First, a confession: This series of concurrent research findings was surfaced in the most recent issue of the Klick Wire, a weekly mHealth newsletter. That said, the conclusion that Klick drew from these separate news stories–that the most effective behavior change coaching is personal–is one that I have long believed in, and one that formed the basis of the digital health coaching product I worked on when I was with HealthMedia/J&J. Certainly the goals and purpose of behavior change coaching need to be personal for it to be effective. Other personal details–like extending the coaching beyond purely health-related issues–can also help. And now a series of new findings suggests even more importance for personalizing health behavior.
The first new evidence for the efficacy of a personal approach to behavior change coaching is a market judgment rather than a research finding. San Francisco-based Lantern, a tech company that helps people build self-help programs for emotional well-being, secured $17 million in new funding. Lantern offers coaching that is both personal to the end user and scripted differently depending on the specific emotional issues the user wants to address. As Klick pointed out, their successful fundraising suggests market faith in their potential to help users.
The second is research results from a study of CPAP users (continuous positive airway pressure, or “that Darth Vader sounding machine” to help people with breathing obstructions sleep). This study found that triggering coaching messages based on specific data patterns reduced the amount of manual effort needed for live coaches to help patients be adherent with the machine. People were just as likely to use the CPAP with the automated personalized messages as with a phone call. This is good news because it suggests that some of the personal touch can be achieved with technology, making coaching more affordable and scalable.
Finally, in a study of diabetes patients, researchers found that messages sent via a simple learning algorithm could help patients be more physically active (Hochberg et al., 2016). The learning algorithm sent messages with positive or negative feedback based on the patient’s previous activity as measured by a fitness tracker. The personalized messages were more effective at increasing people’s exercise than neutral, non-personalized reminders sent at the same time intervals.
The ability of technology to help personalize behavior change interventions remains incredibly promising, in my opinion. From fMRI results suggesting that personalized coaching triggers self-relevant brain activity (Chua et al., 2011) to eye saccade data showing greater attentiveness when images are personalized to be “like me,” the research in support of a personalized approach is strong. It’s my belief that personalized coaching is also a more positive user experience, which means people may be more likely to engage with it. In fact, self-determination theory would suggest that the more a coaching experience seems to fit an individual’s experiences and preferences, the more it would support a sense of relatedness and the more people will participate.
All of that said, we’re still learning how to build a great personalized coaching experience via technology. I think the future will bring enormous improvements as we figure out how to weave in data from people’s trackers, devices, consumer behavior, and self-reports (in a non-creepy, non-invasive way, of course, which is in itself a huge puzzle to untangle). But for now, it’s worth thinking about how we can make each small interaction around health behavior feel just a little more personal.
References
Chua, H. F., Ho, S. S., Jasinska, A. J., Polk, T. A., Welsh, R. C., Liberzon, I., & Strecher, V. J. (2011). Self-related neural response to tailored smoking-cessation messages predicts quitting. Nature Neuroscience, 14(4), 426.
Hochberg, I., Feraru, G., Kozdoba, M., Manor, S., Tennenholtz, M., & Yom-Tov, E. (2016). Encouraging physical activity in diabetes patients through automatic personalized feedback via reinforcement learning improves glycemic control. Diabetes Care, 39, e1-e2.