My family is slowly morphing into a FitBit commercial. My sister was the initial domino, and we gave my mom a push soon after with this year’s mother’s day gift. While I am not a “FitBitter” myself (the lack of red underline assures me this is in fact a word), I use my Garmin running watch regularly. In addition, my brother and my bikes have often been expensively outfitted with devices to track speed (lower than desired) and miles (less than needed). Our ability to observe those steps and miles and then compete against ourselves and each other has become a kind of familial substitute for a personal trainer.
But let’s consider a contrast case, an industry where FitBitting (again, no red) is seen as less of a threat by insiders. This past Tuesday, I was invited to speak at the NorCal FPA conference. This professional society hosts a wonderful conference of over 600 financial advisors primarily located in the Bay Area. This year’s event was graced by keynote speakers like Christina Romer, Obama’s former chairman to the counsel of economic advisor, and Liz Ann Sonders, the Chief Investment Strategist for Charles Schwab.
In the lead-up to the event, I wanted to know how members of the advising world saw their ‘robo-advisor’ competitors. For those of you less familiar, the robo-advisor model seeks to take the place of a flesh and blood financial advisor. These apps guide your allocation strategy and make decisions based on your expressed goals. The movement is picking up steam. My friend Christie, a Google employee, was recently guided by her employer to Betterment for financial advising. She also uses Learnvest to find financial advice catered to younger women.
One of the common themes across these conversations was that even if the financial advice piece is equivalent, there is a kind of ‘hard conversation’ that is better handled by a real person. In other words, if my advisor wants me to increase my savings amount and bugs me every time we get together until it happens, I might be more likely to do so than if I just a click away from opting out of this choice. Shame and guilt is powerful, and seemingly best delivered live.
Point well taken, but how do I make sense of the growth of FitBit (tracking of behavior, structured goals) and the ways in which these devices seem to be eating into the personal training market? Would the personal trainer of 5 years ago also say that a watch cannot drive behavioral change like they can in a one-to-one meet up?
Putting aside the financial advising market, I want to think for a moment about what automation might have to look like for it to drive the kind of outcomes typically reserved for relational intervention. For the sake of a challenge, let’s take an incredibly complicated area where I have significant interest, albeit little expertise… mental health. As Jerry Seinfeld says, “The less you know about a field, the better your odds. Dumb boldness is the best way to approach a new challenge.” So, here we go.
On Saturday night, I attended the wedding of a friend and former student, Josh Jackson. Sitting next to me was a local psychiatrist. Having participated in some counseling myself, and also serving as a board member of the Pine Rest Foundation, one of the country’s largest stand-alone behavioral health hospitals, I find the field fascinating and incredibly important. Ultimately, my table mate and I ended up getting into an extended conversation on the differences between the various different schools of counseling.
From this conversation and a few other things I have read, I understand the major schools of thought to include psychoanalytic (Freudian and others in response to this movement), cognitive behavioral therapy (CBT), mindfulness therapy, and Dialectical Behavioral Therapy (DBT) for more severe cases. A very simplified way to understand the difference between the first three camps (summarized in this podcast Invisibilia) is that psychoanalysis assumes one’s underlying thoughts have some kind of meaning that should be teased out (e.g. “This is really about your mother or the trauma of early family life”), CBT does not assume their meaning and suggests these automatic thoughts should be put to reason (e.g. “What is the evidence for the feeling you are bad at relationships?”), and mindfulness argues that there is not much pragmatic benefit of focusing on these feelings, regardless of their truth. In practice, a psychoanalytic approach might work at identifying those feelings and determining their corresponding meaning. A CBT approach might get the patient to look at the underlying thoughts, move them outside of their head to assess their truth, and then to look at identify specific non-adaptive behaviors. A mindfulness approach would likely spend significant time teaching the patient the mental skill of observing these thoughts as they come to mind but then letting them fade away without given them power or control.
So what would a FitBit or “robo-advisor” model applied to psychological health look like in a world of Apple Watches, integrated devices and quantified self?
In an interview on Harvard Business Review’s podcast, Evernote CEO Phil Libin argued that the biggest difference in a post-Apple watch world is that companies have to design with the end-user’s needs directly in mind. The designer can not longer just think about how the apps works on the phone, but instead must consider behavior in which technology integrates with the user in multiple places (phone, watch, computer, tablet, glasses? car?). The question then shifts to how data can be collected and insight delivered at multiple different user touch points to drive change. At the core of this user-, or human-centered design approach is letting the “needs, wants, and limitations of end users of a product, service or process are given extensive attention at each stage of the design process.”
So, if I am walking around with a tablet and computer in my bag, a phone in my pocket, and watch on my wrist, all digitally linked together, how might this network be utilized to drive mental health?
Let’s first consider the information that can be captured.
A few newsletters ago, I ran an analysis of my email to observe the times of day, month, and year when I was emailing more frequently. In this work, I discovered a spike of emails in the Spring of 2013 and then a significant drop-off in the next month. I hypothesized that this was a likely a signal or trace of my own burnout. I know that sometimes when I feel overwhelmed, trudging through an expanding list of email can be emotionally draining. It usually makes me realize the number of people I feel like I am letting down, or makes salient the tasks on which I am behind. And so, when I am stressed, I often drop the emails sent off and fall further behind. This observed spike followed by the drop-off was a sign for me that I was feeling this kind of stress and anxiety.
What about these other traces I might capture, and what they might mean…
- The difference between a “haha” and a “ha ha” as a signal of how one is doing
- The frequency by which someone shows up to their appointments late or on time
- The pattern by which one checks social media, and the types of photos one observes and likes
- Physiological data as a sign of how often one is exercising
- Physiological traces of one’s blood pressure and other signals of underlying anxiety
The more we interact with devices in ways that can be tracked, trended and visualized, the richer our potential insights on what is ‘really going on.’ The areas above are behavioral (text, meetings, email) and physiological (heart rate), but you could easily add to this the quick capture of a psychological response of how one is feeling on a 1–5 scale. Pulled together, you start to have a pretty interesting set of potential data.
What this all circles around is the power of of N=1 trial. This spring, Nature ran a wonderful article on what this “person over time” level of analysis might deliver in regards to insights for medical treatment. Here is a brief summary of the idea:
Within these traces, you can find meaning, even if it doesn’t rest easily visible on the surface. By looking at trends over time (N=1 approach), similarity to other people (comparison across multiple N=1 analyses), or providing data amenable to expert interpretation (data delivered to a local psychoanalyst), you can start to discover some interesting psychological insights… perhaps a nice complement to the therapist couch. The challenge becomes how to best capture, interpret, and act on these insights.
As I started to think the what this would look like in practice, I sketched out a hypothetical process document that you can see below. In this approach, a person’s response to stimuli leads to a variety of different psychological, physiological and behavioral responses, complex in their interrelationship. As these responses are tracked, aggregated, and trended, they prompt questions about whether these thoughts are true (the CBT approach), meaningful (the psychoanlytic approach) or helpful to focus on (the mindfulness approach). Through automated or professional interpretation, what might result is a kind of behavioral nudge. Someone whose signs point to burnout might receive a recommendation of more sleep. Someone whose email causes them stress might find a kind of scheduled time to focus on the task at hand. Someone who needs to grow in mindfulness might see specific behaviors crafted into their schedule — five minutes at the end of the day for meditation, or an email delivered to their inbox as a prompt for a journaled reflection.
For some of you, this world is hard to imagine. It might seem more a product of Hollywood than something coming to a phone near you. Others of you might see this technology as creepy. The recent popularity of movies like Ex-Machina hint that we have a kind of fear on the growth of ‘intelligent technology’ and what that means for human-technology interaction. What is touch to deny however is that the above model is no longer science fiction.
I am a bit skeptical of those who see a utopian technological future and also those who pine for some pre-technological age. That said, I hope we can to grow more sophisticated in our designing technology that moves beyond the aggregation of information to drive us in humanistically-oriented directions. If making strides in that direction helps people make better decisions about exercise, financial planning, and even grow in their psychological health, then that is a world I could get behind.