AR’s journey into healthcare has followed a familiar path: slow initial uptake constrained by regulation and cost, then faster adoption once efficacy evidence and cleared products reached a tipping point. In 2026, AR isn’t experimental in clinical settings anymore — it’s procured technology with clearance dossiers, published outcome data, and genuine line items in capital expenditure budgets.

Surgical Navigation: Precision at the Point of Care

The most consequential AR application in healthcare is surgical navigation — overlaying imaging data, instrument tracking, and anatomical models directly into the surgeon’s visual field during procedures.

Stryker Mako is the category benchmark for orthopaedic surgery. The robotic surgical system integrates CT-derived 3D bone models with real-time AR guidance for knee and hip arthroplasty. It pre-plans implant positioning digitally, then provides haptic and visual constraints during surgery to keep the surgeon within the planned boundaries. A 2024 meta-analysis in the Journal of Arthroplasty covering 14 studies found Mako-assisted procedures delivered significantly better implant alignment accuracy and reduced early revision rates. Stryker reports over one million Mako procedures performed globally.

Medivis SurgicalAR takes a different approach — FDA-cleared mixed reality software running on Microsoft HoloLens 2. Surgeons see DICOM-imported 3D anatomy (CT, MRI, or angiography data) overlaid spatially onto the patient. Neurosurgery, orthopaedics, and vascular surgery are the primary use cases. The platform’s spatial registration capability keeps the overlay anchored to the patient as the surgical team moves around the table. Medivis received 510(k) clearance in 2022 and has since expanded into over 40 US hospital systems.

Vein Visualisation: AccuVein

AccuVein makes handheld and mounted devices that project a real-time map of subcutaneous veins onto the patient’s skin surface using near-infrared light and AR projection. The problem it solves is significant: peripheral IV placement fails on the first attempt in around 40% of adults, rising to over 60% in difficult-access patients including the elderly, obese, and paediatric populations. Published data shows a 3.5x improvement in first-stick success rates. The device is now used in over 10,000 hospitals and clinics across 100 countries, making it one of the most widely deployed AR medical devices in the world.

Medical Training: Fundamental Surgery

Medical simulation has been transformed by immersive platforms that cut the cost and scheduling constraints of cadaveric and mannequin-based training. Fundamental Surgery is a VR and AR surgical training platform simulating procedural skills including laparoscopic technique, regional anaesthesia, and vascular access, paired with haptic feedback hardware to replicate instrument resistance.

A peer-reviewed study published in Surgical Endoscopy found trainees using Fundamental Surgery showed equivalent procedural competency to those trained on physical simulators — at a fraction of the cost per session. The platform is now formally integrated into postgraduate medical curricula at NHS trusts and US residency programmes.

Beyond the operating theatre, AR is proving useful in patient-facing applications. Platforms including Echopixel and Touch Surgery (now part of Medtronic) let clinicians show patients 3D visualisations of their own anatomy — derived from their own imaging data — to explain diagnoses and planned procedures. Studies consistently show that patients who receive visual explanations of their condition demonstrate better understanding, higher satisfaction scores, and better adherence to pre- and post-operative instructions.

For complex procedures with significant risk, enhanced informed consent supported by AR visualisation is increasingly documented in clinical records as a risk management measure. It’s a small shift in process that makes a real difference.

The Regulatory Picture

Medical AR sits within the frameworks governing software as a medical device (SaMD). In the US, the FDA’s Digital Health Centre of Excellence oversees 510(k) and De Novo pathways for AR medical software. Whether a tool needs Class II or Class III clearance depends on intended use — anything that assists a clinician in making a diagnosis or performing a procedure typically requires 510(k) clearance or PMA approval.

In the UK and Europe, the Medical Device Regulation (EU MDR 2017/745) similarly classifies clinical AR software as a medical device, requiring conformity assessment through a Notified Body before CE marking. For UK Conformity Assessment post-Brexit, the MHRA governs the equivalent process under UK MDR 2002 (as amended). The regulatory pathway adds 18 to 36 months to development timelines for most clinical applications — which is the primary reason consumer-grade AR hardware hasn’t simply been repurposed for surgical use.

For procurement teams in NHS trusts or private healthcare groups, the critical due diligence question is whether the intended clinical application is covered by the vendor’s cleared indication. Using a cleared device outside its cleared indication is off-label use, with real liability implications.

ROI Evidence

Cost-effectiveness data is maturing. A 2025 analysis from the Nuffield Trust examining AR-assisted surgery in the NHS found that the reduction in revision procedures from improved implant accuracy — using Mako data — produced a net saving of approximately £2,800 per primary knee arthroplasty over a five-year horizon. That more than offsets the system’s capital cost when annualised across procedure volumes.

For training applications, the economics are more immediately visible. Virtual procedural training costs around 80% less per trainee hour than cadaveric simulation once hardware is amortised. For AccuVein-type vein visualisation, the ROI centres on reduced failed IV attempts — each carrying direct supply costs, nursing time, and patient satisfaction metrics that increasingly affect NHS value-based care contracts.

What Comes Next

The near-term development curve in medical AR involves two convergences: spatial registration accuracy improving to sub-millimetre levels needed for microsurgery, and real-time AI analysis being integrated into the AR overlay to flag anatomical anomalies, predict tissue planes, or alert to instrument proximity to critical structures. Both are in active clinical trials in 2026. The infrastructure being built today — cleared platforms, trained clinicians, NHS procurement pathways — is what will absorb those capabilities when they arrive.