The Missing Layer: Why Better Measurement Isn't Enough to Change Behavior
Digital health has become extraordinarily good at measuring. Continuous glucose monitors, sleep and recovery wearables, diagnostics that are faster and increasingly continuous. But measurement was never the hard problem. The hard problem happens in the 30 seconds between seeing the data and deciding what to do with it.
We've solved measurement. We haven't solved action
A glucose spike doesn't stop someone from eating the same food again tomorrow. A poor sleep score doesn't make them go to bed earlier. A warning about inactivity doesn't, on its own, create the motivation or context needed to move. We have multiplied the volume of available health data, and yet adherence in chronic disease remains stubbornly low: the World Health Organization estimates, in its landmark report "Adherence to Long-Term Therapies: Evidence for Action" (WHO, 2003), that adherence to long-term therapies in developed countries averages only around 50% — a figure that two decades of sensor innovation have failed to move substantially.
The law of attrition: why digital engagement fades
This phenomenon even has a name in the digital health literature: the law of attrition, described by Gunther Eysenbach in his seminal paper "The Law of Attrition" (Journal of Medical Internet Research, 2005). Eysenbach documented a pattern that repeats across virtually every digital health intervention: a significant share of users abandon active use of the tool long before reaching the expected clinical outcome, regardless of how precise or sophisticated the underlying sensor is.
More sensors, more precision, and higher measurement frequency have failed to reverse this pattern because they target the wrong bottleneck: the constraint isn't data quality — it's what happens after the person sees the data.
Are we building systems that help people change, or simply better systems for documenting that they did not?
Why information alone rarely changes behavior
Behavioral science has been clear on this point for decades. The COM-B model, proposed by Michie, van Stralen and West in "The Behaviour Change Wheel: A New Method for Characterising and Designing Behaviour Change Interventions" (Implementation Science, 2011), establishes that behavior change requires three conditions simultaneously: capability, opportunity, and motivation. Information — a data point, an alert, a number on a screen — can at best contribute a fraction of capability. It doesn't create opportunity or motivation on its own, and without all three conditions present at once, behavior doesn't change, no matter how precise the data that triggered it.
Timing matters: just-in-time adaptive interventions
The same intervention can fail at 9am and work at 6pm, because readiness to change is not constant — it varies with context, emotional state, and time of day. The Just-in-Time Adaptive Interventions (JITAI) framework, formalized by Nahum-Shani et al. in Annals of Behavioral Medicine (2018), describes exactly this need: intervening not when the system has data available, but when the person is actually in a position to act on it.
This connects directly to Prochaska and DiClemente's Transtheoretical Model of behavior change (Psychotherapy: Theory, Research and Practice, 1982), which describes change as a staged process — precontemplation, contemplation, preparation, action, maintenance — rather than a binary event that a single alert can trigger on its own.
The strategic opportunity: from sensors to behavioral infrastructure
For investors, providers, and digital health founders, this carries a direct strategic implication. Sensors are becoming a commodity: increasingly precise, increasingly cheap, and increasingly difficult to defend as a sustainable competitive advantage. The defensible clinical and commercial value likely no longer sits in collecting more data, but in building the layer that translates that data into action at the moment it actually matters: interpreting what's relevant, intervening at the right moment, understanding each person's readiness to change, and combining technology with human support when context requires it.
At IKI Health Group, we are building exactly this layer: translating daily data on sleep, activity, nutrition, stress, and emotional wellbeing into actionable clinical indicators, designed so professionals can intervene at the moment intervention actually has a real chance of driving change — not simply when the data becomes available.
The next breakthrough in digital health will probably not come from another sensor. It will come from building the missing layer between data and action.
The question for the industry is no longer how much data we can capture, but what clinical and behavioral infrastructure we need to build so that data translates into real change. Where do you believe the real defensible value will sit over the next decade: better sensors and diagnostics, AI interpretation, or behavior change and adherence infrastructure?