Why data doesn't change behavior: the science of adherence

"The industry has solved measurement — CGMs, wearables, diagnostics all work. What nobody has solved is the 30 seconds between seeing your glucose spike and deciding what to eat anyway." This comment, raised in a recent debate about longevity investment, points precisely at a field behavioral science has studied for decades: why having the right data almost never translates into acting differently.

This isn't a new intuition, nor one unique to digital health. It is, in fact, one of the most consistent and well-documented findings in behavioral research and public health. Here's what the evidence actually says about this gap between knowing and doing, and which approaches have shown they can close it, even partially.

The "law of attrition" in digital health

In 2005, Gunther Eysenbach described in the Journal of Medical Internet Research what he called the "law of attrition": in virtually any digital health intervention, a substantial share of users stop engaging shortly after starting, regardless of the tool's quality. Twenty years later, with far more apps, wearables, and CGMs on the market, this pattern is still consistently documented across the digital health literature.

The WHO's treatment adherence figure — around 50% for chronic disease in developed countries, published in its landmark 2003 report — also hasn't meaningfully improved in two decades, despite a proliferation of monitoring technology. That's a telling data point: if more available data were sufficient to change behavior, this figure should have improved. It hasn't.

What behavior change science has actually shown

The COM-B model, developed by Susan Michie, Maartje van Stralen, and Robert West and published in Implementation Science (2011) as part of their "Behaviour Change Wheel," is today one of the most widely used frameworks in health behavior research. Its central claim: sustained behavior requires Capability, Opportunity, and Motivation simultaneously. A clinical dashboard or a glucose alert provides, at best, Capability — knowing what's happening. It rarely changes Opportunity (the physical and social context at the moment of deciding) or Motivation in that exact instant. That's why information alone, without addressing the other two components, tends to have a limited and often temporary effect.

On top of that sits Prochaska and DiClemente's Transtheoretical Model (1983), which shows that people don't change a habit linearly — they move through distinct stages of readiness. The same clinical recommendation — "sleep more," "walk 30 minutes a day" — may be useless for someone who isn't even considering change, and perfectly actionable for someone already in the action phase. Treating every patient with the same message at the same moment ignores this variable entirely.

Timing matters more than data: just-in-time interventions

Science actually has a name for the exact problem raised in the original debate: Just-in-Time Adaptive Interventions (JITAIs), formalized by Nahum-Shani et al. in Annals of Behavioral Medicine (2018). Their premise is that behavioral support should adapt in real time to a person's internal state and context — not be offered as a static panel someone has to go check. This framework maps directly onto the "30 seconds" problem, and it's the scientific basis for why a well-designed, contextual alert can have more clinical impact than a thorough but passive dashboard.

Coaching as the execution layer

A systematic review by Kivelä, Elo, Kyngäs, and Kääriäinen, published in Patient Education and Counseling (2014), examined the effect of health coaching on adult patients with chronic disease. Its conclusion: coaching produces modest but consistent improvements in self-management behaviors — physical activity, diet, treatment adherence — compared to usual care without that support. It's not a miracle solution, but it's among the few interventions with repeated evidence that sustained human interaction over time moves behavior where information alone does not.

Designing the context, not just informing

Richard Thaler and Cass Sunstein, in their book Nudge (2008 — with Thaler later awarded the Nobel Memorial Prize in Economic Sciences largely for this body of work), argue that people's decisions depend heavily on how the context in which they're made is designed — "choice architecture" — not only on the information available. Applied to health: it isn't enough to show someone their glucose level; what matters is which option appears first, when the message arrives, and how much friction exists between the alert and the recommended action.

Motivation isn't a feature you can bolt onto a dashboard. It has to be designed into how and when the information reaches someone, not just how accurate that information is.

What this means for clinical infrastructure

Taken together, these four lines of evidence — the law of attrition, the COM-B model, JITAIs, and the effect of coaching — point to the same conclusion: closing the gap between data and action doesn't require more sensors. It requires a layer of interpretation and support that acts at the right moment and in the right format for each person. That's the difference between a system that measures and a system that actually helps someone decide.

Is your organization investing in collecting more data, or in building the layer that translates that data into a concrete action at the moment it actually matters?

Sources cited

• Eysenbach G. "The Law of Attrition." Journal of Medical Internet Research, 2005.
• World Health Organization. "Adherence to Long-Term Therapies: Evidence for Action." 2003.
• Michie S, van Stralen MM, West R. "The Behaviour Change Wheel: A New Method for Characterising and Designing Behaviour Change Interventions." Implementation Science, 2011.
• Prochaska JO, DiClemente CC. "Stages and Processes of Self-Change of Smoking: Toward an Integrative Model of Change." Journal of Consulting and Clinical Psychology, 1983.
• Nahum-Shani I, Smith SN, Spring BJ, et al. "Just-in-Time Adaptive Interventions (JITAIs) in Mobile Health: Key Components and Design Principles for Ongoing Health Behavior Support." Annals of Behavioral Medicine, 2018.
• Kivelä K, Elo S, Kyngäs H, Kääriäinen M. "The Effects of Health Coaching on Adult Patients with Chronic Diseases: A Systematic Review." Patient Education and Counseling, 2014.
• Thaler RH, Sunstein CR. Nudge: Improving Decisions About Health, Wealth, and Happiness. 2008.