Data-driven FMEA approach for hazard identification and risk evaluation in digital health
This method uses a structured risk assessment technique called Failure Modes and Effects Analysis (FMEA) to find and rank potential risks in digital health systems. It looks at different types of healthcare data—like clinical records, operational data, and patient-reported information—and different data formats such as text, images, and videos. Experts and research were used to score each risk based on how serious it is, how likely it is to happen, and how easy it is to detect, resulting in a priority score for each. The method helps improve the safety and reliability of digital health systems, especially when using new technologies like AI and blockchain.
At a glance
Use when
Assessing safety and reliability risks in digital health systems, especially those using AI, blockchain, or multi-modal data; during design, implementation, or audit phases of health IT systems
Avoid when
Rapid, resource-limited risk screening is needed without expert involvement; when real-time risk monitoring is required rather than prospective analysis
Inputs
Healthcare data systems involving clinical, operational, and patient-reported data in text, image, tabular, audio, or video formats; expert input; literature on digital health risks
Outputs
Prioritized list of failure modes with Risk Priority Numbers (RPNs); identified high-risk areas; recommendations for improving data integrity, security, and interoperability
How it works
The data-driven FMEA approach systematically identifies and evaluates failure modes in digital health systems across three healthcare data categories (clinical, operational, patient-reported) and five data modalities (text, image, tabular, audio, video). Each failure mode is assessed using severity, occurrence, and detectability criteria derived from expert consultation and literature review, enabling calculation of a Risk Priority Number (RPN) for quantitative risk prioritization. The method supports hazard identification and risk mitigation in complex digital health environments, particularly those integrating AI and blockchain technologies, by highlighting critical vulnerabilities such as unauthorized access, data corruption, transmission errors, and privacy breaches.
- Project
- INSAFEDARE
- Funding
- Horizon Europe
- Project status
- Ongoing
- HTA domains
- Safety, Organisational aspects, Aspects Beyond HTA
- Assumptions
- Expert judgments accurately reflect real-world risk levels; failure modes can be adequately scored on severity, occurrence, and detectability; digital health systems are sufficiently documented to enable systematic analysis
- Strengths
- Comprehensive coverage across data types and modalities; quantitative prioritization using RPN; expert- and literature-informed assessment; supports proactive risk management in emerging tech environments
- Limitations
- Dependence on expert subjectivity; may not capture rare or unforeseen failure modes; limited validation in real-time operational settings; RPN scoring may oversimplify complex risk interactions
- Also known as
- Data-driven FMEA for digital health, INSAFEDARE FMEA method
Questions this answers
- › What are the potential failure modes in digital health data systems?
- › How can risks in clinical, operational, and patient-reported data be systematically evaluated?
- › Which data modalities (e.g., text, image, video) are most vulnerable to specific hazards?
- › How can severity, occurrence, and detectability of risks be quantified in digital health?
- › What are the highest-priority risks threatening patient safety and data integrity?
- › How can digital health systems be made more secure and resilient against emerging threats?
References & sources
Related methods
Similar by meaning
- Framework for Assessment of the Use of FAIR Principles for Health Data in Digital Health Devices Regulation
- Framework for Digital Health Equity
- SPIFD (Structured Process to Identify Fit-For-Purpose Data)
- FIFARMA MCDA framework for healthcare decision-making
- Multiple Criteria Decision Analysis for Health Care Decision Making (MCDA)
Beta record. Generated from the primary source via AI extraction and independent audit, pending final human review.