The Titration

Why DiabeXO exists

DiabeXO exists to make the invisible progress of type 2 diabetes visible early enough to change its course. It is not meant to be another app that collects data.

Why “XO”. On a naval ship, the XO is the Executive Officer, the captain’s second in command. The XO runs the ship day to day: keeps the watch, keeps the log, sees that every station is ready and brings the captain what he needs to decide. The captain commands. In DiabeXO, the patient is the Captain and the app is the XO: it watches, records, reminds and prepares, and the decisions stay with the Captain and his care team.

It comes from three decades of lived care. Vinayak has been a carer to a family of diabetics since his teens, and has managed his own type 2 diabetes for about thirty years.

The app store is full of diabetes apps, and he has used many. None satisfied him, because they collect readings without understanding the person, the condition or the conversation between patient and clinician. Wearables, a powerful computer in every pocket and language models now make something better possible: an ecosystem, not an app.

Why now, and why him. Vinayak first had the idea about fifteen years ago, when the first smartwatch showed what a body-worn computer could become. The technology was not ready then; he was healthier. Now the tools have arrived. Someone who does not write code can build a working app in days, and the cost of starting has collapsed. He is older and his health is compromised, but he carries what the younger man lacked: three decades of evidence of what textbooks miss, where the hidden harms lie, and what a ten-minute appointment fails to see. His conviction is stronger for it.

The first name. The idea came in a dream, soon after he began wearing the first Apple Watch, and he called it Diabetonomy: diabetes and autonomy, the patient governing his own ship, and also the economy of diabetes, what it costs and who pays. It was ahead of its time. Then came the pandemic, and years of his own health taking first place. DiabeXO kept Diabetonomy’s first meaning and gave it a crew. The old name survives as the destination: DiabeXO is the ship, Diabetonomy the voyage’s end, and every step taken with the XO is a step towards autonomy over one’s own diabetes. The years away were not lost: they gave the hazard log much of what it now holds, written from experience rather than imagined.

The titration

Type 2 diabetes works like the titration we all studied in school chemistry lab work: the reaction begins with the first drop, long before the indicator changes colour. Medicine usually notices the colour change, which is the complication. DiabeXO exists to show the drops.

Acid is added to alkali drop by drop. An indicator turns colour only when the reaction is complete. Vinayak’s chemistry teacher made the point that matters: the process starts with the first drop. It is only invisible until then.

In type 2, metabolic dysfunction is already under way when diagnosis is made and treatment begins. From there it can take many routes: damage to blood vessels, nerves, eyes and kidneys, and in the worst cases gangrene. Some people never reach the endpoint and some do. Nobody knows in advance which, and that is why the drops must be watched.

The aim is not to frighten anyone. It is to make each person a responsible, informed Captain of their own condition, who can see the drops and act on them in time.

Nine convictions

Nine convictions from lived experience define what DiabeXO must do. Each one names a gap that current care and current apps leave open.

1. Understanding. Most patients are never helped to understand the condition they live with.
2. Communication. Patient and clinician rarely share the same picture. Ten minutes are spent guessing.
3. Multi-modal care. Medicines are one part of treatment. Food, movement, sleep, stress and monitoring matter as much.
4. A pathway into it. There is little community or primary-care help to bring a newly diagnosed person into a multi-modal plan and adapt it as life changes.
5. Glycaemic control matters from the first drop. The titration starts at diagnosis, not at the first complication.
6. Awareness without fear. A patient who knows the possible routes can take responsibility. Fear makes people look away.
7. One person, many conditions. Rules written for one condition often contradict another. Crohn’s disease limits fibre while type 2 advice demands it. The person, not the textbook, must be the unit of care.
8. Context changes the treatment. The same person responds differently in different worlds. Heat speeds insulin absorption, so a dose that suits a UK winter can act like a larger one in an Indian summer. Ancestry shapes risk: people of South Asian heritage develop type 2 earlier and at lower body weight. Climate, season, country, travel and fasting all leave their mark.
9. Every remedy is a medicine. Ayurvedic, homeopathic, herbal and over-the-counter products act on the body and can harm it. The danger is not tradition itself but what is unknown: unlisted ingredients, heavy metals absorbed from soil, and undeclared steroids added to products sold as natural. Vinayak has seen friends seriously harmed this way. Prescribed medicines have side effects too, but theirs are known, listed and treatable. The XO records every remedy like any medicine, asks what is in it, and makes sure the clinician sees the full list.

Ten principles

Every build of DiabeXO answers to ten principles. A feature that breaks one does not ship.

1. The Captain decides; the XO keeps watch. The XO observes, records, reminds and prepares. It never suggests doses or treatment changes. When it prepares for a clinician, the clinician decides.
2. The XO learns the person. General knowledge is the starting point. The person’s own evidence outranks it, and no rule is applied that their conditions contradict.
3. Every drop is recorded the same way. Each measurement carries what it is, its value, unit, time, source and context (climate, season, place, travel, fasting).
4. Every figure carries its source. Figures for a clinician or regulator come from a reference source, such as the manufacturer’s own export.
5. Nothing depends on real-time glucose. Alarms stay with the sensor maker’s app. The sensor maker’s app and a carer’s sharing are never disturbed.
6. Works with or without a sensor. Most people with type 2 have no continuous glucose monitor. The XO must be useful from finger-pricks, HbA1c and daily life alone.
7. Data stays with the person. It lives on their phone. Nothing leaves without their consent, each time.
8. Neutral and portable. No device maker is favoured. Official data routes come first. The core logic is shared so iPhone and Android run the same tested rules.
9. Free for patients. DiabeXO is a giving back. Any funding comes through the NHS, never from the patient.
10. Only answers travel. When many people’s experience is studied, the question goes to each phone and only combined answers come back, with consent for each study.

The map: from invisible to visible

Each route the titration can take has early, measurable drops long before the endpoint. Most are already in the NHS annual care processes, yet only about half of people with type 2 diabetes in England receive all eight each year (National Diabetes Audit). DiabeXO’s core screen will be this map: each route, its drops, when each was last measured and which way it is drifting.

The map also shows the drops that have been stopped: a falling ACR, a stable eye grade, a better HbA1c. Awareness without fear means showing progress as clearly as risk.

Checks and balances

A system that watches health must itself be watched. Five checks keep DiabeXO honest, and each one leaves evidence for the safety case.

1. Safety before features. Every way the XO could cause harm is entered in the hazard log with its controls before the feature ships. The log holds 44 hazards so far, including the wrong insulin in a pen, ketoacidosis with normal glucose and remedies with unknown ingredients.
2. Prediction against reality. Every forecast is scored against what actually happened. The gap is shown, not hidden: a surprise is news, and a model that is often wrong says so.
3. The data checked against its source. Readings arriving by any route are reconciled with the manufacturer’s own record, against a pass mark set before the comparison starts.
4. Cause tested, not assumed. Patterns found in the data are only suspects. Planned experiments, with control days, turn a suspect into evidence.
5. People over software. The clinician decides on treatment and the Captain decides on his own life. A clinical safety officer signs off before the XO reaches anyone else. Disagreements between letters and the patient’s account are recorded and settled from the source, never silently resolved.

The regulatory path follows from these checks. A locked, validated core does the reckoning, with an agreed plan for how it may change. Language models translate and explain; they never decide.

The crew

The parts of this system that are not software are people. No one builds it alone.

• Founder and first CaptainVinayak Supekar: thirty years of lived care; the first user.
• GP adviserA practising GP gives the primary-care view of the clinic summary.
• Sought: clinical lead and independent reviewerA consultant diabetologist to set clinical content, and an independent one, from outside the founder’s own care, to audit it.
• Sought: clinical safety officerTo sign off the safety case (DCB0129) before DiabeXO reaches others.
• Sought: research sponsorAn NHS or university sponsor for Sea Trials, with ethics approval through them.
• Sought: NHS partnerAn integrated care board or Health Innovation Network, for the route into services.

Learning from many

When many people use DiabeXO, their combined experience could show which treatments are well tolerated by whom, and which conditions travel together. Nothing of this kind will happen without research ethics approval, a published protocol, consent for each study, and analysis in which the question travels to each phone and only combined answers come back. Real-world data shows associations, not causes; findings must be tested before they change care.

Outlasting the founder

Vinayak’s aim is to be the trigger and to arrange the material in advance, so the chain reaction continues without him. The reaction must be a reactor, not a bomb: sustained and controlled, with the checks and balances as its control rods.

1. The reasoning written down. The founding statement, hazard log, decision record and evidence log let someone who never met him understand every choice.
2. A home that keeps it for patients. Its legal form is being settled with advisers. Whatever it is, it will hold DiabeXO for patients and keep it free for them.
3. Other hands on the controls. A clinical lead, a developer who owns the code, and a patient lead to follow the founder.
4. Standards others already speak. FHIR for data, NICE for evidence, NHS secure data environments for research.

What we do not know yet

These are open questions, each to be settled by evidence or by the right person.

• How much data does a personal model need before its forecasts beat a straight line? Months, or a year with its seasons?
• Which drops matter most for each person? The map lists them all; the weighting is unknown.
• Where exactly does DiabeXO sit in the MHRA’s classification once the clinic summary and the map are built?
• How is a multi-modal plan best co-written with a clinician who has ten minutes?
• Which open research datasets may be used, under which licences, for the general starting picture?

The endpoint DiabeXO works to prevent is the colour change. Each step moves another drop from invisible to visible.