Medical Device Link.

Originally Published IVD Technology June 2004

REGULATIONS & STANDARDS

Dan Olivier is president of Certified Software Solutions Inc. (San Diego). Marta Chase is vice president of quality assurance at Bayer Diagnostics (Tarrytown, NY). The authors can be reached HERE and HERE, respectively.

Design controls have been mandated by the FDA quality system regulation (QSR) since 1998, and by ISO 9001 since 1994.^1,2 Design controls are also discussed in ISO 13485:2003, which defines quality management systems for medical device manufacturers.³ The FDA and ISO definitions of design control are similar. For example, this similarity is reflected in the documentation requirements for design control (see Table I). Since FDA’s QSR and the ISO standards are essentially the same, a single design-control process can satisfy both requirements.

Although design control is a well-established requirement, best practices for applying design controls to IVDs continue to evolve. In general, applying design controls to IVDs is more complex than other medical devices. This is because IVDs require integrating additional disciplines. For most medical devices, design controls involve integrating hardware engineering (i.e., electrical and mechanical) and software engineering. 

For IVDs, in addition to these two disciplines, reagent design must be included. A flow diagram shows the complexities of this parallel development of reagents, instruments, and software in IVDs (see Figure 1). This article will address practices for applying design controls to IVDs and include lessons learned on optimizing this process.

Design and Development Plan 

The design and development plan specifies the activities that will support the IVD design efforts as well as the responsibilities for those activities. As a project nears completion, the design and development plan should be updated further so that it is consistent with the actual activities conducted during the development process. This is especially important since FDA uses the plan as the basis for inspecting the design history file (DHF). 

For IVDs, the design and development plan also coordinates the schedule and resources for parallel development of the reagents, instruments, and analysis software, integration and validation testing, and preparing the manufacturing facilities. Defining these tasks in the plan is the easy part. The difficulties arise in the parallel implementation of these tasks.

Design Input: Requirements 

Table I. Documentation requirements for IVD design control (click to enlarge).

The design input is a collection of the inputs that define an IVD. While the inputs may come from many sources, they should be formally captured in specifications such as a system or product requirements document. The final definition of the design input should include not only the functional requirements but also the safety requirements that are identified in a safety risk analysis. Such a risk analysis should address the essential safety requirements that are defined in Annex I of the IVD Directive.⁴ These requirements are the minimum safety requirements that ensure product safety.

Getting the system requirements correctly defined is an essential first step for any device development process. As the reagents, instruments, and software depend on each other, the requirements for IVDs are closely integrated. Any change to one of these components may affect the other two. 

An optimum design could theoretically isolate the dependencies among the reagents, instruments, and software. However, in practice, this can be achieved to only a limited degree. To ensure that each of these IVD components works together, early and frequent prototypes are essential. 

In addition, the system requirements must consider manufacturability. A design that cannot be produced economically and repeatedly will not be acceptable for market release.

Safety Risk Management

FDA’s QSR and ISO 13485 require safety risk management, which is an important part of the design of IVDs.⁵ A thorough and effective risk management process can anticipate and prevent possible error conditions and can ensure that if device failures occur, they do not result in adverse patient safety consequences. The risk management process includes identifying potential scenarios that could directly or indirectly result in patient injuries (e.g., false-positive or false-negative results). The process also specifies requirements that prevent the occurrence of risk or reduce risk to an acceptable level. 

ISO 14971, the ISO standard for risk management in medical devices, provides a framework for the risk management process. This standard describes a total product life-cycle process, from planning through the maintenance phases. The risk management process begins with identifying potential hazards at the design input phase. Annex B of ISO 14971 provides a list of several factors that should be considered as part of an IVD safety risk analysis, such as carry-over effects, specimen identification errors, and problems related to the stability of specimens. 

Risk analysis is only the beginning of the risk management process. Appropriate safety requirements must be identified for potential safety hazards as well as implemented and verified. Typical safety requirements for IVDs include the use of controls, quality assurance checks, and clear labeling of the intended use of the device. A thorough and effective risk management process is an essential activity for anticipating and preventing possible error conditions and ensuring that if device failures occur, the results of these failures do not result in adverse patient safety consequences. 

Design Output: Specifications 

Figure 1. A flow diagram shows the parallel development of reagents, instruments, and software in IVDs.

At the design output phase, the significant differences in the design and development process for IVDs can be seen. As stated above, IVD design involves the reagents, the instruments that will analyze the treated samples, and the software that performs the final analysis of diagnostic results. This phase generates the instrument design specifications, the software architecture design specifications, and the reagent formulation requirements. 

Although some independent design may be allowed at the design output stage, the different design disciplines should continue to coordinate their efforts. It is important to know early whether slight changes in the reagent formulation, instrument sensitivity, or software algorithms may affect the other components. To the extent that the system requirements are correct and well defined, independent design by the different disciplines may proceed with a high probability of success.

Design Review 

Design review includes reviewing specifications produced throughout the IVD design process, as well as a series of milestone reviews required by the design and development plan. Specification reviews identify design errors prior to implementation or integration. Milestone reviews ensure that acceptable progress has been realized before progressing to the next stage, and communicate design architecture and intent to the entire design team. Records of the design review meetings are retained as part of the DHF, and the minutes document the reviews and any decisions.

Design reviews are an excellent way to identify errors early in the process for all types of design activities. For IVD projects, the benefits of effective design reviews are heightened as the costs of design failure are quite high (the design of the instrument, reagents, and software may all be adversely effected by a single design error).

Design Output: Implementation 

Table II. Challenges to the IVD design process (click to enlarge).

At the implementation phase, the design of the IVD is completed: the reagent formulation is finalized, the hardware is constructed, and the software code is written. As with the other stages of the design process, it is important not only to complete the activity but also to define qualitative criteria for acceptability to minimize errors. 

Since the implementation of the design of the product is finalized at this phase, it is also important to have established practices to minimize defects. Such practices include techniques that prevent likely error conditions, such as effective reviews, automated tools that detect difficult-to-locate errors, and adherence to guidelines that have been designed to prevent errors.

Design Verification and Validation

Design verification includes the test and review activities that are conducted throughout the IVD design process to ensure the design is correct. Such verification activities include component tests for the instrument, unit tests for software, and boundary studies for testing the accuracy and effectiveness of reagents. Thorough design verification provides increased confidence in the successful execution of validation testing and in the final product’s quality. A traceability matrix is developed to document the relationship between the testing and the requirements documents.

Design validation is the final series of testing that is conducted to support a product release decision. Validation testing uses the actual IVD equipment in a simulated real-world environment to exercise safety risk controls and demonstrate that user requirements have been satisfied. 

Problems that are commonly found in IVDs include errors experienced at the limits of performance specifications, errors caused by unexpected operator entry sequences or variability in reagent performance, and instrument reliability problems. Examining these errors or similar types of errors that are found during the development process should guide the focus and depth of validation activities. Once the validation testing has been completed and outstanding defects are reviewed such that no critical safety problems exist, a product release decision can be made.

Design Transfer and Changes

Design transfer involves transferring the engineering design to the production team. In order to reduce the time to market, design transfer activities must be planned and conducted throughout the IVD design process. Up-front coordination with the manufacturing department can minimize schedule delays. Doing so ensures that the appropriate equipment has been procured, process validation activities have been conducted, and adequate material and personnel resources are available to begin production runs.

Once the design is completed, the product life cycle has begun. If the device is successful, numerous changes (e.g., corrections, enhancements) will be requested. Product maintenance procedures must be established to support design controls for new product releases. The change control process must include procedures for evaluating, implementing, and validating every new release.

Lessons Learned for IVDs

Because of its complexity, the IVD design and development process must be carefully managed to prevent significant delays. An effective way to prevent costly delays is to analyze historical projects and lessons learned by other manufacturers (see Table III). If such lessons can become the basis for refining existing development processes, significant problems can be avoided, and productivity of the processes can be enhanced. 

Optimizing the Design

Table III. Lessons learned about the IVD design process (click to enlarge).

The fact that IVDs require interactions between three subsystems (i.e., reagents, instruments, and software) offers challenges as well as opportunities for design optimization. Each of the hardware and software subsystems can be designed in a generic manner to support the use of different assays. For example, an instrument can be designed with generic functions; therefore, by using a controlled script, the instrument can process many different operation sequences to support multiple assays. In the same way, the analysis software can be modified to support different assays by changing the algorithms for analyzing the results from any instrument run. The reagent formulation processes can also be defined in a generic manner to facilitate the reuse of predefined raw materials and generic formulation steps to produce new reagents as needed for new assays. 

Designing these subsystems such that generic features can be programmed using script files for future assays provides flexibility for addressing a wide range of diagnostic tests. However, providing this capability requires significant up-front design planning and a thorough validation process. Investing in a robust, up-front design for an IVD can provide the ability to support future changes and process future assays.

Conclusion

The design and development of IVDs present unique challenges. If not properly managed and integrated, the complexities involved in designing IVDs can become overwhelming. The design controls as defined by FDA and ISO provide a practical process for addressing these complexities, a process which not only satisfies regulatory requirements but also addresses business needs.

In addition to the sequence of activities that are performed, a good design process must also include a qualitative analysis of the design decisions made throughout the process, an area where many design processes fall short. This article has offered some strategies on how to address the challenges of IVD design, lessons learned on how to avoid problems experienced by other manufacturers, and a discussion on how a robust and flexible design for each subsystem of an IVD can yield significant reductions in maintenance costs and a significant capability to produce future assays.

References

1. “Quality System Regulation,” Code of Federal Regulations, 21 CFR Part 20.

2. “Quality Management Systems: Requirements,” ISO 9001:2000 (Geneva: International Organization for Standardization).

3. “Medical Devices: Quality Management Systems; Requirements for Regulatory Purposes,” ISO 13485 (Geneva: International Organization for Standardization, 2003).

4. “Directive 98/79/EC of the European Parliament and of the Council of 27 October 1998 on In Vitro Diagnostic Medical Devices,” Official Journal of the European Communities L331 (1998): 1–37.

5. “Medical Devices: Risk Management; Application of Risk Management to Medical Devices,” ISO 14971 (Geneva: International Organization for Standardization, 2000). 

6. “Clinical Laboratory Improvement Amendments (CLIA),” Code of Federal Regulations, 42 CFR Part 493.

7. “The Health Insurance Portability and Accountability Act (HIPAA) of 1996,” Code of Federal Regulations, 45 CFR Parts 160 and 164. 

Copyright ©2004 IVD Technology