Is there an app for that?  Without a human coach, it may not matter

By: Dr. Carolyn Bradner Jasik, MD, Chief Medical Officer

The rush toward automating human workflows in digital health is in full swing.  Tech companies across the industry are seeking out functions traditionally performed by physicians, nurses, administrators, health coaches, and other healthcare professionals, and working to replace them with AI-led review, text alerts and notifications, or other algorithm-driven processes. There are areas where this approach is already showing significant promise - the use of AI in radiology is already demonstrating it lessens the burden for doctors, and improves diagnostic accuracy.

But in other areas, tech-only solutions are failing to deliver on outcomes. This is especially true when it comes to driving the sustainable behavior change necessary to combat chronic disease. To achieve this type of change, we need the combination of technology and human coaching; automation alone won’t succeed.

A CEO of a digital health company recently said, “The human touch is important when things are serious,” he said. “But we don’t need a coach to tell us to eat broccoli if we can do it in an app.” Unfortunately, the clinical evidence contradicts this claim. Application and device-only solutions like MyFitnessPal and FitBit suffer from diminishing engagement. They have also failed to demonstrate sustained weight loss as stand-alone solutions. Waiting until things are serious for a coach to reach out is akin to waiting to install a smoke detector in your house until after the fire has already started -- the purpose of a coaching intervention should be to prevent the bad health outcome in the first place.

Decades of research in behavioral science is clear - building the confidence to make and sustain changes requires consistent human support from real people with whom an individual has a relationship, and to whom they feel accountable.  Tech solutions try to mirror this with AI-based bot coaching, automated reminders, or human coaching by a pool of staff. But when it comes to delivering outcomes, there is just no substitute for an ongoing, one-on-one connection. People don’t feel accountable to an app or anonymous (or unfamiliar) coach.  Interventions that replace a human coach with a bot or an automated notification - or even having random coaches with whom a patient isn’t familiar -- have not delivered outcomes at the same level as those that use a single, longitudinal coach.

In the rush to optimize efficiency, companies are moving too quickly to devalue the impact of ongoing contact -- and established trust -- with credentialed, human healthcare professionals.  The emotional, social, and physical barriers to behavior change are unique to each individual and are often not apparent to the person themselves. Like society, people absorb overly simplified labels for their behavior such as “I am just lazy” or “I don’t have time” or “I forget.”   If solving this problem were as easy as setting a reminder to eat our veggies, everyone would be doing it. The process requires respecting the nuanced circumstances of each user, peeling away the onion to identify barriers, setting goals, and following up. This is possible through a long-term therapeutic relationship between a coach and patient.

Forty percent of Americans are now obese. We need to deliver behavior change at scale in the most efficient and cost-effective manner.  With more than 200,000 users who have completed our program, we’ve worked to develop a deep understanding of when an automated nudge is sufficient, and when a message is better delivered by a real person who the participants knows.

We also need to empower people change their lifestyle, take their medications, & adhere to complex care plans. All of these require a personal investment. Only proactive education and support from a familiar coach and a group can prevent adverse health incidents before they ever occur.  This should be the goal of any chronic care prevention or management program, technology-enabled or otherwise.  Anything else shortchanges the patient, and denies them a chance at meaningful long-term change.

But that does not mean that technology does not play a role. The truth is that though there are evidence-based models for behavior change in the in-person world, these solutions are stunted by logistics and cost. In addition to providing scale and lowering barriers to access, technology can also connect people to their coach, and their own health,in a more intimate way through more frequent touchpoints. Technology can leverage data to point coaches to the right person at the right time.  But when we translate these interventions to the virtual world, we can’t strip out the core human components that make them successful.

And yet every week there is a new app or device that attempts behavior change with promises of AI-driven coaching. Whether it’s a device that talks to the user while they check their blood pressure, automated reminders to take meds, or a bot coach, these solutions don’t have the clinical studies to back up the claim that their solution is better, or most cost-effective,than existing in-person programs.  

If we are going to replace proven interventions that have been well-studied in the in-person world with automated solutions, we need the data to show they are as effective or more.  Until then, it’s just hype. Our patients deserve more than a machine peppering them with reminders to eat veggies or check their blood pressure.

Health coaching is much more than an emergency alert system. It is the backbone on which sustainable behavior change is built. That sustainable change is the difference between someone finding a ping on their phone another annoyance in a life full of them, and empowering an individual to free themselves from the grip of a disease that can dominate their day-to-day. For any program that deigns to put patients -- and outcomes -- first, the choice is clear.