From Multiomics to Clinical Intelligence: Volv Global at Frontiers Health 2026
Volv Global at Frontiers Health's panel on AI & biomarker science translating complex biological signals into cli...
People with rare and difficult-to-diagnose diseases can wait years for a correct diagnosis, and many carry no accurate code in any health system. That gap falls hardest on those with rarer conditions, atypical presentations, and access to fewer specialists.
inTrigue research applies Volv Global’s machine learning to de-identified, population-scale data, estimating how many people may carry a disease that was never coded, and where they sit in the system.
Why finding patients matters
Patient data captures only a fraction of the real disease population in rare and difficult-to-diagnose conditions because coding systems were not built to capture the full complexity of rare phenotypes. Those patients not captured remain absent from the clinical programmes and commercial strategies built for them.
When the diagnosed population understates the full one, every model built on it carries the error forward: trial assumptions, market access dossiers, and commercial forecasts.
inTrigue replaces that uncertain count with a real, characterised population, grounding every decision in evidence that holds up to scrutiny.
One of he most common reasons clinical trials in rare disease fail recruitment is that the patient population was smaller than assumed. Patients who exist are missed simply because they sit outside what standard coding captures.
inTrigue finds them before recruitment stalls, widening the recruitable pool.
Knowing the real patient population before protocol lock, indication selection, or launch planning removes one of the most common sources of late-stage programme failure.
Engaging inTrigue earlier carries more credible evidence into every stage that follows, from regulatory review to commercial launch.
Where inTrigue creates value
Mapping the biomarkers and phenotypes that define the eligible cohort, de-risking trial design and building evidence regulators and HTA bodies will scrutinise.
Grounding patient counts on real numbers, not diagnostic proxies, to strengthen regulatory, HTA and market access arguments.
Mapping biomarkers and trajectories across the patient population to surface the phenotypic complexity that shapes medical communication.
Finding where the patients actually are, letting launch targeting and field deployment begin on real numbers.
How inTrigue works
inTrigue applies machine learning to population-scale real-world data, including electronic health records, claims data, primary care records, and ambulatory data, to surface patients whose data patterns are consistent with the clinical signatures of rare and difficult-to-diagnose diseases in patients who have never been correctly coded.
Every finding is validated with clinical experts before being delivered. The methodology is fully explainable: the clinical reasoning behind each flag can be examined, challenged, and published.
What inTrigue makes possible
In research studies the estimated population was consistently larger than coded data suggested: 2.5 to 4.6 times the diagnosed count in hypertrophic cardiomyopathy, and 3.2 times the sponsor’s forecast in neuroendocrine tumours. In acute hepatic porphyria, working retrospectively from a database with no confirmed diagnoses, a blinded clinician judged 90% of the highest-ranked candidate records plausible.
In each case, the true patient population was larger than the coded data suggested, and many of the patients found had been waiting years for a diagnosis they had not received.
Pharmaceutical teams partner with Volv Global on inTrigue research to test feasibility assumptions against real-world populations. The intelligence generated in each engagement is shared with our partners and is publishable, so the findings compound beyond the project itself.
Read our case studies
What people ask
inTrigue applies machine learning to population-scale real-world data to detect the clinical signatures of rare and difficult-to-diagnose diseases in patients who carry no relevant diagnostic code. It does not search for codes; it searches for patterns. Those patterns are validated with clinical experts, producing population-level estimates with a transparent evidence trail, reviewed with clinical experts.
When the coded population understates the real one, every decision built on it carries the error. Trial assumptions, market access submissions, and commercial forecasts all rest on a patient count. If that count is wrong, the consequences compound across development, regulatory review, and launch.
The patients left out of the count also pay a personal cost: years without a correct diagnosis, without access to the right treatment, and without the option to participate in a trial designed for their condition. That cost falls on a real person, whether or not they appear in the coded data.
Clinical trials in rare disease frequently stall because the recruitable population was smaller than assumed. inTrigue finds the patients that conventional feasibility misses, widening the recruitable pool and de-risking the programme. In HCM, it found between 2.5 and 4.6 times more patients than the diagnosed count. In NETs, 3.2 times the sponsor's prediction. Beyond recruitment, the wider population also produces stronger real-world evidence, richer natural history data, and a more credible comparator for regulatory and HTA review.
inTrigue works with real-world data accessed through Volv Global's trusted data partners, covering more than 400 million patient records across the United States, United Kingdom, Germany, and beyond. No identifiable patient data is obtained at any point. All work is conducted under applicable privacy and regulatory frameworks, and Volv Global does not hold patient data.
Standard approaches search for patients who already carry the relevant diagnostic code. inTrigue detects clinical signatures in patients with no code at all, applying a proprietary machine learning methodology validated with clinical experts across population-scale data and multiple geographies. Many machine learning approaches to patient-finding are built and validated on a single population and lose accuracy when applied elsewhere; inTrigue's methodology is KOL-validated across data environments and geographies.
Every model is fully explainable: the clinical reasoning behind each flag is transparent and publishable. The knowledge generated compounds; each engagement adds to a methodology that grows more precise as it is applied across diseases, data environments, and geographies. Where inTrigue studies the undiagnosed population; inCare, the diagnosed population not yet treated.
Shaping the future today
Volv Global at Frontiers Health's panel on AI & biomarker science translating complex biological signals into cli...
Volv Global CEO Christopher Rudolf will speak at the DREAMS Neuromuscular Disease AI Webinar
Volv Global at the 25th Annual NEALS Meeting, presenting new research on early identification of ALS using advanced M...
Presenting new research on whether a machine learning model can flag which patients with CAP are most likely to progr...
The next useful thing AI does in solid tumour care is about timing, not detection. The patient must reach the right t...
Results for the model flagging early ARDS risk in pneumonia will be presented at ERS Congress 2026.
Volv Global at ERS Congress 2026, presenting new research in collaboration with CSL Behring: Can We Catch ARDS Before...
AI Innovation in Real-World Evidence & Data Analytics Volv Global has been selected as a finalist in the 2026 Fie...
Volv Global applied machine learning to 24M UK primary care records, surfacing ~1,200 undiagnosed GEP-NET patients, 5...
GEP-NETs are among the most diagnostically elusive tumours in oncology. This case study shows how machine learning ap...
ASCO 2026 produced landmark oncology data. The question no one asked from the podium: can clinical infrastructure get...
Drive For Change white paper: three paradigm shifts to unlock orphan drug development