in silico surgical trials unit

RCT-grade evidence for surgery.

Building world-class digital surgical twins. Clinician led. The future of surgical trials.

We are the leading centre for in silico trials and target trial emulation in surgery. Expert surgeons co-design the question and deep AI builds what no human team could construct alone: detailed digital surgical twins, complex probabilistic pathways, and degradation models creating clinical reality at scale. The result is a powered, pre-registered surgical trial in under six months, with regulatory grade results.

Led by Professor Aneel Bhangu, Professor of Surgery and Consultant Colorectal Surgeon, University of Birmingham
→ how it works view focus areas
time to regulatory-grade output
Under six months
From scoping through pre-registration to peer-reviewed, regulatory-grade output; a timeline that conventional surgical trials cannot approach.
cost versus conventional RCT
Approximately 10% of the cost
The direct cost of a VISTA in silico trial is a fraction of a conventional surgical RCT, with scope and budget agreed before any data are accessed.
effective sample size
10 to 100 times larger
The effective sample reflects the full range of real-world surgical practice across a health system, rather than the restricted eligibility of a conventional trial.
01 what we do

In silico surgical RCTs.
Faster and cheaper
than conventional trials.

the method
Clinician led. Deep computational AI.
A full trial protocol covering population, intervention, comparator, and outcomes, written and pre-registered before any data are touched. Expert clinicians define the question and select the inputs. Deep computational AI then constructs the detailed digital surgical twin cohort and builds the mathematical architecture of the trial: probability distributions, clinical effect pathways, and reality degradation layers, at a scale and complexity impossible to assemble by hand. Where observational data independently supports it, target trial emulation provides a corroborating estimate.
the data
Best-in-class. For every question.
Each study begins with a rigorous search for the highest-quality available data: published RCTs, national audit outputs, registry datasets, and administrative records, drawn from wherever the best evidence exists. The data are not fixed in advance. They are selected specifically for the question, the population, and the health system in view. Results that cut against the primary finding are reported in full.
conventional RCT
time8 years average from design to HTA decision
cost£3m to £8m direct costs, excluding opportunity cost of delayed adoption
sample300–800 patients at selected specialist centres, strict eligibility
populationTrial-eligible patients, not representative of NHS practice
system fitSingle country; results applied globally regardless of context
VISTA in silico RCT
timeMonths, not years from scoping to regulatory-grade output
costA fraction of a conventional trial, scoped to the question, not a fixed price
sample1,000–100,000 patients reflecting the full range of surgical practice
populationReal NHS patients; the population who will actually receive the technology
system fitAny health system; calibrated to local populations, practice patterns, and costs

Working from device-specific data, VISTA generates regulatory-level submissions from pilot data through to full in silico RCT outputs, without the need for a decade-long, multi-million pound conventional trial. Patient care and surgical technology are moving too fast for the evidence base to wait that long. The infrastructure now exists to match the pace of clinical innovation with evidence of equivalent rigour.

02 process

From question to
evidence.

phase activity detail timeline
01 / scope Priority setting Structured workshop to define the decision problem, the evidence gap, and the regulatory endpoint. We identify whether the question is answerable from available data before any commitment is made. Budget agreed. wk 1–2
02 / design Trial protocol Full protocol written in PICO format: population, intervention, comparator, outcomes, with follow-up windows and analysis plan specified in full. Pre-registered before data access. Expert clinical review at every stage. wk 3–6
03 / generate In silico analysis Best-in-class data inputs are assembled by expert clinical review. Deep AI models construct the detailed digital surgical twin cohort and build the mathematical structures: probability distributions, clinical effect pathways, and reality degradation layers, that define the simulation architecture. This is where the computation happens that no human team could perform by hand. mo 2–5
04 / validate Benchmarking Results tested against the best available clinical evidence and known real-world outcome distributions. Pre-specified sensitivity analyses run and reported in full, including those that cut against the primary finding. Clinical plausibility reviewed by expert surgeons before any output is finalised. mo 4–6
05 / translate Regulatory output Results formatted for regulatory submission, health technology assessment, or peer-reviewed publication as agreed. Full documentation package to publication standard. Reporting follows pre-registered analysis plan throughout. mo 5–6
04 foundation

Infrastructure built
over a decade.

01 data
Best-in-class data. Wherever it exists.
Every study draws on the highest-quality available data for that specific question: published trials, audit outputs, national and international registry datasets, and real-world records. The selection is clinically led and question-specific, not constrained to any single source or system.
02 AI
Deep computational AI
Deep computational models construct the mathematical architecture of the in silico trial: detailed digital surgical twin cohorts, probability distributions, clinical effect pathways, and reality degradation structures of a complexity that no human team could assemble by hand. The AI is the analytical engine; the clinical team is the driver.
03 knowledge
Fifteen years of running surgical RCTs
Over fifteen years of designing, running, and publishing surgical RCTs at multicentre scale gives VISTA the operational knowledge to calibrate every in silico analysis against clinical reality. We know where conventional trials fail, and we know what regulators actually need from a submission.
04 expertise
Surgical leadership. Regulatory fluency. Methodological depth.
Surgical clinical leaders, epidemiologists, and regulatory specialists within a single unit. The triangulation of expert clinical judgement, deep AI capability, and health technology assessment knowledge in this form does not exist elsewhere.
Note

The proprietary AI models, the fifteen-year RCT knowledge base, and the trusted clinical leadership underlying VISTA represent a unique infrastructure that cannot be replicated. The combination of surgeon expertise, deep computational AI, and question-specific data selection produces something genuinely new: evidence that is faster, larger, and more health-system relevant than anything conventional trials can deliver.

Validation

Our methodology has been validated against recent multicentre surgical RCTs. The in silico outputs sit within the confidence intervals of the randomised evidence; not as a coincidence, but as a design requirement. Every VISTA study is benchmarked against the best available trial data before any finding is reported.

The data opportunity

The world already has more surgical data than it will ever fully use. Ninety per cent of it sits in registries, audit systems, and administrative records: collected, structured, and never interrogated at the scale the questions demand. VISTA exists to change that. Better utilisation of existing data, not decade-long trials waiting for answers that are already within reach.

04 scope of work

Surgery and health
technology assessment.

Surgical procedures
Comparative effectiveness of operative approaches across all major surgical specialties, including robotic versus laparoscopic surgery, open versus minimally invasive repair, and oncological resection techniques. Applicable wherever a procedural comparison exists and conventional trial recruitment is not feasible at the required scale.
e.g. robotic versus laparoscopic rectal cancer resection
Surgical and implantable devices
Post-market comparative evidence for surgical stapling, mesh, fixation, vascular, and orthopaedic implant devices. Generating regulatory-grade submissions to support commercial differentiation and HTA evaluation where prospective RCT data cannot be obtained at the required scale or within a commercially viable timeframe.
e.g. EVAR versus open repair; hernia mesh comparative effectiveness; joint replacement platforms
Health economic analysis
Cost-effectiveness, budget impact, and value-of-information analyses generated alongside clinical endpoints within the same in silico framework. Outputs calibrated to the health system and decision context in view, structured for technology appraisal and commissioning submissions.
e.g. cost per quality-adjusted life year for robotic surgical platforms across NHS trusts
Health system productivity
Target trial emulation applied to workforce, capacity, and pathway questions in surgical and wider health systems. Evaluating the productivity impact of innovations, care pathway redesign, and staffing interventions at the population level, where prospective trials are neither feasible nor funded.
e.g. population-level productivity impact of expanding robotic surgical capacity
05 current pipeline

Technology assessment pipeline.

Robotic and minimally invasive surgery
01Robotic versus laparoscopic resection in oesophagogastric cancer
02Inguinal hernia repair: open versus minimally invasive surgery
03Robotic surgery in high-performance centres: a step-wedge productivity analysis
Surgical devices and wound management
04Advanced wound dressings in gastrointestinal surgery
05Small bites fascial closure in emergency laparotomy
06ICG fluorescence imaging across country-level surgical populations
Surgical decision-making
07Appendicitis: antibiotics versus operative management
06 global reach

Health-system appropriate.
Any country.

The VISTA framework applies to any health system with appropriate data infrastructure: NHS, European, Gulf, USA, Australasian, and beyond. Every analysis is calibrated to the populations, practice patterns, costs, and regulatory context of the health system in question. Evidence that is relevant to the decision being made, not evidence from a different country applied by assumption.

07 about

The people
behind the method.

VISTA is led by clinical academics and methodologists with direct experience of designing and running surgical RCTs at multicentre scale. We are not a data analytics company that has read the literature. We are the people who wrote parts of it.

Director
Professor Aneel Bhangu
Statistician
Omar Omar
Data Analyst
John Obafunsho Abiola
Research Fellows
Appointments pending

A start-up company housed at the Surgical Data Institute, University of Birmingham, UK.

08 get in touch

The trial you need.
Without the wait.

Whether you have a product approaching NICE evaluation, a device requiring post-market evidence, a health system that needs locally-relevant data, or a research question a conventional trial cannot answer; every project begins with an honest conversation about what is feasible, what the data will support, and what it will cost.

→ contact us
How a project begins
01A structured scoping conversation to define the decision problem, the evidence gap, and whether the question is answerable from available data. No commitment is required at this stage.
02A written feasibility summary with an indicative scope, timeline, and budget range, reviewed and agreed before any formal engagement begins.
03A full trial protocol written to pre-registration standard, with analysis plan locked before data access. Every project is documented to regulatory and publication standard throughout.