Contents
In June 2026, MyTwin Lab built a first prototype of the MyTwin Engine: the layer of the digital twin that turns a person’s data into health scores. Test results, wearable data and lifestyle habits go in; a biological age, scores for the main organ systems, risk projections and what-if simulations come out.
The principle is simple: wherever a scientifically validated model exists, the engine uses it rather than inventing its own. And every score says how far it can be trusted. The prototype was tested on fictional profiles only, with no patient data.
What the engine computes
Biological age. From a routine blood test, the engine computes Phenotypic Age, or PhenoAge: a measure that combines chronological age with nine common biomarkers, such as albumin, glucose, creatinine and C-reactive protein. In the original study, on US population data, each additional year of Phenotypic Age was associated with a 9% higher risk of death, independently of chronological age.
Organ-system scores. A view of the main systems of the body (cardiovascular, respiratory, musculoskeletal), each built on the markers and clinical criteria that medicine already uses for it.
Risk projections. How a risk may evolve over the coming years. Fracture risk, for instance, relies on QFracture, a model derived and validated on several million patients in UK primary care.
What-if simulations. What would change under a concrete decision: losing weight, starting a treatment, changing one’s lifestyle (see below).
A score is an estimate, not a forecast: our article on what a risk score really means explains the difference. None of these scores is a diagnosis; they are meant to be read with a doctor.
A reliability level on every score
Health scores rarely say how much they can be trusted. The engine does, for each of them, with one of four levels: a clinically validated model used as published; an estimate grounded in validated literature; a composite index of our own, not clinically validated; or no score at all, when the data or the evidence isn’t there yet.
- ValidatedA clinically validated model, used as publishedFracture risk with QFracture
- EstimateGrounded in validated literature and risk factorsBiological age with PhenoAge, a prognostic measure rather than a diagnosis
- CompositeA wellness index of our own, not clinically validatedThe overall Health Score
- Not yetNot enough data or evidence for a reliable scoreCancer: no composite score until screening data supports one
The last level matters as much as the first. Rather than guess, the engine leaves a domain empty: there is no composite cancer score, for instance, until screening data can support one.
What-if: simulating a change before making it
A simulation changes the person’s own biomarkers the way an intervention would, then runs them back through the same models. The result depends on where the person starts: lowering a high cholesterol level moves the projection more than lowering one that is already low.
The size of each effect comes from published trials. For a statin, cardiovascular risk is lowered using the effect measured by the Cholesterol Treatment Trialists’ meta-analysis (a 22% lower rate of major vascular events per 1 mmol/L of LDL cholesterol lowered). For weight loss, diabetes risk follows the Diabetes Prevention Program, where each kilogram lost was associated with a 16% lower risk.
These are evidence-based projections, not promises. They use trial averages, not yet the person’s own response: a statin may lower one person’s cholesterol more than another’s. Repeated tests over time are what will close that gap.
From prototype to MyTwin Longevity
The prototype ran on a handful of fictional profiles, from a very active 52-year-old executive to a 67-year-old with high blood pressure, to check that the scores, their reliability levels and the simulations held together. It made no claim about real patients, and none is made here.
Two months later, the engine went into a product: MyTwin Longevity, where biological age, organ-system scores, risk projections and what-if simulations are read with a doctor specialised in prevention and longevity. More models will be added as the data and the evidence allow.
Frequently asked questions
Was the MyTwin Engine tested on real patients?
No. The June 2026 prototype was tested on fictional profiles only, with no patient data. It checked that the scores, their reliability levels and the simulations held together; it makes no claim about clinical performance.
Does a score from the engine replace a diagnosis?
No. Each score is an estimate built on a published model, with its reliability level shown next to it. It is meant to be read with a doctor, never instead of one.
Sources
Sources · 4
- 01Liu Z. et al., A new aging measure captures morbidity and mortality risk across diverse subpopulations from NHANES IV, PLOS Medicine, 2018
- 02Hippisley-Cox J., Coupland C., Derivation and validation of updated QFracture algorithm to predict risk of osteoporotic fracture in primary care in the United Kingdom, BMJ, 2012
- 03Cholesterol Treatment Trialists’ Collaboration, Efficacy and safety of more intensive lowering of LDL cholesterol, The Lancet, 2010
- 04Hamman R. F. et al., Effect of weight loss with lifestyle intervention on risk of diabetes, Diabetes Care, 2006
This news is provided for information only. The technologies it describes are at research or pilot stage, and none of them replaces advice, diagnosis or treatment from a healthcare professional.
