The GNL Podcast, recorded at DICE 2026

Episode 42, seven voices from DICE 2026

A quiet corner off the main hall at Croke Park, a recorder, and ten minutes with whoever had just walked off stage. Ireland had spent the day launching a national diabetes strategy, and the people who will have to turn that document into actual care kept passing through. So we caught them, one after another, and asked the same thing each time: what changes now, and what are you still waiting for?

Ask Grace

Want to know how a bounded, evidence-only assistant like Grace differs from the chatbot that learned the open internet, the exact line Derek O’Keefe drew on stage in Galway?

Recorded on location at DICE 2026, the Diabetes Ireland Conference and Exhibition, Croke Park, Dublin, on 21 May 2026. This is the technology and drugs strand of the GNL Podcast. Podcast hub

Listen

Episode 42 cover, DICE 2026 highlights with Diabetes Ireland. The Glucose Never Lies Podcast

Listen on Apple Podcasts Listen on Spotify Listen on Buzzsprout

Available on Buzzsprout, Apple Podcasts, and Spotify. Recorded at DICE 2026 with seven speakers from across Irish diabetes care (full list below). Host: John Pemberton. Director of Creativity: Anjanee Kohli. These are the independent views of the speakers and the host; the episode is not affiliated with or endorsed by Diabetes Ireland or any company named in the conversations.

Why this episode exists

If you live with diabetes, you already know the distance between what a guideline says you are entitled to and what your local service can actually put in front of you. A pump you qualify for on paper, a psychologist who is named in the plan but does not exist in the clinic, a box of test strips you can have in principle but have to chase in practice. That gap is where most of the frustration lives, and closing it is harder than writing the plan that promises to.

At DICE 2026, Ireland launched a national diabetes strategy that puts numbers behind the staffing, the technology, and the psychology that a modern service needs. A strategy is a beginning, not an outcome. So rather than report the headline, this episode collects the people who have to deliver it, and asks each of them the same honest question: now that it is written down, what actually changes, and what is still missing?

In this episode

The format is deliberately fast. Seven ten-minute conversations, recorded back to back between the conference sessions, each with a person who carries a different part of the system: two dietitians, a psychologist, a clinical-trials professor, a pregnancy consultant, a physician engineer building AI, and the diabetologist who set up Ireland’s technology network. Heard together they form a single argument about what good diabetes care now requires.

The thread that runs through all seven is the same one that runs through GNL: technology and evidence only matter when they reach the person, in language they can use, at the moment they need it. The day kept circling one tension, which the closing callout names.

Episode chapters

00:00, Welcome from DICE, Croke Park
00:11, Sinead Powell, Diabetes Ireland, Language Matters and emotional wellbeing
09:00, Cathy Breen, National Diabetes Programme, Language Matters in the consultation
11:55, Kate Gaievska, Diabetes Ireland, psychology written into the strategy
16:20, Fidelma Dunne, University of Galway, clinical trials and real-world data
27:12, Christine Newman, Galway, pregnancy and diabetes technology
32:48, Derek O’Keefe, University Hospital Galway, AI in diabetes care
50:05, Tomas Griffin, Galway, the DTN Ireland model and access to technology

Chapter times are approximate and will be aligned to the final edit.

Listen to the full episode

Key themes

Language is part of the treatment, not the wrapping around it

The first three conversations all circled the same point from different angles: how a clinician speaks to a person shapes whether that person comes back. Sinead Powell, who has worked in diabetes for thirty years, described moving from a medical model that listed complications to a way of working that starts with “you made it here today, thank you for coming.” She has built an emotional wellbeing programme for people with type 2 diabetes called COPE (Connect, Overcome, Prioritise and Empower), because she kept meeting people who knew what they were meant to do and were stuck on the inner monologue of why they had not. Cathy Breen sharpened the idea: Language Matters is less about hunting for the perfect word and more about whether you are judging the person or sitting alongside them. Get that right, she argued, and someone will tell you when a word lands badly. Kate Gaievska then carried it into the strategy itself, where psychology has, for the first time, been recognised by the Irish Department of Health as part of the plan.

Evidence is how good care gets funded

Fidelma Dunne, professor of translational clinical trials at the University of Galway, made the case that research is not the icing on the cake of a diabetes service; it is part of the recipe. Randomised controlled trials remain the highest grade of evidence, but they are slow and costly, and she is just as interested in what well-built real-world datasets can do. With large CGM records you can take a single person and find a matched period where the only thing that differs is the intervention, which gets you close to a controlled comparison from observational data. She also made the point that the work does not stop when a trial ends: it was post-trial surveillance, not the original study, that surfaced the increased fracture risk with the older glitazone drugs. Her closing figure is the one worth keeping, on her account: services that deliver clinical trials tend to return roughly ten pounds of benefit for every pound invested, and a small local service evaluation can reveal savings that are well worth writing up.

AI that does the 80 percent, so the clinician can do the 20

The conversation closest to GNL’s own work was with Derek O’Keefe, a physician engineer in Galway who trained first in electronic engineering and then in medicine. Faced with a diabetes educator whose GLP-1 teaching list had jumped from twenty people a week to two hundred, his team built an assistant called Mochara (the Irish for “my friend”). The design choice is the part that matters: rather than let a general model loose, they loaded only the approved, regulator-marked manufacturer documents and trusted health-service material, and let the assistant answer strictly from that. When John asked the obvious follow-up, whether it was retrieval from a bounded source and whether it carried a philosophy on top of the evidence, Derek was clear that they kept it to the information itself, because heavy interpretive rules made the output unpredictable. In their trials the avatar educated people as well as the nurse and the doctor did, and they gave a reason that should make every service think: they were not embarrassed to ask it a simple question. The framing he returned to again and again is the one GNL builds on too.

Derek O’Keefe’s framing, and the one GNL Grace is built around: a bounded, evidence-only assistant can carry the routine information load, which frees the clinician for the part only a clinician can do. Grace works the same way, trained on type 1 diabetes evidence alone, graded A to D, with no opinion poured on top.

Access to technology is a model you can copy

Two clinicians showed what closing the gap actually looks like on the ground. Christine Newman, a pregnancy and diabetes consultant in Galway, described her practice shifting from around eighty per cent of women on injections in 2020 to roughly seventy per cent now starting or entering pregnancy on hybrid closed-loop systems, with her remaining wish list being free glucose strips for women with gestational diabetes and the funding to bring care to the patient rather than the other way round. Tomas Griffin, who trained in Leicester before setting up the Diabetes Technology Network Ireland, has spent three years travelling the country to show clinicians and people with diabetes that pumps are available and allowed. The result he is proudest of is structural: insulin pumps are now being started in every tier of Irish hospital, including the smaller Model 2 and Model 3 units that were once told to send everyone to the big centre. Access, in other words, turned out to be a model you could carry from town to town, not a budget line you waited for.

A plan on paper, no budget attached yet. Almost every speaker landed on the same tension. The strategy names the staffing, the psychology, and the technology, and that gives advocates something concrete to point at. What it does not yet carry is the money to deliver it. As Kate Gaievska put it, what they have is the vision and the goal, not the promised funding; and as Sinead Powell noted of her own programme, the research exists but the money to scale it does not. The strategy is the lever. The next budget is where it either moves or does not.

Practical exploration

For people living with type 1 diabetes and their families

A strategy day in another country is more useful to you than it sounds, because the arguments made in Dublin are the arguments worth making to your own care team.

If a technology is named as standard in a plan or guideline, that is something you can reasonably ask your team about: a pump, hybrid closed loop, CGM in pregnancy. Knowing it exists is the first step to being offered it.
If the words used in a consultation do not sit well with you, you are allowed to say so. The clinicians in this episode would rather you did; Language Matters cuts both ways.
A bounded assistant like Grace exists for exactly the questions you did not want to interrupt the clinic to ask. It answers from type 1 diabetes evidence only, and it is in your pocket at 2am when the clinic is closed.
The emotional load of living with diabetes is part of the condition, not a weakness on top of it. Programmes like the ones described here exist because that load is real and shared.

For clinicians and educators

The episode is a quiet argument that the soft skills and the hard technology are the same project, and that you can move faster than you think.

Language Matters is a clinical skill, not a courtesy. Whether you are judging or sitting alongside the person tends to matter more than the exact term you reach for.
A small service evaluation on data you already hold (time in range, insulin, a simple supportive nudge) can reveal a result worth writing up. You do not need a randomised controlled trial to start.
A bounded, retrieval-based assistant trained only on approved evidence can carry the routine teaching load, as Mochara did, which frees you for the edge cases and the relationship. The design discipline is to resist pouring interpretation on top of the source.
The technology-network model travels. Showing a Model 2 or Model 3 unit that a pump start is within reach changed what those units did. Access can be taught, not just funded.

The seven voices

Each conversation stands on its own. Open any of the speakers below for who they are and what they brought to the day.

Sinead Powell, Diabetes Ireland

An Education and Support Coordinator and dietitian with thirty years in diabetes care, Sinead has watched the field move from a complications-first medical model to person-centred support. She built COPE, an emotional and mental-health wellbeing programme for people with type 2 diabetes that uses light cognitive-behavioural tools and, above all, the shared space of a group. Her honest constraint: the research is there, the funding to scale it is not.

Cathy Breen, National Diabetes Programme

A clinical specialist dietitian in diabetes and obesity and a dietetic lead for the National Diabetes Programme, Cathy helped land Language Matters in the day’s policy recommendations. Her contribution was practical: person-first language where the person wants it, attention to how you are with someone rather than just the words, and a deliberate move away from a history of focusing on numbers towards health and quality of life.

Kate Gaievska, Diabetes Ireland

A psychologist by background and clinical manager for advocacy and research at Diabetes Ireland, Kate spoke to one of the strategy’s biggest wins: psychology recognised by the Department of Health as part of diabetes care. She was candid about the implementation challenge, that recognition and even funded posts do not instantly produce psychologists with diabetes expertise, and that the honest plan for now is to keep advocating. She has lived with diabetes herself since 1987.

Fidelma Dunne, University of Galway

Professor of translational clinical trials and director of both the Institute for Clinical Trials and the National Clinical Trials Network in Diabetes, Fidelma made the case for research as a core part of care delivery. She values randomised trials and real-world data alike, champions industry, academia and clinicians working together, and reminded listeners that post-trial surveillance is how some of the most important safety signals are found.

Christine Newman, Galway

A diabetes consultant working across pregnancy services in Galway and beyond, Christine described a fast shift towards hybrid closed-loop therapy in pregnancy and the practical barriers that remain, including the cost of glucose strips for women with gestational diabetes. Her priority is bringing consistent, technology-enabled care to women wherever they live, rather than concentrating it in one centre.

Derek O’Keefe, University Hospital Galway

A consultant physician and professor of medical device technology, Derek is a self-described physician engineer who trained in engineering before medicine. He built Mochara, a bounded, retrieval-based education assistant trained only on approved manufacturer and health-service material, and has trialled it against nurse and doctor teaching. He is equally interested in ambient documentation that gives clinicians their time back, and in intelligent, dynamic triage that books the next appointment by need rather than by calendar.

Tomas Griffin, Galway

A consultant diabetologist splitting his time between hospital and community care, Tomas trained in Leicester before co-founding the Diabetes Technology Network Ireland. For three years the network has travelled the country, running events with industry and local teams to show that insulin pumps are available and allowed. The measure he points to is that pump starts now happen across every tier of Irish hospital, not just the largest centres.