What’s the missing product layer between wearables and clinically usable insight?

For over a decade, the promise of digital health wearables has been remarkably consistent: more data will lead to better health.

Heart rate variability, sleep stages, activity levels, readiness scores, stress indicator - you name it. Today’s watches, rings and patches generate a continuous stream of physiological signals that would have been unimaginable just a few years ago. From a technical standpoint, this is a success story. From a human and clinical standpoint, the picture is more complicated.

A recent npj Digital Medicine article by Mahajan & Gilbert asks an interesting question: what happens when the volume, frequency and fragmentation of health data begin to undermine. rather than support, wellbeing? And what can we do to avoid this?

Their analysis highlights a growing gap between data availability and meaningful health decision making. That gap represents both a risk for the sector and a significant opportunity for healthtech innovators.

The Wearables Paradox: Information Without Insight

The current wearable ecosystem is optimised for measurement, not interpretation.

Users are presented with dashboards full of metrics, trends and scores, often without adequate context, prioritisation or explanation. For motivated, data literate individuals, this can already be challenging. For the average user, it can quickly become overwhelming.

The paper points to several emerging consequences:

• Cognitive overload from constant health monitoring

• Anxiety and hyper-vigilance around normal physiological variation

• Maladaptive behaviours, such as over-training, compulsive tracking or misinterpretation of benign signals

• Decision paralysis, where more data leads to less clarity

Crucially, these effects are not edge cases. They arise precisely because wearables are working as designed: continuously, passively and at scale.

Fragmentation Makes the Problem Worse

This challenge is compounded by fragmentation across devices and platforms.

Data is scattered across watches, rings, phones, patches and portals, each with its own app, terminology, visual language and scoring logic. Even for experienced users, aligning multiple data streams over time and interpreting them in the right physiological or behavioural context is non-trivial.

And this is before we consider the clinical interface.

Very little of this data ever reaches healthcare professionals in a form that is usable within the real world constraints, such as a 10 - 15 minute consultation. When it does, it often arrives unfiltered, poorly contextualised or disconnected from clinical decision making pathways.

The result is a widening disconnect between consumer facing digital health and clinical care.

A Shift in Focus: From Raw Signals to Meaningful Mediation

To that end, Mahajan & Gilbert explore a potential solution that is gaining traction across the sector: AI-mediated “health guardians” or companions.

Rather than exposing users directly to raw data streams, these systems sit between individuals and their wearable data. Their role is not to generate more metrics, but to:

• Filter noise from signal

• Contextualise data across time, behaviour and physiology

• Prioritise what actually matters now

• Translate complex patterns into actionable, proportionate insights

In other words, to act as an intelligent intermediary rather than a passive dashboard.

Google’s Personal Health LLM (PH-LLM) is one early example cited in this space. By fine-tuning large language models on wearable data, PH-LLM aims to transform weeks of sleep, activity and physiological signals into tailored coaching insights, rather than charts and scores.

This marks a subtle but important shift.

Why This Is a Real Opportunity for Innovators

From an innovation perspective, this space remains relatively underdeveloped.

Most wearables companies still compete on:

• Sensor accuracy

• Battery life

• Device form factor

• Incremental feature additions

Far fewer are addressing the harder problem of human centred interpretation, despite this being where user trust, long term engagement and clinical relevance will ultimately be won or lost.

There is a clear opportunity for innovators to focus on:

• Intelligent aggregation across devices and data sources

• Behaviourally informed feedback that avoids anxiety and over-medicalisation

• Context-aware systems that adapt to user goals, health status and life stage

• Clinician-aligned outputs that support, rather than burden, healthcare workflows

This is not simply a UX problem. It is a socio-technical challenge that sits at the intersection of AI, behavioural science, clinical practice and regulation.

The Regulatory and Trust Dimension

As these “health guardian” systems evolve, they also raise important regulatory and governance questions:

• At what point does an AI companion become a medical device?

• How should risk be assessed when outputs influence behaviour rather than diagnosis?

• What transparency obligations apply to filtered or prioritised health insights?

• How should responsibility be allocated when AI systems mediate health decisions?

Innovators operating in this space will need to navigate medical device regulation, AI governance frameworks, data protection laws and emerging expectations around explainability and human oversight - often simultaneously.

This is precisely where many promising products stall: not because the technology is insufficient but because regulatory strategy, clinical integration and human factors have not been addressed early enough.

Looking Ahead

The wearables sector is entering a more mature phase.

The shift from “more data” to “better decisions” will define the next generation of digital health products. I am a firm believer that those who succeed will be the ones who treat interpretation, context and human experience as core product features.

For innovators, this represents a genuine gap in the market. For healthcare systems, it offers a path toward more usable, scalable digital health integration. And for regulators and advisors, it signals the need for more nuanced, forward-looking approaches to AI-mediated health technologies.