condition and cope with its symptoms. She began taking medication for it, which has successfully reduced her manic highs and depressive lows. Now, two years later, she is looking for work as a graphic designer.


“She's doing great,” says her therapist, Joe Parker. “I think it's because she had this really strong support system and was able to get the right treatment early on.”

While this is just one person's story, and certainly not typical of most people who suffer from this disorder, it should serve as an example of what can be possible. It also underscores the importance of early detection and treatment in mental health care.

Fortunately, there are signs that providers are increasingly using data to better assess their patients' situations and to identify trends that might indicate self-harm or other serious problems.


Innovations in data collection and analysis are helping providers make powerful interventions to help more people, especially those who are most at risk for suicide.

Let's take a look at some of the innovative ways providers are using data to better serve their patients with mood disorders.

Google Trends can predict mental health trends related to suicide rates

A team of researchers in the UK recently analyzed Google search patterns in the United States and the UK to determine how they related to suicide rates. They found that when certain phrases like "how to kill yourself" or "I want to die" were trending on Google, it correlated with an increase in actual suicides.

Their research suggests that cities where these phrases were searched more frequently saw an increase in suicide rates, but that it took a little time – about two months – for the connection to appear.


More and more leaders and agencies within the health care industry are recognizing the value of visual data. These types of analyses can show mental health trends and help providers better assess their patients' situations, identifying trends that might indicate self-harm or other serious problems.


This type of data can also help providers understand whether their outreach efforts are reaching people who need help most, and identify important differences in how conditions affect different groups around the world. Google's “What women want” report, for example, indicated that searches related to anxiety, depression, and menopause are on the rise.

“It gives us great hope to know that people are searching for help with these conditions online – it means they are potentially reaching out for assistance, too,” says Katherine Hull Fliflet, Google's vice president of global marketing solutions. “Searches demonstrate how people are feeling, what they're doing, and the challenges they're facing.”

Knowing this can help providers find new ways to talk with their patients about mental health concerns and encourage them to seek support when needed.

Real-time data collection brings treatment into everyday life

Just like wearable devices can collect information about your daily physical activity and location, data can now be gathered about how someone takes their medication.

For people with bipolar disorder, one of the most challenging aspects of treatment is taking their medications as prescribed. The problem is that they can feel normal between manic highs and depressive lows, so it's easy to forget—or fail to connect—that these highs and lows might be related to their medications.


A new technology called the Personalized Real-time Intervention for Bipolar Disorder (PRINTO) will allow people with bipolar disorder to track their moods, sleep patterns, and medication use as they go about their lives. PRINTO is a mobile phone app that lets patients keep track of their daily routines and moods, which can help them "connect the dots" between how they feel and what they are doing. The data from PRINTO can be shared with a patient's doctor to allow him or her to offer guidance on-the-go.


Insight Center Innovating for Value in Health Care Sponsored by Medtronic A collaboration of the editors of Harvard Business Review and the New England Journal of Medicine, exploring best practices for improving patient outcomes while reducing costs.


It's important to remember that not all mental health conditions are treated equally. When I worked in a hospital, everyone rehabilitating from an injury was treated the same way—the focus was more on the condition of the injured body part than on the whole person. But mental health is different.


The more we can do to recognize and understand how people are feeling both as individuals and as a community, the better our chances of offering targeted solutions that help those who need it most. In order to do that, we have to take a two-pronged approach to mental health data: using it to understand real-time trends and uncovering insights, as well as using it to tailor treatments.


The more we can learn about how people are feeling both as individuals and as a community, the better our chances of offering targeted solutions that help those who need it most.

Both efforts are important because there is still so much we don't know about the brain. For example, many doctors still think of addiction as a moral failing rather than a disease. Although mental health providers can now prescribe medication to treat depression, bipolar disorder, and other conditions like they would any other illness (a practice that has helped reduce stigma), it's still not common for primary care doctors to monitor patients with serious mental illness. And even when they do, their support usually stops once the patient starts feeling better, leaving patients vulnerable during future episodes of depression, mania, or psychosis.


As we continue to collect real-time data on the brain and behaviors that accompany mental health conditions, we need to better understand what it all means. Only then can we use what we learn to personalize treatments and provide the right care at the right time.