Video: Quality Management according to ISO 13485 | Duration: 2708s | Summary: Quality Management according to ISO 13485 | Chapters: Welcome and Introduction (90.595005s), NSF Services Overview (158.31500499999999s), QMS Process Approach (269.605005s), Process Models (354.13998s), ISO-MDR Relationship (555.025005s), Global QMS Requirements (732.745005s), Regulatory Change Management (872.475005s), System Integration Strategy (1025.115005s), Process Integration Strategies (1209.085005s), Practical Implementation Examples (1387.145005s), Economic Operators Coordination (1596.015005s), Implementation Roadmap (1826.855005s), Closing and Resources (2082.105005s)
Transcript for "Quality Management according to ISO 13485":
So hello, and welcome to this keynote on ISO thirty four eighty five and, beyond. The ISO is the central, for a strong medical device quality management system, but it's not the complete regulatory picture. Medical device organizations also need to meet legal requirements in, different markets. Let me first introduce, myself. My name is Eamon Ganini. I'm a senior consultant in medical devices at NSF with a background in biomedical engineering and hands on experience in regulatory affairs, quality management system. I have worked with several device companies, medical device companies on quality management systems, audits, technical documentation, different regulatory projects. I'm also a qualified lead auditor with more than, ten years of audit experience. So before moving into the... Today's topic, I would like to briefly introduce NSF. NSF has been working to improve human and planet health. This mission is supported by specialists from many different, scientific and technical disciplines around the world. For the life sciences, industry, this means that we combine technical knowledge, regulatory experience, and practical implementation, for supporting. Our objective is not only to identify compliance gaps, but also to help organization find solutions, network, and daily operations. This mission is supported by a broad international presence. NSF has offices, laboratories, and experts in more than 25 countries. This international network allows us to support both global organizations and companies entering real markets. Within this global network, NSF provides a broad range of services across the life sciences industry. So NSF supports life science organization from the identification of an issue through to practical implementations. Our services include consulting, auditing, remediation, outsourcing, training, and as well clinical services. We work across medical devices and IVDs, also pharmaceutical and biotechnology products, combination products, and as well as clinical research. Depending on the situation, support may range from a focused gap assessment or an internal audit to the design and implementation of complete regulatory and quality processes. So with that short introduction to NSF, let us now, start with our main topic for today. A quality management system has to support the complete product life cycle. This starts with regulatory strategy and product development. It continues then through technical documentation. We need to define suppliers. We need to define our production and distribution. It also includes after going into the market complaints, post market, surveillance and improvement information. So from one area must reach them the next area. For example, campaign data may change the risk file, the technical evaluation, or the labeling. It can also influence own suppliers, which we have. So this is the route, we will take today. First, we look at the process approach because processes are the basic building block of a quality management system. Then we look at the ISO as a regulated core and the additional requirements from MDR and other markets. After that after that, we will, look at recent developments. In the second half, we will move more into the integration of several management systems. Finally, then we will discuss the governance and practical implementation of those management systems. So let us begin with the process, approach. Some of you may know the total model. Some, may not know. I would like to emphasize on this model. The model is a simple tool for understanding and controlling, process. Let us use complaint handling, for example. The input for a complaint, process is not just an email from the customer. We need several informations. We need the device itself. We need the lot number, the serial number, event date, some information which are necessary so that we can work on the process. The output is more than a closed complaint. It may include interfaces to other, decisions we have to make, customer response, maybe a copper or some supplier actions. It can also lead to adjusting risk management files. The other questions around the total are equally important. So to understand a complete process, we need we need to understand who is responsible for for the process, which people are working in the process, what kind of training is necessary. So... And furthermore, which system is used? Are we using a tool to work on our complaint handling? So what kind of escalation time do we have? How do we know whether the process works? Useful indicators could be overview of complaints, reporting compliance, repeated failure modes, and investigation time. This model help us find, gaps at the interfaces. Once a process is defined, then we we move and... To the plan do check at, method. So the PDCA is a simple logic behind the living quality management system. We have it we have it overall for the complete thirteen four eight five, but we have it also in different, processes itself. So in the planning phase, we identify requirements. We identify risk and objectives. We define our scope. In the doing phase, we perform the controlled activities and create records. In the check phase, we review through through orders. We review through KPIs. We review through complaints and post market data, the output of our doing. And in the act phase, we make correction, implement probably CAPAs or update risks, documentation, or probably also the product itself. For example, if you... If complaint trending shows an increase in one failure mode, the check phase should not end with a chart showing us, okay, we have here in trend. The company needs then to investigate the course. The results may be different. It can be a supplier cover. It can be a design change or an update to production controls. After the implementation, the company checks the effectiveness, whether this complaint will actually decreases. So we have this continuous cycle with a plan, do check, and act. So the thirteen four eight five gives us a foundation... Foundation. In the next slide, we will look at the core of this ISO, the relationship with the MDR and the global regulatory overlays. The value of the standard is that it controls the complete life cycle from development, purchasing, production, delivery, servicing, and feedback have to work together. Risk influenced and the level of control. Supply of a critical implant, for example, or a critical component than an implant will need more qualification and more monitoring than a supplier of standard, purchasing equipment like screws, A process where the results cannot be fully verified later may need a validation, but this has to be also considered for suppliers. If there are processes which need validation, the manufacturing to, monitor also those processes. So documented evidence is also important if a process output, for example, design change or complaint assessment cannot be reconstructed and the organization cannot show that that process was controlled. Records and documentation is key in our field. So to summarize, so thirteen four eight five is not a product file. It's a operating system that allows the organization to create and maintain reliable product evidence over time. So then we are coming to the connection between the ISO and MDR. They are very close connected. Right? They do have different targets. The ISO expands how to build and maintain the management system. The MDR is the law, defines legal responsibilities and specific outcomes, such as clinical evaluation plan, report, same for post market surveillance and vigilance. The example complaint handling as I'm continuing with that one. The ISO gives us tools such as responsibility, the procedure, the tools we are using, the records we need to, provide data analysis, interfaces to other, processes. The MDR tells us that the manufacturer has to fulfill this reporting deadline, for example. So we are getting more information from the MDR, which we need to put into our quality management system. So this is why thirteen four eight five certificate is valuable. It does not automatically prove full MDI compliance. The certification covers the management system within its scope. Product specific and market specific obligations still have to be demonstrated to show that we are fulfilling the law and the regulations. So this slide shows, what the MDR adds to the operational QMS. Take regulatory strategies first, classification, conformity assessment, and change assessment must be controlled activities. The clinical evaluation is not finished when the initial report is, approved. The PMS, post market surveillance, the PMCF information may require an update. Same data may also affect risk management or the instruction for use. So we are in continuous plan to check x cycle, and this is what the quality man... Management systems, provide. So the main point is that the MDR expands the output that the quality management system has to deliver reliable. These are must also connect across the complete, product life cycle. The same principle applies globally, but each jurisdiction has a different, regulatory, overlay. So the standards, as mentioned, gives companies a common QMS language, but each market keep its own legal requirement. In The EU, we have the MDR, which already mentioned specific topics. In The United States, we have the quality management system regulation, now incorporates more the thirteen four eight five by reference. This creates more alignment, but FDA specific reporting, records, inspection, and enforcement still apply. So, also, inspectors from the FDA is looking more into another approach doing... Through this quality management system regulation. Canada, for example, require the medical device single audit, program to be able, to market devices in Canada. The audit is based on the ISO, but it includes country specific regulatory task. UK has its own framework, including registration and The UK responsible person. The internal process structure can be shared, but each market has its own entry condition. One complaint process may be used globally for each of those market, but the reporting decision and the deadline depend on the market. So recent changes show why regulatory change management is a real quality management system process. The FDA QMSR became effective on the February 2. Companies need to review the internal audit criteria, regular... Regulatory reference, training, and inspection readiness because FDA inspects... Inspector did also change their approach. They can now look into the audit... Internal audit report. They can look more into the management review and make based on that conclusions. The MDCD guidance for PMS gives practical guidance on different post market surveillance topics. It emphasized data sources, analysis, and the links between PMS, risk management, clinical evaluation, and and CAPA. Standards monitoring is also important. It changed standard, may affect test methods, acceptance criteria, risk controls, and technical documentation. The new revised ISO 9,001, which was published this, September or this month, it strengthens topics such as leadership, quality culture, accountability, opportunities, and alignment with other management, system standards. For med tech organizations, these ideas are useful when the QMS needs to work across several functions and risk. So for every change, the practical question is which product, process, record, systems, and also which people are affected by this change. So the ISO remains still the core quality management standard for medical device organizations. However, the operating environment around the quality management system is becoming broader. The company may also need to manage information security, environmental performance, which is more relevant nowadays, regarding tenders in some countries, which are requesting environmental quality management systems, which has been applied in house. Occupational health and safety are also may be covered, or they're responsible to use, AI. These topics are increasingly connecting with normal quality processes. So the first step is to understand the complete management system landscape. The next step is to identify which management mechanisms can be shared. Issue is not the number of standards we have. The issue starts when every standards create separate systems, and we are building more and more processes, more and more documentation. So those separate management systems can create unnecessary complexity. Imagine that, document control has one procedure for quality, another for information security, and a filter for environmental management. So employees then need to decide which procedure applies. The same supplier could be audited several times by a different team, which normally is not so, effective. Kappa could sit in different systems with different owners, and this makes the processes unnecessary complex, and it also reduces the effective output. So the next section moves more from this problem to a method how we can, tackle, several management systems. So the risk is different for these, different management systems, but the logic behind this is similar. As we have for thirteen four eight five, we focus more on product product quality and patient safety. The ISO 9,001 look more broadly at the enterprise quality management system and performance of processes. The ISO 27,001 focuses on confidentiality, integrity, and availability of information, and, 42,001 focuses on responsible AI governance. The logic behind this is similar. The technical controls are different. For example, a sterilization validation cannot be replaced by an information security assessment, but the company can still use common mechanisms for roles, competences, supplier oversight, the planning of audits, management reviews, which can be aligned also when you are having different management systems in house. So as I mentioned, one internal audit program can cover a supplier management process. The audit team can include quality expert, information security expert, for example. The process is reviewed once, but the specialist criteria means separate and so that we can fulfill, or complete oversight related to all standards we are trying to apply. So this leads them to the design principle of the integrated quality management system. So we shall begin with one real process. This is a business process, then we add controls that the process needs. Let us use change control as an example. The business process is the same. From start to finish, someone proposes a change. The change is assessed, approved, implemented, and checked. That that that is the bottom layer we are looking at. The next layer contains controls that apply to almost every change. So we need an owner, a reason for the change, effective documents, responsibilities, due dates, training, and evidence that the change was completed. Furthermore, the next layer, we are at the domain controls. The design change may require risk management updates, verification, maybe an updated clinical evaluation or a regulatory assessment. In comparison, a software change may require testing or some cybersecurity assessment. A change to AI tool may require new performance checks and oversight and controls, or a supplier change may require requalification and a quality agreement update. Once clear workflow with common controls and specialist models edit when they are relevant. So... But not every process should be integrated in the same way. Document control is a good example for full integration. The basic questions are the same. If you are using the thirteen four eight five, you're using the other standards. Who approves the document? Which version is the current one? Who needs access, and how long is the record kept? One company wide process normally works well. But supplier management is a better handling with modules. The workflow is common, but the criteria differ. A component supplier needs product and process control. Cloud provider needs security and continuity controls. For AI, we may also need transparency about data, model changes, and the monitoring. If you look on the right path, the clinical evaluation is different. It requires specific competences, evidence, and legal outputs. It should remain a dedicated process even it shares... If it shares document control, kappa management review as a interway... Interface with the wider system. It's it's tackling a specific MDR part so that we are putting that in a specific domain. The decision over is practical. Same workflow and evidence logic means integrate, same workflow, but different criterias, means use modules, different hazards or legal outcome means keep the process domain specific and not combine it. So so this is the target architecture that results from the previous rule I'm... Was talking about. At the bottom, there are the business processes, development, purchasing, production, distribution, and post market activities. These are the process where the real work happens. Above them are regulatory and specialist modules. For example, the development process may use an MDR module for clinical evidence. A QMS r module views specific expectations and a security model for connected software. Shared controls sits across the organization, Document control, training, supplier management, CAPA, and all this should not be rebuilt for every module. There's enough if we have one module. Leadership then sets priorities, provides sources, and review overall performance. The architecture is not necessary one... For one software platform. There's one there's one agreed logic for processes, ownership, decisions, and evidence. Good. Now we are moving now to... From from this concept to practical, examples. I already started with the topic supplier management. So the life cycle of a supplier management is always similar. Define the need, which we need to purchase, which we need to acquire, access the criticality, qualify the supplier, agree requirements, monitor the performance, manage changes, and we evaluate the supplier. But the evidence depends on the service. For a critical component supplier, we may review process capability, validation, traceability, and, change notification. For a calibration service, we may focus on accreditation, competence, scope, and certify, the quality or the quality of the qualification. For a cloud provider, we may focus on access control, incident notification, backups, business continuity. And for an AI provider, we may also need information about k... Training data, performance monitoring, and changes to the AI model. The workflow and process owner can be common. The qualification checklist, acceptance criteria for those different suppliers, and the monitoring indicator should match the risk. Yeah. And a second example is, for example, the management review, where many data sources must lead to one set of management results. Management review is where the integrated system should support the real outcome. Consider a recurring supplier problem. Quality KPI show an increase in incoming defects. Complaints show the same failure mood in the field. There are couples which are overdue, and the supplier has already announced a process change. Looking at each item separately may hide the full risk and the full scope of, this issue. An integrated review brings the information together. So it can be a classical management review, or it can be monitoring boards, which has been, done together with the management. After that, the management can then decide whether additional resources are needed, whether the supplier should be escalated, whether a design alternative is required, or whether regulatory action is, necessary. Specialist topics may still need a separate technical, review. For example, the cybersecurity team or may analyze a vulnerability in detail. The final risk, resource needs, and action plan should then feed onto this management review. The purpose is not one very long meeting. The purpose is just one clear decision process. So, also, I want to, highlight the processes related to other economic operators. These operators may not need a full thirteen four eight five satellite system. However, they still need controlled ways of working, and this process approach we were talking at the beginning is, key for that. For example, an importer must verify key conformity elements before placing a device on the market. Checklist can make these consistent. Storage and transport conditions need defined responsibilities and records. Complaint and recalls need a register and a clear escalation rule to the manufacturer. So here, the manufacturer and the distributor and importer needs to work together on a common process. For example, if a customer reports a serious problem, the distributor must know who to contact, what information to collect, and how quickly to to escalate, those information. The scope of the process should match the legal role, the value of the QMS principles, and the same, clear responsibility, consistent execution, evidence, and the escalation. So, complaint handling shows how, one shared intake can support different responsibilities. In this case, we are looking at, the distributor. So imagine that a hospital reports a device stopped working during use, and the distributor receives the first call. The distributor should not simply write forward, unstructured email. It has... The information has to be collected. Right? Which device is it? Is it... So which which serial number or batch number? Which event date? What kind of patient impact, did we have? Where's the current device location? And what kind of other evidence are available? The case is sent to us and, and the complaint, a nonconformity device, a potential serious incidence, or what kind of, case do we have here? The manufacturer receives the information and performs the technical and regular... Regulatory assessment. The authorized representative, the distributor, may need to be informed, and an authority report may be required. So the information has to come back then to the customer. The investigation can lead them to CAPA, to supplier action, to an risk update. And here, the distributor and the manufacturer needs to work together. And, therefore, a clear process also, going over, to to entities, to, to to different economic operators, is key. And, for example, this case related to a complaint handling and feedback from the market. Good. In the last section, we look at governance, the implementation sequence, and practical starting points if we have, integrated or to go into an integrated quality management system. The executive management remains still responsible for the resources, for priorities, and the strategic decisions made. QMS governance or the QMR coordinate the overall system, but should not became the owner of every process. I think this is clear. The process owner is responsible for end to end performance. For example, can be that the purchasing department may own the supplier process, while quality defines medical device criteria. IT department defines security criteria, and regulatory affairs department defines maybe market specific requirements, which are relevant if products are purchased. So domain experts therefore support the process on the IT or data owners responsible for system access, access, data integrity, digital controls. Regulatory experts define escalation criteria and market obligation. So this model prevents two common problems, quality owning everything or no one owning the interfaces. So the implementation sequence is important. If we start by matching SOPs, we may create a cleaner document set without fixing the operating model. Firstly, we indentify all relevant, requirements. Then after that, we map this actual end to end process, including the interfaces to each other. The third one is to build a matrix that links requirements to controls and evidence. This shows duplications and, gaps. After that, we assign one process owner and the necessary domain's builds. Coordinate assurance through forming audits, monitoring KPIs, looking for, preventive actions, and, feeding those in data analysis and in the end to end management review. Only then, the organization should digitalize workflows or consolidate procedures. For example, the training. Before combining training SOPs, define one competence model, one ownership structure, and one source of training evidence. A document can then follow the agreed process. The roadmap can begin with a few practical and, yeah, low risk actions. The first step should be practical and manageable. You should create a requirement inventory and select one, two shared processes. Let's start, simple with document control, training, supplier management, change control. Those are good candidates, because duplication is usually visible. In the second phase, we build the integration matrix and harmonize forms or records. For example, one supplier form can contain a common section, an an additional section for product quality, security, or AI. In the third phase, we test the design. One, our integrated audit, combined selected KPI reporting. So the pilot will show whether rules, criteria, and evidence are really clear. We should start small, learn from the pilot, and then expand. This is usually more effective than announcing a complete QMS transformation, at once. This is not only related to different standards. It can also be if we're having we're having, a manufacturer who is not doing only medical devices, but also different branches of products which are related to other areas. So so let me close with the main message. The IRO thirty four eighty five is the regulated QMS backbone. We have regulatory overlays, which must be translated into these processes, records, and evidence. So one QMS does not mean one procedure for every requirement, so we need to integrate common workflows, but preserve the risk specific controls or specific domains and requirement. Use process ownership, monologue controls, and coordinated assurance. And in the end, start with an asset... Architecture and, governance. So we as an NSF can support organizations in moving from the initial assessment to a QMS that works in a daily operations. The objective is a practical quality management system that people can use and understand and manage. It should not be an overburden just to create documents. So thank you for your time and attention. If you would like to discuss our quality management architecture or implementation needs, please, contact us. Furthermore, I would also like to highlight the NSF Medical Device Summit in Hamburg in November. The event offers, practical discussions with industry experts on current regulatory quality and technology topics in, med tech industry. You can find the complete agenda and registration details by scanning the QR code. If you would like to explore, further, regulatory and quality topics, NSF offers additional keynotes with practical guidance and real world examples. The upcoming sessions are available also through web QR code or the link shown on the slide. Thank you very much for your time and attention. I hope this webinar has provided some useful, ideas for developing a practical quality management that goes beyond formal compliance. If you have further questions, feel free to contact us. Thank you.