Diabetes intelligence
A phrase is worth nothing. These are the measurements.
Wellovue reads your own record — glucose, meals, activity, labs — and reports what it can measure in it, how sure it is, and what it could not account for. Where it finds an association, it offers you a way to test it rather than asking you to believe it.
What it measures
Four questions, asked of your data.
Glucose patterns. What your readings do across a day and across a month: waking glucose, how much it varies between days, and how much of the time you spend inside your target range.
Meal response. What happens after you eat — how high glucose went, how long it took to peak, how many minutes it spent above your target range, and when it came back. Measured on each meal’s own curve and then averaged, so meals that peak at different times are not flattened into one shape that never happened.
Activity effect. Whether meals followed by movement were followed by a different response from meals that were not, with both groups shown side by side rather than reduced to one number.
HbA1c trend. The direction of your lab results over a longer window than the rest of the screen uses, because a quarterly test has nothing to say inside a month.
The part most tools skip
An association is a question, not an answer.
If meals followed by a walk look different from meals without one, that is a real observation and a weak one. Those two groups differ in more than the walk: what you ate, when, how you slept, what kind of day it was. Wellovue says so, in the finding itself.
So the finding comes with an offer. You run a short experiment — the same meal on six days, three of them followed by a walk, chosen at random rather than by how you feel. The platform writes down its prediction before the experiment starts, and the prediction cannot be edited afterwards. When you finish, the outcome is recorded beside it.
That is the whole idea. A system that scores itself after the fact can always be right. One that writes down what it expects, then cannot touch it, produces the only number worth anything — how often it was right about you specifically.
How sure it is
Thin data produces a thin answer, and says so.
Every finding carries the number of observations behind it, a confidence, and a list of what it does not account for. When there is not enough data, you are told there is not enough data and what to record to change that. It is a worse-looking answer and a more useful one.
Findings are produced by a statistical model. A language model is allowed to phrase one into a readable sentence; it is never allowed to create one, soften one, or fill a gap where the record is thin. How this works goes through that in order.
Boundaries
What it will not do, whatever the data says.
Wellovue is not a medical device. It does not diagnose, prescribe, calculate insulin doses, or take the place of clinical care. Some subjects are gated behind a conversation with your clinician no matter how clear a pattern looks.
Never, under any circumstances
- Adjust your medication, or tell you to
- Calculate an insulin dose
- Diagnose a complication
- Advise you during a hypo or a hyper
- Tell you to stop taking something
Only with a clinician
- Medication timing
- Fasting protocols
- Major diet changes
- Exercise changes if you carry cardiovascular risk
Where to go next
By what you are living with.
Type 2 diabetes and prediabetes describe what the engine reads in each kind of record. The clinician report is what comes out of it at the end of thirty or ninety days.