The simplest way to represent a person's health state is as a label: healthy, at risk, affected. Labels are legible and easy to store. They are also lossy. The moment you assign a binary label, you discard information about where in the underlying biological distribution that person sits, how far they are from a threshold, and which direction they are moving.
Continuous coordinates preserve that information. Instead of asking whether a person crosses a threshold, a coordinate-based framework asks: where does this person's biological state currently sit in a high-dimensional space, how does that position compare to their own baseline and to reference populations, and what is the uncertainty on that estimate? That question is harder to answer, but the answer is more actionable for research and, eventually, for understanding health trajectories.
The infrastructure implications
Moving from labels to coordinates requires different data structures, different validation approaches, and different communication practices. Continuous values need uncertainty bounds, not just point estimates. Comparisons across people need normalization for demographic and measurement variation. Interpretation requires calibration to what coordinate distances actually mean biologically.
None of this is insurmountable, but it requires deliberate design from the start. Retrofitting continuous outputs onto systems built for binary labels is harder than building coordinate-native infrastructure from the beginning. That is part of why Cytognosis is starting with the infrastructure layer rather than rushing to the application layer.
Open notebook
Our thinking on how to define, validate, and communicate health-state coordinates continues to develop. This page is part of our open notebook and will be updated as the science progresses.