Your smartphone, your therapist?

All over the world, mental health is a growing challenge, but is still treated as a second class condition by most healthcare systems.

In the United States, about 60 million people are affected by mental health issues every year, and about 90% of them suffer from anxiety and depression. Anxiety and depression are all around us and take many different forms - including depression triggered by life events like divorce or moving to a new city, postpartum depression, which affects over 30% of new mothers, PTSD for military veterans, or co-morbid depression prevalent with a majority of health disease and diabetes patients.

Despite being widespread, there is still a stigma around mental health. Talking about depression remains a challenge. The onus is on us - as leaders of government and industry, we need to create support structures in our organizations to recognize and support mental health.

60% of patients receive no treatment at all

Nonetheless, our healthcare systems are not ready for this challenge. We cannot train enough psychiatrists or hire enough therapists to ensure access to high-quality care without technology playing an important role. For example, in the US, approximately 60% of those with mental health illness received no mental health services in the last year, and wait times to see a psychiatrist average between 2 months (cities) and 8 months (rural counties).

Inadequate care results in people showing up to the Emergency Room - mood disorders such as depression are the third most common cause of hospitalization in the U.S. for both youth and adults aged 18 to 44. Doctors give most of the care in anxiety and depression, but typically end up relying on medication due to lack of other options.

We believe the combination of intelligent sensing, mobile phones and high-quality clinical care will help elevate mental health to a 1st class citizen in our national healthcare systems. Below are three pieces to this complex puzzle that we believe will help to improve care, without runaway costs.

1. Determine when users need help using better sensing and machine learning

A fundamental challenge in anxiety and depression is sensing how patients are doing - unlike measuring a certain form of hemaglobin in diabetes, or using an electrocardiogram (ECG) for heart disease, most data in clinical practices today is self-reported by the user. Even that information is collected infrequently. Fortunately, recent advances in mobile sensing technology are a boon for more frequent collection of objective behavioral data. More importantly, such data allows us to understand how mental health behavior changes are reflected in everyday behavior.

In academic settings, smartphones have been used to learn about social networks, characterize human mobility patterns, measure physical activity and health, and estimate the spread of potential contagious airborne diseases. Our research at MIT, as well as the work of other academic researchers such as David Mohr and Andrew Campbell demonstrates the vast potential for the application of mobile sensing to mental health.

For example, it was hard to detect when patients were isolating themselves due to their depressive state. However, with these digital crumbs of information, we can detect such an event much more easily. In reality though, the behavioral patterns indicating a patient’s state are much more complex. Depressive symptoms like lethargy and anhedonia are seen in a person’s behavior - they may have challenges getting to work, school or social activities (movement patterns), the way they interact with others changes (communication patterns), their sleep cycle is disrupted (smartphone usage). Millions of data points about a person’s behavior captured using their smartphone, help us model and predict their health state.

At my company Ginger.io, we been able to screen over half a million people with anxiety and depression, and and have gathered over 600 million hours of data from them across many behavioral dimensions. Such a large and rich data set allows us to use machine learning to detect complex patterns of behaviors that predict when a user may need help. To create the best predictive models, we have built an analytics pipeline that has allowed us to explore over 2000 different models as our approaches evolved. These behavioral patterns are not universal and our approach allow us to customize predictions based on individual user behaviors. We are able to continuously improve our models and learn from errors, because they are used by clinicians and health coaches to provide timely interventions every day.

Understanding the state of the patient allows us to personalize care and intervene at the most appropriate time. Further, we may be able to intervene early and keep patients from reaching a severe state. Such timely care will not only keep populations healthier, but also helps reduce healthcare costs.

2. Your mobile phone as a digital therapist

Over 90% of Americans report that their smartphone is within an arm’s length distance (~3 feet) most of the time. Our generation wants an effortless, on-demand consumer experience in healthcare - just as we use our smartphones to balance our bank accounts or order groceries, we want to get better from the freedom of our couch.

Some parts of the healthcare establishment sees this change in consumer expectations and behavior as a burden - better screening and on-demand access require more clinicians and can increase healthcare costs.

But we believe that is a limiting perspective - as consumers change their behavior, they also desire care in forms that go beyond the 50 minute therapist appointment every other week. We have an incredible opportunity to design interventions and tools to help users where they seek help today - likely on subreddits and Facebook, and not in the doctor’s waiting room.

Ginger.io users learn how to build skills based on clinically validated strategies such as cognitive behavior therapy, mindfulness, and dialectical behavior therapy. These approaches have been shown to be as effective as medication. Our Calm Down Kit is a practical implementation of dialectical behavior therapy - and one of our most popular features, which includes access to a 24x7 nurse hotline. However, six times as many users would rather watch puppy videos to relax than speak to a trained professional nurse.

We invest in scientific validation of these new approaches through research studies and trials. For long-term success, it's critical that these new classes of digital interventions are held to the same scientific bar as existing therapies and medications for symptom improvement. This will ensure we create new efficiencies and scale access in digital mental health.

3. Keeping trained health professionals in the equation

In our drive to improve care and accessibility for patients with depression and anxiety, we also recognize the fundamental importance and value of trained mental health professionals. Mental health professionals, from primary care providers and psychiatrists prescribing medications, to therapists and behavioral coaches providing therapeutic relief, are critical to the success of the patient. The challenge we faced was how to include human care in a scalable manner.

One successful approach is a collaborative care model, where a care manager, or Ginger.io health coach, is able to connect a patient to the right sort of health care provider at the right time, thus playing a pivotal role in balancing patient needs with available resources.

Furthermore, our behavioral care team has continuous access to patient reports that bring together vital information collected through the Ginger.io app to better understand patients’ mental health. Rather than just an occasional trip to a therapist, data provides continuous insights on a patient’s state of mind.

As we develop such novel healthcare technologies, it’s important to design for user privacy. It’s important that users opt-in get and direct value from such tools, which in digital health translated to getting quality care.

As long as patients’ needs lie at the heart of new digital tools, they can give us a much-needed revolution in mental healthcare. Data has the power to make healthcare personalised and proactive, to scale our clinical resources, and help us address an unacceptable shortfall in today’s healthcare system.