Author of the newly published book, Reading Our Minds: The Rise of Big Data Psychiatry (Spring 2021), clinical psychiatrist and pain physician Daniel Barron, MD, PhD, speaks to Psycom Pro Executive Editor Angie Drakulich about how he hopes to reshape the field of psychiatry with quantitative approaches. In the fall, Dr. Barron will join Brigham and Women’s Hospital as Medical Director of its new Interventional Pain Psychiatry Program, and as a faculty member at Harvard Medical School. Dr. Barron earned his medical degree and completed his psychiatry residency at Yale University; he is currently completing a Pain Medicine fellowship at the University of Washington.

Here, he talks about his ongoing research into data-driven psychiatry. Read the edited transcript* or listen to the full conversation on our Psycast playlist below.

Annual Mental Health Screenings

Reading Our Minds: The Rise of Big Data Psychiatry

Psycom Pro: In your book’s opening letter to the reader, you point out that patients are commonly subjected to temperature, weight, and blood pressure reads, and, when necessary, imaging and lab workups – but mental health is missing from this regimented battery of assessments. Why do you think mental health – as a checkup of sorts – is not addressed as part of standard healthcare in the US?

Dr. Barron: I think some of this is changing already. For example, JCAHO (The Joint Commission) is one of the regulatory bodies across the United States that accredits hospitals to allow them to take care of patients, and they have already started requiring certain depression and suicide questionnaires for every patient evaluation. And that’s how we were trained in Connecticut.

However, the depression and suicide screening questionnaires that we use are symptom-based, which is to say, very different from quantitative objective measures like temperature or weight. The field is trying to figure out what may prove useful in the workup of a patient. And one of the things I was thinking through in my book was what sorts of quantitative measures might be helpful in patient evaluation. Once we decide what to measure, who pays for those measurements becomes a question of whether or not the measures prove to be useful in terms of patient outcomes.

In the coming months, Psycom Pro and its sister site Psycom will be focusing on current mental health screenings in the US, where the gaps lie, and how we can finally put mental health first. See what our expert roundtable has to say.

Personal Data Tracking and Patients

Psycom Pro: Some of the digital solutions in your book include personal data tracking, for example, monitoring social media in a patient with depression, tracking GPS location for someone with a substance use disorder, or watching movement patterns in a child with ADHD. How do you see this data collection benefiting clinicians and how do you think patients will respond in terms of privacy?

Dr. Barron: Every time a patient comes into the clinic or emergency room, that whole interaction is a data-gathering procedure. Clinicians are asking patients questions related to potential diagnoses or treatments, and they’re observing the patient. All of this is about data and the reason why that’s important is because good clinical decisions require really good data – so the longer it takes you to get the necessary data, or rather higher quality data, the longer that interaction will be. So these sorts of technologies, the social media and GPS activity, could be very helpful forms of clinical data in terms of being able to detect or measure the efficacy of a treatment.

Regarding patients, I think they will be understandably concerned about data sharing and privacy, especially now during the COVID pandemic when a lot of online data has been in the news, including who’s following our search histories. These are common forms of marketing surveillance, and then there is security surveillance which the governments do – so incorporating healthcare surveillance into this could help to sculpt the future.

Using Big Data to Make a Diagnosis

Psycom Pro: You talked about using data for treatment outcomes and clinical decision-making. Your book also provides examples of using data to flesh out a diagnosis or differentiate a diagnosis. Could you share one example that resonated with you?

Dr. Barron: Sure, one of the more fun examples is not specific to psychiatry but provides a framework for understanding where I think psychiatry might go in the future.

About 100 years ago or so, there was a condition called dropsy. A patient’s hand or leg would all of a sudden become very swollen. Today, we call this edema, but it took a while for clinicians to drill into dropsy to understand what was going on.

As it turns out, there are many different ways you can end up with dropsy, which is essentially a symptom or clinical sign – something you would observe clinically. The presence of dropsy was useful to indicate that there might be an underlying problem but in terms of actually treating an underlying disease process (such as heart, liver, or kidney disease), the presence of dropsy was not very helpful.

The only way that the medical field understood dropsy was by taking very specific quantitative measures of different organs and being able to separate which organ might have the problem. That’s exactly what happens today in medicine. A patient shows up reporting swelling, a clinician will perform a battery of tests and based on those tests, will home in on which organ system might be involved.

In psychiatry, there have been decades of effort trying to better understand and describe with language, the sorts of phenomena that our patients report. I think we’ve exhausted the utility of that approach and there’s a large conversation right now over how best to organize diagnoses. In the same way dropsy was tackled, progress in mental illness diagnoses won’t happen until we develop quantitative measures to better understand and measure those diseases.

An example specific to psychiatry would be an affective disorder. Depression and mania are often seen as opposite ends of the affective spectrum, with mania as having way too much energy and depression as having low energy. You can see that across a couple of different axes.  So right now, we ask patients about their symptoms but we really don’t have quantitative ways to assess things like energy activity – and to me, that would be pretty low-hanging fruit. We should be able to measure psychomotor activity for instance to have a testable hypothesis.

The Guided Clinical Interview

Psycom Pro: You noted that the psychiatry field has perhaps exhausted language as a way of obtaining a diagnosis. Professionals in the behavioral and mental health field have screening tools and questionnaires but they also use interviewing and conversations to get to the root cause of a patient’s condition. Your research and suggested approach focuses on straight data, with less talking. Is that an accurate description?

Dr. Barron: I guess I wouldn’t view it necessarily as less talking but more as guided talking.

Clinicians in every specialty still meet with patients in person, speak to their patients, and are empathetic toward their patients.

Take an oncologist, for example – even if all of their treatment decisions are based on molecular profiles of tumors, which is to say that the real actionable data that happens in those sorts of treatment conversations isn’t based on language, but rather it’s based on a tissue sample. Because of the molecular profiles, what conversations take place are very different and more precise. That’s where I see psychiatry going – using tools and instruments to guide conversations with patients. The goal of writing my book wasn’t to say that conversation was not important but that we have other forms of data that might prove more helpful.

When Pain and Mental Health Overlap

Psycom Pro: Speaking of different fields, you are a trained psychiatrist and a pain physician. You likely see a lot of patients with overlapping physical and mental health disorders, such as depression combined with musculoskeletal or nerve pain. How do you see a data approach unpacking these complex comorbid cases?

Dr. Barron: The greater the complexity, the more we need data to help guide our decisions.

One of the general misconceptions about big data is that it adds complexity, whereas I think it actually helps reduce complexity by helping us focus on what is the most useful information for a specific decision. So, in pain medicine, the complexity is very similar to psychiatry because pain physicians haven’t developed quantitative measures to diagnose or stage disease progression. The burden remains on the patient to be able to describe in words what they’re feeling and to remember the sense of that pain over months, if not years, in terms of development of a pain condition.

Having access to tools and instruments that could help us better quantify the stages of these illnesses is what I think will be really helpful. Curiously enough, I think the same instruments that I described in the book being related to psychiatry would also prove to be very helpful in pain medicine.

(More on chronic pain disorders and mental health overlap, and pain and PTSD

Adding Big Data to Your Psychiatry Practice

Psycom Pro: Let’s shift gears and talk about time management. As you know, there is a shortage of psychiatrists and mental health professionals in the US today. If your concept were to move forward, additional tools and time for analysis would be added to the treatment decision-making process.  How do you see that fitting into the clinician’s already overloaded schedule?

Dr. Barron: I think a large part of the clinical burden is not knowing where to direct one’s attention. Consider again the case of dropsy – a patient comes to your clinic reporting symptoms of dropsy. Without specific tests and tools to work up what was the cause of the dropsy, that patient interaction would be much longer, much more onerous and most likely, less effective.

Having better ways of approaching and understanding data would, in my mind, only make clinicians more efficient and more effective in what they’re already doing.

See also, how digital mental health is reshaping clinical psychology and how to start a practice.

Part of your question suggests having access to this sort of care. The tools I was describing benefit from the fact that almost everyone has a smartphone, which means almost everyone has access to the measuring devices we’ve been talking about. More patients may be able to produce data that their clinicians can then process. Of course, that’s if my work proves to be clinically useful – a lot of this is in the research phase right now but I feel like the promise is there.

Psycom Pro: What about costs in terms of payers taking this on? Even with regard to smartphones, they are not accessible to every patient population, such as those living in rural or low-income economic areas. Any insight or ideas on who may take on the costs of data-based treatment?

Dr. Barron: I think it’s a pretty straightforward question if you view it in terms of efficacy. If a tool proves to be really useful and good at doing a particular function, then it’s going to be reimbursed.

Costs are directly related to patient outcomes and so, if these digital tools prove to be useful enough that they positively affect patient outcomes, I think insurance companies will very much be interested in that.

The Future of Data Science

Psycom Pro: I’d like to highlight a quote from your book that looks forward and backward:

Psychiatry is in a similar position to where cardiology was after Roosevelt’s death… It remains unclear which data will help unravel the complexity of mental disorders and, therefore, inform novel treatments. ….”

Could you elaborate on this?

Dr. Barron: Overall, everyone agrees there’s a lot of progress to be made and that there are many things that can be improved in mental health care. So I think the question isn’t so much whether psychiatry must improve but how this improvement may be best realized. Even though we’re spending millions of dollars and putting out thousands of publications, we still haven’t found a better way to guide clinical practice. Clinical psychiatry occurs in very much the same way as it did 100 years ago, where a clinician will sit down and talk to a patient and based on that conversation, develop a treatment plan.

Cardiology and other disciplines of medicine have progressed by having access to quantitative measures that they’re able to study, evaluate, and demonstrate whether or not they’re helpful to patient outcomes. Psychiatry is only now beginning to adopt that sort of model at large, in part because these technologies (eg, data tracking on a phone in your pocket) simply didn’t exist even a decade ago.

The hope is that we can begin to think in different ways, make use of the technologies we have, and bring these technologies into clinical practice to benefit patient care.

Psycom Pro: Before we close, if you could create your ideal diagnostic toolbox for mental health utilizing big data and other methods, are there are a few things that you absolutely have to have in there?

Dr. Barron: I’m actually trying to design a platform to use in the clinic that I’ll be starting in the fall. What I like about using smartphones as a healthcare data collection device is that once you have the ability to collect, say, accelerometer data, you also have the ability to collect search history or social media data, or geolocation, etc.

Another thing that I’ve been spending a lot of time trying to think through is how to collect the sort of conversational data that is already being created at every client interaction.

As I mentioned, clinical interaction is a data-gathering procedure. When I’m observing a patient, I’m looking for specific forms of data right like how they’re moving, how they’re speaking, etc, that I then include in my clinical reasoning. Having a way to video record that clinical interaction and process that conversational data is something I started researching as a resident and I’m working on the first paper now (check back here for links to Dr. Barron’s papers). But, if I could pick another assessment tool to have right now, it would be a functioning video camera and microphone that can adequately track a patient’s face during a clinical interaction.

*Transcript edited for style and clarity

About Daniel Barron, MD, PhD

Daniel Barron, MD, PhD

Daniel Barron, MD, PhD

More about Dr. Barron and his book, Reading Our Minds: The Rise of Big Data Psychiatry, on Follow  him at Twitter@Daniel_ _Barron





Last Updated: Jun 10, 2021