If your strategic radar for emerging medical AI categories includes the established frontiers (clinical decision support, imaging AI, generative AI in workflows) but not medical digital twins — dynamic, data-linked computational representations of individual patients that update over time and support forecasting, simulation, and treatment optimisation — the 2015-2025 publication and patent trajectory shows the category warrants explicit strategic attention. Research publications on medical digital twins surged from near-zero in 2015 to 372 in 2025. Patent filings in healthcare digital twins (CPC class G16H) increased from 30 in 2016 to 4,926 in 2024 (the 2025 figure of 4,087 partially reflects publication lag). In a randomised controlled trial of 150 diabetes patients on Twin Health's Whole Body Digital Twin platform, 71% achieved healthy blood sugar levels (HbA1c <6.5%) within twelve months while safely reducing their intake of blood sugar–lowering medications. The category is forming with both research depth and emerging clinical evidence.

The publication trajectory is one of the steeper growth curves in the medical AI literature. Annual publication counts:

The cumulative effect: by 2024-2025, approximately 17,000-18,000 healthcare digital twin patents have been filed. The intellectual property landscape in the field is being established rapidly.

The prescriptive question for organisations setting strategy: what is the actual capability today, what is the likely trajectory, and what should be done now?

The current capability has substantial limits. A 2025 scoping review in npj Digital Medicine (Sadée et al., 2025) assessed 149 human digital twin studies published between 2017 and 2024. The review found that only 12.1% (18 studies) satisfied the National Academies of Sciences, Engineering, and Medicine (NASEM) definition of a digital twin. The NASEM definition requires three elements: personalisation, dynamic updating, and predictive capability. Only 19% of systems were tested in real healthcare environments. The conceptual rigour gap is real: most "digital twin" research does not meet the technical definition of a digital twin.

The data observation: the publication and patent volume substantially exceeds the rigorous-definition research base. The category is forming rapidly in terms of activity; the conceptual maturity is catching up more slowly.

The clinical trial evidence is the brighter signal. Two trial categories show 2025 progress:

Oncology trials: A pilot trial in prostate cancer using adaptive therapy concluded in 2025 with significantly increased survival (Zhang et al., 2022; the trial completed reporting in 2025). New trials extended the approach to breast cancer (Mayo Clinic phase II) and ovarian cancer (ACTOv phase II RCT, n=80).

Diabetes trial: A randomised controlled trial (n=150) of Twin Health's Whole Body Digital Twin platform found that 71% of participants achieved an HbA1c below 6.5% within twelve months, while safely reducing their intake of blood sugar–lowering medications. The trial outcome is clinically significant: HbA1c <6.5% is the diabetes remission threshold for many patients, and achieving it while reducing medication use is a strong clinical signal.

The trajectory observation: where rigorous medical digital twin research exists, the early clinical results are promising. The 71% diabetes outcome and the prostate cancer survival improvement are not minor signals; they suggest the technology can produce clinically substantial outcomes in well-designed deployments.

Three structural drivers shape the trajectory.

The first driver: the underlying technology is multi-component and benefits from advances across many areas. Medical digital twins require patient data acquisition (continuous glucose monitors, wearables, EHR integration), computational modelling (physiological models, AI-based pattern recognition), prediction infrastructure (forecasting future patient states), and clinical decision support (translating predictions into clinical recommendations). Advances in any component (better sensors, better models, better integration) improve the overall capability. The 2020-2025 period saw progress in all components.

The second driver: chronic disease management is a structurally good fit for the digital twin approach. Diabetes, cardiovascular disease, cancer, and other chronic conditions require ongoing monitoring, adjustment, and intervention over years. Digital twin approaches that maintain dynamic patient models, predict progression, and support treatment decisions match this clinical workflow. The 71% diabetes outcome in the Twin Health trial reflects this fit.

The third driver: industry investment is substantial. The 4,926 patent filings in 2024 reflect investment across multiple firms: Twin Health (diabetes), various oncology-focused firms, EHR vendors building digital twin capabilities into clinical platforms, and major pharmaceutical firms exploring digital twin approaches for drug development. The investment is broad enough to support continued capability development across multiple use cases.

Three prescriptive moves follow for organisations operating in or adjacent to medical digital twins in 2026.

The first prescriptive move: track the high-rigour clinical evidence specifically. The 12.1% of human digital twin studies that meet NASEM criteria represent the credible evidence base. The remaining 87.9% are interesting research but may not represent actual digital twin capability. Strategic planning should engage with the rigorous-evidence subset rather than the aggregate research output.

The second prescriptive move: identify the chronic disease use cases where digital twin approaches have the strongest fit. Diabetes (Twin Health trial), oncology (Mayo Clinic, ACTOv trials), and cardiovascular disease (multiple emerging trials) have early clinical evidence. Organisations operating in these therapeutic areas should engage with digital twin approaches as near-term clinical capability. Organisations operating in acute care or single-encounter settings should engage with digital twins as a longer-horizon capability.

The third prescriptive move: build the data infrastructure that digital twin capability requires. Digital twin platforms depend on rich, continuous patient data: wearable sensor data, continuous glucose monitoring, EHR integration, lifestyle data, genomic data. Health systems and patient-care organisations with strong patient data infrastructure are positioned to deploy digital twin capability. Organisations without this infrastructure face higher barriers to entry. Strategic data infrastructure investment in 2026-2027 enables digital twin deployment in 2028-2030. Alongside it, organisations should weigh how to engage the IP landscape (4,926 patents in 2024) through partnership, licensing, or independent development depending on strategic position, and plan for 24-48 month timelines as the category matures from proof-of-concept to enterprise scale.

The trajectory: through 2026-2028, medical digital twins will likely follow the same pattern as ambient AI scribes (article #121) and other clinical AI categories, moving from current proof-of-concept and limited deployment to enterprise-scale deployment in specific chronic disease use cases. The diabetes use case (Twin Health and successor platforms) is likely to be the first to reach broad clinical deployment, with oncology applications following and cardiovascular and other chronic disease applications maturing later. By 2030, medical digital twins are plausibly an established clinical AI category with substantial penetration in chronic disease management. The 4,926 patents in 2024 and the 71% diabetes outcome in the RCT are both substantial signals that the category is forming now; strategic plans calibrated against this trajectory will be aligned with the visible direction, while plans that treat medical digital twins as a longer-horizon possibility may underestimate the pace of deployment in specific high-fit use cases.