That’s why, in June, patients, families and carers filled a lecture theatre at the Western General for our first Living Well with IBD evening. On 9 September, we’re doing it again, this time to talk about IBD medicines.

Why might one treatment be chosen over another? How do we know whether it is working? What happens if it stops working? Can treatment ever be stopped, and what options are coming next?

Our clinical team will discuss these questions, explain the different types of IBD medicine and leave plenty of time for questions from the room.

For the first time, the event will also be available online. Anyone can register and watch the one-way livestream, wherever they are in the world.

These events are an important part of how we want Edinburgh IBD to work.

Outreach, education and community-building are not optional extras. Informed, connected patients are better placed to participate in decisions about their health. The events also allow our clinical and research teams to hear directly about what matters to patients and families.

Event details

Wednesday 9 September 2026

• Doors open at 18:00 UK time

• Talks and questions from 18:15 to 19:30

• In-person discussion, posters and refreshments until 20:00

• Lecture Theatre EG.02, IGC East Building, Western General Hospital, Edinburgh

• Attend in person or watch the one-way livestream

• Registration is free

Patients, family members, friends and carers are all welcome.

Book your free place through Eventbrite

This is the second event in a growing series.

Future evenings will explore diet, psychological wellbeing, monitoring, surgery, AI and wearables, the gut microbiome, fatigue and intimacy.

We want patients and families to help shape what comes next.

These free events are part of our wider work to build more proactive, predictive and patient-centred IBD care. If you would like to help us continue that work, you can learn more about supporting Edinburgh IBD research.

This time I went physically fit, really fit. And with purpose. A solid training block behind me, a three-day mountain ultra lay ahead.

Until it did not. The mountains had other plans for me, for many others too. Fire. The Willow Fire, to be precise. Not so cute as the name implies. Ruthless heat, smoke, destruction ... human impact high; wildlife impact unimaginable. The world is burning. Terrifying.

A race cancelled, zero complaints. Yet still the same need for ... for what?

For aliveness: vitality, creative energy, beautiful humans.

The kind of aliveness found dancing with strangers in a desert car park at two in the morning. Standing behind the lens, allowing the photographer space to exist again ... not for anyone this time. For me.

The musician, so inspired by a piano concerto so hard that even Horowitz called it unplayable, and the composer had to rewrite the finale. Moved to tears sat beside a 75-year-old stranger who invited me to come and stay when I returned. And returning the next evening to hear the enigmatic Yuja Wang play the Barber again: the same programme, a different set of beguiling encores, the same airless auditorium by the Flatirons ... mesmerised once more. A different woman beside me ... even older this time, no less warm and welcoming. Wondering if Yuja wants to have dinner with me (maybe Edinburgh next month, after her gig ... dream big, haha!).

The physician who works with big data and AI, and the philosopher who thinks deeply about the future not just of medicine but of humanity. Challenged to think deeper.

In the end, will the machines love us back?

It was Adrián Villar Rojas who had the biggest impact on me. A humble man from Argentina, a deep thinker, and one of the most important artists of a generation. A one-ton triceratops skull cast in bronze, living ivy casting reflections from his soul. All of it working within the space my sister is guardian of.

He challenges us to think about how we are merely a conduit for what came before and what comes ... except not. Our imprint (impact is stronger; damn it, all my words are clichés) ... is BIG.

Thanks to Nicola, to AIR — her truly miraculous vision coming to live in the mountains — talking with Adrián and so many others ... feeling the ALIVEness return.

For the writer to come alive, inspired by dinner chatter with published authors.

Here he is, writing. Alone. Alive. Waiting.

The changes that come next: none of us are really prepared. I wonder if we are better going down fighting or blissfully unaware? I find it hard to square. As a techno-optimist I see a world of rapid scientific and medical discovery that will transform my ability as a physician to heal. I use AI in my daily workflows, and increasingly to accelerate my creativity.

And yet, whenever I can, I reach for analogue. Books that can get wet in the bath, the scratch of vinyl and my fountain pen on paper, and the dialogue in my head justifying the need for another camera so I can switch from digital to film. And yes, running free in the mountains with nature, alone or with others.

Writing to think, to explain, to connect.

Music, medicine, art, photography, mountains, movement, beauty, humans and love.

Alive again. Back from the mountains, planning my return.

Feeding off the energy in the mountains, yes. Running too.

try again
try again
try again
try again
again
try again

The patient does not care whether our hard is scientific, operational, commercial or political.

It is the simplest line in the piece. It is also the one that organises everything else.

We spend a lot of time in IBD arguing about which hard matters. Mechanism versus implementation. Trials versus real world. Drug development versus delivery. Basic science versus clinical service. The arguments have merit. They are also, often, a kind of professional comfort. Each camp gets to feel that their hard is the real one.

The patient is not in those arguments.

They are in clinic, with fistulas that will not close, or fatigue no one can explain, or a flare that has arrived three weeks before a wedding. They are watching their twenties narrow before they have really begun. They are doing the maths on whether to start a family.

They do not care whether the answer comes from a biologist, a data scientist, a policymaker or a pharma company.

They care whether it comes.

And whether it comes in time.

That is the test that ultimately matters. Not the elegance of the science. Not the citation count of the paper. Not the shape of the funding stream. Those things may matter inside the system. They are not the measure the patient lives by.

The measure is simpler, and harder.

Did the person in front of you have a better year than the last one?

If yes, you were working on the right hard.

If no, we should be honest enough to ask whether we chose the wrong one, or pursued the right one too slowly and too far from the patient.

That is what I mean by our hard: not the work that lets us sound serious to one another, but the work that leaves the patient with a better year than the last one.

That is the work.

A new patient with Crohn’s disease. Prompt diagnosis. Proper work-up. Early effective therapy. Tight monitoring. The patient responds. Inflammation settles. Life opens up again. Easy mode, relatively speaking.

A patient with ulcerative colitis arrives flaring. We use an oral small molecule early. Symptoms switch off quickly. No steroids. Hospital admission avoided. Deep remission follows. Again, it can feel straightforward.

Of course none of this is truly easy. Any disease that can take bowel, energy, confidence, fertility, sleep and autonomy is not easy. But from where I stand, as an IBD clinician in a well-resourced tertiary centre with access to most of the drugs and most of the monitoring tools, some problems feel tractable.

Then there is hard mode.

Refractory penetrating Crohn’s disease. Multiple admissions. Multiple surgeries. A permanent ileostomy. Draining fistulas. Ongoing active inflammation. Every approved therapy tried, failed or exhausted. No clean path forward.

Or ulcerative colitis with primary sclerosing cholangitis, PSC. Cirrhosis. Liver transplant. Active colitis. Dysplasia, perhaps colorectal cancer. Very hard.

That contrast has been sitting with me recently. Not because I think I have resolved it, but because it forces a more interesting question.

The Geography of Hard

What do we actually mean when we say something is hard in IBD?

The answer depends entirely on where you stand.

Scientifically, hard sits upstream. Why do some people with a similar risk profile develop IBD while others do not? What is the primary pathological defect, or more likely, how many different defects are we still grouping together under one label? Can we define disease by mechanism rather than by phenotype? Which genes, proteins, microbes and metabolites matter most, and when?

For clinicians, hard is more immediate. Right drug, right person, right time sounds neat on a slide. In practice it is still too often an aspiration. Which drug first? What sequence next? How do we implement close biomarker surveillance across a whole population with finite staff, finite time and finite money? What do we do when none of the drugs work, when the therapeutic ceiling still sits stubbornly at 30-40%? How do we manage fistulas, strictures, post-operative recurrence, fatigue, pain, urgency and brain fog, not as separate problems but as one disease experience happening inside one human being?

For patients, hard is different again. It is unpredictability. It is decision fatigue. It is trying to manage work, relationships, travel, intimacy, diet and identity around a disease that may stay quiet for months and then turn without warning. It is being told the inflammation is under control while your life clearly is not. Patients do not experience disease in categories. They experience it as life.

Society has its own version of hard. How do we pay for all of this? How do we design systems that catch disease earlier, treat it better and do not deepen inequity? How do we reduce postcode prescribing and global variation in access to IBD care? And if prevention matters, are we really willing to confront the bigger questions around food systems, antibiotics, smoking, deprivation and the wider environment in which IBD now flourishes?

Industry, too, lives with its own pressures. Some of that pressure is healthy. Competition matters. Good drugs matter. Better trial design matters. But commercial gravity can also pull smart people into the same place at the same time. You can see it in the clustering around TL1A. Competition is good for patients. A crowded procession of similar programmes is not always the same thing as progress.

Choosing Our Hard

Once you see hard from all these angles, another truth appears.

Not every hard is equally important.
Not every hard is equally actionable.
And not every hard belongs to you.

So I come back to the question that organises my working life: how do we improve outcomes for people living with IBD?

It sounds obvious. Maybe too obvious. But it is the lens through which I see clinic, research, teaching, travel and community. It is the question I return to when the field feels noisy, fragmented or overconfident.

Not every hard question belongs to my team. Some hard scientific questions are important and beautiful and worth answering even if they do not change clinical outcomes in the next five years. Some hard belongs to drug developers. Some belongs to public health. Some belongs to policymakers. Some belongs to the people brave enough to think seriously about prevention at population level.

All of that matters.

But it is not all our hard.

Our hard is this: to understand how inflammation behaves over time in real people with IBD, across the full disease course.

Not as a single clinic snapshot.
Longitudinally.

Not only in highly selected trial populations.
Across the full spectrum of patients we actually see.

Because hidden inside those longitudinal patterns is something I suspect we have still not properly grasped: hidden heterogeneity. Patients who look similar on the surface are not always biologically similar underneath. Their risk is not the same. Their trajectory is not the same. Their need for intervention is not the same. And our current labels are often too blunt to guide what matters most.

So the work, as I see it, is to model that risk better. To understand inflammatory behaviour more deeply. To connect biology, trajectory and outcome. And then, crucially, to translate that science into something useful in the real world.

Because science that never leaves the R code or the manuscript is not enough.

I am interested in implementation. In iteration. In AI-enabled tools that can be tested, improved and scaled, tools that can read hidden patterns in longitudinal data, or flag subtle shifts in symptoms and biomarkers before a flare truly takes hold. Used properly, decision support should not replace clinical judgement. It should sharpen it. It should help us act earlier, monitor smarter and make better decisions more consistently. It should move us, even incrementally, closer to what we have been promising for years: right drug, right person, right time.

That is our hard.

There is a wider point here too.

I have watched the global IBD community grow enormously since I started in this field. Bigger in spirit, bigger in ambition, bigger in capability, bigger in reach. This is good news. There are brilliant people working very hard on genuinely important problems.

But effort and progress are not the same thing.

Some hard questions are still being tackled too weakly. Real-world studies with numbers too small to guide practice. Data collected too late. Too much missingness. Experiments that were never really capable of answering the question being asked. Underpowered work is still hard work. But it does not always move the field.

And sometimes the reverse is true. The work is strong, the funding is there, the teams are excellent, but there is too much convergence. Too much clustering around the same target, the same pathway, the same safer bet. The commercial logic is obvious. The scientific opportunity cost is less often discussed. Patients need competition. They also need originality, rigour and the courage to go after questions that are genuinely under-addressed.

We cannot dictate who does what, nor should we try.

But we can think more clearly about where the true unmet need lies.
We can ask whether a question is merely hard, or whether it is important.
We can ask whether we are resourcing it properly.
We can ask whether we are being rigorous enough.
And we can ask, repeatedly, whether the patient at the centre of all this would notice the difference if we succeeded.

Back to the Clinic

Because that is where I end up.

Back in clinic.

Back with the patient with draining fistulas, or PSC and colitis, or disabling fatigue in apparent remission, or the young person whose life trajectory is already narrowing before it has really begun.

That patient does not care whether our hard is scientific, operational, commercial or political.

They care whether next year is better than this year.
They care whether we understand their disease in time.
They care whether we intervene early enough.
They care whether the system around them is intelligent, humane and fair.

That is the joy of this work. The challenge too. And the duty, not just of the clinician, but of all of us who work in and around IBD.

Choose the hard that matters.
Then do it well enough to count.
This is our hard.

The streets were quiet in the way a city can be quiet only on a holiday morning: sun already warming the stone, shutters half open, a few voices drifting out from cafés, the sense that the day had all the time in the world. Below me the city tilted towards the river. Around me were families, couples, groups of friends beginning their day.

And then there was me.

A man away alone, in a good hotel, on a holiday weekend, with no conference badge around his neck, no lecture to give, no dinner to host, no cover story of work.

What surprised me was not the aloneness. It was the ease.

I was not lonely. I was in solitude.

That distinction matters more to me now than it once did.

This was not an escape. It was an intentional decision to find the state shift I needed.

The stretch from ECCO in mid-February to DDW in early May has been one of those compressed phases of life where almost every domain is live at once. Clinics. Research. Team leadership. SENTINEL. Grants. Talks. Writing. Finishing the house. Rebuilding my body after a ridiculous back spasm in January that forced a slowness I did not welcome and could not ignore. Trying to train properly again, not as theatre, but as part of becoming more fully myself.

At home in Edinburgh, much of it has been good. More than good, often. The work is real. My team is alive. The house is slowly becoming the place I want it to be. The body is coming back under me again, with Paul Giblin setting the running, José holding the strength work, and TransRockies in July giving the next block a shape and a reason. But good work and a good life are not the same thing as spaciousness.

Easter arrived with Holly and Cameron making their own plans, a four-day window in the diary, and the sort of weather forecast that makes Edinburgh feel more like duty than delight.

So I left.

Two nights. Direct flights. A dawn departure from Edinburgh on one of the usual budget airlines, with all the familiar indignities that still somehow remain worth it. A sixteenth-century palace converted into a hotel. A city with light, hills, tiled façades, good coffee, proper fish, and the sort of air that changes your breathing before you have quite admitted that you need it.

There was something faintly decadent about going just as the work was quickening pace.

That mattered too.

The team had momentum. The ideas were getting sharper. The tools and data assets we have built were beginning to feel less like promise and more like infrastructure. SENTINEL, in particular, was moving from concept towards implementation at precisely the moment AI is accelerating at a pace medicine has barely begun to absorb. I did not leave because the work had gone flat. I left because it was becoming more alive; and I wanted to meet that next phase with more clarity, not less.

This was not a weekend against the work. It was a weekend in service of it.

In the past, I thought there were only two honest ways to achieve a state shift.

One was dark and expensive. Intensity for its own sake. The wrong kind of romance. Hedonism that felt vivid at the time and left the soul slightly dimmer afterwards.

The other was more socially acceptable, even admirable. Endurance. Long weekends of “active recovery” that were really their own form of extremity: trail running, mountain biking, mountains, exhaustion, and the private satisfaction of pushing beyond what was sensible because it felt cleaner than collapse.

One was obviously self-destructive. The other often looked like health.

Both were driven by the same nervous system.

That is what this weekend showed me most clearly.

I should say this plainly: I am not turning against the endurance line. Quite the opposite. Running has saved me more than once, and the current return to training matters deeply to me. There is something profoundly right in waking early, putting on shoes, and letting a body remember what it is for. There is discipline in it, but also release. Precision. Humility. Joy. Endurance still feels like one of the cleanest ways I know to come back to myself.

But like everything powerful, it can be used well or badly.

Sometimes eating properly, hydrating, running, lifting, mobility work, recovery and sleep all feel part of one coherent rhythm. Sometimes, especially in the current culture of endless optimisation, they can begin to feel like a nagging list of essential things to do just to remain basically functional. Even healthy systems can become strangely airless when they are asked to do too much.

This weekend suggested a third way.

Not grim abstinence, efficiency, and tighter routines. Not darkness and its aftertaste. Not active recovery pushed so far that it becomes another respectable form of escape.

Something else.

Recovery through beauty.

Through architecture and light. Through a room that felt calm the moment I walked into it. Through the quiet competence of people who prepared food, changed sheets, drew curtains, brought coffee, arranged transfers, and asked nothing of me except that I arrive. Through being held, for once, by quality and service rather than by another person or by work. Through a lingering breakfast, a long run in the heat of the afternoon sun, a swim in cold water, the strong hands of a gifted massage therapist, a book by the pool, a camera in my hand, and a cold crisp white with enough depth to hold the meal and the evening.

These are surface details, yes. But not trivial ones.

I spend much of my life holding. Holding patients, teams, decisions, vision, logistics, children, atmosphere, momentum, myself. To be held, even briefly, by competence, beauty, and care is not decadence. It is medicine.

That was the real gift of the weekend.

Not the city itself, exactly.

Not the hotel.

Not even the sun, though that helped and was definitely part of the prescription.

It was the discovery that pleasure does not have to end in collapse.

That solitude does not have to mean absence.

That a man can leave home on his own, stay somewhere beautiful, eat well, run well enough, sleep imperfectly, read, think a lot, write a little, and return more fully himself rather than less.

By the end of the weekend I found myself thinking that not every important trip needs a cover story. Some of us do not recover only by subtracting. Sometimes we recover by choosing better inputs: better light, better rooms, better food, better movement, better silence, better beauty, better ways of feeding an appetite that would otherwise go looking for darker fuel.

That, for me, is the strategic lesson.

The work ahead is too important, and the next month too alive with possibility, to keep asking an exhausted nervous system to improvise its way through. The team is moving. The ideas are sharpening. What we are building in data, algorithms, clinical tools, and implementation is becoming more real by the week. The return to training matters deeply to me too - the full embrace of the endurance lifestyle.

This weekend was not a break from that build. It was part of it.

A city had asked nothing more of me than that I arrive.

And in that simple demand, I found a way of returning to my own life more intact, and more ready for what comes next.

The Conversation We Are Not Having

My biggest concern is not that AI will replace doctors - it won’t. It is that most doctors have no idea what is coming.

At medical congresses like ECCO, we tend to hear two types of AI talk. One is the “clinician update,” often out of date and focused on surface-level chatbots. The other is the “technical deep dive,” a niche presentation on histological scoring algorithms that feels abstract to clinical practice.

Both miss the point. We are ignoring the massive middle ground: what happens when the entire operating system of medicine encounters intelligence this powerful, this fast?

The Problem: We Are Drowning

I open a patient record in clinic and watch the scroll bar sink to a sliver. Ten years of letters. Unstructured. Unindexed. A folder of PDFs and pasted text with no map.

The arithmetic of IBD care is broken. Prevalence is heading toward 1%, but capacity is stagnant. We cannot hire our way out of this.

We have better drugs than ever, but the gap between what is possible and what is delivered remains vast. Why? Because IBD is a disease of longitudinal patterns, yet we treat it in snapshots. We treat in crisis because the signals were buried on page six of the document repository.

The Solution: Three Layers of Intelligence

“AI” is not just chatbots.

For IBD, it is three distinct tools to solve three distinct problems.

• 1. Language Models: Turning text into structure

  Medicine lives in words. Diagnosis dates, phenotypes, and histories are trapped in unstructured letters.

  • The unlock: LLMs can summarise histories, draft referrals, and structure data instantly.

  • The evidence: Randomised trials are already showing LLMs can reduce specialist consultation time by nearly 30% while improving care coordination.

• 2. Computer Vision: Making scoring objective

  We still rely on subjective human assessment for our hardest endpoints.

  • The unlock: Converting endoscopy, histology, and MRI into objective, reproducible data.

  • The evidence: AI scoring of ulcerative colitis video is now predicting outcomes better than the Mayo score. In capsule endoscopy, AI reads can cut review time to four minutes while increasing sensitivity to >94%.

• 3. Predictive Models: Treating trajectories, not snapshots

  Language models predict words, not biological trajectories. That requires maths.

  • The unlock: Systems that ingest routine biomarkers (CRP, calprotectin) and map patients onto dynamic risk trajectories.

  • The Action: In Edinburgh, we are launching a pilot of SENTINEL in March: a system that updates risk every time a blood result lands, flagging the patients who are drifting off course before they flare.

The Risk: The “Deskilling” Trap

Capability is not reliability. The most dangerous failures in medicine are the ones that look confident.

We must be honest about the risk of “deskilling.” Recent data suggests that endoscopists who rely on AI for polyp detection may see their own detection rates drop when the AI is removed.

We are building SENTINEL with a “human-in-the-loop” precisely for this reason. AI should quiet the noise, not replace the signal.

The Timeline

• 1-2 Years: Ambient documentation and AI-generated summaries are standard.

• 3-5 Years: Decision support tools like SENTINEL are in major centres.

• 10 Years: Multimodal prediction (genomics + imaging + biomarkers) guides population health.

Enjoy ECCO. If you see me there, let’s talk about how we build this safely.

I open a patient record in clinic and watch the scroll bar sink to a sliver.

Ten years of letters sit in a list. Unstructured. Unindexed. A folder of PDFs and pasted text with no map. Twenty years of bloods and faecal calprotectin results live in tables that run across multiple screens. You can read them. You cannot see them. The colonoscopy report is on page six of SciStore, our document repository. The histology is somewhere else. The MRI report somewhere else again. Clinic notes written in different templates, by different teams, across different systems.

This is NHS Scotland in 2026. If you work with Epic, and are properly set up in it, you may not recognise this particular version. But the deeper pattern is familiar everywhere: more data than ever, less synthesis than we need. For chronic, lifelong inflammatory bowel disease, this means critical information is buried. A low TPMT result. A drug that was working and forgotten about years later. An uncertain surgical history. Co-morbidities and polypharmacy accumulating over a lifetime of inflammation. The disease events are not a story. They are a pile.

I am not describing bad medicine. I am describing the conditions under which good medicine has become extraordinarily difficult.

The uncomfortable arithmetic

In Lothian, our registry data projected IBD prevalence would reach about one percent of the population by 2028.

One percent does not sound dramatic until you sit with what it implies. A lifelong, relapsing disease. A growing cohort of people needing monitoring, imaging, endoscopy, biologics, surgery, pregnancy counselling, cancer surveillance, mental health support, urgent access when things tip. And the workload is not evenly distributed. Over a decade of Scottish data, around three percent of patients accounted for half of IBD-related bed days. Meanwhile, half of the broader cohort has not been seen in clinic for five years or more, some lost to follow-up, some no longer in need of our input.

We can build more capacity, and we should, and we are. But capacity alone will never keep up. The maths is wrong. And I write this from a specialist centre where IBD is what we live and breathe. Elsewhere around the country, and globally, the gaps are much wider.

An implementation problem

We have better drugs than ever. The new generation of p19 inhibitors offer remarkable safety profiles, excellent efficacy in biologic-naive patients, and superb persistence with no immunogenicity. It is now legitimate to ask whether, in the near future, a strategy of early, effective treatment might get a large proportion of newly diagnosed patients well first time and keep them there. For now, cost remains a limiting factor for many of our best therapies, though this is being offset by biosimilars as originators reach loss of exclusivity.

But this is a future argument, and it applies to new diagnoses.

The vast majority of our prevalent patients, the hundreds of thousands already living with IBD, are not in that situation. They are navigating complex treatment histories, accumulated toxicities, and disease that has been managed reactively for years. For them, the problem is not the absence of effective therapies. It is implementation. The gap between what is possible and what is being delivered remains vast. Even perfect drugs do not help if we cannot see the trajectory early enough to deploy them well.

Timing remains the hardest clinical decision, and timing is often an information problem.

IBD is a disease of longitudinal patterns, not snapshots. A single calprotectin is a data point. A decade is a story. If you cannot see the story, you treat late. You treat in crisis. You call it “aggressive disease” when some of it is, in truth, late detection. Even in leading centres with nominal access to all monitoring tools, most patients remain in reactive, crisis-driven care pathways. A one-size-fits-all approach to monitoring and follow-up is typically implemented. Given that IBD is heterogeneous, dynamic, and lifelong, this approach fails to account for individual risk profiles and is unsustainable in the face of compounding prevalence.

Most of the time, the signals were there. They were just buried.

This is the context in which artificial intelligence (AI) arrives.

The scale of what is coming

I am wary of AI hype in medicine. I am also wary of the opposite reflex: assuming it is all conference theatre because some of it is. What feels different in early 2026 is that capability, investment, and early clinical evidence are moving in the same direction.

In 2025, global AI investment exceeded two hundred billion dollars in venture capital alone, half of all venture funding on the planet. Gartner, the technology research firm, forecasts total AI spending will reach two and a half trillion dollars in 2026. The EU has mobilised two hundred billion euros. Saudi Arabia has launched a hundred-billion-dollar initiative. China has invested forty-seven billion in semiconductor infrastructure. These are industrial commitments on a scale not seen since the building of the internet, and some argue we are in the early stages of a bubble, with valuations detached from revenue.

The tension between genuine transformation and speculative excess is real, and anyone paying attention should hold both in mind.

Medicine is specifically in the crosshairs. Demis Hassabis at DeepMind has made healthcare a central pillar of their research programme. Mustafa Suleyman at Microsoft AI has spoken publicly about building “medical superintelligence,” with Microsoft expecting to spend around $140 billion in capital expenditure this fiscal year building AI infrastructure. OpenAI launched ChatGPT Health in January 2026 after reporting that over 230 million people globally ask health and wellness questions on ChatGPT every week, and that seventy percent of those conversations happen outside normal clinic hours. These are not niche early adopters. These are patients, looking for answers, at scale, right now.

The FDA has now authorised over 1,200 AI-enabled medical devices. In 2015, that number was six.

Whether we engage or not, this is arriving.

The conversation we are not having

My biggest concern is not that AI will replace doctors. It is that most doctors have no idea what is coming.

Thoughtful clinicians, clinician-scientists, real leaders in medicine, are largely absent from this conversation. They see hype and ignore it. Or they find the technical language confusing. Or they have never actually used the tools. Or they are simply too busy. The result is a vacuum. The AI presentations at medical congresses are either hyper-technical niche talks about histological scoring algorithms, or shallow demonstrations of chatbots. There is almost no serious discussion about what happens when the whole system of medicine encounters a technology this powerful, this fast.

Consider a scenario that is no longer hypothetical. Within the next year or two, most doctors will be able to pipe a patient’s complete data into an AI model and receive a structured differential, a treatment plan, a risk assessment. Many are already doing this informally. This is not currently safe or acceptable in most systems without strict governance, but the technical capability is already here, and informal use is widespread.

More importantly, patients are already doing it. They are uploading their results, challenging their doctors, and making decisions based on AI-generated recommendations. This is happening whether we prepare for it or not.

The free versions of these tools are not the frontier. Only around five to six percent of ChatGPT’s 900 million weekly users pay for a subscription. The paid models, with structured memory, deep research, and persistent project context, are substantially more capable. And the models not yet released are more capable again. The shift is no longer incremental. It arrives in step-changes, every few months.

We need serious, informed dialogue about this. Not panic. Not dismissal. Practical discussion about where it helps, where it fails, what it does to skills, and what it does to equity.

Three kinds of AI that matter for IBD

When people say “AI” they often mean chatbots. That is the visible part of the iceberg. In IBD, the real opportunity is broader. It sits in three distinct but complementary domains.

1. Language models: turning text into structure. IBD generates enormous volumes of unstructured text. Diagnosis dates, Montreal classification, drug history, adverse events, previous surgeries, endoscopy impressions, histology nuance, the patient’s own account of their life: much of the clinical record lives in language that no system currently organises. Language models can summarise long histories into usable narratives, extract phenotype from clinic letters and reports, draft plans and patient communications, and coordinate tasks and follow-up. This is where AI is already changing workflows, because it meets medicine where much of its complexity lives: in words.

A concrete example: Tao and colleagues published a randomised controlled trial in Nature Medicine (January 2026) testing an LLM chatbot that performed pre-assessment before specialist consultation. Over 2,000 patients across 24 medical specialties were randomised. Before seeing a specialist, patients in the intervention group interacted with the chatbot, which took a structured history, generated preliminary diagnostic suggestions, and produced a referral report. Specialist consultation time fell by 28.7%, physician-rated care coordination more than doubled, and patients reported significantly better communication. The chatbot did not replace the specialist. It did the preparatory work that primary care often cannot do at scale, freeing the clinician to focus on the decision that matters. That is not a demonstration. That is a randomised trial, in a top-tier journal.

There is a broader dimension too. For patients with IBD, the need for trustworthy, hyper-personalised guidance is enormous: drug counselling, what a calprotectin result means, vaccination schedules, cancer surveillance timelines, dietary advice during a flare, when to worry, when not to. Well-trained language models can excel at this kind of signposting, delivering the right information to the right person at the right time. But accuracy is non-negotiable. A hallucination in a reassurance message about symptoms could delay someone seeking urgent care. The opportunity and the risk sit side by side.

But there is a trap. Language models can be persuasive even when they are wrong. Bedi and colleagues (Nature Medicine, 2026) introduced MedHELM, an evaluation framework for medical LLMs, and found that only a small fraction of published evaluations use real-world electronic health record data. If you have not tested a model in the mess of actual clinical records, you have not tested it. And this is hard to do, because most institutions remain understandably cautious about running AI models against real patient data.

2. Computer vision: making our images objective. AI in colonoscopy is already well advanced for polyp detection and characterisation, with sensitivity exceeding 95% in some systems. But the biggest long-term prize for IBD specifically is different. It is turning the core objective assessments of disease activity into something reproducible, scalable, and trajectory-aware. That means endoscopy, histology, and cross-sectional imaging.

In endoscopy, Stidham and colleagues (Gastroenterology, 2024) used ulcerative colitis trial videos to generate a computer-vision cumulative disease score that captures global mucosal burden and appears more sensitive to change than conventional Mayo endoscopic scoring. Cai and colleagues (Clinical Gastroenterology and Hepatology, 2026) trained models to quantify mucosal ulceration severity in Crohn’s disease from endoscopic video, aiming to reduce subjectivity and improve endpoint reliability. In capsule endoscopy, AI-assisted reading has demonstrated sensitivity above 94% for Crohn’s lesions versus around 60% for expert reviewers, while cutting reading time to under four minutes.

In histopathology, AI-powered neutrophil detection on whole-slide images has enabled the PICaSSO Histologic Remission Index. A PHRI score of zero was associated with substantially fewer poor outcomes over the subsequent twelve months, moving histology from confirmation to prognosis.

And imaging matters because Crohn’s is a transmural disease. Deep learning and radiomics applied to CT enterography have outperformed radiologists for detecting and quantifying intestinal fibrosis, a clinically crucial feature that humans still struggle to assess reliably. Intestinal ultrasound is another area where AI could have transformative impact: automated measurement of bowel wall thickness and vascularity could lower the skill threshold for IUS, enabling wider deployment and more consistent monitoring.

Computer vision in IBD is still behind language models in maturity. It is harder to validate and harder to deploy across different centres and image qualities. But it may be the bigger unlock, because it targets the heart of treat-to-target care: objective inflammation and tissue damage.

3. Predictive models: IBD as trajectories, not snapshots. Inflammation rises quietly. Symptoms lag. Biomarkers are stochastic. A single calprotectin is a data point. A decade of calprotectins is a story.

Language models are not built to predict what number comes next in a noisy, longitudinal biological system with missingness, measurement error, and latent heterogeneity (hidden subgroups of disease behaviour that standard models do not capture). That needs mathematics. The label matters less than the outcome.

In our group, Nathan Constantine-Cooke led work using latent class mixed models to identify distinct longitudinal trajectories of faecal calprotectin and CRP across two large European cohorts. The study, published in Clinical Gastroenterology and Hepatology this year, identified eight FC and eight CRP trajectory patterns: rapid remitters, delayed remitters, relapsing-remitting courses, and non-remitters. These were independently replicated in Danish national data. Victor Velasco-Pardo then built landmaRk, an R software package that combines landmarking with latent class mixed models to enable dynamic risk prediction. The key innovation is that the model updates each time a new lab result arrives, learning which trajectory a patient is on and recalculating risk in real time.

This is what powers SENTINEL. A clinical decision support system that ingests routine lab data, places each patient on their trajectory, and produces a daily, risk-prioritised view of the caseload. The aim is alerts that prompt proactive action before the crisis. The MRC-funded pilot launches in Lothian in March, with a human in the loop at every stage, so we can test, iterate, and build confidence before scaling.

How we build matters

There is a fourth dimension that does not get talked about enough: the way AI changes the act of building.

Nathan, who has built the core analytical infrastructure and much of SENTINEL’s front-end tooling, puts it concretely: “Over the past couple of months, we’ve written 55,000 lines of code, equivalent to eight engineers working on it for fifteen months. I could write all of the code we’ve generated, and I review it, but it would obviously have been much slower if I had manually written it all.”

That is not automation. It is an extremely talented data scientist working with a frontier model as a pair-programmer, combined with ruthless testing and iteration. The bottleneck in clinical AI is not usually the idea, though there are plenty of bad ones. It is implementation: getting the right data in, getting the output into the workflow, making it safe, equitable, and usable when real life is messy.

In healthcare, building faster is not the win. Building safely into workflow is.

The deeper promise: science moving faster

Beyond clinical care, AI is beginning to reshape the science that underpins it:

• AlphaFold has mapped the three-dimensional structure of virtually every known protein. For IBD, this opens new routes to understanding the immune pathways that drive mucosal inflammation and identifying drug targets with structural precision

• Multi-omics approaches integrating genomic, proteomic, and metabolomic data are generating datasets whose complexity exceeds human interpretive capacity. In IBD, where disease heterogeneity has frustrated decades of biomarker research, AI offers the first plausible route to actionable biological stratification at scale

• The microbiome remains one of the great unresolved questions in IBD. For years, the promise has been constrained by the difficulty of making sense of high-dimensional ecological data. AI-driven analytical approaches may finally begin to change that

• Drug discovery pipelines for inflammatory diseases are being compressed through protein structure prediction, large-scale screening, and generative chemistry, with direct relevance to the next generation of IBD therapeutics

The honest counterweight

Capability and reliability are not the same thing. The most dangerous failures in medicine are the ones that look confident.

Berzin and Topol, writing in The Lancet, introduced a formulation worth keeping close: deskilling, mis-skilling, never-skilling. In gastroenterology, this is already more than theoretical. Budzyń and colleagues (Lancet Gastroenterology & Hepatology, 2025) found that experienced endoscopists who routinely used AI for polyp detection showed a significant decline in adenoma detection rates when AI support was removed, dropping from around 28% to 22%. That is not an argument against AI. It is an argument for designing it properly: training strategies, AI-off intervals, and clear boundaries around what is safe to automate and what must remain a human skill.

SENTINEL starts as a Lothian pilot precisely for this reason. The tool makes key data visible and meaningful at the point of care, and identifies patients at risk based on population-wide clustering of routine labs. But we test, iterate, and retest, always with a clinician in the loop, before we scale.

There are harder questions too. Who pays for this? In an NHS that struggles to fund existing services, the economics of deploying AI-powered clinical tools are far from resolved. And will health inequalities widen before they narrow? The digital divide is real. Patients with the best access to technology, the highest health literacy, and the most engaged clinicians will benefit first.

If we are not deliberate about equity, AI will entrench the disparities it has the potential to close.

A sense of the timeline

Think about a young person starting medical school this autumn. By the time they finish medical school in five or six years, let alone specialty training another six or seven years after that, every aspect of clinical practice will look different.

On an aggressive timeline, within one to two years ambient AI documentation is standard, most referral letters and clinic summaries are AI-generated, and AI-powered health assistants are handling a significant proportion of day-to-day clinical queries from patients. Within three to five years, decision support tools like SENTINEL are deployed across major centres, AI-scored endoscopy and histology are embedded in clinical trials, and most routine clinical decision-making involves an AI layer that synthesises patient data, suggests differentials, and proposes treatment plans for clinician review. Within a decade, multi-modal predictive systems integrating biomarkers, imaging, wearables, and genomics are guiding treatment selection and monitoring strategy at population scale.

On a more conservative timeline, double all of those estimates. It still means that a trainee starting now will qualify into a profession that has been transformed more fundamentally than at any other point in the history of modern medicine.

Either way, the conversation needs to start now.

What comes next

This is the first in a small, evolving series on AI in medicine through an IBD lens. The landscape (and my schedule) moves too fast for a fixed itinerary, but the territory is clear: language models in workflow, computer vision in endoscopy and pathology, predictive modelling and proactive care, wearables, and the human questions of attention, agency, and equity.

There is a short film in post-production that goes deeper into a simple idea: what happens when we introduce a third intelligence into the clinical encounter, a system that helps both doctor and patient see the signal through the noise.

The goal is not a colder, more automated medicine. It is the opposite: quieter systems, earlier signals, fewer crises, and more time for the conversations that matter.

Simone Weil wrote that attention is the rarest and purest form of generosity. That is what I want these tools to protect.

Glossary

Artificial intelligence (AI): A broad term for computer systems that can perform tasks typically requiring human intelligence, including learning from data, recognising patterns, and making predictions.

Large language model (LLM): An AI system trained on vast amounts of text that can generate, summarise, and interpret language. ChatGPT, Claude, and Gemini are examples.

Foundation model: A large AI model trained on broad data that can be adapted to many tasks. GPT-5, Gemini, and Claude are foundation models.

Frontier model: The most capable AI models currently available, typically accessible through paid subscriptions or enterprise agreements. Significantly more capable than free-tier versions.

Computer vision: AI that can interpret images and video, enabling applications like automated endoscopic scoring and histological analysis.

Superintelligence: A hypothetical AI system that exceeds human cognitive ability across all domains. Currently theoretical, but increasingly discussed as a near-term aspiration by major AI companies.

AGI (Artificial General Intelligence): AI that can perform any intellectual task a human can. Distinct from current AI, which excels at specific tasks. No consensus exists on whether or when this will be achieved.

Model training: The process of teaching an AI system by exposing it to large datasets so it can learn patterns and make predictions.

Hallucination: When an AI model generates information that sounds plausible but is factually incorrect. A significant safety concern in medical applications.

Latent class mixed model (LCMM): A statistical method that identifies hidden subgroups within longitudinal data. Used in our work to discover distinct inflammatory trajectory patterns in IBD.

Landmarking: A statistical technique for dynamic prediction that updates risk estimates each time new data arrives, rather than relying on a single baseline assessment.

References

1. Jones G-R, Lyons M, Plevris N, et al. IBD prevalence in Lothian, Scotland, derived by capture-recapture methodology. Gut. 2019;68(11):1953-1960.

2. Lyons M, Plevris N, et al. Patterns of emergency admission for IBD patients over the last 10 years in Lothian, Scotland: a retrospective prevalent cohort analysis. Aliment Pharmacol Ther. 2022 Jul;56(1):67-76. doi: 10.1111/apt.16867.

3. Tao X, Zhou S, Ding K, et al. An LLM chatbot to facilitate primary-to-specialist care transitions: a randomized controlled trial. Nat Med. 2026 doi:10.1038/s41591-025-04176-7.

4. Bedi S, et al. A holistic assessment of the reliability of machine learning systems in clinical settings. Nat Med. 2026. doi:10.1038/s41591-025-04151-2.

5. Berzin TM, Topol EJ. Preserving clinical skills in the era of AI. Lancet. 2026;407(10475):347-349. doi:10.1016/S0140-6736(25)02075-6.

6. Stidham RW, Ghanem LR, Fletcher JG, Bruining DH. Artificial intelligence-enabled clinical trials in inflammatory bowel disease: automating and enhancing disease assessment and study management. Gastroenterology. 2025;169(3):427-443.

7. Stidham RW, Cai L, et al. Using computer vision to improve endoscopic disease quantification in therapeutic clinical trials of ulcerative colitis. Gastroenterology. 2024;166(1):155-167.e2.

8. Cai L, Stidham RW, et al. Artificial intelligence for quantifying endoscopic mucosal ulceration severity in Crohn’s disease. Clin Gastroenterol Hepatol. 2026;24(2):337-346.

9. Constantine-Cooke N, Vestergaard MV, Plevris N, et al. Large-scale clustering of longitudinal fecal calprotectin and CRP profiles in IBD. Clin Gastroenterol Hepatol. 2025 Dec 29:S1542-3565(25)01069-9. doi: 10.1016/j.cgh.2025.12.022.

10. Plevris N, Lees CW. Disease monitoring in inflammatory bowel disease: evolving principles and possibilities. Gastroenterology. 2022 Apr;162(5):1456-1475.e1. doi: 10.1053/j.gastro.2022.01.024.

11. Hirten RP, Danieletto M, Sanchez-Mayor M, et al. Physiological data collected from wearable devices identify and predict inflammatory bowel disease flares. Gastroenterology. 2025;168(5):939-951.e5

12. Jumper J, et al. Highly accurate protein structure prediction with AlphaFold. Nature. 2021;596:583-589.

13. Elford AT, Constantine-Cooke N, Jenkinson PW, et al. Twenty-year trends in colectomy rates and advanced therapy prescribing in Lothian, Scotland. Aliment Pharmacol Ther. 2025 Oct;62(7):699-709. doi: 10.1111/apt.70240.

14. Budzyń K, Romańczyk M, Kitala D, et al. Endoscopist deskilling risk after exposure to artificial intelligence for polyp detection: a multicentre, observational study. Lancet Gastroenterol Hepatol. 2025;10(10):896-903.15.

In eight days I turn fifty.

I am writing this from my desk in Edinburgh on a winter morning. Bach is playing on vinyl. The light outside is grey and honest in the way that February here always is: no performance, no flattery, just stone and rain and the slow promise that the days are getting longer even when it doesn’t feel like it.

I have been thinking about focus. Not productivity. Not the morning routine that will change your life. I mean the deeper kind: knowing what matters, pointing yourself at it, and staying there long enough for it to become real.

I have not always been good at this.

I spent my thirties and forties moving between versions of intensity: clinical medicine, research, endurance sport. Each one producing something I am proud of, each one leaving something else unattended. That is the subtle cost of focus. Not the missed dinners. The parts of yourself that quietly atrophy while you are busy building elsewhere.

Somewhere in my late forties, patterns I had carried for decades became impossible to ignore. The effect was clarifying. I stopped fighting a version of myself that was never going to work, and I started asking a question I had never properly asked.

What happens when you stop going all-in on one thing at a time and try to hold all of it?

Because that, I think, is what this stage of life asks of you. Not a crisis. Not a cliff edge. A reckoning with the full inventory of who you are, all at once.

I am a clinician-scientist who cares for people with inflammatory bowel disease. I am a father to two young adults, step-father and grandfather besides. I am a runner who has not yet given up on the mountains. And I am someone who needs the quiet as much as the intensity.

None of these is optional. None of these is the whole picture. And none of these will survive if I do what I have always done: go all-in on one and let the others fade.

The project of the next decade is integration.

Not balance — I have always hated that word. Balance implies equal weight, carefully distributed. It implies stillness. Life is not still. It oscillates. Some weeks the work takes everything. Some evenings the loneliness arrives uninvited. Some days the body demands attention the mind would rather not give.

Integration is different. It means none of it goes to zero. It means the runner is still there on the weeks when the scientist is working late. It means the father is present even when the leader is exhausted. It means you design your recovery before the collapse, not after.

High performance requires imbalance. To build something meaningful, to care for patients deeply, to lead a team, you have to go all in. But you have to build the rhythm that lets you sustain it. The goal is not equilibrium. The goal is a rhythm you can live inside.

For me, right now, that rhythm is still taking shape. It involves protecting the mornings, training even when the body would rather not, and leaving space for the quiet that lets everything else work. Small architecture, not grand plans.

There is something else I have been thinking about, and I find it harder to say.

People who lead and build and carry responsibility are often remarkably alone in midlife. Not all of them. But enough. The ones who find themselves sitting with the gap between professional visibility and personal invisibility. The ones who are witnessed intensely — in the lecture hall, in the clinic, on the platform — and then not at all.

An audience is not a witness. A follower is not a witness. A witness is someone who sees the cost of the work, not just the result.

Travelling for work, I once watched men gather at a barber on a Friday evening, not for a haircut, but for connection. They sat together. They talked. They were present with each other in a way that required no agenda, no pretence. It was infrastructure, built into the rhythm of the week. We have lost something like this, or perhaps we never built it outside of the pub.

Leaders don’t need more followers. We need more witnesses.

That same week I will be in Stockholm for ECCO, the largest inflammatory bowel disease meeting in the world. I have attended every year for two decades. Two decades that saw my children grow up, relationships change shape, and a career that has rebuilt itself more than once. Deep friendships in this community, developed and deepened over the years, and every year new ones.

This year I am not speaking, which is unusual, and I notice myself negotiating with what that means. But walking into that conference surrounded by colleagues and friends, many of whom are reading this now, I am reminded that the work is not the talk. The work is the care and curiosity and collaboration that brought me here. The talk is just the visible part.

The work continues whether anyone is watching or not.

Steve Jobs gave that commencement speech when he too had just turned fifty, after his cancer diagnosis. He talked about death as a clarifying force. The mirror question: if today were your last day, would you be doing what you are about to do?

I used to hear that as a motivational line. I hear it differently now. At fifty, you stop being abstract about mortality. You see friends get diagnoses. You notice your own body negotiating with you in ways it never did before. I am thankfully only recovering from a humbling back injury, not anything worse, but even that has been enough to sharpen the question. And it stops being motivational and starts being existential.

Am I spending my days on the things I believe in?

Not every day. Not perfectly. But enough of them. The patients are real. The science matters. The team believes. The words I write occasionally reach someone who needs them. The mountains are still there when I am ready.

I don’t know what fifty looks like yet. I will find out in eight days.

But I know what I want it to feel like. A life that is inhabited rather than performed. Where the hard days are in service of something I believe in. Where the loneliness, when it comes, is space rather than void. Where I am building things that matter now and into the future, not chasing legacy, but doing work that earns the right to last.

Care for people who trust me. Run so far into the mountains that everything else dissolves and the only battle left is to keep going, stay in the moment, eat, endure. Photograph strangers in honest light. Write what is true, even when it is uncomfortable. Design recovery before collapse. Build the rhythm, not the monument.

I am still becoming.

And if you are reading this on a quiet evening, wondering whether anyone else feels this way:

You are not the only one.

If you have IBD and you’re under my care in Edinburgh, I’ll ask you for it. Regularly, for years. Some people find it weird. Most eventually find it reassuring. Because it gives us something symptoms and clinic appointments can’t: an objective readout of what’s happening in the gut wall.

In 2005, we started using a single lab-based ELISA assay for faecal calprotectin (FC) testing at the Edinburgh IBD Unit. By 2009, it was embedded in routine care. Two decades later, we’ve run hundreds of thousands of FC assays in total (now tens of thousands per year), built a research programme on top of that data stream, and we’re using it to set the scene for what comes next, including a follow-up post on SENTINEL.

TL;DR

• FC let us stop guessing about inflammation, at scale, across a whole health board.

• The “250 rule” made FC usable in real life (trend is the art; thresholds are the language).

• PREdiCCt suggests the “grey zone” (50–250) isn’t benign — and SENTINEL is how we unlock the true power of trends.

What we’ll cover

• Why FC solved a problem that scopes and symptoms alone couldn’t

• The Edinburgh studies (triage → prognosis → treat-to-target → trajectories)

• The single simplest message that made FC usable at scale (and why SENTINEL matters)

The Problem We Were Trying to Solve

IBD is defined by inflammation, and while symptoms can be brutal, they’re not a reliably calibrated proxy for what’s happening in the bowel wall. Sometimes symptoms and inflammation track perfectly. Sometimes they don’t: a patient can feel fine while inflammation smoulders; another can feel terrible from functional symptoms despite quiescent disease.

For decades, the only way to know for sure was to look inside.

Colonoscopy remains the gold standard, but you can’t scope everyone every time. It’s invasive, expensive, and capacity-limited. Cross-sectional imaging with MRI helps, particularly for small bowel Crohn’s disease. And increasingly, intestinal ultrasound (IUS) is gaining worldwide traction as a non-invasive monitoring tool that can be performed at the point of care. But we needed something simpler, cheaper, and repeatable. A window into the gut that didn’t require a camera or a scanner.

Faecal calprotectin offered exactly that. A protein released by neutrophils, measurable in stool, reflecting the degree of intestinal inflammation with remarkable fidelity. The technology existed. The question was: could we make it work in real clinical practice?

First Steps: Separating IBD from IBS

The first clinical challenge was triage. Young patients present with abdominal pain and altered bowel habit all the time. Most have irritable bowel syndrome. A minority have IBD. Identifying which is which, quickly and accurately, matters enormously.

Back in the late 2000s we deliberately focused on ages 16–50: once you get beyond that, the triage question isn’t just “IBS vs IBD” but also “could this be cancer?”, and most patients end up scoped anyway. These days we also have qFIT in the pathway. In practice, if FC <50 μg/g and qFIT <10 on two samples, the negative predictive value for serious organic disease is around 99%. Not yet a fully formalised pathway, but already useful for deciding who needs a scope urgently.

Nick Kennedy, my first clinical research fellow to tackle this problem, led a study that would become foundational. Nick is now a consultant gastroenterologist in Exeter, leading major research programmes there. Using data from the Edinburgh Faecal Calprotectin Registry, we examined 895 patients aged 16-50 presenting to gastroenterology services for the first time.

The results were striking. At a threshold of 50 μg/g:

• Sensitivity for IBD: 97%

• Negative predictive value: 99%

When combined with the absence of alarm symptoms, the NPV reached 100%. Not a single patient with FC below 50 and no alarm features had IBD.

The clinical translation was immediate: if your FC is low and you don’t have red flags, you almost certainly don’t have IBD. Colonoscopy can be safely avoided. During our study period alone, this would have prevented 200 unnecessary endoscopies, saving £88,000 in direct costs and, more importantly, sparing patients procedures they didn’t need.

This paper, published in the Journal of Crohn’s and Colitis in 2015, helped establish FC as a triage tool worldwide.[1]

Beyond Diagnosis: FC and Disease Progression

But diagnosis was just the beginning. The bigger question was whether FC could tell us something about the future. Not just whether someone has IBD, but how their disease would behave.

Nick Kennedy tackled this next. In a cohort of 918 Crohn’s disease patients followed for over 4,000 patient-years, we asked a simple question: does FC predict disease progression?

This was a deliberately pragmatic analysis: we anchored on a single FC value (the first non-diagnostic sample we had for each patient), rather than a trend, because that’s what we could do at scale at the time.

It did. Convincingly.

The optimal threshold was 115 μg/g, lower than the 250 μg/g commonly cited at the time. Patients above this threshold had a hazard ratio of 2.4 for disease progression, defined as a change in Montreal behaviour classification, surgical resection, or hospitalisation for severe flare.[2]

Crucially, this association held even in patients without symptoms. FC was capturing subclinical inflammation that would eventually declare itself, but might be treatable before it did.

The clinical message was clear: don’t just treat symptoms. Treat inflammation, because if you want to prevent bowel damage, you need to know the inflammation is there, even when the patient feels well.

Normalisation as a Treatment Target

If elevated FC predicts bad outcomes, does normalising it predict good ones?

Nik Plevris, another clinical research fellow who is now a consultant gastroenterologist right here in Edinburgh (we got to keep one Nik; wish we could have kept both!), led the work that answered this question.

We looked at 375 newly diagnosed Crohn’s disease patients with elevated FC at diagnosis and asked: what happens if we get FC below 250 μg/g within the first year?

The answer: everything improves.

Patients who normalised their FC had:

• 64% reduction in disease progression (HR 0.36)

• 78% reduction in Montreal behaviour change (HR 0.22)

• 67% reduction in hospitalisation (HR 0.33)

• 61% reduction in surgery (HR 0.39)

At 7 years, the cumulative progression rate was 29.9% in normalisers versus 60.5% in non-normalisers. This wasn’t a subtle effect. It was transformative.

The paper, published in Clinical Gastroenterology and Hepatology in 2021, provided hard evidence for what we had suspected: FC normalisation should be a treatment target. It supported the “treat-to-target” approach that was gaining momentum internationally.[3]

Put simply: get FC down early, and you change the slope of the whole disease curve.

One question remained: does FC work for isolated small bowel Crohn’s disease, where MRI is the gold standard? Gareth Jones, another research fellow who is now a consultant gastroenterologist and mucosal immunologist in Glasgow (maybe we can lure him back across the M8 one day!), examined 119 MRE scans matched to FC measurements within 30 days. FC correlated well with MRI findings: the optimal threshold for detecting severe ileal inflammation was 145 μg/g, remarkably close to the 115 μg/g threshold we’d identified for predicting disease progression. And FC outperformed CRP for detecting MRI-confirmed disease activity.[4]

So what? If the bowel is smouldering in the ileum, FC usually tells you, and it tends to tell you better than CRP.

IUS is increasingly complementing both colonoscopy and MRI for monitoring, particularly in UC and colonic Crohn’s. It’s non-invasive, can be done at point of care, and is gaining huge traction globally. But FC and IUS are complementary rather than competing tools. Different jobs, both valuable.

How We Made FC Work in Practice

Having a good biomarker is one thing. Making it work in routine care across an entire health board is another. Our success in Lothian came down to a few very practical choices.

Same assay, always. We’ve used the same lab-based ELISA platform since 2005. Assay variability between different manufacturers is a real problem in FC testing. By sticking with one platform, highly optimised in our biochemistry lab at Western General Hospital, we’ve avoided the confusion that comes from switching between systems with different thresholds.

Education at every level. This starts with the conversation at diagnosis, which I never outsource. I explain to every patient: IBD is progressive. It’s the inflammation that drives the progression. There’s a disconnect between symptoms and inflammation. And FC is the best way for us to monitor what’s really going on.

A patient-facing script (feel free to steal this)

“Symptoms matter, but in IBD they don’t always tell us what the inflammation is doing. Faecal calprotectin is our ‘smoke alarm’ for gut inflammation. If it’s low, that’s reassuring. If it’s rising, it often means trouble is brewing before you feel it. The goal isn’t obsessing over one number; it’s watching the pattern over time and acting early.”

And yes: accepting FC monitoring means giving us a scoop of your poo regularly, but it’s the price we pay for staying ahead of the disease.

The Edinburgh rule of 250

For well over a decade we taught everyone — doctors, nurses, junior staff, pharmacists, dietitians, administrators, and yes, even surgeons — that 250 μg/g is the level where you take action.

It sounds almost too simple, but that simplicity mattered. It stopped dithering, reduced variation, and gave patients a clear message.

Of course the real art is the trend, because what “normal” looks like is highly individual. But trend-following is hard to do consistently in real life when you’re usually staring at the last 2–3 results. SENTINEL is designed to unlock the true power here.

Practical logistics. Kits are ready at every clinic, handed out by staff who know how to talk patients through the process. Samples are collected at home (FC is stable at room temperature for around 3–4 days, which helps), then dropped at the GP surgery where they go on the sample van to our central biochemistry lab.

Risk-stratified frequency. Higher risk patients check every 2 months. Medium risk every 4-6 months. Lower risk annually. FC at every clinical interaction. Additional samples at key timepoints: after treatment initiation, change, or cessation; during suspected flare; when checking therapeutic drug levels.

Cost. The FC ELISA costs about £10-12 per assay in our lab. A bargain for the information it provides.

We trialled home testing with IBDoc and CalproSmart during the pandemic. It works, but the results are less reliable than lab ELISA, more expensive, and don’t flow automatically into our labs database. For now, the postal kit to GP to central lab pathway remains our standard.

The result? The proportion of our IBD population undergoing FC monitoring has climbed steadily year on year, with a mirrored reduction in sigmoidoscopy and colonoscopy for disease monitoring. Patients are empowered to ask about their FC results. It’s become part of the culture of IBD care in Edinburgh.[8]

The biomarker was easy. The behaviour change was the hard part, and that’s what scales.

PREdiCCt: A More Ambitious Target?

PREdiCCt is a prospective cohort study of 2,629 IBD patients across 47 UK centres. We designed it to identify factors associated with disease flare in patients in self-reported clinical remission.

The FC findings, published in Gut in January 2026, challenge current practice.[9]

We stratified patients into three groups:

• FC <50 μg/g (52% of patients)

• FC 50-250 μg/g (30%)

• FC >250 μg/g (18%)

These were patients who told us they felt well. Nearly one in five still had FC >250 at baseline, and many more sat in the “grey zone” between 50 and 250.

PREdiCCt: the “grey zone” isn’t grey

Even in patients who feel well, FC 50–250 μg/g carried more than double the 2-year risk of objective flare compared with FC <50.

• FC <50: 8% objective flare at 2 years

• FC 50–250: 20%

• FC >250: 26%

In adjusted analyses, FC 50–250 had HR 1.98 and FC >250 had HR 3.25 for objective flare (vs FC <50).

What does this mean for practice? It suggests that patients who consistently track below 50 μg/g behave differently. They have genuinely quiescent disease. This is a highly desirable place to be.

But should FC <50 become “the new treatment target”? Here we need to be careful. We must not let the perfect be the enemy of the good. For many patients, getting to FC <50 simply isn’t achievable with current therapies. Setting an unattainable target risks therapeutic nihilism and patient demoralisation.

We also know that lower FC thresholds correlate with histological healing in UC, with values around 70-75 μg/g associated with histological remission in some studies. So there’s biological plausibility to aiming lower than 250. But the clinical reality is that 250 remains a sensible, achievable target for most patients, and getting there still delivers major benefits.

The message from PREdiCCt is not “250 was wrong”. It’s that “lower is better if you can get there”, and that the 50-250 range is not the same as being below 50. The “grey zone” is a risk signal, not a shrug.

Unlocking the Trend at Scale: The Lothian IBD Registry

If thresholds are the language, then trajectories are the story. But you can’t do trajectory thinking without the right infrastructure.

In Edinburgh, we’ve been fortunate to build something unusual: a comprehensive, validated, population-based IBD registry linked to every FC measurement performed in the region.

Gareth Jones led the publication describing this resource in Gut in 2019. The Lothian IBD Registry (LIBDR) covers nearly 900,000 people, 16.5% of Scotland’s population, with case ascertainment completeness of 94.3%.[5]

By 2018, IBD prevalence in Lothian had reached 784 per 100,000: 1 in 125 people. We projected this would rise to 1 in 98 by 2028. The ageing of our IBD population presents new challenges: by 2028, patients over 40 will outnumber our entire current caseload.

Recently, we’ve fully integrated LIBDR with Lothian’s DataLoch, giving us an even richer research resource linking IBD data with the broader health and social care data ecosystem.

The registry gave us the foundation for everything that followed. Every FC measurement, linked to every patient, across every clinical encounter.

So what? It turned FC from a test into a dataset, and it meant we could start asking better questions than “what’s your number today?”

Enter Data Science: Modelling Longitudinal Inflammation

By the early 2020s, we had two decades of FC data. Tens of thousands of measurements. But we were still thinking about FC in static terms: one measurement at a time, threshold-based decision making.

Nathan Constantine-Cooke changed that.

Nathan joined us as a data scientist, first as a PhD student and now as a postdoctoral researcher. His question was elegantly simple: instead of looking at individual FC values, what if we modelled the trajectories? Could we identify patterns of inflammatory behaviour that capture disease heterogeneity better than traditional classification?

Critical to this work has been our collaboration with Catalina Vallejos, who co-supervised Nathan’s PhD and now effectively co-supervises all the data scientists in my group. Her statistical expertise has been fundamental to everything we’ve done in this space.

Using latent class mixed models (LCMMs), Nathan analysed 2,856 FC measurements from 356 newly diagnosed Crohn’s disease patients.

Four distinct trajectory clusters emerged:

• Early rapid responders (26%): sharp FC decrease immediately after diagnosis, sustained remission (46% received early biologic therapy)

• Persistently elevated (54%): FC remains high throughout follow-up, never dropping below 250 μg/g (only 18% received early biologics)

• Delayed responders (16%): improvement, but not until about a year after diagnosis

• Loss of response (4%): initial improvement followed by recurrence, the classic secondary loss-of-response pattern

This was no longer about single FC values. It was about understanding disease behaviour over time. The paper, published in Clinical Gastroenterology and Hepatology in 2023, introduced a new way of thinking about IBD phenotyping.[6]

If you’re trying to practice treat-to-target properly, this matters, because trajectories capture the behaviour of the disease, not just the last result.

Expanding the Framework: FC and CRP Together

Nathan’s next step was to expand this approach: larger cohorts, longer follow-up, and integration of CRP alongside FC.

This work involved a major international collaboration with Professor Tine Jess, who established the PREDICT centre for the molecular prediction of IBD in Denmark. Marie Vestergaard led the Danish analysis. The collaboration has been so productive that I’m now formally an adjunct professor at Aalborg University.

The results, published in Clinical Gastroenterology and Hepatology in December 2025, analysed 1,036 Scottish patients in the FC discovery cohort with replication in 7,880 Danish patients, and separately, 1,838 Scottish patients in the CRP discovery cohort with replication in 10,041 Danish patients. Eight distinct inflammatory trajectory clusters emerged, reproducible across both cohorts.[7]

A crucial finding: FC and CRP trajectories don’t always align. Only 72% of patients in the most favourable FC cluster were also in low-CRP clusters. This makes biological sense. FC reflects mucosal inflammation while CRP captures systemic response. But it has profound implications for clinical decision-making.

So what? FC and CRP are telling you different things. If you only listen to one, you can miss the plot.

The Impact: Changing Outcomes in Lothian

Has any of this made a difference? The population-level data suggests it has.

Colectomy rates in ulcerative colitis have fallen dramatically. From 2.48 per 100 UC patients in 2004 to just 0.22 per 100 in 2023, a 91% reduction. Advanced therapy prescribing increased in parallel, with a critical inflection point in 2013 when both trends accelerated.[10]

We’ve seen a similar direction of travel in Crohn’s disease since adopting a more top-down approach for newly diagnosed moderate-to-severe cases around 2015, with falling surgical rates over time. We’re currently working on an updated analysis, but the earlier dataset is already published.[12]

IBD-related hospitalisations have declined too. Age-standardised admission rates fell from 39.4 per 100,000 population in 2010 to 25.5 in 2019, a 36% reduction, with an average annual decrease of 3%.[11]

Let’s be clear about what’s driving this. The improved outcomes are largely down to new drugs in both CD and UC, and using them better. Top-down therapy in CD over the last decade has been transformative. We have more effective treatments and we’re deploying them earlier and more aggressively.

But treat-to-target with FC has undoubtedly changed the game too.

It’s given us the objective feedback loop we needed to know whether our treatments are actually working at the mucosal level. Correlation isn’t causation, and multiple factors contribute. But the temporal alignment is striking: as we’ve deployed FC-guided monitoring, escalated treatment earlier, and targeted inflammation rather than symptoms, hard outcomes have improved.

New drugs moved the ceiling. FC helped us stop walking into it.

What We’ve Learned

Twenty years of FC work in Edinburgh has taught us several things:

1. Lower is better, but 250 is still a sensible target for most. FC <50 μg/g represents genuinely quiescent disease, but achieving this isn’t always possible. Getting below 250 still delivers major benefits. Don’t let perfection be the enemy of good.

2. Trajectories matter more than snapshots. A single FC value is useful; a pattern over time is transformative.

3. Symptoms and inflammation diverge. Patients can feel well while inflamed, and feel terrible without inflammation. FC provides the objective ground truth.

4. Early effective therapy works. Patients who achieve FC normalisation early have dramatically better outcomes.

5. Implementation matters. Same assay, educated patients, practical logistics, and a culture of monitoring throughout the clinical team.

6. Infrastructure enables insight. LIBDR, with every FC linked to every patient, made this entire body of work possible.

The People Behind the Programme

This story is also a people story.

It starts with clinical fellows asking practical questions (Nick Kennedy on triage and prognosis; Nik Plevris on early normalisation; Gareth Jones on imaging correlation and the registry; Elliot Gemmell picking up the baton), and it now includes a data science group that lets us see the disease in motion (Nathan Constantine-Cooke and Catalina Vallejos, Victor Padro and Alex Rudge), plus the wider collaborations that let us test whether ideas generalise.

And none of it works without the multidisciplinary clinical system: nurses, pharmacists, dietitians, administrators, lab scientists, and yes, surgeons, all speaking the same simple language when they look at an FC result.

What Comes Next: SENTINEL

All of this has been building toward something.

We’re now building SENTINEL, an AI-driven clinical decision support tool for IBD. It’s the product of everything we’ve learned: two decades of FC data, trajectory modelling, and a clearer understanding of what predicts outcomes.

Here’s the problem SENTINEL is built to solve: even when you know trends matter, the reality of clinic is that you’re often looking at the last two results, in a patient you haven’t met before, on a busy Tuesday afternoon.

SENTINEL is designed to change that.

Imagine an IBD service that can see which patients are quietly losing control, not next week when they call with symptoms, but now, from the biomarker data we already collect. A system that spots drift, flags risk, and nudges the right action before the crisis.

We’ll start by piloting it in NHS Lothian, then iterate and scale with partner centres. I’ll cover the build, the workflow, and what we’re measuring in the next post.

But it’s important to understand that SENTINEL didn’t emerge from nowhere. It’s the culmination of a 20-year journey: from the first FC assays in 2005, through foundational work on diagnosis and prognosis, to longitudinal modelling and population-level outcomes.

The simple stool test turned out to be more important than we ever imagined.

Next in this series: SENTINEL — Building an AI-Driven Future for IBD Care

References

Edinburgh / Lothian work

[1] Kennedy NA, Clark A, Walkden A, et al. Clinical utility and diagnostic accuracy of faecal calprotectin for IBD at first presentation to gastroenterology services in adults aged 16–50 years. J Crohns Colitis. 2015;9(1):41–49. doi:10.1016/j.crohns.2014.07.005.

[2] Kennedy NA, Jones G-R, Plevris N, et al. Association between level of fecal calprotectin and progression of Crohn’s disease. Clin Gastroenterol Hepatol. 2019;17(11):2269–2276.e4. doi:10.1016/j.cgh.2019.02.017.

[3] Plevris N, Fulforth J, Lyons M, et al. Normalization of fecal calprotectin within 12 months of diagnosis is associated with reduced risk of disease progression in Crohn’s disease. Clin Gastroenterol Hepatol. 2021;19(9):1835–1844.e6. doi:10.1016/j.cgh.2020.08.022.

[4] Jones G-R, Fascì-Spurio F, Kennedy NA, et al. Faecal calprotectin and magnetic resonance enterography in ileal Crohn’s disease: correlations between disease activity and long-term follow-up. J Crohns Colitis. 2019;13(4):442–450. doi:10.1093/ecco-jcc/jjy187.

[5] Jones G-R, Lyons M, Plevris N, et al. IBD prevalence in Lothian, Scotland, derived by capture–recapture methodology. Gut. 2019;68(11):1953–1960. doi:10.1136/gutjnl-2019-318936.

[6] Constantine-Cooke N, Monterrubio-Gómez K, Plevris N, et al. Longitudinal fecal calprotectin profiles characterize disease course heterogeneity in Crohn’s disease. Clin Gastroenterol Hepatol. 2023;21(11):2918–2927.e6. doi:10.1016/j.cgh.2023.03.026.

[7] Constantine-Cooke N, Vestergaard MV, Plevris N, et al. Large-scale clustering of longitudinal fecal calprotectin and C-reactive protein profiles in inflammatory bowel disease. Clin Gastroenterol Hepatol. 2025 Dec 29:S1542-3565(25)01069-9. doi:10.1016/j.cgh.2025.12.022.

[8] Plevris N, Lees CW. Disease monitoring in inflammatory bowel disease: evolving principles and possibilities. Gastroenterology. 2022;162(5):1456–1475.e1. doi:10.1053/j.gastro.2022.01.024.

[9] Constantine-Cooke N, Gros B, Plevris N, et al. Associations between demographic, clinical and dietary factors and flares in inflammatory bowel disease: the PREdiCCt prospective cohort study. Gut. 2026 Jan 19:gutjnl-2025-337846. doi:10.1136/gutjnl-2025-337846.

[10] Elford AT, Constantine-Cooke N, Jenkinson PW, et al. Twenty-year trends in colectomy rates and advanced therapy prescribing in Lothian, Scotland. Aliment Pharmacol Ther. 2025;62(7):699–709. doi:10.1111/apt.70240.

[11] Lyons M, Derikx LAAP, Fulforth J, et al. Patterns of emergency admission for IBD patients over the last 10 years in Lothian, Scotland: a retrospective prevalent cohort analysis. Aliment Pharmacol Ther. 2022;56(1):67–76. doi:10.1111/apt.16867.

[12] Jenkinson PW, Plevris N, Siakavellas S, et al. Temporal trends in surgical resection rates and biologic prescribing in Crohn’s disease: a population-based cohort study. J Crohns Colitis. 2020; Advance Access Aug 25, 2020. doi:10.1093/ecco-jcc/jjaa044.

Other key calprotectin literature (summarised in Appendix below)

[13] Tibble J, et al. Gut. 2000;47:506–513.
[14] Van Rheenen PF, et al. BMJ. 2010;341:c3369.
[15] Mosli MH, et al. Am J Gastroenterol. 2015;110(6):802–819.
[16] Costa F, et al. Gut. 2005;54:364–368.
[17] Louis E, et al. Gastroenterology. 2012;142:63–70. (STORI)
[18] Mao R, et al. Inflamm Bowel Dis. 2012;18(10):1894–1899.
[19] Heida A, et al. Inflamm Bowel Dis. 2017;23(6):894–902.
[20] Shi J-T, et al. J Clin Med. 2023;12(3):1206.
[21] Colombel JF, et al. Lancet. 2017;390(10114):2779–2789. (CALM)
[22] Turner D, et al. Gastroenterology. 2021;160:1570–1583. (STRIDE-II)
[23] Wright EK, et al. Gastroenterology. 2015;148:938–947. (POCER)
[24] Boschetti G, et al. Am J Gastroenterol. 2015;110:865–872.
[25] Makanyanga JC, et al. Eur Radiol. 2014;24:277–287.
[26] Wei S-C, et al. Intest Res. 2018;16(4):546.
[27] Smith ES, et al. J Clin Gastroenterol. 2024;58(6):607–613.
[28] Foster AJ, et al. World J Gastroenterol. 2019;25(10):1266–1277.

Appendix: Table 3 (Global landmarks)

For readers who want the wider FC context beyond Edinburgh, here are the headline external studies and thresholds.

This is the first major paper from a large UK prospective cohort designed to answer a deceptively simple question: when people with IBD feel they are in remission, what actually predicts who will flare next?

The Journey

We started PREdiCCt in 2016 with a deliberately ambitious aim: build a deeply phenotyped cohort in routine clinical care, capture diet and lifestyle data prospectively, and link this to clinically meaningful outcomes over years, not months.

We recruited 2,629 people in self-reported remission across 47 NHS centres, collected detailed baseline data including faecal calprotectin and validated dietary questionnaires, and followed them for a median of just over four years.

Two flare definitions, on purpose

IBD flare is not a single concept, and treating it as one is a major source of confusion in both clinical practice and research.

In PREdiCCt, we separated them deliberately:

1. Patient-reported flare
   Defined via monthly questionnaires, using the question: “Do you think your disease has been well controlled in the past 1 month?”

2. Objective flare
   Required symptoms plus treatment escalation and biochemical evidence of inflammation, defined as CRP at least 5 mg/L and/or faecal calprotectin at least 250 µg/g.

Over 24 months, patient-reported flares were common (31%). Objective flares were less frequent (14%). That disconnect is not a footnote. It is one of the central messages of the study, and it sets up what is coming next.

The symptom-inflammation disconnect is not a footnote here. It is one of the central messages.

The early warning signal

Baseline faecal calprotectin was the strongest predictor of future flare across both Crohn’s disease and ulcerative colitis.

Two things matter clinically. First, the signal was dose-responsive: even in people who felt perfectly well at enrolment, higher calprotectin translated into higher subsequent risk. Second, the “grey zone” mattered. Risk was meaningfully increased even in the 50–250 µg/g range, compared with levels below 50.

To make that tangible: in ulcerative colitis, the probability of an objective flare within two years rose from 11% in those with baseline calprotectin below 50 µg/g to 34% in those above 250. Subclinical inflammation carries a real cost, even when symptoms are quiet.

This reinforces a simple point: subclinical inflammation carries a real cost, even when symptoms are quiet. It also raises an important question for treat-to-target practice, where FC thresholds of 250 µg/g are often used pragmatically.

PREdiCCt suggests that more stringent targets may be required for truly low relapse risk.

The diet signal in ulcerative colitis

Diet is where the evidence has often been loud and thin. PREdiCCt brings more signal and less speculation.

In ulcerative colitis, higher habitual total meat intake was associated with roughly double the risk of an objective flare. The absolute two-year risk rose from 12% in the lowest quartile of meat intake to 26% in the highest. When we looked at granular categories, the strongest signals were for unprocessed red and white meat. Fish was not associated with flare risk.

Just as importantly, we need to be explicit about what we did not show. We found no consistent associations between objective flares and ultra-processed foods, dietary fibre, polyunsaturated fatty acids, or alcohol. And we did not see these dietary associations in Crohn’s disease.

What this means

This paper is the opening chapter, not the full story, but it gives us a practical foundation for preventive care.

If we want to prevent inflammatory flares, we need to use biomarkers like calprotectin proactively rather than reactively. The ulcerative colitis meat signal is credible and actionable, and it should now motivate targeted intervention trials rather than sweeping dietary claims. And recognising the difference between symptom flares and inflammatory flares helps patients, clinicians, and researchers interpret signals correctly.

The team and what comes next

This study was a decade in the making and represents a huge collective effort across data scientists, research nurses, and investigators from the UK.

Most importantly, it was made possible by 2,629 people living with IBD who gave their time and trust. This work exists because of you.

PREdiCCt has much more to say. Next come the psychosocial analyses, which sit naturally alongside the symptom-inflammation disconnect highlighted here, and the deep metagenomic analysis of the gut microbiome.

The short version: feeling well is not the same as being at low inflammatory risk. PREdiCCt shows us how to tell the difference

Key points

• We recruited 2,629 people with Crohn’s disease, ulcerative colitis, or IBD-unclassified in self-reported remission from 47 UK centres, and followed them for a median of just over 4 years.

• We deliberately separated symptom flares from inflammatory, treatment-escalating flares by using two endpoints: patient-reported flare and objective flare.

• Over 24 months, patient-reported flares were common (31%), while objective flares were less frequent (14%).

• Baseline faecal calprotectin (FC) was the strongest predictor of both patient-reported and objective flares, including within the 50-250 µg/g range.

• In ulcerative colitis, higher habitual total meat intake was associated with higher risk of objective flare (around doubled risk comparing highest vs lowest quartile).

• We found no consistent associations between objective flare risk and ultra-processed foods, dietary fibre, polyunsaturated fatty acids, or alcohol intake.

• The symptom-inflammation disconnect is not a footnote in PREdiCCt. It is one of the central messages, and it sets up what is coming next, especially the psychosocial work.

Browse the data online

You can browse the full PREdiCCt dataset here thanks to the brilliant Nathan: https://www.constantine-cooke.com/predicct-analysis/

It was a holiday with my daughter, Holly. Not the lying-on-a-beach version. More the walking-until-your-feet-complain version. Temples, markets, boats, street corners, and long conversations with people who live inside the systems we usually talk about from a distance.

I also arrived there already tired. The trip followed a hard work sprint through Q4 2025, the kind that leaves your nervous system running a little too hot even when you finally stop.

Somewhere between Singapore, Cambodia, and Vietnam, the holiday became something else too.

An anthropological field trip, with IBD as the lens.

Not in a grandiose way. In the simple sense that travel lets you see variables, not just scenery.

1. Arrival: displacement as data

We left the Scottish winter and landed in Singapore into heat, humidity, and immediate sensory density.

The shock to the system was not just jet lag. It was the body noticing before the brain catches up. Temperature changes how you breathe. Air conditioning changes your sleep. Noise and light set a tempo. Food availability is different, not only in what exists, but in how constant it is.

I arrived tired, overstimulated, and immediately aware that my body was reading the environment before my mind had words for it.

Travel strips away the illusion that our baseline is universal. If a change in environment can alter how you feel within 24 hours, it is worth asking what a lifetime of environmental mismatch does to a gut that is already prone to inflammation.

Not as a slogan. As a prompt.

2. Three countries, three systems

Singapore, Cambodia, and Vietnam sit close together on the map. They do not feel like they live inside the same operating system.

Singapore is an outlier. It is almost a perfected social experiment in dense urban living. It has to be, because the population is roughly Scotland’s, compressed into a tiny space.

It is extraordinarily organised, efficient, rule-bound. There is a sense of trust in the system, and a low tolerance for things that erode it. Hawker culture is still, for many, genuinely good food at scale. And obesity is strikingly uncommon. It is genuinely unusual to see fat people.

Then Cambodia, based around Siem Reap and the countryside beyond it. Here the difference is not just infrastructure. It is what daily life requires of you.

In rural areas, you take care of everything yourself. Water, food, repairs, work, family, logistics. That simplicity is real, and it can bring a kind of grounded competence that is hard to replicate in a hyper-specialised digital life.

It also comes with struggle. Heat, uncertainty, fewer buffers. Development is not just aspiration, it is relief. People talked openly about corruption too, the sense of potential being slowed by systems that siphon rather than build.

And the history sits there, quietly refusing any simplistic story about “modernity.” The brilliance of the engineers more than a millennium ago, putting water management systems in place that made Angkor possible. The sheer magnificence of Angkor Thom and Angkor Wat is not just beauty. It is proof of capability and long-range thinking.

Between Cambodia and Vietnam, we slowed down properly. We spent three to four days sailing down the Mekong from Phnom Penh to Saigon. River life becomes a moving cross-section of exposures and culture. The way people live, the way commerce flows, the way the language and the food shift. We took trips off the boat onto local islands to meet families and farmers, and to see daily routines up close. It felt like a transition in real time.

Finally Vietnam. Dense, kinetic, slightly chaotic in the best way. A country in motion, with momentum you can feel. Street life, markets, scooters, construction, commerce. People spoke about corruption here too, but with a different tone. Less resignation, more sense that things are moving, improving, tightening.

These countries sit side by side, but they live in different metabolic eras.

And IBD incidence follows development curves, not borders.

3. Diet is part of the story, but it is not the whole story

We did talk about diet, constantly. Not usually as “IBD diet.” More as an entry point into everything else.

Food reveals refrigeration. Refrigeration reveals electricity. Electricity changes what people can store, what they can buy, what is available at night, how supermarkets arrive, how snacking culture appears, how ultra-processed food becomes normal.

In Cambodian villages there is an electricity paradox. It is there, but it is too expensive for many to use beyond basic light at night. Air con for a month can cost $100 to $200, completely unaffordable for most.

Ultra-processed food is not just a dietary exposure. It is a cultural one.

So are antibiotics. Access, misuse, expectation. People mentioned them casually, as tools you reach for. Traditional medicine came up too, not as folklore, but as a parallel system with its own logic and history. The more you listen, the more you realise that “health behaviours” are not behaviours. They are culture and constraint.

One repeated theme, especially in Cambodia and Vietnam, was metabolic disease. Diabetes and obesity in particular, across all ages, especially in the cities. Again, Singapore is the outlier.

IBD is still relatively uncommon in many of these settings, but you can sense where the curve is heading. These are stage 2 countries, in the acceleration phase. Cambodia feels like it is only just entering it.

What I kept thinking was not “what should they do.” It was “what did we normalise without noticing.”

4. Movement, work, and the loss of life skills

In the UK, we try to repair a sedentary system with an hour in the gym. We compartmentalise movement as “exercise,” something you schedule to compensate for the rest of the day.

In much of Southeast Asia, movement is not an activity. It is the background.

In rural Cambodia, nobody “exercises”. They move.

Walking, carrying, lifting, standing, squatting, tending, trading. Daily life has load built into it. Not as virtue, as necessity.

The other contrast is subtler, and it stayed with me. When you have to take care of everything yourself, you retain competence. Practical life skills are not hobbies, they are inheritance. In the modern Western urban experience, we outsource almost everything. Convenience is real, but the trade-off can be a strange fragility. Digital life, loneliness, depression, and hard work that never quite ends.

Not a moral judgement. Just a contrast you can feel when you move between systems.

5. Why this matters for IBD

IBD is not just immune dysregulation. It is the biological cost of modern life, expressed through an immune system that evolved for a different world.

That line can be abused if it becomes a single-cause story, so I mean it carefully. IBD is heterogeneous. Genetics matters. Treatment matters. Social determinants matter. Access to care matters. But it is increasingly hard to look at the global pattern, and at the lived mechanics of urbanisation, and pretend environment is a side issue.

This trip also included work.

In Hanoi University Hospital, I gave a seminar on diet and IBD, and spent time with Dr Hang Viet. Back in Singapore, I had a stimulating roundtable discussion with CJ Oi and colleagues.

One clear theme is that Vietnam is changing fast, and the epidemiology will change with it. I will write a follow-up post specifically on Vietnam, with a deeper dive into the data, the environment, and the shifts over time.

And a flag here too. The PREdiCCt paper is out in Gut on Tuesday, with an accompanying blog post here.

The human thread

I travelled with Holly, who turns 21 in two months. I turn 50 in one.

We settled into a really good rhythm together, having not spent this much time side by side for a long time. I am back to work properly on Monday. She is back to university, forensic psychology.

Cameron chose to stay behind, instead visiting a friend in freezing Stockholm during the university break. History and politics. A very different kind of field trip.

Travel does not give answers. It sharpens questions.

This region has also left me with a growing list. Malaysia, Indonesia and Thailand next, I hope. Back to China. More of India. South America, where I still need to tell the stories from last summer. And Africa too. I keep thinking about Nigeria, and what is changing there, and how fast.

IBD is global. Our thinking needs to be too.

Between 20 Dec and 3 Jan I published a cluster of posts: one personal piece, then a four-part year-end review and look-ahead. If you were off-grid (or your NHS inbox was in holiday triage mode), these probably landed straight in the backlog.

This is the catch-up post.

Pick your route

If you’re mainly in clinic: Part 1 → Part 3
If you’re mainly in research: Part 2 → Part 4
If you run services / pathways: Part 1 → Part 4 → Part 3

The posts

0) The Life I’m Living (Dec 21)
A personal reflection written in the days after my inaugural lecture. Not IBD content. On the lives we don’t choose and the one we’re actually living.

1) The Great Remodelling of IBD Care
The therapeutic hierarchy shifted. IL-23 p19 moved up. Comparative data finally arrived. Sequencing became the story.

2) The Papers That Changed Practice
Fistula biology, strictures, TL1A crowding, global epidemiology, paediatric trials, appendicectomy, and the work that made 2025 feel like a hinge year.

3) What to Look Forward to in 2026
ABTECT maintenance, DUET combination therapy, VERDICT treat-to-target. The readouts that will define the year.

4) Predict • Personalise • Prevent
The Edinburgh IBD Research Group year in review. SENTINEL launching this month. PREdiCCt outputs. The team, the methods, and what we’re building next.

The through-line

Sequencing is now the clinical question.

Not “does it work?” but “what works first, for whom, for how long; and what do we do when it doesn’t?”

One request

If you catch up on any of these, I’d genuinely like one thing: what did you disagree with?

That’s the fastest way Atomic IBD stays sharp.

Any trial timelines and endpoints are worth verifying with primary sources before you put them in a slide deck.

For now, I’m switching off

Currently somewhere on the Mekong with Holly, travelling from Cambodia into Vietnam. Later this trip: a talk on diet and IBD at Hanoi University Medical Centre, time in their endoscopy unit and IBD clinic, and a stop in Singapore to meet colleagues before flying home.

Reflections to come. But not yet.

“We are, genuinely, in the disease-modifying era of IBD. The task now is to combine science, strategy and trust so that most people, most of the time, can live the long, full lives they deserve.”

Why This Matters Now

IBD care is entering a new era. Complexity has overtaken capacity.

In 2025, the field shifted. Therapy sequencing became the story. The hierarchy changed under the weight of comparative evidence. The question stopped being “what works?” and became “what works first, for whom, for how long, and can we deliver it in real clinics?”

Science is moving fast. Clinics are not.

That gap between what science makes possible and what health systems can reliably deliver is exactly where our work sits.

Our group’s mission is simple to say and hard to do: predict, personalise, and prevent inflammatory bowel disease. We are building data-driven tools that learn from the signals already present in routine care, combine them with prospective multimodal cohorts, and translate that into practical, compassionate, world-leading care.

This is a review of what we built in 2025, who built it, what we learned, and the 2026 plan. The year we turn platform into practice at scale.

What We Are Building

We are building a transformative IBD medicine programme designed to improve lives wherever people live. Through research. Through compassionate care. Through scalable digital translation.

At its centre is SENTINEL.

SENTINEL is an AI-driven, clinician-supervised decision support tool designed to shift IBD care from reactive and crisis-driven to proactive and personalised. It runs alongside the EHR. It turns routine inflammation signals into daily, pathway-aligned action prompts for clinical teams. It learns each patient’s unique inflammatory pattern from routine CRP and faecal calprotectin results, predicting deterioration weeks before symptoms develop.

It has one rule: usefulness first.

The system generates daily risk-prioritised lists and trajectory-aware prompts for nurse-led triage and treatment optimisation. Critically, it works from labs alone. Benefits accrue from day one even if patients never engage with additional features. Equitable and operationally robust.

Around SENTINEL sits an ecosystem:

• The Lothian IBD Registry: approximately 10,000 patients, 20+ years of longitudinal data, 20 years of faecal calprotectin results all measured on the same laboratory assay. A data asset that cannot be easily replicated. Now fully integrated with DataLoch.

• Danish national registry data through our collaboration with PREDICT at Aalborg University, validating methods across 70,000+ patients.

• PREdiCCt, our UK-wide prospective cohort of 2,629 participants across 49 NHS sites, capturing deep baseline phenotyping with longitudinal follow-up. Built for prediction, prevention, and real biological insight into why flares happen.

• A broad collaborative network: nationally, and globally - across epidemiology (GIVES-21), genetics (UK IBD Bioresource and International IBD Genetics), microbiome and diet (IBD-response, PREdiCCt), metabolomics (Nightingale Health), multi-omics (IBD-OPEN), translational clinical research (PANTS, PRED4, CLARITY, VIP, PROFILE), and policy (the Lancet Commission on Global IBD).

• And Atomic IBD. My effort to translate the fast-moving science of IBD into signal, clarity, and hope for clinicians, patients, and partners worldwide.

The Foundations

I started gastroenterology training in 2003. I’ve been a consultant in NHS Lothian since 2009.

I am a clinician first.

The clinic is where this entire programme comes from. Not as an abstract “digital health” idea, but as a response to real pain points: inflammation rising quietly in the background, patients becoming unwell before we can get to them, helplines overwhelmed, capacity stretched, decisions increasingly complex.

When I began in Edinburgh, too much of IBD care meant cycling steroids to control flares and accepting too many surgeries while chronic inflammation quietly accrued harm over decades. Today, we can set people up for success from day one. Prompt diagnosis. Full assessment. Effective therapy early rather than asking patients to “earn” their way to it.

Edinburgh is unusually well placed for what comes next. World-class biomedical data science at the Institute of Genetics and Cancer, paired with one of the most complete longitudinal IBD real-world cohorts globally.

2025: Platform Year

This was a year defined by turning what we’d built over decades into something deployable.

Infrastructure

The Lothian IBD Registry is not just a dataset. It’s a living clinical history of IBD in an entire region.

In 2025, the integration with DataLoch matured into something that can support real translation: the pathway from routine data to analytics to action in clinical workflows. This matters because modern IBD care has outgrown the old model of waiting for symptoms to declare themselves.

If we can learn from routine longitudinal biomarkers, we can create earlier, calmer interventions. Fewer crises. Fewer steroids. Fewer emergency admissions. More trust in care.

Methods

We’ve been building an approach to IBD that treats inflammation as a trajectory, not a snapshot.

In 2025, that trajectory view matured into a platform capability. Nathan Constantine-Cooke led the work mapping inflammatory trajectories using faecal calprotectin and CRP. Using latent class mixed models, we identified eight distinct FC and eight CRP trajectory patterns: rapid remitters, delayed remitters, relapsing-remitting disease, and non-remitters. These were independently replicated in Danish national registry data with Professor Tine Jess at PREDICT.

The paper was published last week in Clin Gastro Hep.

Victor Velasco-Pardo built landmaRk, an R software package that combines landmarking with latent class mixed models. Tested, validated, and ready for operationalisation. Stats paper online soon.

This is what powers SENTINEL. A system that learns each patient’s pattern and uses new lab results as updates, not noise.

Translation

SENTINEL moved from prototype thinking to genuine deployment readiness. A clinician portal. Lab ingestion. Workflows. A design philosophy that puts equity first.

The MVP is functional. CRP and faecal calprotectin ingestion operational. Phenotypic inputs from the registry integrated. Live wearable feed incorporated in the demo environment.

This is not AI for AI’s sake. It is service redesign powered by predictive analytics.

PREdiCCt

PREdiCCt shifted into output mode. The first paper, accepted by Gut, shows baseline faecal calprotectin predicts objective flares even when patients feel well. A UC-specific association between higher meat intake and roughly double objective-flare risk.

The second paper, presented at ECCO 2025, found that depression and lack of exercise predict objective flares in ulcerative colitis. Signals that these factors belong in relapse-risk assessment alongside biomarkers.

A major metagenomics paper is forthcoming.

Publications

Our work continues to drive real changes in clinical practice worldwide.

The Nature paper from Hracs et al. and the GIVES-21 consortium documented IBD’s global evolution across epidemiologic stages. Ebert et al. in the American Journal of Gastroenterology delivered the first phenome-wide association study across more than 1,500 comorbidities. Gros et al. on ftofacitinib versus vedolizumab in UC (JCC). Twenty-year trends in ulcerative colitis: colectomy rates and advanced therapy prescribing in Lothian published in AP&T.

Trials

Edinburgh remains a major centre for IBD clinical trials, both commercial and investigator-led.

Commercial studies currently recruiting include TARGET (AbbVie phase 2 platform combination study in CD), VICTREVA (Takeda vedolizumab plus upadacitinib combination in CD), VECTORS (vedolizumab treat-to-target with intestinal ultrasound in early CD), and ICONIC (Janssen phase 3 oral IL-23 icotrokinra in UC).

Non-commercial studies include OPEN-IBD, IBD-RESPONSE, ACCESSO, OCEAN, GONDOMAR, and the IBD Bioresource.

The future of IBD care will not be built in one type of study. It will be built at the intersection of real-world outcomes, rigorous trials, mechanistic insight, and care pathways that can scale.

Atomic IBD exists because science only helps patients when it becomes practice.

The platform continued to grow: 4,174+ newsletter subscribers, 21,200+ followers on X, 2,950+ YouTube subscribers. The year-end review series captured the field’s inflection point (Part 1 - Part 2 - Part 3). Two premium video productions filmed at Elstree Studios (very excited to release these soon). 30 to 40 Atomic IBD Uncut episodes. Regular atomic essays.

For our group, Atomic IBD is not separate from research. It is part of the translation pipeline.

The People

No serious programme is the work of one person.

Multidisciplinary. Co-located. Aligned.

The Research Group

Our post-doctoral data scientists sit at the heart of what makes our approach distinctive. Nathan Constantine-Cooke leads SENTINEL technically and pioneered our inflammatory trajectory modelling. Victor Velasco-Pardo created the landmaRk software and leads our cumulative inflammatory burden work. Alex Rudge drives forward our analytical capabilities. This team sits jointly across our group and Dr Catalina Vallejos’ biomedical data science group.

Our clinical fellows bridge the gap between data science and patient care. Clara Ramos-Belinchon works with Victor on cumulative inflammatory burden scoring. Elliot Gemmell advances our clinical data science capabilities. Alex Elford and Beatriz Gros have been prolific in producing real-world evidence. Nik Plevris, now consultant locally, provides the clinical-academic bridge that defines our approach.

Our research operations team makes large platforms real. Debbie Alexander leads, with Claire McQuillian and Louise Taylor. Asim Azfer runs our laboratory science.

Edinburgh IBD Clinical-Academic Community

Research in Edinburgh extends well beyond any single group.

Jonny Blackwell is local PI for OPEN-IBD. Shahida Din leads Accesso locally with her fellow Amir and runs a programme on colitis-associated cancer and Crohn’s fibrosis with local collaborators. Rahul Kalla leads FATE-CD at the Royal Infirmary, using PET imaging to study fibrosis in Crohn’s disease with his fellow Lex. Eleanor Watson (at WGH) and Richard Russell (at the Sick Kids Hospital) lead clinical implementation for SENTINEL. Ian Arnott, clinical director for gastroenterology, has supported IBD research in Edinburgh for over two decades.

This breadth of clinical-academic activity, all under one regional service, is rare. It makes Edinburgh a powerful environment for IBD research.

The Wider Clinical Team

None of this happens without the clinical infrastructure that supports it.

Our IBD specialist nurses and pharmacists are the backbone of patient care and trial delivery. Our administrators keep the machinery running. Our surgeons, radiologists, and pathologists contribute expertise that shapes every aspect of our work.

Challenges

A serious review has to name the friction.

Building in the NHS means building through complexity. Data governance, workflows, and service realities are not side issues. They are the core. Translation is mostly not about brilliant ideas. It’s about relentless operational execution.

Prediction is only useful when it changes action. The hardest step is not modelling. It’s the last mile: presenting risk in a way that clinicians trust, that nurses can use daily, and that improves outcomes without increasing burden.

Capacity is the hidden limiting reagent. That’s why “better drugs” is only part of the story. The real innovation is systems that help us deliver what we already know works.

2026: Delivery Year

This next year is about consolidation and delivery.

SENTINEL Goes Live

January 2026 marks the launch. Our objective: flip approximately 30% of nurse-helpline workload from reactive calls to proactive outreach based on predicted risk.

What success looks like:

• A shift from reactive helpline firefighting to proactive outreach

• A daily, risk-prioritised view of the caseload

• Trajectory-aware prompts aligned to pathways

• A learning system that improves through supervised use

This is the year we turn predictive models into an embedded clinical service.

Scale Readiness

Several major NHS boards beyond Lothian are committed as early adopter sites. We are building reproducible pipelines to ensure outputs are rapidly usable by others. Doing the hard work of evaluation, safety, and regulatory readiness.

Intentional Leadership

My personal focus: leadership, vision, and funding. More intentional travel and speaking. Fewer low and mid-tier engagements. Building SENTINEL and Atomic IBD as platforms that can scale impact far beyond what any individual can achieve in clinic.

The Lancet Commission on Global IBD reports at ECCO in February. A milestone in articulating the global challenge and the solutions we are building.

Work With Us

Come build with us.

We want collaborators who have cohorts or registries that can validate generalisability. Who want to co-design pathway-aligned decision support. Who can help us test the ethics, safety, and fairness of predictive systems in the real world.

We want partners who care about scalable multi-omics, remote monitoring, and pragmatic evaluation. Who want patient-centred innovation that works even when engagement is variable.

We want patients excited about the direction of travel. SENTINEL is designed to ensure that care finds you before crises occur.

Join Us

We are hiring.

• Advanced Clinical Fellow, Edinburgh IBD Unit. World-leading centre, cutting-edge research, advanced training in IBD and data science. Application deadline: 5 January 2026. How to apply

• AI4BI CDT PhD Studentship (Project 24). Fully funded 4-year programme for September 2026 entry: SENTINEL, multi-modal dynamic prediction for proactive, personalised IBD care. Application deadline: 20 January 2026. Project details | How to apply

Looking Forward

2025 was the year we built the platform. 2026 is the year we prove what it does.

A world-class NHS service. Long-run real-world data. Serious statistical modelling. Prospective multimodal cohorts. A high-energy team. Pointing in the same direction.

IBD is heterogeneous, dynamic, and unpredictable. For many it spans six or seven decades.

The task of modern IBD care is to hold optimism for the newly diagnosed while maintaining empathy for those living with IBD for 40, 50, 60+ years. SENTINEL represents our commitment to that task. Using the data we already collect, the methods we have developed, and the tools we are building.

This is the disease-modifying era of IBD. The task now is to combine science, strategy, and trust. And that is exactly what we are doing.

Connect

Professor Charlie W. Lees University of Edinburgh | NHS Lothian | UKRI Future Leaders Fellow

X: @charlie_lees Newsletter: charlielees.substack.com YouTube: @charlie_lees Research Group: Lees Group at IGC

Edinburgh IBD Research Group | Institute of Genetics and Cancer | University of Edinburgh January 2026

2026 is different. It’s not about another crowded mechanism. It’s about proof: that we can go deeper, go safer, go faster, and still deliver it in real clinics with real capacity constraints.

Not “what works?” but what lasts, what scales, and what changes outcomes.

The 2026 dashboard

Three readouts will define the year.

1. ABTECT maintenance (UC): Does obefazimod hold? Topline expected Q2 2026.

2. DUET-UC / DUET-CD: Does advanced combination therapy get its first rubber stamp? Primary endpoints at Week 48.

3. VERDICT: Will treat-to-target finally tell us how deep “deep” needs to be? UK 96-week follow-up completes March 2026.

Everything else matters too, but these are the anchors.

1. ABTECT maintenance: the “safe oral” bet gets its verdict

The story so far is clean.

In July 2025, Abivax reported positive Phase 3 induction results from ABTECT-1 and ABTECT-2 in moderate-to-severely active UC. The 50mg once-daily dose met primary and key secondary endpoints in a large, globally representative programme.

But induction is the trailer. Maintenance is the film.

What we’re watching for

Topline maintenance results expected Q2 2026. The questions are simple:

• Durability: Does the signal persist across a year?

• Safety: Does it stay boring? (Boring is good.)

• Positioning: Does it become an oral option for all lines of therapy?

Abivax has been explicit: if maintenance is positive, they plan to submit for FDA/EMA filings in H2 2026. That makes this the year of the data, not the year of the drug on shelves.

I wonder if we also see a major pharma acquisition next year?

2. DUET-UC / DUET-CD: can combination therapy break the ceiling?

If ABTECT is about a new therapy, DUET is about combining existing drugs.

The DUET programme tests a fixed combination of guselkumab + golimumab against monotherapy and placebo in UC and Crohn’s. The trial design tells you what Janssen thinks the problem is: not “we need another monotherapy” but “we need to stop pretending monotherapy is enough for everyone.”

DUET-UC (Phase 2, 577 participants) includes guselkumab, golimumab, and combination arms. DUET-CD has Week 48 primary endpoints with clinical remission and endoscopic response, plus an optional long-term extension.

The real question: is ACT (advanced Combination Therapy) a concept or a category?

Combination therapy has been slowly shifting from “wild idea” to “quiet reality,” particularly in refractory disease. But it’s still treated as exceptional, expensive, risky, hard to justify.

A positive DUET signal would be the first genuine rubber stamp for ACT as a category. Not just a clever case report.

If DUET hits, it won’t just be a result. It will be a permission structure.

Can we start already?

This is the uncomfortable bit.

If the combination is licensed, it becomes legible to payers, guidelines, and governance. If not, we’re stuck where clinicians improvise, systems resist, and patients experience postcode biology.

There are two possible futures:

Future A: combo becomes a product. DUET’s co-formulated combination pushes us toward a world where companies make ACT easy. One injection, one label, one reimbursement pathway.

Future B: combo becomes a practice. The biosimilar era makes combination therapy harder to ignore. People will ask whether low-cost pairs could deliver most of the benefit. The obvious thought experiment: biosimilar ustekinumab + biosimilar adalimumab.

But a plausible idea is not evidence. A cheap idea is not automatically a safe idea.

If DUET is positive, the right next step isn’t “everyone start mixing biologics.” It’s: build the trials and systems that make ACT safe, scalable, and equitable.

If DUET is negative? Still progress. It tells us that “more blockade” isn’t automatically “more healing” … or less dramatically that gus + goli is the wrong combination.

Either way, the field learns.

3. VERDICT: defining the target, not just chasing it

VERDICT is one of the most strategically important trials running in IBD.

It randomises patients with active UC to different treatment targets:

• Corticosteroid-free symptomatic remission

• Symptomatic + endoscopic remission

• Symptomatic + endoscopic + histologic remission

Funded through Takeda, using a structured algorithm with vedolizumab as the core escalation pathway, it’s trying to answer the question the field keeps circling:

Does deliberately pushing all the way to histological healing change the natural history?

We have interim analyses to Week 48, and a clear anchor: 96-week follow-up completes 2026; results hopefully at UEG in Barcelona.

The implication is operational, not ideological

VERDICT isn’t just an important data gap, nor just a philosophical argument about histology. It’s a resource question.

If histology meaningfully reduces complications:

• Biopsies become routine targets, not optional extras

• Colonoscopy (and pathology) capacity becomes a bottleneck

• AI-enabled standardisation becomes necessary

• Treat-to-target becomes even more demanding

If histology doesn’t add value beyond endoscopic remission:

• We stop over-measuring and over-treating

• We simplify targets without lowering ambition

Either outcome is clarifying. Clarity is rare.

The next wave gathering steam

The headline readouts matter. But the real work of 2026 is that the pipeline keeps laying track. Three threads to watch. Not because they’ll finish in 2026, but because they become unavoidable.

TL1A: the class that refuses to be ignored

Merck has stated that tulisokibart is in Phase 3 trials for UC and Crohn’s, while expanding into other immune-mediated diseases. TL1A is no longer “interesting biology.” It’s an industrial-scale bet.

Then there’s Spyre. SPY002 is a half-life-extended anti-TL1A antibody, now in the Phase 2 SKYLINE-UC platform trial. Spyre has guided that induction data are expected in 2026.

That’s the kind of milestone that changes the feel of a class.

Oral IL-23 receptor blockade: icotrokinra is the one to watch

J&J has initiated ICONIC-UC Phase 3 (adults and adolescents) and ICONIC-CD Phase 2b/3 following Phase 2b ANTHEM-UC results.

This matters because it pushes the IL-23 axis toward the holy grail: biologic-like efficacy, pill-like scalability.

Even without Phase 3 readouts in 2026, the implications start now: recruitment pace, trial design choices, and the first credible signs of gut translation at scale.

Other 2026 highlights

The Lancet Commission on Global IBD

ECCO has flagged the official launch at ECCO’26 (Stockholm, Feb 18-21, 2026).

The epidemiology is no longer a background slide. It is the story.

PREdiCCt: from signals to full manuscripts

PREdiCCt has been ten years of my life.

A UK prospective study of 2,629 people with IBD in remission (47 NHS centres, 2016–2020; median follow-up just over four years). The first paper, accepted in GUT and due online imminently, shows baseline faecal calprotectin predicts objective flares even when patients feel well (in UC, about 11% vs 34% two-year risk across low vs high calprotectin).

It also finds a UC-specific association between higher meat intake and roughly double objective-flare risk, not seen in Crohn’s.

The second PREdiCCt paper, presented at ECCO 2025, moves away from diet and looks at behavioural and mental health predictors. The key finding is that in ulcerative colitis, depression and lack of exercise predict objective flares. This is not a “mind over matter” story; it is a signal that these factors belong in relapse-risk assessment, alongside biomarkers.

There is also a major multi-omics layer coming.

We now have 2,000 gut metagenomes fully sequenced, through QC, and moving towards analysis. That should let us test how microbiome features interact with baseline inflammation, diet and behaviour in predicting relapse, using prospective follow-up rather than retrospective associations.

What to expect on the ground

This is the part we talk about least. It isn’t glamorous. But it’s the real constraint.

More patients. Everywhere.

IBD has a compounding problem: once prevalence is high, it stays high because people live with the disease for decades. Meanwhile, incidence is accelerating globally.

A recent Nature review (see part two) describes the global evolution across epidemiological stages. The forward-look literature is blunt: the burden is rising through 2025-2045.

More patients, more complexity, more therapy lines, more monitoring, more contact points.

Treat-to-target becomes the default

The field has agreed on the principle. VERDICT is testing the depth.

Whichever way VERDICT lands, the direction is obvious: more calprotectin, more scopes, more biopsies, more imaging, more treatment adjustments, more documentation, more shared decision-making.

It’s not “extra.” It’s the job.

And it will stress systems built for a lower-prevalence era. This is where monitoring systems and workflows that scale stop being a shiny add-on and become basic infrastructure.

A tidal wave of comparative effectiveness and RWE

It’s coming.

As the therapeutic menu expands, we’ll see pragmatic head-to-heads, registry comparisons, payer-driven analyses, biologic class positioning studies, safety-focused real-world datasets.

Useful. But they won’t deliver the one thing everyone asks for.

Don’t expect RWE to hand you “the” sequencing answer.

Because sequencing isn’t one question. It’s fifty questions, each conditional on phenotype, risk, prior exposure, time horizon, patient preference, and system capacity.

RWE will help with durability patterns, safety in broader populations, what happens outside pristine trial protocols.

RWE will not reliably answer: what you should do next in this patient, what you should have done first, what the counterfactual would have been.

Sequencing will remain a clinical craft. Increasingly informed, but never fully automated.

Closing

If 2025 was the year we rebuilt the therapeutic map, 2026 is the year we test whether that map can guide us somewhere better.

• ABTECT maintenance asks if a new oral mechanism can be durable and boring-safe.

• DUET asks if ACT can truly break the ceiling, and if it can, whether we can afford to ignore it.

• VERDICT asks what treat-to-target really means when you follow it all the way down to histology.

And underneath: more IBD, more everywhere, and a care model that has to evolve as fast as the science.

2026 won’t be a victory lap. It will be a year of consequences.

That’s a good thing.

This second post is the evidence base that made 2025 feel like a hinge year. It spans system-level epidemiology, deep tissue biology that finally gives names to fistula tracts and strictures, and trial data that changes what we say to patients next week. One important note before we start: this list is not exhaustive, and any omissions are my oversight.

Let’s get into it.

The global map

IBD is no longer “rising”. It is embedding.

The Nature macro-epidemiology paper from the Calgary and Hong Kong groups (Kaplan and Ng) is the clearest framing I have seen for what comes next: a predictable evolution across epidemiologic stages, and the uncomfortable arithmetic of compounding prevalence. Incidence may plateau, but total burden keeps climbing as patients live longer and health systems carry IBD across decades.

The point is not panic, it is planning.

Reference: Hracs L et al. Nature. 2025. Global evolution of inflammatory bowel disease across epidemiologic stages.

Fistulas, finally

Perianal Crohn’s is not a side quest. It is a phenotype with its own biology and its own evidence gaps.

Two things happened in 2025.

First, the deep tissue work got sharper: spatial atlases are starting to treat fistula tracts and strictures as structured niches, not just damage.

The deep spatial multi-omics study from Alison Simmons’s lab in Oxford builds a high-resolution atlas of 68 intestinal fistulae and shows that fistula tracts are not “just inflammation”, they have a structured, layered biology. They identify a fistula-associated stromal (FAS) fibroblast programme arranged in concentric zones (from proliferative, lumen-adjacent cells through an active lesion core to an outer pro-fibrotic zone), with tightly linked immune niches and extracellular matrix remodelling that likely stabilises and perpetuates the tract.

A key take-home is that common cellular pathways appear to underpin fistulae across anatomical locations, and that targeting fibroblast behaviour (selectively limiting invasive/fibrotic programmes while preserving repair) may be central if we want better medical options for fistulising Crohn’s disease.

Second, the clinical world started to converge around definitions that make trials possible.

A Lancet Gastroenterology and Hepatology systematic review and consensus paper tackled isolated perianal Crohn’s disease head-on, proposing diagnostic criteria and practical management advice, then testing the approach in a prospective cohort of 50 “idiopathic fistula” patients where 12% met criteria. That is exactly the kind of work that turns an awkward clinic label into something you can recruit, stratify, and study.

This is also where TopClass matters.

The consortium is explicitly trying to standardise how we define, image, phenotype, and treat perianal Crohn’s across centres and countries, so that trials are comparable and registries speak the same language.

If you cannot define it, you cannot trial it.
If you cannot trial it, you cannot fix it.

References: McGregor C et al. Nature. 2025. Spatial fibroblast niches define Crohn’s fistulae.; Hanna et al. Lancet Gastroenterol Hepatol. 2025. Isolated perianal Crohn’s disease consensus and criteria.; TopClass Consortium (international perianal Crohn’s collaboration).

Fibrosis gets a cell type

Strictures and fistulas have always been where medical therapy has struggled in Crohn’s disease.

The 2025 shift is that fibrosis is starting to look like a tractable network: spatial modules, specific fibroblast programs, immune crosstalk, and the first real bridges from genetic risk to tissue behaviour. The stricturing Crohn’s spatial paper (Boston heavyweights here) is the kind of work that makes anti-fibrotic development feel less speculative and more like a sequence of solvable steps.

It does not give us an anti-stricture drug tomorrow, but it tells us what we would even be targeting.

Reference: Kong W et al. Nature Genetics. 2025. Single-cell and spatial transcriptomics of stricturing Crohn’s disease highlights a fibrosis-associated network.

IL-23 p19, briefly

Part One already covered this in detail, so I am not repeating it here.

The short version is that 2025 produced enough comparative and program-level data that IL-23 p19 moved from “promising” to “default starting point” for many Crohn’s pathways, and it did so because of outcomes and logistics, not hype.

Key 2025 publications and programs:

1. Janssen: GALAXI (Crohn’s, phase 3 program)

2. Janssen: GRAVITI and ASTRO (subcutaneous induction and maintenance)

3. Eli Lilly: mirikizumab program extensions, including paediatrics (see SHINE-1 below)

4. AbbVie: risankizumab UC program, including COMMAND (maintenance) (noting the shift toward class-level positioning)

TL1A becomes crowded

If IL-23 is now established, TL1A feels like the platform mechanism that companies are trying to line up behind.

Key 2025 TL1A datasets to know:

• Roche: afimkibart, TUSCANY-2 (ulcerative colitis). Phase 2b, dose-ranging, treat-through design with 1-year follow-up.

• Merck (via Prometheus): tulisokibart, APOLLO-CD (Crohn’s disease). Phase 2a induction with an endoscopic response signal and biomarker stratification.

• Sanofi and Teva: duvakitug, RELIEVE UCCD (ulcerative colitis and Crohn’s). Basket-style induction study with strong endoscopic response in Crohn’s at week 14.

A second wave is also taking shape, but it is mostly product engineering rather than new biology: longer half-life, fewer injections, and potentially different ways into the same axis.

That is what is driving the TL1A race: multiple credible datasets, multiple corporate homes, and a very fast move into late-stage development.

References: Merck/Prometheus. Tulisokibart APOLLO-CD. Lancet Gastroenterol Hepatol. 2025; Pfizer/Roivant/Roche. Afimkibart TUSCANY-2. Lancet Gastroenterol Hepatol. 2025.

A toxin, not “the microbiome”

The microbiome story is often too blurry to act on.

A Science paper from a heavily China-based collaboration (Nanjing and beyond) sharpened it into something much more concrete: an Aeromonas variant producing aerolysin that promotes susceptibility to ulcerative colitis.

This matters because it pushes us toward a more testable framing:

• Not “dysbiosis” as a vibe

• But specific microbial functions and specific toxins that can be detected, quantified, and potentially neutralised

If the next era of microbiome therapeutics works, I suspect it will look more like infectious diseases than probiotics.

Reference: Jiang Z et al. Science. 2025. An Aeromonas variant that produces aerolysin promotes susceptibility to ulcerative colitis.

Paediatrics needs its own evidence

Paediatric IBD is not just small adults. It is earlier disease biology, different growth and development trade-offs, and a narrower evidence base.

Two Lancet Gastroenterology and Hepatology papers help here.

Mirikizumab in paediatric UC (SHINE-1) enrolled across 19 centres in Canada, Israel, Japan, South Korea, and the USA. Week 12 outcomes included clinical response (69.2%), clinical remission (38.5%), and endoscopic remission (53.8%) by the modified Mayo score, with week 52 signals that justify larger work.

VEDOKIDS followed vedolizumab maintenance in children over 54 weeks. The interpretation was straightforward: vedolizumab maintenance appears safe and efficacious, with higher efficacy in UC than Crohn’s disease.

We should treat these as what they are: scaffolding for the next generation of paediatric trials, and a reminder that off-label is still doing too much heavy lifting.

References: Eli Lilly. SHINE-1. Lancet Gastroenterol Hepatol. 2025; Atia O et al. VEDOKIDS. Lancet Gastroenterol Hepatol. 2025

What “early” means in Crohn’s now

Early therapy is a phrase that used to mean anything.

Last year we got clarity on what early really means in Crohn’s disease. It means at the point of diagnosis. In PROFILE, the median time from diagnosis to enrolment was around 12 days. It is hard to read that and keep calling a two-year delay “early.”

Then there is earlier-than-early, prevention.

Groups like INTERCEPT are explicitly building towards identifying and intervening in pre-IBD states, which is a different ambition entirely. This is important work, gathering steam. Great to see meaningful funding behind this effort.

And there is a third “early” that matters clinically: post-operative Crohn’s disease, when the macroscopic disease has been resected and you have a clean window. REPREVIO showed that vedolizumab started within weeks of ileocolonic resection reduced endoscopic recurrence versus placebo.

Finally, still on vedolizumab, the LOVE-CD study asks the simple but important question: does “early vs late” Crohn’s change what we see with the same drug? It is a helpful framing, even before you argue about specific definitions.

References: D’Haens et al. REPREVIO. Lancet Gastroenterology and Hepatology, 2025.
LOVE-CD Study Group. Vedolizumab in early and late Crohn’s disease. Lancet Gastroenterology and Hepatology, 2026 online.

Trials are going global

Global access gaps are real, but global trial participation is widening too.

ENLIGHT UC evaluated etrasimod in patients with moderately to severely active UC in East Asia, and concluded it was effective and well tolerated as an oral induction and maintenance treatment in that population. Funding came from Everest Medicines, a sign of how regional development partners are now central to “global” IBD drug development.

A related Lancet Gastroenterology & Hepatology commentary makes the complementary point: even as trial footprints widen, whole regions remain under-represented, and the Middle East and north Africa (MENA) is a clear example. The authors argue that if we want “global” trials to produce genuinely generalisable evidence and fair access pathways, sponsors need to design for regional participation up front, with transparent reporting of who was enrolled and where.

It is a useful check against complacency. And a reminder that the next wave of evidence will be geographically broader than the last.

Reference: Wu K et al. ENLIGHT UC. and Mohammad Shebab et al, both Lancet Gastroenterol Hepatol. 2025.

Appendicectomy in ulcerative colitis still feels disruptive

Appendicectomy keeps returning as a serious idea, not a fringe curiosity, because the data refuse to go away.

ACCURE is the landmark here: appendicectomy plus standard therapy versus standard therapy alone for maintenance of remission in ulcerative colitis. It is the kind of trial that forces you to treat the appendix as immune tissue with influence, not a historical leftover.

Alongside that, the COSTA programme is pushing the question in a slightly different direction, appendicectomy as a potential induction strategy in harder-to-treat UC cohorts. Even if you remain skeptical, it is now a coherent research line, not a one-off.

References. ACCURE Study Group. Appendicectomy plus standard therapy versus standard therapy. Lancet Gastroenterology and Hepatology, 2025. Visser et al. COSTA appendectomy abstract (DOP053). ECCO 2025.

Diet returns with real trial architecture

We have had decades of diet opinions and far fewer diet trials that stand up in clinic.

ADDapt is one of the exceptions. Presented by Aaron Bancil at ECCO 2025, it was a properly powered, randomised, double-blind, placebo-controlled design using a low-emulsifier diet and a re-supplementation challenge. Whatever your prior, this is the direction diet research needs to go.

Reference. Bancil et al. Emulsifier restriction is an effective therapy for active Crohn’s disease (ADDapt). ECCO 2025, DOP097.

Acute severe UC is still where we need to be humble

Two themes stood out for me: optimisation of what we already do, and urgency for better rescue strategies.

PREDICT-UC is important even though it sits at the very end of 2024, because it gives real-world clarity on a question that comes up in every on-call week: is a first dose of infliximab 10 mg/kg better than 5 mg/kg in steroid-refractory acute severe UC? The answer in that trial was no. But for sicker patients with lower albumins the higher dose is still the default.

On the pipeline side, the idea of adding upadacitinib earlier in the admission is gaining momentum. ACCESO is leading the way: upa (or placebo) plus standard intravenous steroids on day one of admission, designed to test whether earlier combination reduces the need for subsequent rescue and surgery.

Reference. Choy et al. PREDICT-UC. Lancet Gastroenterology and Hepatology, 2024.
ACCESO trial registration (upadacitinib plus IV steroids).

Looking sideways

Some of the most interesting IBD thinking still comes from watching adjacent fields.

• TYK2 is still the clean-oral dream for many. Takeda’s phase 3 plaque psoriasis data for zasocitinib (TAK-279) is not an IBD result, but it is a signal that highly selective TYK2 inhibition is maturing fast. And IBD is next.

• Bioelectronic medicine is inching from concept to clinical reality. A randomised, double-blind, sham-controlled trial of vagus nerve stimulation in rheumatoid arthritis just published in Nature Medicine is a serious piece of evidence for the “inflammatory reflex” as a modifiable system.

• Early-onset colorectal cancer trends continue to be a background alarm bell. For IBD teams, it is another reason to stay disciplined about inflammation control and surveillance quality.

Reference. Takeda newsroom. Zasocitinib (TAK-279) phase 3 psoriasis update. 2025.
Koopman et al. Vagus nerve stimulation for rheumatoid arthritis. Nature Medicine, 2025.
Sung et al. Colorectal cancer incidence trends in younger versus older adults. Lancet Oncology, 2025.

Honourable mentions and near misses

Obefazimod (Abivax) deserves to be on any 2025 list because the phase 3 induction programme hit its endpoints and kept the oral innovation drumbeat going. Maintenance is still the real punchline.

SOR102 is an oral bispecific inhibitor of TNF and IL-23 tested in a first-in-human phase 1 study in healthy participants and patients with UC. By day 42, symptomatic remission was reported in 43% of the once-daily group and 56% of the twice-daily group, versus none in placebo, with a favourable safety profile supporting development.

MORF-057 is a good example of a “near miss” story that can still be valuable. Public reporting around EMERALD-2 highlights numerically higher remission with a placebo response that was higher than expected, and the strategic pivot toward combination approaches is already visible in trial registries pairing MORF-057 with mirikizumab.

References. Abivax press release. ABTECT-1 and ABTECT-2 induction. 2025. Sorriso Pharmaceuticals. SOR102 phase 1. Lancet Gastroenterol Hepatol. Epub 2025; issue Jan 2026. UEG Week 2025 programme. EMERALD-2 results session. ClinicalTrials.gov. MORF-057 plus mirikizumab combination study.

A human pause

Before Part Three, one moment of stillness.

Brian Feagan died recently, and it is hard to overstate how much modern IBD trial culture and clinical standards were shaped by his work and his insistence on progress and clarity.

And a thank you, sincerely, to the people who make these papers real: researchers and basic scientists, gastroenterologists and surgeons, IBD nurses and research nurses, pharmacists, trial coordinators, and the industry and digital teams who build the tools.

Most of all, to patients, who keep showing up, taking risks, and teaching the rest of us what matters.

2025 rebuilt the therapeutic hierarchy.

If the last decade was defined by the proliferation of biologic options, this year was defined by comparative data that clarifies how to use them. We did not just get new drugs; we got head-to-head evidence that challenged old defaults. The long-standing dominance of anti-TNF therapies as the undisputed first-line biologic standard faced its most significant erosion to date, driven by superiority rather than hype.

Sequencing became the story.

Crohn’s disease: IL-23 p19 becomes first choice

The p19 signal is now hard to ignore.

GALAXI (guselkumab vs ustekinumab) and SEQUENCE (risankizumab vs ustekinumab) did not just show non-inferiority; they showed that targeting IL-23 alone delivers stronger endoscopic outcomes than blocking IL-12 and IL-23 together.

In North America, the Middle East, and parts of Europe such as Germany, IL-23 therapy has displaced anti-TNF as the first-line standard for Crohn’s disease, after a period of drift toward first-line ustekinumab drawn by superior drug persistence (no meaningful immunogenicity) and fewer sequencing problems compared with anti-TNF.

Now the first-line move in Crohn’s is to risankizumab or guselkumab.

Induction without an infusion

Subcutaneous induction has arrived.

GRAVITI and ASTRO showed that fully subcutaneous induction with guselkumab is possible and highly effective in both Crohn’s disease and ulcerative colitis.

This matters for patients who want autonomy and for systems trying to decongest infusion centres, and it strengthens the “p19 first” argument for providers and payers who weigh logistics alongside outcomes.

It changes the practical conversation.

Biosimilars set a new cost floor

The revolution is not evenly distributed.

The arrival of biosimilar ustekinumab has created a price floor we have not seen before, with seven FDA-approved biosimilars entering the US market in 2025.

Europe is already leaning into this, having built experience with biosimilars a decade earlier.

For cost-sensitive systems, biosimilar ustekinumab and the established biosimilar adalimumab remain formidable gatekeepers.

After more than a decade, vedolizumab remains without a biosimilar.

Top-down Crohn’s strategy reaches scale

Top-down is finally being implemented.

I have argued for top-down therapy in Crohn’s disease since 2010, and it has been standard of care in Edinburgh for a decade once biosimilars made it cost-effective.

PROFILE, published in 2024, helped settle the debate: hitting hard early drives deeper remission and prevents complications.

In 2025, that evidence translated into wider implementation.

The debate is now about execution.

Crohn’s sequencing in 2025

The hierarchy is clearer than it has been for years.

Phenotype still matters, but the evidence now points to distinct starting choices for distinct problems. This is how I currently think about sequencing in Crohn’s disease.

1. Severe, fistulising, or extensive disease: infliximab plus azathioprine is still the heavyweight champion. Start here, stabilise, and keep the option to switch to subcutaneous infliximab and drop azathioprine later.

2. Moderate-to-severe luminal disease: a p19 inhibitor is now the evidence-based first choice, with better endoscopy, safety and persistence.

3. When cost drives the pathway: biosimilar ustekinumab or adalimumab remain excellent drugs. They are no longer the ceiling of efficacy, but they are a very high floor.

Biosimilar vedolizumab would be welcome, especially in light of the new LOVE-CD data, and VICTRIVA will teach us whether vedolizumab with an upadacitinib induction blast does the job.

Ulcerative colitis: more options, less consensus

UC pathways remain variable.

There are now many effective drugs, and pathways differ across the world and within regions, driven by cost, familiarity and guidelines. The result is a wide spread of “standard” practice, even among specialist centres.

Consensus sequencing remains limited.

· Biologics: infliximab (IV and subcutaneous), adalimumab, golimumab, vedolizumab, ustekinumab, mirikizumab, risankizumab, guselkumab.

· Small molecules: ozanimod, etrasimod, tofacitinib, filgotinib, upadacitinib

· Conventional options: azathioprine (for steroid-dependent UC when needed); don’t forget that many patients only ever need 5-ASA; optimize before moving on.

Small molecules in UC: clearer niches and one new signal

The post-tofacitinib landscape is settling.

In Europe, filgotinib has found its niche for moderate disease: cheaper, simple dosing, good efficacy, excellent persistence, and a favourable safety profile. Upadacitinib sits above it for moderate-to-severe UC, and increasingly for acute severe ulcerative colitis (ASUC). The ACESO trial is recruiting across the UK, testing whether adding upadacitinib to IV corticosteroids early on day one can reduce the need for rescue therapy or colectomy. S1P receptor modulators (ozanimod and etrasimod) remain effective and well-tolerated, but still lack a clear positioning statement in a crowded field.

Obefazimod is the new entrant to watch.

The ABTECT phase 3 induction results presented at UEG in October showed significant and clinically meaningful remission rates across a broad population, including patients who had failed multiple prior advanced therapies (including JAK inhibitors). It is a first-in-class oral miR-124 enhancer.

Maintenance data are eagerly awaited.

UC sequencing: still more opinion than algorithm

Every expert will tell you their opinion, often shaped by local access and precedent.

Options are excellent for clinicians and patients, but the practical requirement is to know the pros and cons of each drug and to be able to discuss them in shared decision-making. I am still not sure we know the right time to start advanced therapy in UC, and I see both delayed starts and arguably early starts.

Timing remains the hardest decision.

Global access: the widening gap

It is worth stepping back.

For much of the world, these choices do not exist. In many low- and middle-income countries, anti-TNF biosimilars are the ceiling, not the floor. In India, generic tofacitinib has flooded the market and become a de facto first-line advanced therapy, driven by cost and availability rather than sequencing logic.

The access gap is widening.

Prediction and personalisation

We still largely prescribe by phenotype and hope.

The promise of biomarkers to guide drug selection remains largely unfulfilled in 2025, though work continues.

More on this in a future post.

This needs its own discussion.

Comparative evidence: what to trust

We still need more comparative trials.

Network meta-analyses can be useful for hypothesis generation and to satisfy health technology assessment bodies, but they are not a substitute for real comparative data.

Real-world comparative effectiveness studies also need caution: very large datasets, for example TriNetX, cannot fully correct hidden bias and confounding even with propensity matching.

Expect big publications to point in different directions.

Blinded head-to-head studies still matter most.

Surgery: appendicectomy in UC

Appendicectomy returned to the UC conversation.

The ACCURE trial published in Lancet Gastroenterology & Hepatology showed that laparoscopic appendicectomy added to standard medical therapy significantly reduces clinical relapse at 12 months. COSTA data reinforced the signal, showing over 20% induction of remission compared with medical therapy alone. This may have been the most unexpected therapeutic development of the year.

More to come, including how we might identify who will respond.

Looking ahead

The next challenge is strategy.

What we really need are innovative trials that enable head-to-head comparisons, sequencing, add-ons and true combinations. Emerging concepts around combination strategies, such as DUET trial concepts and “induction blasting” with JAKs, suggest that for ceiling-hitters, 2026 might be the year of deliberate combinations.

But that is for another post.

Next: top publications of 2025, then a 2026 preview.

Photo of me in the Grey Corries by Colin Henderson

“There was another life that I might have had, but I am having this one.”
Kazuo Ishiguro

There are moments, usually quiet ones, when that line lands with unexpected force.

For me, it landed in the days after my inaugural lecture: standing in front of a packed room in Edinburgh, speaking about the science and the stories that shaped my career, and then, just hours later, finding myself alone in a small hotel room in Treviso, tired, overstimulated, and wondering how I’d arrived at this particular version of my life.

This was never the plan.
Maybe there never was a plan, only a series of choices, accidents, impulses, and sliding-door moments that carried me here.

The straight-line careers you read about do not exist, and mine certainly doesn’t.
There were years of chaos and instability.
There were years of clarity.
There were the times I burned brightly and the times I burned out.
There were wrong turns and almost-right turns.

And there were the rare moments, lecturing in Brazil, running in the Pentlands, the completion of PREdiCCt, a patient thanking me quietly after clinic, when everything aligned and I felt certain I was doing the work I was meant to do.

But the truth I’ve been circling recently is this:
We don’t live one life; we live through the aftermath of all the lives we didn’t choose.

The one I might have had.
The one that slipped away.
The one that broke.
The one that asked too much.
The one that was almost mine.

And then the one I’m living now.

The inaugural lecture was a milestone, but it was also a mirror.
It forced a reckoning: what exactly have I built? What am I still building? What do I want the next decade to feel like?

No one prepares you for the emotional hangover that follows a night like that.
You walk off stage into applause, warmth, old friends, colleagues, pride.

Then you wake the next day in a different city. Venice for me, which softened the landing.
And the day after that, Treviso, where the only thing waiting was a hard bed, an overheated room, and the sudden weight of your own thoughts.

This is the part of leadership no one sees.
The nights alone.
The travel fog.
The crash after the high.
The way success can sharpen the silhouette of loneliness.
The strange tension between being known professionally and unseen personally.

But somewhere between the Grand Canal and the night streets of Treviso, something shifted.
Maybe it was the fatigue. Maybe it was the overstimulation.
Maybe it was the recognition that I’ve been running on fumes for too long.
Maybe it was simply time.

I realised I need to make space.
To clear the noise.
To finally deal with the open loops: admin, finances, logistics, unanswered emails, and unfinished projects; the many pieces of life that drain energy just by existing unresolved in the background.

Because only when the noise quietens can you hear the deeper questions:
How do I actually want to live?
Who do I want to be outside of work?
What kind of man do I want to become in the second half of life?

Right now, it feels like I’m standing at the threshold between chapters.
Not lost, but not fully defined.
Not stuck, but not yet free.
Not unhappy, but not fully aligned.

And yet, Ishiguro’s line brings a kind of peace.

“There was another life that I might have had, but I am having this one.”

This is the one shaped by my patients, my children, my research, my failures, my resilience, the cities I’ve travelled through, the talks I’ve given, the colleagues who’ve held me, the friends who’ve stayed.

The one shaped by late-night messages from my son, by Holly asking if we can hike in Cambodia, by the team who believe in Sentinel, by the students I teach, and by those quiet moments in clinic when a patient says, “Thank you for listening.”

This is the one shaped by the aftermath of everything that didn’t happen.
And somehow, that makes it enough.

The path forward isn’t linear. It never was.
But it feels clearer now.

Do the work.
Clear the noise.
Rebuild the body.
Strengthen the mind.
Return to the things that make me feel alive: running, mountains, the Pentlands, Copenhagen in the summer, Aspen in winter, photographing strangers in alleyways, writing, creating, building something meaningful with the people I trust.

I don’t know exactly where this version of my life leads, but I know this:
I’m still becoming.

And for the first time in a long time, that feels like hope rather than regret.

Subscribe now

Today, we can set people up for success from day one. That means prompt diagnosis, full assessment, and — especially in Crohn’s — starting effective therapy early rather than asking patients to “earn” their way to it. It means treat‑to‑target and careful, objective monitoring so we can follow the arc of inflammation over years and decades, not just weeks and months.

In this short Atomic IBD video (~5 min), I cover:

• Why early, effective therapy changes a lifetime trajectory.

• Moving beyond symptom‑only management to treat‑to‑target.

• Using faecal calprotectin/CRP, intestinal ultrasound, colonoscopy and imaging to track disease.

• The need for safe, durable drugs — and truly frictionless, passive monitoring tools.

• Holding optimism for the newly diagnosed, and empathy for those living with IBD for 40, 50, 60+ years — and learning from their stories.

We are, genuinely, in the disease‑modifying era of IBD. The task now is to combine science, strategy and trust so that most people, most of the time can live the long, full lives they deserve.

Watch the video above, and share your experiences below — there are clues in every story about how we can do better today.

I outline our Edinburgh work using longitudinal faecal calprotectin and CRP to model inflammatory behaviour over time; at service level and for the individual patient.

This is the analytics engine behind Sentinel, our AI-driven dynamic clinical decision support tool.

We’re preparing a pilot now, with first patient in January.