fenome turns your lab results into a personal, plain-language story of your biomarkers. The friendly sentence is the easy part. Underneath it sits decades of published laboratory-medicine science and a deterministic engine that decides what to say, so the read is honest, repeatable, and something you and your doctor can trust. Here is how it works, and why we built it this way.
Every lab report measures you against a reference range, the band printed beside your result. It looks like a verdict, but it is a population statistic: the middle 95% of a large group of healthy people. By construction, about 1 in 20 perfectly healthy people fall outside it, and being inside it was never a promise that a value is right, or even usual, for you.
That creates two quiet blind spots. A result can sit comfortably inside the "normal" band yet be genuinely unusual for your own body. And a result can sit just outside it yet be entirely ordinary for you. The most meaningful movements in your own numbers often happen entirely inside the population band, invisible against a ruler built from strangers.
The range describes where most people sat. It was never a target drawn for you.
This is well established in laboratory medicine, going back to the foundational work on how reference intervals are derived (Doyle and Bunch, 2023). fenome still always shows you that standard band, because it is the shared, validated starting point. It just does not stop there.
Think of resting heart rate. An endurance runner might sit near 50, someone else near 70, and both are perfectly healthy. The population "normal" is the whole crowd blurred together, but your own normal is far narrower than the crowd's. The measure of this is the index of individuality, first described by Harris in 1974.
The striking, repeatedly confirmed finding is that the majority of common blood markers have marked individuality: your day-to-day variation is smaller than the spread across people (Petersen and Fraser, 1999). For those markers, the printed range is a loose fit. A value can be "normal" for the crowd and clearly unusual for you, or near the edge and perfectly usual for you. The better reference, then, is not the population band. It is you, measured against yourself over time.
When your own band is far narrower than the crowd's, the crowd's ruler hides what is happening to you.
Two of your results are never identical, even when nothing has truly changed. The lab instrument has its own imprecision, and your body naturally varies day to day with sleep, food, hydration, stress, and time of day. It is like stepping on a bathroom scale twice and getting slightly different numbers. You did not change in thirty seconds; the scale just twitched.
Laboratory medicine has a precise tool for this: the reference change value, the smallest difference between two of your own results that is unlikely to be explained by noise alone (Fraser and Harris, 1989).
fenome uses the RCV to gate every trend. If a move between your readings is within the noise, the engine calls it stable and stays calm. Only a change that clears the noise floor is described as moving. This is the rule that keeps an ordinary wobble from ever being dressed up as a worrying trend.
Put those ideas together and a clear direction emerges. Because most markers fit you more tightly than they fit the crowd, and because real change is best read against your own prior results, the most useful reference is your own baseline: where your results have actually sat over time. As your history builds, fenome learns the value your readings usually settle around and shows where a new one sits beside your own pattern.
We are deliberate, and deliberately honest, about what this is:
A baseline is started, not set. One reading is a starting point. It forms over several comparable tests and only becomes a real personal estimate with enough history. We say started, forming, settled, never "your number" on day one.
Your baseline is not your "normal". A personal baseline can sit above or below the healthy range, so it can never excuse an out-of-range value. Your status is always judged against the standard reference range; the baseline is personal context layered on top, never a replacement for it.
It is self-knowledge, not a verdict. fenome describes where your marker tends to sit and how it is moving. It does not diagnose, predict, or claim to find disease. Real-world data from different labs widens the uncertainty rather than hiding it, and the honest promise is a clearer picture of yourself over time, then a word with your clinician about anything you want to act on.
It is tempting to hand your numbers to a large language model and ask it for a nice paragraph. We deliberately do not. Every status, every trend, and the calm signal you see are computed by a pure deterministic engine: plain, transparent code that turns the science above into a result. The same readings always produce the same answer, and every answer traces back to a formula, not a guess.
So we give AI the two jobs it is genuinely good at, and nothing more. It reads your report (turning a PDF or a photo into structured readings), and it puts the engine's result into warm, plain language. It never decides the signal.
That restatement is checked before you ever see it: every figure in the sentence must match the engine's own result, and the direction must agree. If anything fails the check, fenome falls back to a plain, engine-written line that is guaranteed to be faithful. The friendly voice is welcome; it just never gets to decide what is true.
The numbers are the easy part. The meaning is the hard part, and it is too important to leave to a guess.
The noise floor and the individuality of every marker are not our opinion. They come from the EFLM Biological Variation Database, a European clinical reference that pools thousands of datasets from hundreds of peer-reviewed studies, each quality-graded. fenome reads its within-person and between-person variation straight from that source, so each marker's sense of "how much does this naturally bounce around" is grounded in the published literature, not tuned by hand.
We are equally clear about the limits. fenome's current trend math uses a simple, symmetric form of the RCV; the precision-medicine evidence for dense personal baselines is promising but still early; and the evidence that any of this changes behaviour is genuinely thin. So we promise what is well supported, honest self-knowledge, and nothing more. The science is the soul of the engine, not a poster claim.