GenAI in medical device post-market surveillance | HCLTech
Life Sciences and Healthcare

GenAI and its potential for medical device post-market surveillance

Learn how GenAI can transform medical device post-market surveillance by automating data retrieval and analysis, enabling stakeholders to make swift, informed decisions.
 
5 minutes read
Amit Mahadeo Ramning

Author

Amit Mahadeo Ramning
Practice Leader- MedTech
5 minutes read
Share
GenAI and its potential for medical device post-market surveillance

Introduction

Post-market surveillance of medical devices is crucial for monitoring their safety and effectiveness after they are released to the market. Post-market surveillance for medical devices includes collecting, storing, evaluating and assessing incidents arising from their use to continuously monitor the risk-benefit ratio and implement necessary measures to mitigate adverse events.

Post-market surveillance stakeholders

Post-market surveillance involves reporting incidents from stakeholders such as patients, healthcare professionals (including nurses and technicians) and internal personnel like manufacturing and service engineers. While patients and healthcare professionals interact with the exterior parts of medical devices, manufacturing and service engineers are trained and authorized to interact with the interior parts. Therefore, a comprehensive approach requires evaluating incidents reported by both external users and internal engineers.

Evaluation and assessment process

Collaboration among product experts and stakeholders from various functions including R&D, quality assurance and regulatory affairs (QARA), manufacturing and service, is critical for incident evaluation. This step helps all stakeholders understand the nuances of design and process documentation.

Post-market surveillance ecosystem

The below diagram provides a detailed overview of the high-level steps in the post-market surveillance ecosystem. Additionally, the key areas where GenAI can be integrated are marked with the 'AI' symbol.

 

Post-market surveillance ecosystem

GenAI can significantly enhance several key steps in this process, as illustrated below:

Step 1: Complaint intake and workflow triggering

Stakeholders involved: Service engineers

After receiving a complaint or adverse event report, service engineers may be deployed to investigate and resolve the issue. These engineers often refer to product service manuals and may seek technical support from the service, manufacturing and R&D departments. Key documents referenced during this phase include the service manual, product architecture, sub-system architecture, detailed design descriptions and error codes.

GenAI can parse through these documents to:

  • Identify the subsystem, hardware or software associated with the incident 
  • Outline progressive troubleshooting steps to isolate the problem 
  • Determine necessary spare parts to restore regular operation

This allows service engineers to operate more autonomously, reduce their dependency on higher-level technical support and speed up incident resolution.

Step 2: Detailed assessment, evaluation and root cause investigation

Stakeholders involved: Risk engineers

Product literature referenced: Product risk management file

Risk management for medical devices helps manufacturers identify potential hazards and assess the associated risks so that steps can be taken to reduce those risks.

The product risk management file contains the following risk metrics:

  • List of various hazards 
  • Event descriptions 
  • List of hazardous situations 
  • Probability of occurrence of hazardous situations (P1 value) 
  • List of harms 
  • Probability of hazardous situations leading to harm (P2 value) 
  • Severity of harms (S) 
  • Risk control measures (Mitigations)

Risk engineers gather information from the product risk management file to assess potential risks, identify existing protective features and document failure modes.

GenAI can read through the risk management file and assist by:

  • Identifying hazardous situations and their likelihood of occurrence 
  • Assessing the severity and risk of harm to patients or users 
  • Mapping incidents to existing failure modes and risk mitigations 
  • Highlighting gaps that necessitate new mitigations and documentation updates

Additional stakeholders involved: Troubleshooting engineers, system, hardware and software engineers

Product literature referenced:

  1. Service manual: Contains detailed instructions on servicing of the medical devices to assist service engineers. 
  2. Detailed troubleshooting instructions: Contains detailed workflows, instructions, tips and tricks for troubleshooting and identifying the root cause of product/hardware failure. 
  3. Circuit schematics: Contains detailed schematics of an electrical circuit design. 
  4. Sub-system architecture: Contains detailed architecture of the individual sub-systems within the product. 
  5. System-level technical description: Contains a detailed technical description of the overall medical device product.
  6. Software architecture: Contains detailed technical architecture of the medical device product software. 
  7. Detailed design documentation: Contains detailed technical documentation of the electrical circuit schematics.

GenAI's role: It can streamline the investigation process by,

  • Identifying key components related to the incident 
  • Providing detailed troubleshooting guidance and spare parts identification 
  • Facilitating rapid data retrieval from various documentation sources

This integration of GenAI tools enables stakeholders to focus on decision-making and action rather than manual data retrieval, significantly enhancing the efficiency and effectiveness of post-market surveillance efforts.

Conclusion

GenAI has the potential to revolutionize medical device post-market surveillance by automating data retrieval and analysis to help stakeholders make quick but informed decisions. While GenAI is not a replacement for human expertise, it is a valuable tool in supporting stakeholders' capabilities in the surveillance ecosystem.

Share On