Long Product Lifecycles.
An implant placed today may need outcome data collected for ten or fifteen years. Very few other industries commit to data collection on that horizon.
INDUSTRY
Medical Devices
The sector where the evidence work starts at approval rather than ending there. Surveillance, clinical follow-up, registry commitments and real-world outcomes run for as long as the device is on the market — often decades. This page covers how that work is structured, what governs it, and where the data sits.
Discuss Your Use CaseTHE SECTOR TODAY
What Defines The Medical Device Sector
Medical devices cover hardware, diagnostics and software under the same regulatory perimeter. Risk classification determines the evidence, pathway and surveillance expected after launch.
Three structural features shape how device organisations operate:
Distributed Evidence.
The manufacturer holds the device data. The hospital holds the clinical outcome. The patient holds the quality-of-life answer. No single party has the full picture, which makes evidence work fundamentally an integration problem.
Software Has Become The Device.
A growing share of the sector isn't hardware at all. Software as a Medical Device is regulated in its own right, which puts version control, change management and model behaviour directly inside the regulatory perimeter.
Company shapes vary — large diversified manufacturers, single-product specialists, and software-first companies that reached device regulation from the technology side rather than the clinical one. The obligations converge regardless of where a company started.
THE LIFECYCLE AFTER APPROVAL
What Happens Once A Device Is On The Market
Pre-market work gets most of the attention. The larger operational load sits after it.
Post-Market Surveillance (PMS)
An active, planned system for collecting and reviewing field experience. It draws on complaints, service records, literature, registry data, similar-device data and user feedback. Under EU MDR this is a documented system with a plan per device, not an ad-hoc activity.
Post-Market Clinical Follow-Up (PMCF)
The clinical arm of surveillance. PMCF confirms safety and performance across the expected lifetime, checks for emerging risks, and tests whether the original clinical evidence still holds in real use. Methods range from registries and surveys to structured clinical investigations.
Vigilance And Incident Reporting
Serious incidents and field safety corrective actions reported to competent authorities within defined windows. In the US, medical device reports go to FDA and appear in the MAUDE database.
Periodic Reporting
Depending on class and jurisdiction, manufacturers produce periodic safety update reports or post-market surveillance reports summarising what surveillance found and what was done about it.
Registries
Long-running data collections, sometimes manufacturer-run and sometimes national or society-run, such as joint replacement or cardiac device registries. Registry participation is increasingly a condition of market access rather than a voluntary activity.
Real-World Evidence (RWE)
Data generated outside a controlled study — routine care, device telemetry, claims, patient-reported outcomes. RWE now supports label expansions, reimbursement submissions and surveillance obligations, which raises the bar on how it is collected and traced.
THE REGULATORY FRAME
The Standards And Regulations That Shape The Work
EU MDR — Regulation (EU) 2017/745
EU devices
Documented PMS system and per-device plan, PMCF within ongoing clinical evaluation, periodic safety update reporting for higher classes, trend reporting, UDI and EUDAMED registration
EU IVDR — Regulation (EU) 2017/746
EU diagnostics
Parallel obligations with performance evaluation and post-market performance follow-up
MDCG guidance
EU interpretive
Templates and expectations for PMCF plans and evaluation reports, clinical evaluation and equivalence claims
FDA QMSR
US quality systems
Alignment of US quality system expectations with ISO 13485
21 CFR Part 803
US
Medical device reporting of adverse events and malfunctions
Section 522 studies
US
FDA-ordered post-market surveillance studies for specified devices
ISO 13485
Global
Quality management systems for device organisations
ISO 14971
Global
Risk management across the device lifecycle
ISO 14155
Global
Good clinical practice for clinical investigations of devices in humans
21 CFR Part 11
US
Electronic records and signatures — audit trails, access control, record integrity
IEC 62304
Global
Software lifecycle processes for medical device software
GDPR
EU/UK
Lawful basis, consent, minimisation and subject rights over clinical and outcome data
The practical consequence: any system holding surveillance, clinical follow-up or registry data has to carry audit history, controlled access, versioning and traceable lineage from source to report. These are design constraints, not features added later.
WHERE THE DATA LIVES
The Systems Landscape
Quality And Complaint Systems (eQMS)
Complaints, non-conformances, CAPA, document control. Usually the system of record for anything that becomes a regulatory event.
Clinical And Evidence Systems
EDC for structured studies, ePRO/eCOA for patient-reported outcomes, registry platforms for longitudinal capture.
Device And Telemetry Data
Readings, logs, usage and performance data, held in a manufacturer cloud or on the device itself.
Hospital Systems
EHR, PACS, laboratory and theatre systems, where clinical outcome context actually sits. Access is governed by the hospital, not the manufacturer.
Safety And Vigilance Databases
Adverse event capture and regulatory submission.
Commercial Systems
CRM, field service and training platforms used by representatives and clinical specialists.
Analytics And Reporting
Warehouses and BI layers where evidence is assembled.
The gap between these systems is where most manual effort goes. Device telemetry rarely reaches the clinical record. Registry data rarely reaches the complaint system. Field observations rarely reach evidence teams in structured form.
WHAT'S CHANGING
Five Shifts Reshaping How Device Organisations Work
01
Evidence Expectations Rose, And Stayed Up
The EU regulatory reset raised what counts as sufficient clinical evidence and reduced reliance on equivalence claims to older devices. Programmes that were surveys and literature reviews are now structured data collection efforts.
02
AI-Enabled Devices Changed The Change-Control Question
When a device includes a model that could be updated, the regulatory question shifts from whether this version is safe to how future versions will be governed. Predetermined change control plans turn model updates into a planned pathway rather than a new submission each time.
03
AI Governance Now Sits Alongside Device Regulation
AI systems embedded in regulated devices attract obligations under emerging AI-specific regulation in addition to device law — risk management, data governance, logging, human oversight and transparency.
04
Real-World Evidence Became Regulatory Currency
RWE now supports surveillance, market access and reimbursement rather than just publication. That raises the standard on provenance: capture, transformation, review and reporting all need to be traceable.
05
Connected Devices Moved Collection Into Routine Care
Devices transmitting continuously change surveillance from periodic sampling to something closer to continuous monitoring. The technical problem becomes volume, signal quality and knowing which changes warrant attention.
COMMON FRICTION POINTS
Where The Work Usually Breaks Down
Evidence Data Fragmented Across Tools.
Surveillance inputs in spreadsheets, complaints in the eQMS, PMCF responses in a survey tool, registry data with the sites. Reports rebuilt manually every cycle.
Follow-Up That Depends On People Remembering.
Twelve-month, twenty-four-month and multi-year intervals tracked by calendar reminder. Missed windows surface later as evidence gaps.
Device Data That Never Reaches Clinicians.
Telemetry sitting in a vendor portal while the clinical team works in the EHR.
Programmes Rebuilt Per Product.
A registry configured for one device and one region that cannot be reconfigured for the next, so the second programme costs what the first one did.
Traceability Reconstructed After The Fact.
Provenance assembled at audit time rather than captured as work happens — expensive, and hard to defend.
AI Pilots That Stall At Validation.
A model performs well but nobody can show version history, review decisions or oversight, so it never leaves pilot stage.
BUILD WITH CONFIDENCE
Build Your Next Medical Device AI Platform
Nirmitee Health builds evidence, device-data and clinical software for medical device organisations. Healthcare is the only industry we serve.