MEDICINE BETWEEN APPOINTMENTS

4 days ago
5 min read

Hand wearing smart watch with text Lord of the Rings?

Wearables capture a continuous stream of health data. AI finds the signals to remake our understanding of human wellbeing. Pharma, medtech and insurers are building business models around this new learning loop.

Join the team?

Before this week’s long-read, we thought you might like to get involved in the AI content we make. Join our webinar on 11 August to discover why CHAP - an open protocol for collaboration between humans and AI agents - is gaining traction.

Come along and ask questions like: Why will CHAP accelerate AI adoption in regulated industries? How does CHAP deliver an audit trail of agent-human interactions? Or just come along for the briefing.

Now – to this week's long read. 👇

In June, LillyDirect began giving its GLP-1 patients easier access to the Oura Ring – the sleek Finnish wearable.

Three weeks later, Eli Lilly bought a stake in Oura itself.

So why is one of the world's largest drugmakers taking such an interest in techno jewellery? Well - a prescription is delivered in a moment. And while its results unfold over months, the detail of that unfolding is unseen by the drugmaker.

Imagine if you could measure the impact continuously. The insights might be stellar. And wearables can supply that continuous signal. While AI makes interpreting it simple and affordable.

For instance, tirzepatide produced average weight loss of around a fifth of body weight in its landmark trial. The drug plainly works. But staying on treatment remains a struggle for many: one large US analysis found only 32% of people taking GLP-1 medicines for obesity still on them at 13 months.

And when a withdrawal trial switched participants off the drug, 82.5% regained at least a quarter of the weight they'd lost within a year.

That's a chronic condition doing what chronic conditions do – improving under treatment, returning without it. So the prize is persistence. And persistence draws on everything around the medicine: side effects managed, doses taken, routines held, help arriving before a wobble becomes a reason to stop.

Which means the months between appointments matter clinically to the patient. And they matter commercially to the company that makes the medicine.

So, back to the ring.

A month before the LillyDirect deal, Oura launched GLP-1 Insights, letting members log doses, side effects and weight beside the sleep, activity, stress and temperature the ring already records.

By the time Lilly invested, more than 100,000 members had logged GLP-1 use.

The result is a continuous record of the months the clinic rarely sees – and a daily channel for support. For now, that channel is Oura's own app: the LillyDirect partnership shares no user data.

This is quite the innovation. But the learning loop it is based on is a proven route to success.

Motor insurers have run it for two decades: a black box under the dashboard, reporting speed, braking and the hour of the day. As telematics spread in Britain, casualty rates among 17 to 19-year-old drivers fell 35% between 2011 and 2017 – more than twice the decline among drivers overall.

Health insurers have also got into data monitoring: Vitality members with the Apple Watch clocked 34% more tracked active days a month in an independent evaluation.

Continuous measurement can change behaviour, and therefore reduce risk, when it plays into a decision, creates an incentive or opens up timely support.

The value is the connection between the signal, its detection and the response.

A black box handles a handful of variables. A ring, watch or glucose monitor produces a far noisier stream – sleep, heart rate, temperature, activity, glucose, medication timing, symptoms, personal history.

AI changes the economics by lowering the cost of signal detection.

Models compare each person with their own baseline and compress months of readings into a trajectory. A slightly worse week of sleep, a resting heart rate sitting above baseline, logged doses stretching apart: each a weak signal on its own. But together? An early sign that something is changing. The surfacing of the few changes that may deserve a closer look – long before anyone's behaviour diverts from the lower risk path.

The clinician, coach or underwriter still owns the consequential decision. What the model changes is when their attention is called, and what evidence arrives with the case.

Which is why the pieces are now being bought and bolted together.

Lilly isn't building this alone. LillyDirect has been adding nutrition, behaviour-change and pharmacy partners around its medicines. Novo Nordisk is assembling its own remote-monitoring and adherence partnerships, down to connected insulin pens. Pharma is starting to surround the prescription with services that help the treatment work in ordinary life.

Medtech has gone further, into the data itself.

Dexcom put £75m into Oura and linked glucose readings with the ring's biometrics. Abbott has invested in Whoop, whose clinical arm has been selected for a Medicare programme in the US – a route into reimbursed, technology-enabled care.

The arrangements take different forms – some connect data, some add distribution, support or reimbursement – but they point the same way: towards the period in which a product becomes a real-world result. No single deal builds the whole loop. Each buys a piece of it.

So why now? Because the ingredients to build the learning loops have arrived together. Consumer wearables now have scale and years of personal history. Over-the-counter glucose monitors add a richer physiological stream. AI makes those streams usable at scale. And direct-to-consumer platforms, wearable apps and reimbursed care pathways give a useful signal somewhere to go.

Until recently, these things lived apart – wearables showing data to their owners, predictive models living in research programmes, care arriving through scheduled appointments.

The recent investments are connecting the sensor, the inference and the response.

Beyond health, many regulated products are delivered through formal, episodic decisions while the conditions determining their impacts change continuously. The same architecture can notice the change, bring evidence to an accountable person and support action between formal touchpoints.

And, of course, the closer the system gets to action, the more governance becomes part of the product.

Much consumer wearable data in the US sits outside HIPAA, though not outside regulation altogether – one reason the careful boundaries in these first deals matter.

A continuous service needs clear limits: what's collected, which inferences are allowed, who owns an alert, what happens automatically and when a person must approve the next step. Consent, validation and workflow ownership belong in the design from day one.

None of the current pharma–wearable deals proves better long-term outcomes yet.

What they show is the stack coming into view: continuous signals, usable inference and a governed route to respond.

The appointment isn't going anywhere. What's changing is how much useful medicine can happen before the next one.

MEDICINE BETWEEN APPOINTMENTS

Recent Posts

brightbeam.com

© Brightbeam 2025

Read our: Privacy Policy