A Reality Check on Mobile Mental Health Apps with Dr. John Torous — The 5Q
In the 5Q, Tincture sits down with leaders to discuss their day to day work and share their perspectives on healthcare, medicine, and progress.
John Torous, MD is a staff psychiatrist and fellow in clinical informatics at Beth Israel Deaconess Medical Center, Harvard Medical School. He serves as editor-in-chief for JMIR Mental Health and currently leads the American Psychiatric Association’s workgroup on the evaluation of commercial smartphone apps, co-chairs the Massachusetts Psychiatric Society’s Health IT Committee, and serves as co-director of the BIDMC Psychiatry Digital Care, Assessment, Research, and Education (D-CARE) program.
You are a man who wears many hats: Psychiatrist, Researcher, Editor, Technologist. Across your work, what is the central problem are you trying to solve?
I am trying to improve patients’ mental health outcomes. The United States’ suicide rate recently reached a 30-year high. Worldwide, mental health conditions sadly remain one of the most disabling and costly illnesses. There is an increasing realization that we need to look for new solutions and research new tools to help connect individuals with high quality care — and the simple math that we don’t have enough mental health providers to deliver that care solely face to face or in person. Thus technology solutions like smartphones, wearables, and virtual reality have garnered much attention as potentially useful tools to improve mental health outcomes.
However, it is not as easy as just building an app (just like offering running shoes to people will not make them marathon runners) and there is a need to approach technology for mental health like we would with any other new medication, therapy, or test — with careful science, early patient involvement, focused clinical studies, and transparent data. Creating well validated technology solutions for mental health, based on a foundation of trust, offers the chance to improve clinical outcomes but is a large task — yet one the field is capable of.
How close (or far) are we from a “black bag” of vetted apps for behavioral health care providers to prescribe to patients?
There are some useful apps currently available today that many patients would benefit from — and there are also some frightening ones that no one should ever use. Jennifer Nicholas of the Black Dog Institute in Australia published a paper looking at the quality of apps for bipolar disorder on the Apple and Android app stores. The results were not encouraging and to use the words from her paper, “the content of currently available apps for bipolar disorder is not in line with practice guidelines or established self-management principles.” She even found one app telling patients to drink hard alcohol if they had trouble falling asleep. Researchers have revealed similar findings for apps offering to help with substance abuse, suicide, anxiety, and depression.
We know there are many great apps out there, but there are a lot more unhelpful and even dangerous ones out there too. To face this challenge, we created a task force with the American Psychiatric Association focused on app evaluation. Rating these apps is tricky because they can update and change so quickly (often every month) and thus we are currently focusing on creating algorithms and flow sheets that both patients and clinicians alike can use to ensure they make informed and educated decisions about using an app. So while we don’t yet have a black bag of vetted apps, we hope we are close offering tools to help others make the optimal decisions about what they want to put in their own black bag of apps.
Can you comment on the recent surge of interest in personal analytics platforms that claim to turn passively collected smartphone data into actionable insights? From what you’ve seen, is the technology behind today’s products up to the task?
Passive data holds exciting potential for all of healthcare, and especially mental health. The idea is so simple that it is almost alluring. Our phones have a myriad of sensors in them that record information such as location, voice, ambient light etc, and that information may also tell us something about mental health. For example, someone suffering from a depressive episode perhaps may not leave the house or make as many phone calls as when they are feeling better — and the phone can automatically pick up those signals. Perhaps the phone then automatically calls the patient’s therapist and ensures help is on the way.
Unfortunately it is not that easy and research has shown that passive data is a vast new frontier for mental health that we are only just learning about. In part because diseases like depression manifest so differently for each person, it is hard to know what passive data means for an individual’s mental health (Maybe the person did not leave their home because they are having friends over or finishing a home repair project).
More clinical research on passive data is being published and the results are interesting but still not conclusive. In some ways, this is a problem of trying to fit modern data into older paradigms and definitions of mental illness. Perhaps a better way forward for the field is to stop trying to fit squares into circles and rather treat passive data with the scientific respect it is due. Let’s conduct the clinical studies and learn what it really means and what it can really tell us.
While many will claim that today they can transform passive data into clinical insights, they are often selling something and details on how they do such are always a little too vague. But there are several promising research efforts that are taking a more scientific approach to passive data (e.g. http://www.monarca-project.eu/ or https://www.hsph.harvard.edu/onnela-lab/ or http://www.cbits.northwestern.edu/ or http://www.mh4mh.org/ or our own group at Beth Israel Deaconess Medical Center http://www.psych.digital) and following progress here will offer a better picture of where passive data for mental health is currently at.
For many patients, access to mental health care remains a challenge. Can online or mobile-driven cognitive behavioral therapy (CBT) programs work if designed correctly? Looking back over the last couple of years, is the industry moving on the right track to deliver these solutions?
We know that psychological treatments like cognitive behavioral therapy (CBT) are some of the most valuable treatments that the mental health field can offer. Yet accessing these treatments can be difficult and thus the potential of online and mobile CBT has been well discussed. There is actually a strong clinical evidence base for computer based CBT and many studies have shown it achieves results on par with in person therapy. But there is a catch. In clinical studies of computer based CBT, research subjects are often encouraged to complete all the sessions of therapy and often have many phone or even in person check-ins to keep them motivated.
In real life, people may start a CBT program but without extra support they rarely make it through and instead quickly lose interest. There is less research on CBT delivered with a smartphone although some studies have shown encouraging results. We are still learning how to design these programs and apps correctly. For example, the features that may be the most popular in some therapy-based apps may not actually be the most helpful. Currently we are seeing many efforts to take what works for in person therapy and translate it directly onto an app. The next wave of therapy apps will be smarter and work to utilize the unique features of mobile technology while designing around its limitations.
What advice or guidance would you offer an entrepreneur trying to start a digital behavioral healthcare company today?
There is tremendous potential for digital behavioral health. Pioneering entrepreneurs who have been first to market have shown us how difficult this space can be. Appreciating that difficulty, instead of ignoring it, can lay the foundation for successful efforts in the digital behavioral health space. For example, consider all the various stakeholders that a successful app needs to appeal to such as patients, clinicians, policy makers, hospital leadership, payers, etc.
While it is important to design around patients needs — keeping the needs of how the technology will integrate into the healthcare system is also critical. Also, as the space becomes more crowded and more companies enter, offering clinical data to show that the technology solution actually works will become increasingly valuable. That said, mental health is a vast space and there are many still unfilled niches that may benefit from the efforts of bold new companies willing to take risks and do the science.