Real-World Evidence & Post-Market Surveillance for Medical Wearables: A Practical Guide

Getting regulatory clearance is the starting line, not the finish line. Once your medical wearable is on the market — worn by real patients, generating real data, in real clinical environments — the regulatory obligation shifts from proving it is safe and effective to proving it stays safe and effective. This is post-market surveillance (PMS), and for medical wearable manufacturers, it is the single most under-invested phase of the product lifecycle.

This article provides a practical guide to building a PMS system for medical wearables, covering ISO 13485 and EU MDR requirements, real-world evidence (RWE) collection strategies, vigilance reporting, and the OEM-specific challenge of PMS when the brand owner and the manufacturer are different entities.

1. Why PMS Matters: Market Access Is Not a One-Time Event

Post-market surveillance is not a box-ticking exercise. It is the system that keeps your product on the market. Under EU MDR 2017/745, every medical device manufacturer must maintain a PMS system that is proportionate to the risk class and appropriate for the device type. The PMS system feeds into the technical documentation, the clinical evaluation report (CER), the periodic safety update report (PSUR), and the post-market clinical follow-up (PMCF) plan — all of which are living documents that must be updated throughout the device’s lifetime.

The consequences of an inadequate PMS system are severe. A Notified Body audit that finds gaps in PMS can result in certificate suspension — which means your product cannot be sold in the EU until the gaps are closed. For a wearable OEM serving multiple brand partners across multiple markets, a single PMS failure can cascade into a multi-product, multi-market crisis.

Beyond compliance, a well-designed PMS system is a competitive asset. The real-world data your devices generate — usage patterns, performance drift, adverse event signals, user feedback — is the most valuable input to your next-generation product definition. Companies that treat PMS as a regulatory burden miss the opportunity to turn post-market data into product leadership.

2. ISO 13485 and EU MDR: The PMS Regulatory Framework

ISO 13485:2016, the quality management system standard for medical devices, establishes PMS as a mandatory process under Clause 8.2.1 (Feedback) and Clause 8.5.1 (Improvement). The standard requires manufacturers to establish a documented procedure for collecting, analyzing, and acting on post-market data. Critically, ISO 13485 does not prescribe the content of the PMS plan — it requires that the process exists and is effective.

EU MDR 2017/745 goes much further. Article 83 requires a PMS system that is an integral part of the manufacturer’s quality management system. Article 84 mandates a PMS plan for each device. Article 85 requires a PMS report (for Class I devices) or a periodic safety update report (PSUR, for Class IIa and above). Article 86 mandates post-market clinical follow-up (PMCF). Annex III spells out the minimum content of the PMS plan:

• A proactive and systematic process to collect data on device quality, performance, and safety
• Methods to assess the collected data, including statistical methods
• Indicators and threshold values for the risk-benefit analysis
• Methods to investigate complaints and analyze market experience
• Tools to identify trends and detect signals of new or increased risks
• Communication protocols with competent authorities, Notified Bodies, and economic operators
• References to procedures for CAPA, vigilance, and field safety corrective actions
• A schedule for the PSUR and PMCF report

For wearable medical devices, the PMS plan must address device-specific risks: sensor degradation over time, algorithm performance drift, cybersecurity vulnerabilities discovered post-market, and use-related hazards that only emerge in real-world settings (e.g., skin irritation from prolonged wear, inaccurate readings during exercise, or user misinterpretation of alerts).

3. PMS Data Sources for Medical Wearables

A robust PMS system draws from multiple data streams. For medical wearables, the available data sources are richer than for traditional medical devices because the device itself generates continuous, high-frequency data:

Complaints and adverse event reports: The most direct signal of product problems. Every complaint must be logged, triaged, investigated, and trended. For wearable devices, common complaint categories include skin reactions, battery failure, connectivity issues, and inaccurate readings. Each complaint must be assessed for reportability under vigilance requirements.

Device telemetry and usage data: Wearables generate a continuous stream of data on device performance, user behavior, and algorithm outputs. This is the most underutilized PMS data source. Device telemetry can reveal: battery degradation patterns, sensor calibration drift, firmware version distribution, and real-world algorithm performance (e.g., the proportion of ECG recordings that are unclassifiable). For connected wearables, over-the-air firmware update success rates and app crash logs are also critical PMS inputs.

Return and repair data: Every returned device is a PMS goldmine. A systematic teardown and root cause analysis of returned units can identify failure modes that were not anticipated during design verification. For wearable OEMs, this data is essential for supplier quality management — if a specific batch of PPG sensors shows elevated failure rates, the OEM must trace the issue to the component supplier and implement corrective action.

Clinical literature and registry data: PMS is not limited to your own data. Published clinical studies, adverse event databases (e.g., FDA MAUDE, EU EUDAMED), and disease registries can reveal signals relevant to your device. If a competitor’s wearable with similar PPG technology is associated with a specific adverse event, that is a PMS signal for your device.

Customer feedback and social media: App store reviews, social media posts, and customer support transcripts are unstructured but valuable PMS data. Natural language processing tools can surface emerging issues — such as a pattern of users reporting “battery swelling” — before they appear in formal complaint channels.

4. Real-World Evidence (RWE): Three Collection Pathways

Real-world evidence is clinical evidence derived from the analysis of real-world data — data collected outside the controlled environment of a clinical trial. For medical wearables, RWE serves three purposes: supporting regulatory submissions (e.g., PMCF), informing product improvement, and generating commercial evidence for payers, providers, and brand partners.

Three RWE collection pathways are particularly relevant for wearable devices:

Pathway 1: Device-Generated Telemetry Data

This is the most accessible and scalable RWE source for connected wearables. Every PPG waveform, ECG strip, SpO₂ reading, and activity log is a data point. With proper consent management and data governance, device telemetry can support: algorithm performance validation (comparing real-world sensitivity/specificity to clinical trial results), usage pattern analysis (how often and for how long do users actually wear the device), and population-level health insights (e.g., seasonal variation in resting heart rate). The key challenge is data quality — real-world data is noisier than clinical trial data, and the absence of a reference standard for most readings means you must design validation sub-studies.

Pathway 2: Registry Studies

A registry is a prospective observational study that collects defined data on all patients using a specific device or diagnosed with a specific condition. For wearable manufacturers, registries are particularly powerful for: demonstrating long-term safety and performance (e.g., two-year skin tolerance data), validating clinical outcomes (e.g., does continuous SpO₂ monitoring reduce hospital readmission rates for COPD patients), and supporting health economic arguments (e.g., cost-effectiveness of wearable-based remote monitoring vs. in-clinic follow-up). Registries require significant upfront investment in protocol design, site recruitment, and data management, but they generate the highest-quality RWE.

Pathway 3: Claims and EHR Data

For wearables integrated into healthcare delivery — such as remote patient monitoring platforms reimbursed by payers — claims data and electronic health records provide a rich RWE source. This data can answer questions about: device utilization in real-world practice, association between device use and clinical outcomes (e.g., hospital readmission rates), and healthcare resource utilization. The limitation is that claims data is not collected for research purposes — confounding, coding errors, and missing data are significant challenges.

5. Vigilance: What to Report, When, and to Whom

Vigilance is the most time-sensitive component of PMS. A failure to report a reportable event within the required timeframe is a regulatory violation, regardless of whether the event caused patient harm. Both FDA and EU MDR have specific reporting requirements, and the timelines are short.

EU MDR Vigilance (Articles 87-90):

• Serious public health threat: Report immediately, no later than 2 days after becoming aware.
• Death or unanticipated serious deterioration in health: Report immediately, no later than 10 days after becoming aware.
• Other serious incidents: Report no later than 15 days after becoming aware.
• Trend reporting: Any statistically significant increase in the frequency or severity of non-serious incidents or expected side effects must be reported as a trend. There is no fixed threshold — the manufacturer must define its own statistical method and threshold for trend detection in the PMS plan.

FDA MDR (21 CFR Part 803):

• Death or serious injury: Report within 30 days of becoming aware.
• Malfunction that could cause death or serious injury if it recurred: Report within 30 days.
• 5-day report: Events requiring remedial action to prevent an unreasonable risk of substantial harm, or events for which FDA has made a written request.

Wearable-specific vigilance triggers: Skin burns from battery thermal runaway, severe skin reactions to sensor materials, algorithm false negatives leading to missed clinical events, cybersecurity breaches exposing patient data, and firmware update failures bricking devices are all reportable events for wearable medical devices. For more on cybersecurity requirements, see our guide on wearable cybersecurity under FDA guidance.

6. Trend Reporting and CAPA: Closing the Loop

Detecting a signal is only half the job. The other half is acting on it. The CAPA (Corrective and Preventive Action) process is the mechanism by which PMS signals translate into design improvements, manufacturing changes, labeling updates, or field safety corrective actions.

Trend analysis methodology: The PMS plan must specify the statistical method for trend detection. Common approaches include: control charts (Shewhart, CUSUM) for monitoring complaint rates, Poisson regression for adverse event rates, and time-series analysis for device performance metrics. The threshold for triggering a CAPA investigation should be defined prospectively — for example, “a complaint rate exceeding the upper 95% confidence limit of the historical baseline for two consecutive months triggers a CAPA.”

CAPA workflow for wearable devices:

1. Signal detection: A trend is identified through PMS data analysis — e.g., an increase in ECG algorithm unclassifiable rates in a specific firmware version.
2. Investigation: Root cause analysis involving engineering, clinical, and quality teams. For algorithm issues, this may require re-analysis of raw sensor data from affected devices.
3. Corrective action: The fix — e.g., a firmware update that improves signal processing, a manufacturing change to improve sensor consistency, or a labeling update to clarify usage instructions.
4. Verification of effectiveness: Post-implementation monitoring to confirm the corrective action resolved the issue — e.g., unclassifiable rates return to baseline within 30 days of firmware update deployment.
5. Preventive action: System-level changes to prevent recurrence — e.g., adding the failure mode to the design FMEA, updating the supplier quality agreement, or enhancing the pre-release testing protocol.

7. OEM PMS: Who Does What When the Brand and Manufacturer Are Different

In the OEM model, PMS responsibilities are split between the legal manufacturer (who holds the regulatory clearance) and the contract manufacturer or brand partner. The split is rarely clean, and the most common PMS failures in OEM relationships arise from ambiguity about who is responsible for what.

Typical PMS responsibility split:

PMS Activity	Legal Manufacturer (Brand Partner)	OEM (Contract Manufacturer)
Complaint handling and triage	Primary — receives, logs, investigates	Supports — provides device analysis, failure investigation
Vigilance reporting	Primary — assesses reportability, files reports	Must notify manufacturer within 24 hours of becoming aware
Trend analysis	Primary — owns the statistical methodology	Provides production and service data
CAPA implementation	Owns the CAPA process	Executes manufacturing and design changes
PSUR / PMCF reports	Primary — authors and submits	Provides data inputs
Device telemetry analysis	Primary if the connected platform is theirs	May own the embedded firmware and sensor data

Critical contract provisions for OEM PMS: The quality agreement between the OEM and the brand partner must explicitly address: (1) the maximum time the OEM has to notify the brand partner of a potential reportable event (24 hours is standard), (2) the OEM’s obligation to provide device-level data for PSUR and PMCF reports, (3) the OEM’s responsibility for field safety corrective actions (including recall logistics), and (4) data ownership and access rights after the partnership ends. For more on quality agreements between OEMs and brand partners, see our guide on medical wearable OEM quality agreement checklists.

8. Data-Driven Product Iteration: Using RWE to Define the Next Generation

The most sophisticated wearable OEMs treat PMS not as a cost center but as a product strategy function. The real-world data your devices generate contains the answers to the most important product questions: Which features are actually used? Where does performance degrade first? What do users complain about most? What clinical outcomes are being achieved?

Practical examples of RWE-driven product iteration:

• Battery life: Device telemetry reveals that 80% of users charge their device every 3 days, not the 7-day interval claimed in marketing. The next-generation product targets a 5-day battery life with a larger cell, and the marketing claim is adjusted to “up to 5 days” with a footnote on real-world usage.
• Sensor placement: Complaint data shows a consistent pattern of inaccurate heart rate readings during running. Root cause analysis identifies motion artifact from wrist-based PPG at high cadence. The next-generation product adds an accelerometer-based motion compensation algorithm, and the user manual is updated to recommend the device be worn higher on the forearm during exercise.
• Algorithm thresholds: PMS data shows the AFib detection algorithm has a higher false-positive rate in users under 40 than in the clinical trial population (which skewed older). The PCCP is invoked to retrain the algorithm on a more age-diverse dataset, and the updated model is deployed via a firmware update without a new 510(k). For more on algorithm regulatory pathways, see our guide on SaMD and AI algorithm submission pathways.

9. Downloadable Resources: PMS Plan Template + Adverse Event Decision Flowchart

To help you build or strengthen your PMS system, we have prepared two resources:

1. PMS Plan Template: A structured document covering data sources, statistical methods, threshold values, CAPA procedures, and reporting schedules — aligned with EU MDR Annex III and ISO 13485 requirements.
2. Adverse Event Decision Flowchart: A one-page visual guide to assess whether an event is reportable under EU MDR and FDA MDR, with timelines and escalation paths.

Contact our team at jine@xdunmedical.com to request these documents.

Conclusion

Post-market surveillance is where the regulatory rubber meets the clinical road. For medical wearable manufacturers, PMS is simultaneously the most operationally demanding and the most strategically valuable phase of the product lifecycle. The PMS system must be designed before the first device ships, resourced throughout the product lifetime, and continuously improved based on the data it generates. For OEMs serving multiple brand partners, the PMS responsibility split must be defined in the quality agreement with zero ambiguity — because when a reportable event occurs, the clock starts ticking, and “we thought you were handling it” is not a defense.

The wearable manufacturers who invest in PMS as a strategic capability — not just a regulatory obligation — will be the ones who turn post-market data into product leadership, regulatory confidence, and commercial differentiation. For more on the full regulatory pathway for medical wearables, see our guides on ISO 10993 biocompatibility testing and SaMD FDA submission pathways.