Director’s Blog: Look who is getting into mental health research
In the U.S., biomedical research funding has been estimated at $117 billion, following a rough law of thirds: about one-third government (mostly NIH), slightly more than one-third pharmaceutical companies, and the remainder a mix of biotech, foundations, and philanthropy.1 Support for research on mental disorders looks a little different. As pharmaceutical companies invest less in this area, government (especially NIMH) has become a larger fraction of the funding pool. Now there is a surprising new player in mental health research that is just beginning to emerge from the private sector.
This summer I was invited to Apple, Google (now Alphabet), IBM, and Intel. Why are technology companies inviting the NIMH director to visit? At first, I assumed their interest was purely financial. It’s hardly surprising that companies with large cash reserves have discovered a trillion dollar market (health care is now approaching 20 percent of the U.S. gross domestic product and mental health care is a significant part of those costs).2 But I quickly discovered two other factors that are driving tech companies into biomedical and mental health research. One is big data. As genomics, imaging, and large health care studies generate terabytes of data daily, companies that know how to extract knowledge from data have become essential partners for progress towards new diagnostics and therapeutics. The data analytics from tech companies are becoming part of the engine of biomedical research. The other is the promise of technology to change health care, shifting it from episodic to continuous, from reactive to proactive, from physician-centered to patient-centered. Even beyond wearable devices and online cognitive training, technology can offer information and interventions where and when someone needs it. Tech companies are realizing that mental health is, in their parlance, an excellent “use case.” Just as important, online health care (especially mental health care) creates data that can serve to improve quality, including monitoring the fidelity of psychotherapy.3 In the future, when we think of the private sector and health research, we may be thinking of Apple and IBM more than Lilly and Pfizer.
Here are two fascinating previews of this new world I noted during my travels last week.
One was the publication of results from a collaboration between Columbia University and IBM.4 The team, led by Gillinder Bedi and Cheryl Corcoran, was looking for a biomarker to predict which clinically high-risk youth would convert to psychosis over a two- to three-year follow up period from an initial interview. Rather than depend on a protein in blood or a brain scan, they used an innovative big data approach to analyze the speech from the initial interview. The approach, developed by Guillermo Cecci at IBM, maps semantic coherence and speech complexity as a window into the earliest stages of disorganized thought. While analysis of previous clinical features have yielded, at best, 80 percent prediction, this automated analysis of unstructured speech was reported to be 100 percent accurate for identifying who would convert to psychosis during the follow up period. This is a small study (34 participants with 5 developing psychosis), but it serves as a preview of what we might see as the power of technology is applied to provide objective measures of behavior and cognition.
Also last week I visited Ginger.io, an internet start-up housed on the nineteenth floor of an office building in San Francisco. The founders, Anmol Madan and Karan Singh, took me on a quick tour of their smartphone app, which tracks mood and anxiety to “deliver support to the right people at the right time.” Already used by thousands of people and adopted by several research groups, Ginger.io looks at everything from sleep and activity to social interaction and self-report to quantify mood. Their approach has enormous potential not only for research on mood and anxiety but for the development of interventions that can be deployed globally. While the app has mostly been used to link patients to providers, imagine a future when the app will empower patients with tools to become their own providers.
My summer tour of tech companies, large and small, left me with one unexpected conclusion. While the focus of wearable technology and online apps has thus far mostly been for managing heart disease and diabetes, the tech approach may be best suited for mental health. The biomarkers for depression and psychosis and post-traumatic stress disorder are likely to be objective measures of cognition and behavior, which can be collected by smartphones. Some of our most effective interventions are psychosocial treatments that can be delivered or extended by smartphones and tablets. Most important, the sensors and the interventions can be integrated into a closed loop so that care is continuous and iterative. Increasing symptoms, suicidal impulses, and paranoid thoughts lead immediately to an intervention. Population-based studies have shown that less than half of people with mental illness seek care. And workforce studies have shown that 55 percent of counties have no mental health care provider. Technology is not the answer to all problems, but it may help those with mental illness even more than those with other chronic, serious medical conditions.
1 Moses H 3rd et al. The anatomy of medical research: US and international comparisons. JAMA. 2015 Jan 13;313(2):174-89. doi: 10.1001/jama.2014.15939.
2 Moses H 3rd et al. The anatomy of health care in the United States. JAMA. 2013 Nov 13;310(18):1947-63. doi: 10.1001/jama.2013.281425.
3 Atkins DC et al. Scaling up the evaluation of psychotherapy: evaluating motivational interviewing fidelity via statistical text classification. Implement Sci. 2014 Apr 24;9:49. doi: 10.1186/1748-5908-9-49.
4 Bedi G et al. Automated analysis of free speech predicts psychosis onset in high-risk youths. NPJ Schizophrenia. Doi:10.1038/npjschz.2015.30.