That’s right, frameworks.

CQC have been here before. Back in 2013, they implemented sector-specific frameworks.

But it didn’t work.

Things like safeguarding and governance were assessed using different key lines of enquiry (KLOEs) and different prompts depending on whether you were a hospital, a care home, or a GP practice. This made it harder to assess quality across sectors, and providers operating across multiple service types (like an NHS trust running both acute and community services) faced different assessment criteria for essentially the same governance principles.

However, coming out of the COVID era, where they had suspended routine inspections, they faced a significant backlog. They needed to do something to improve efficiencies. They argued that maintaining separate frameworks for each sector was resource-intensive as they had to use different inspection teams who required different training.

So in 2021, the CQC announced plans to move towards a single assessment framework.

KLOEs were out.

Quality statements were in.

They introduced one set of quality statements and one scoring methodology. This was intended to make it easier to collect and compare data across all provider types, allowing them to move towards continuous monitoring and data-driven assessment rather than periodic inspections.

It launched in 2023.

It wasn’t long until they acknowledged that it wasn’t working.

In trying to make one framework fit everyone, they made it fit no one particularly well.

And as Professor Sir Mike Richards put it,

one size does not fit all.

Working within a CQC-regulated digital health provider, I saw this in practice.

Independent reviews by Dr Penny Dash and Professor Sir Mike Richards examined what had gone wrong and what needed to change.

Fast forward to today, and the CQC have launched their latest assessment frameworks.

And guess what?

KLOEs are back in.

Quality statements are out.

The five key questions remain: Safe, Effective, Caring, Responsive, Well-led.

The Richards review specifically recommended returning to KLOEs but keeping the five key questions, which have “stood the test of time”.

A significant change from both previous assessment types is that the scoring has been replaced by specialist inspectors using professional judgement.

Also included are detailed rating characteristics. They describe what Outstanding, Good, Requires Improvement, and Inadequate look like, significantly removing the guesswork.

Here’s what stood out to me in the draft framework for primary care and community services, and what it means for digital health providers.

Firstly, digital care has been called out.

The framework explicitly mentions remote consultations, online assessments, hybrid care, and digital exclusion.

It also refers to technology, and will assess whether “the technology used to deliver care meets people’s needs appropriately.”

AI is mentioned too, and whether its use in clinical care pathways is “suitable for the intended purpose, secure, up-to-date and used properly.” If you’re using or planning to use AI in triage, clinical decision support, or pathway design, this will be directly assessable at your next inspection.

Cyber security is a CQC matter now. Under Management of Risk, CQC will assess whether cyber security is treated as a strategic risk and whether controls are routinely tested.

Prevention gets a proper mention too. Population health data, health inequalities, and proactive approaches to keeping people well.

The framework has called out contemporary, proactive risk management:

Governance in healthcare tends to be dominated by retrospective data collection - incident reports, complaint reviews and audit cycles that tell you what went wrong three months ago. This framework is explicitly saying that good governance means watching the data in real time, spotting patterns before they become incidents, and using intelligence to prevent harm rather than just responding to it.

For digital health providers, this is a welcome move. Real-time dashboards, pathway analytics and outcome monitoring are all tools already in use. The framework is finally catching up with how well-governed digital services actually operate.

The bit I think needs work:

Infection control and technology assurance have been bundled together under one KLOE called “Safe environments and infection prevention and control.”

Every other KLOE in the framework maps more or less to a team or function. You can look at the title and say “this person owns this.”

But the scope of this KLOE spans your facilities manager (gas, electrical, fire safety), your estates team (premises), your business continuity lead (environmental risks like flooding), your engineering team (digital systems), and your InfoSec team or clinical safety officer (technology assurance). That’s at least five different owners across completely different disciplines, under one heading.

Bundling technology assurance with infection prevention and control risks it being treated as a sub-topic, when it’s actually a big enough area of risk that it could be its own domain.

For a digital health provider, technology assurance is critical. Everything is built upon technology - clinical pathways, consultation environment, patient records, prescribing systems, triage logic, the list goes on.

Treating that as a bullet point under the same KLOE as legionella testing doesn’t reflect the scale or complexity of the risk.

My take?

Technology assurance deserves its own KLOE with its own rating characteristics - covering clinical system safety, AI assurance, data integrity, interoperability, and digital inclusion.

And then there’s the question of who’s doing the inspecting.

Will we see inspectors who genuinely understand both a GP practice and a digital care provider? Or will the CQC further divide inspector specialism within this sector?

The digital health sector needs inspectors who truly understand what strong technology governance looks like, and, more importantly, can recognise when it’s lacking. The experts in this space still remain within the sector itself. The CQC should actively recruit from within the digital health sector, or partner with it, to ensure their inspectors are adequately trained to inspect digital providers.

What’s next?

The frameworks are being finalised, and consultation is open until 12th June.

I’ll be responding to the consultation, and if any of this affects how you deliver care, I’d urge you to as well.

You can provide feedback here

I write about building clinical products in digital health - the messy intersection of medicine, regulation, and product design. If this resonates, follow along for more, and visit clinicalproduct.uk.

Subscribe now

Further reading:

Draft sector-specific frameworks

Review of CQC’s single assessment framework and its implementation - Professor Sir Mike Richards

Review into the operational effectiveness of the Care Quality Commission: full report - Dr Penelope Dash

That’s the formal position of the World Health Organisation, which classifies it as a chronic, relapsing condition in the ICD-11. The American Medical Association recognised it as such in 2013. So did Canada, Portugal, Ireland, Germany, and Italy - which in October 2025 became the first European country to enshrine obesity as a chronic disease in law. The European Parliament has recognised it. The World Obesity Federation has recognised it.

The UK has not.

Despite having one of the highest obesity rates in Europe and an estimated NHS cost burden of £6.5 billion a year, the UK government still does not formally classify obesity as a disease. Yet a growing number of UK digital health providers are building their entire business model around GLP-1 prescribing - and in doing so, implicitly positioning themselves as chronic disease managers. They use the language of long-term care. They reference clinical pathways and ongoing treatment.

Then they design a dispensing journey geared towards first-order discounts to acquire patients and call it a clinical service.

The GLP-1 gold rush has produced a wave of digital prescribing companies calling themselves healthcare providers. But look closely and what you’ll see is a market competing on price to get patients through the door - switching incentives, discount codes and a race to the first prescription. This doesn’t sound very much like chronic disease management to me.

If obesity is truly chronic - and the global clinical consensus says it is - where are the long-term care pathways? Where’s the titration monitoring, the maintenance planning, the escalation route when someone reports side effects at 11pm on a Sunday?

Many digital health platforms treat prescribing as the finish line. Patient completes consultation, clinician reviews, prescription issued. Done. Next patient. This might have been ok for acute treatments, but GLP-1s is not an acute treatment, and prescribing is only the starting line. The real clinical product challenge begins the moment that first injection pen arrives at someone’s door.

If your product doesn’t handle what happens after the first prescription, you haven’t built a treatment pathway, you’ve built a funnel that happens to involve a POM.

The titration problem is a product problem

In the real-world, GLP-1 titration isn’t a simple linear escalation.

Patients experience nausea. They lose their appetite so dramatically it frightens them. They get constipation that no one warned them about. Some tolerate the jump from 0.5mg to 1mg without issue. Others need to hold at a dose for an extra month, or step back down entirely.

Every dose change is a patient and clinical decision point, but most products have put it on autopilot. This isn’t a prescribing decision that happens once. It’s a series of clinical micro-decisions that happen over months, often driven by patient-reported outcomes rather than clinician-initiated reviews, and the product needs to support every branch in that decision tree.

Most platforms don’t. They build a reorder flow, which is essentially a repeat prescription dressed up as ongoing care, and call it a pathway. But a reorder flow can’t distinguish between a patient who’s thriving at 1mg and a patient who’s suffering through side effects because they don’t know it’s ok to slow down.

And when the product isn’t there for those moments, the patient doesn’t complain, they just leave.

The retention cliff nobody talks about

The number that should worry every digital health company offering GLP-1s is the drop-off rate.

The real-world data on long term GLP-1 compliance is quite a contrast to the clinical trial data. Clinical trials report adherence above 85%, but in 2024, a study in the Journal of Managed Care & Specialty Pharmacy found that only 32.3% of commercially insured patients without diabetes were still on their GLP-1 at one year. A population-based Danish study of 77,310 people starting semaglutide for weight loss found that more than half had stopped within twelve months, and a Prime Therapeutics analysis of long-term adherence found that just 8.1% of patients persisted for three full years.

When you look at the reasons people stop, very few are “the medication didn’t work.” One American study found that 47.6% of patients who discontinued injectable semaglutide or tirzepatide cited financial reasons. Patients who experienced adverse effects were 9% more likely to discontinue within the first year. But for the rest, unmet expectations, inadequate support, side effects they weren’t prepared for, confusion about next steps, or just a gap between prescriptions where they lost momentum.

There’s enough evidence to show that the medicine works. So what isn’t working? In digital health, it’s the product. The entire end to end pathway isn’t working well enough for patients to retain them.

When a patient pauses their treatment because of nausea and nobody checks in for three weeks, that shouldn’t just go down as a side effect problem. There’s a missed opportunity to support the patient, and in a world where every paused subscription triggers a retention email, there’s no excuse for digital health to not do the same, but through a clinical lens.

When a patient pauses their medication for a few weeks due to a holiday, and the system locks them out of the app during that period, that’s not a clinical safety feature, it’s a product experience that hasn’t considered normal life.

The platforms that will win on retention aren’t the ones offering the cheapest first month, they’re the ones building product experiences patients actually stay engaged with, and where the administrative side flexes to suit them.

There's also a need to educate patients that obesity is a chronic, relapsing, progressive disease. It's common for patients to be focused on the number on the scales, and rely on the medication alone to deliver their weight loss goal. This means when the weight loss slows, or they reach a plateau, they assume the treatment has stopped working, and that can lead to relapse. The platforms that build this education into the journey, not as a one-off explainer at onboarding, but as something reinforced throughout treatment, will lead to patients who are better equipped to manage their disease, and its ups and downs.

So what would it actually take to build a pathway worth staying on?

What ‘good’ looks like

A chronic condition demands a product built around what happens after the treatment begins.

It starts with structured check-ins timed around the moments we know patients are most likely to struggle - not tied to the reorder cycle. And these shouldn’t be generic either. If your platform holds data on thousands of patients, you know that a 42-year-old woman starting semaglutide has a different side effect profile to a 58-year-old man switching from tirzepatide. You know when nausea typically peaks for her demographic, which dose escalation is most likely to cause problems, and at what point in the journey she’s statistically most at risk of dropping off. The product should be using that data to personalise when and how it checks in, not sending the same templated nudge to everyone at day 14. And definitely not waiting until the patient has already struggled through it alone, or worse, quietly stopped.

It means ongoing contraindication screening - not just at the point of first prescribing, but throughout treatment as a patient’s health profile evolves.

It means giving patients a language for what they’re experiencing. Most people starting GLP-1s have never injected anything before. They don’t know whether their nausea is ‘normal week-two nausea’ or a signal to pause. A well-designed pathway provides context, not just a free-text box asking “any side effects?”

It means building clinical logic that allows for maintaining doses and step-downs without requiring a full clinician consultation every time. The titration journey needs to be flexible, and the product needs to support that without creating safety gaps.

It means designing for the patient who goes quiet. The one who stops logging weight, stops opening the app, or doesn’t reorder. In most platforms, that patient churns. In a well-designed pathway, that silence is a clinical sign and the product should respond to it.

And critically, it means having a plan for what happens when the medication stops. Whether that’s a planned discontinuation, a switch to a different medication, or a transition to maintenance - the off-ramp needs to be as carefully designed as the on-ramp. However, most platforms have no answer for this at all.

Elements of this already exist across the market. Some platforms have built structured side effect check-ins, whilst others use patient-reported outcomes to inform dose titrations. Some manage the maintenance phase, though often as a 1:1 clinical or coaching conversation rather than something designed into the product itself. The building blocks are there, what's rare is seeing them joined up into a single, coherent pathway.

The real competitive moat

The GLP-1 market is enormous and growing fast. Every digital health company wants a piece of it, but the barrier to entry on prescribing is low. Any CQC or GPhC-registered provider with a prescriber can join in.

In a market racing to the bottom on price, the companies still standing in five years won’t be the ones who offered the cheapest medication. They’ll be the ones who understood that engagement - real, sustained, clinically meaningful engagement - is the competitive advantage no discount code can replicate.

It requires thinking about treatment as a longitudinal relationship, not a transaction. It requires product teams who understand that the most important screen in the journey isn’t the checkout - it’s the one a patient sees at week six when they’re wondering whether to carry on.

If your obesity product roadmap ends at dispensing, you haven’t built a clinical product, you’ve built a funnel.

The sector has committed the language of chronic disease management. Now it needs to build products that actually deliver it. The companies that do will earn something no discount code can buy - the trust of patients who stay.

I write about building clinical products in digital health - the messy intersection of medicine, regulation, and product design. If this resonates, follow along for more.

Subscribe now

Curious about clinical product?

Clinical Product is under-defined and poorly understood. My site clinicalproduct.uk exists to change that - with practical, experience-based guidance for clinicians moving into product and teams building regulated healthcare technology.

👋🏻 Hey there! This is the first post in the Translation Series, exploring clinical product at its most effective.

Subscribe now

I talk a lot about clinical product people needing to speak the language of the functions around them. But I realised I’ve never actually shown what that looks like.

So here’s the same clinical concern, framed three different ways for three different audiences.

The scenario

You’re working on a treatment pathway for a digital health product. During review, you spot that the current flow lets a patient select a medication without completing a contraindication check. It’s an edge case, but you feel it’s a real clinical safety gap and you want to get it fixed.

Here’s how the conversation changes depending on who you’re talking to.

Talking to engineering

What you might be tempted to say:

“This is a clinical risk. We need to add a contraindication check before the medication selection step.”

What lands better:

"There's a missing validation step. If a user reaches medication selection via the deep link, they bypass the contraindication screen. It affects about 3% of sessions. A fix could be a check on the screening step, if it's not complete, redirect back. Happy to work together on finding a solution."

Why this works: you’ve described the system behaviour, not the clinical concept. You’ve given them a route to reproduce it, a scope estimate, and a rough shape of the solution. Engineers don’t need you to explain why contraindications matter, they need to understand what’s broken, how users get there, and what “fixed” looks like.

Talking to commercial

What you might be tempted to say:

“We’ve found a clinical safety issue that needs fixing before we scale this pathway.”

What lands better:

“There’s an edge case where about 3% of users can reach checkout without completing a required safety screen. If we launch the paid campaign next month without fixing it, we’re looking at a percentage of orders that might need manual clinical review or cancellation. That’ll mean increased ops cost and higher refund rate. The eng fix is small and it’s in hand.”

Why this works: you’ve connected the clinical gap to money, scale risk, and operational cost. Commercial teams aren’t indifferent to patient safety, but their mental model runs on acquisition, conversion, and retention.

Talking to product

What you might be tempted to say:

“I’ve flagged a clinical risk in the treatment pathway. Can we prioritise a fix?”

What lands better:

“There’s a gap in the pathway that affects about 3% of users - they can reach medication selection without completing screening. I’d suggest we prioritise it this sprint because fixing it actually improves the completion rate for the screening step, which should reduce downstream drop-off. The eng work is scoped and small.”

Why this works: you’ve given the product manager a reason to prioritise it - better conversion. You’ve also signalled that it’s scoped, which means it won’t blow up the sprint. Product managers are constantly prioritising, so make it easy for them to say yes.

The pattern

What’s consistent across all three is that the clinical judgement doesn’t change.

What changes is the frame. You’re choosing the words, metrics, and implications that match how each person thinks about their work.

The mistake I see most often in clinical product is treating “I’ve raised the risk” as the end of the job. It’s not. Your job is to get the right thing built, and that means making it easy for the people around you to act.

This post is a companion to the Translation Series on working across functions in clinical product.

👉 Read the full post here

👋🏻 Hey there! This is the fifth post in the Foundation Series, exploring what clinical product actually is.

Subscribe now

The fifth post in the Foundation Series is live, and this one gets into something most people in the role feel but rarely talk about openly.

👉 Read the full post here

Clinical product can block releases. That’s not a side effect of the role - it’s the enforcement mechanism. Without it, clinical input becomes advisory, your suggestions get weighed against speed, and speed usually wins.

I dig into the tension of being embedded in a team you genuinely care about, while also being the person who might stop that team from shipping. What good blocking actually looks like (it’s not saying no to everything), the cost of not blocking when you should have, and how you build the credibility that makes blocking work.

There’s also a bit at the end for organisations: if the role you’re hiring for doesn’t have the authority to block a release, you’re not hiring for clinical product, you’re hiring a clinical advisor.

👋🏻 Hey there! This is the fourth post in the Foundation Series, exploring what clinical product actually is.

Subscribe now

The fourth post in the Foundation Series is live.

👉 Read the full post here

This one tackles something I see go wrong a lot: treating clinical product as a consulting function. Someone parachutes in, reviews something, gives an opinion, and leaves. It feels like clinical input is covered. In practice, it creates late-stage blockers, diffuse accountability, and clinical review that becomes a tick-box exercise.

I make the case for embedded clinical product - being part of the team from discovery, not just reviewing at the end - and get into why clinical judgment can’t work the same way as regulatory or legal review. There’s no checklist that tells you whether a user flow makes clinical sense. That kind of thinking needs to be in the room when decisions are being made.

I also cover the practical challenge honestly: embedding takes time, it requires organisations to actually structure the role that way, and one clinical product manager spread across six teams isn’t embedded in any of them.

If you’re building clinical product capability or trying to figure out why it isn’t working the way you expected, this one’s worth a read.

👋🏻 Hey there! This is the third post in the Foundation Series, exploring what clinical product actually is.

Subscribe now

The third post in the Foundation Series is up on the site.

👉 Read the full post here

This one gets into what it actually means when your job description says “accountable for clinical safety.” Most organisations know they need someone carrying that accountability - they just don’t always realise what that person needs to do it well, or they’re not willing to provide it.

I break down what you’re really accountable for (clinical accuracy, safety, appropriateness), what that looks like in practice, and the difference between responsibility and accountability - especially if you’re solo, which many clinical product people are.

I also talk about the weight of it. You’re making decisions that affect patients at scale, often with incomplete information and people waiting for you to say yes. You will get things wrong. The question is whether your mistakes are defensible.

There’s a practical framing in there too - three questions I use to pressure-test whether something is ready to ship.

👋🏻 Hey there! This is the second post in the Foundation Series, exploring what clinical product actually is.

Subscribe now

The second post in the Foundation Series is live on the site.

👉 Read the full post here

Last time I talked about the gap between what clinical product roles require and what’s available to support the people doing them. This one tackles the obvious next question - what does clinical product actually do?

I’ve broken it down into five core things: setting clinical direction, embedding clinical judgement across the whole development lifecycle, designing for the full clinical ecosystem (not just isolated users), owning clinical risk in day-to-day product work, and being the point of clinical accountability when partnerships and scale come into play.

I also get into what clinical product is not - because the role gets confused with SME work, compliance, QA, and medical affairs more often than it should. The distinctions matter, especially when organisations are hiring for these roles without a shared understanding of what they’re actually asking for.

If you’ve ever had to explain why your role isn’t just “the clinician who reviews things before release,” this one might resonate.

👋🏻 Hey there! This is the first post in the Foundation Series, exploring what clinical product actually is.

Subscribe now

I’ve just kicked off a new series on clinicalproduct.uk called the Foundation Series, exploring what clinical product actually is and why the gap between what these roles require and what support exists is so wide.

👉 Read the full first post here

The first post digs into something I think a lot of people in this space will recognise - being the only clinician on a product team, expected to make senior-level calls on clinical risk, often without the supervision structures that exist in traditional clinical roles.

I talk about what that was like when I started out, the internal debates between your clinician brain and your product brain, and why I think it’s time we start properly defining this discipline.

It’s the first of several posts covering everything from clinical sign-off to how clinical decisions actually get made inside product teams.

If this sounds like your world, subscribe here to get notified when new posts drop - or follow along on the site.

I’m a Clinical Director in digital health. My journey into clinical product wasn’t planned - it started with curiosity during an early clinical role in a private health tech organisation, where I kept being drawn to the decisions that sat between patient safety and product delivery.

Over time, that curiosity became a career. I’ve spent years building clinical product leadership from the ground up - shaping teams that can develop clinical products where healthcare, technology, and delivery all collide. I’ve embedded clinical product as a function within a broader product organisation, making sure clinical expertise isn’t an afterthought but sits at the centre of decision-making. The kind of decision-making that’s continuous, time-critical, and rarely straightforward.

Along the way, I kept running into the same problems. Clinical product is poorly defined and even more poorly understood. There’s no guide for when or how to implement it within an organisation. And when teams do hire for it, they expect senior-level judgement without providing senior-level support - while almost nothing exists externally to bridge that gap.

That’s why I started writing. To clarify what clinical product work actually involves, how it looks different across contexts, and what good practice means when you’re living it day to day.

A discipline outpacing its own support

Clinical product roles are growing fast across digital health, which is a good thing. But the supply of people who can do this work well hasn’t kept pace. So organisations hire smart clinicians into product roles and expect them to figure it out - own the clinical risk, make the hard calls, advise on regulatory grey areas - often as the sole clinical voice in the room.

In clinical practice, you’d never carry that level of responsibility without years of structured support first. In clinical product, those layers mostly don’t exist yet. No curriculum. No competency framework. No training that prepares you for making clinical judgements where speed, scope, and commercial viability are always part of the conversation.

The debates no one hears

The part that surprised me most wasn’t the clinical complexity. It was the constant internal calibration - the clinician in me, the product thinker, and the person who understood the business needed to survive for any of this to matter.

Should I block this release? Is this risk tolerable or am I being cautious? How do I explain what I’m worried about in terms an engineer can act on? When do I say “I don’t know” without losing trust?

These aren’t questions with clean answers. They’re judgement calls, made under pressure, with incomplete information, dozens of times a week - often with no one to debrief with afterwards.

Why this exists

I kept seeing the same challenges amongst the emerging Clinical Product Manager role in health tech. Organisations trying to figure out where clinical ownership sits and leaders who wanted to build the product but didn’t know what “good” looked like in a clinical setting.

That gap is what this site is about. I write from my experience as a Clinical Director in digital health, where I’ve built Clinical Product as a function within a broader product organisation.

This is an honest account of what clinical product work actually involves - the decisions, the tensions, and the things nobody warns you about. If you’re working in this space, building a team, or trying to understand where clinical thinking fits in your organisation, this is for you.

I write in two places.

• Longer, in-depth articles live on my site —> https://www.clinicalproduct.uk

• Shorter posts, weekly takes, and updates on new articles live here on Substack.

Subscribe to whichever suits how you like to read - or both.

Ways to connect:

✉️ Email: hello@clinicalproduct.uk

💼 LinkedIn: Connect here