by Sarah Downing
Medical device manufacturers must take a variety of requirements into consideration as they strive to provide safe and reliable products. Stemming from customer needs, industry regulations, and technical specifications, requirements can quickly seem to proliferate to an unmanageable degree.
Controlling for variation on too many requirements unnecessarily raises business costs and adds time during the design process. Selecting the wrong items to control can lead to production issues like scrap, rework, and returns, with the end result being unhappy customers, recalls, or even a consent decree.
In deciding what to control, product risk management offers some direction, as it assesses failure modes and mitigations. However, this method does not always account for sources of variation following risk mitigation controls, nor does it include user wants and needs, aside from safety.
A more complete solution for creating compliant product and process designs focuses on identifying the essential requirements, key product characteristics, and critical to quality items for a medical device. Once you have developed a meaningful list of workable requirements, you can use it to promote robust designs and for variation management.
Creating a list of meaningful requirements begins with an understanding of the concepts of key product characteristics (KPCs), key process parameters (KPPs), and critical to quality items (CTQs):
Identifying KPCs, KPPs, and CTQs helps you focus on the critical or essential requirements and supports a risk-based approach for product safety, efficacy, and process efficiency. The ultimate purpose and benefit is to create robust product and process designs that are insensitive to random variation and to reduce the amount of variation in production. Variation is a key contributor to waste and driver of factory costs (3).
The traditional concept for product monitoring and acceptance is that any product or component produced within the identified specifications or engineering tolerance is identified as good product (see Figure 1). In many applications, this concept is appropriate.
However, in other instances where the noticeable consistent performance of a product is deemed highly important to the customer, the theory of Taguchi’s loss function may be applied (see Figure 2).
In this scenario the loss is identified as a function of deviation from the target, or identified performance condition; therefore, more “on-target” product and process performances mean less loss (4).
Complete elimination of variation is not practical. However, you can understand and manage the largest causes of variability by adopting a systematic approach through proper identification and control of KPCs and KPPs.
To define the KPCs, KPPs, and CTQs for a medical device, you must identify and stratify the needs and wants of the customer based on level of criticality. First, identify all customer and product requirements and then use tools to filter what is essential (see Figure 3). The outcome is a reduced set of requirements and specifications that you can easily manage for product and process design and production controls.
Treat the process of defining KPCs and KPPs as iterative and continue to consider your list throughout the entire life of a product. Aside from using KPCs in managing process variation, you can also incorporate them into quality planning to complement other tools and techniques, including:
Early definition is imperative and should be performed in conjunction with the assessment for critical requirements.
Since KPCs are the link between functional customer needs and physical realizations on the final product, their identification should come from top-level customer requirements traced to lower-level design documentation. You can ultimately achieve this through identification of the critical requirements.
The flow diagram in Figure 4 and the instructions that follow identify an approach to defining what’s critical and key.
Step1: Define and understand the voice of the customer.
Inputs should, at a minimum, come from the intent of the product, voice of customer responses, and use cases. Include all groups of customers, such as end users, production, and organizations that modify the product or regulatory bodies.
To better understand the level of importance of each customer requirement and to help segregate users’ needs from wants, use concepts presented in the Kano model (5).
Step 2: Filter for critical requirements.
Once you have fully defined customer requirements, you can use filtering activities to begin the critical item identification process.
Quality function deployment (QFD) and critical to X (CTX) flowdown are methods that convert customer needs into specific features. During critical to X (CTX) flowdown, or the flagging process, the critical requirements are identified. Critical to X is a way to flag a requirement that is most important to customer satisfaction and business success, where X is a defined attribute such as safety, function, reliability, testability, manufacturability, serviceability, etc.
Once all categories of X are identified, assess each requirement per the following logic:
Would a failure to meet the requirement result in a failure to meet X?
Requirement: The customer needs a product that is sterile.
- Would a failure in sterility result in a failure to be safe?
- Would a failure in sterility result in a failure to sell the product (stopping sales) in the regulated environment?
- Would a failure in sterility result in a failure of the product to function?
If the answer is yes, flag the requirement as CTX based on the function of the device.
Note that you can use the tools presented in this section at different points in the process.
Step 3: Identify critically associated system- and component-level requirements.
Once you have tagged customer requirements as CTX, you can use the quality function deployment (QFD) model to identify the critically associated system and lower-level product requirements.
Quality function deployment is a four-phased process that takes customer needs and translates them into lower-level requirements. In this approach, you will use only the relationship matrix portion of the house of quality to identify the critically tracing requirements.
Review each CTX-flagged, or critical, customer requirement for tracing to its related system-level product requirement. Then analyze the tracing link and rank it per the relationship. See Figure 5.
After ranking, flag for criticality all product requirements identified with a strong relationship. Although important for design control purposes, requirements with medium and weak relationships do not need further consideration for this portion of the assessment.
Use the same process for lower-level component and feature requirements, where appropriate. This will provide a complete list of all critical items: requirements, features, and specifications.
You can view requirements from the perspective of customer needs, product needs, and component needs. Depending on the product, analyzing KPCs in a multilayered process may be appropriate as well. Therefore, the KPCs may be initially identified at the unit of use (assembly) level and then decomposed to lower levels during the development process.
Identification of KPCs at an assembly/system level will help define KPPs during the assembly/manufacturing process:
1. Identify initial list of assembly-level KPCs.
Assembly-level key product characteristics (AKPC) are defined as key characteristics at the unit-of-use level and are based on what the user needs from an operational perspective.
This may mean that an AKPC is identified at the assembled-unit level but the specific feature(s) creating the AKPC may exist at the component level. Therefore, the AKPC is only a placeholder during the development process.
The practical importance of the AKPC is for decision documentation, tracing purposes, and identifying the essential or critical requirements. This will help raise a flag if changes are made to KPCs during the development process.
Identifying KPCs at the unit-of-use level is also important because there may not be a feature at the component level that would provide the information needed for variation management.
A review of the critical requirements should identify grouping to form an initial list of assembly-/system-level KPCs. Use affinity diagrams to help organize groupings.
Once you have grouped the critical requirements, you can conduct additional filtering. Because the ultimate goal is to identify KPCs subject to variation during manufacturing, the requirements should additionally be reviewed for relation to product use features. Some requirements may be identified as CTX, but because they are not specifically defining a product feature they may be filtered from consideration for KPC identification. These types of requirements, however, could still be critical to the device, as in a safety requirement that is not related to an actual feature. Consider, for example, the requirement that “product shall be free from visible particulates.”
At this point, you should have an initial list of assembly-level KPCs.
2. Define lower-level and final KPCs.
The (A)KPCs have been defined and are the key output variables of the assembly process. However, this is not the only level (assembly vs. component vs. process) at which KPCs may be monitored, or where variation will occur. In some cases the component, or raw material, may require KPC identification for supplier monitoring purposes.
Where appropriate, decompose the (A)KPCs to identify lower-level KPCs for design features using the following steps:
a) Identify any design outputs that may contribute to the (A)KPCs.
b) Evaluate and rank KPC design outputs for potential variation.
At this point in the process, the number of design outputs associated with the (A)KPCs is too large to gain the benefits of KPC monitoring. Evaluate the list of design outputs for the potential variation that you anticipate will occur.
Figure 6 presents an example of a ranking system.
Consider several factors:
i. Amount of variation relative to feature tolerance
ii. Importance of individual feature on AKPC
iii. Historical in-field or production data
d) Calculate the risk and determine the KPCs.
You can calculate a number for the overall risk to satisfying the (A)KPC by multiplying the “potential for variation” and the “effect on the (A)KPC.” When determining KPCs always consider the feasibility of the inspection method and alternatives.
3. Identify KPCs for design for manufacture and assembly (DFMA).
Identifying KPCs that will ensure that the product is designed such that it can be easily and efficiently manufactured and assembled reduces product costs, increases the ability to meet specifications, and reduces the time to market.
To appropriately control the process, understanding the relationship between process variables, or parameters, and the product results is vital. Once you have identified the KPCs, follow these steps to define KPPs for manufacturing and/or assembly process relationships to the KPC:
1. 1. Define the key points in the process.
Evaluate the KPCs against a process flow diagram to help identify the points in the process that create, affect, or offer an opportunity to inspect a KPC.
2. Identify key parameters.
Identify and evaluate all applicable process parameters associated with the key points in the process. Process parameters are any settings throughout the process that may affect the product.
The first part of the evaluation should consider the effects of variation on the KPC as well as the feasibility to monitor and control the parameter.
3. Set the key parameters.
Once you have identified and selected all key process parameters, perform experimentation to optimize the process settings based on the output performance of the KPC. Use these settings for process monitoring to ensure consistent product output.
The CTX flowdown process identifies all of the critical requirements that must be realized on the product, verified for quality of design, and used during production for lot acceptance for verification of quality of conformance.
Now you have your list of KPCs and KPPs, but what do you do about those items that are critical to the quality or safety of the design but are not at risk of variation within the specification and where traditional monitoring concepts may be more appropriate? This is where you can use the CTX requirements that you previously defined but that did not turn into KPCs. These become your CTQs.
As you use this approach to identify critical requirements, understanding the major differences between key product characteristics and critical to quality items is helpful, as the two are not always synonymous.
Critical items are features or characteristics of a product that require a high level of attention to ensure that all requirements are achieved. For these items, special planning is undertaken, such as supplier involvement, process capability studies, and reliability verification (6). All requirements and specifications identified as critical through the CTX process are considered “critical items.”
Following from the definition of CTQ items presented earlier, the determining factor for identifying a critical to quality (or safety) item is the risk associated with the failure, or a requirement that has been created to mitigate a high-risk failure. Therefore, CTQ items can also be identified through requirements tracing to the product risk management documentation. In-process monitoring of CTQ items provides assurance that the feature conforms to specifications.
As stated previously, a KPC is a feature whose variation, within specification, significantly influences the user-defined performance of the product. Therefore, KPCs can be, but do not have to be, safety related.
Therefore, as illustrated in Figure 7, KPCs and critical to quality/safety items can exist and overlap as follows:
However, for a medical device, Figure 8 may be a more appropriate model. This will be determined by the needs of the user for the product.
One of the purposes of identifying KPCs and KPPs is to ensure optimized product and process performance through monitoring activities. Depending on the part level (system vs. component) at which the KPC resides, the monitoring activities may take place at any point in the manufacturing process and should be documented in the overall quality plan.
Methods of process monitoring should include establishing control charts and control plans once you have identified KPCs and KPPs. If you have identified a process variable or parameter as a KPP, implement statistical process control charting, depending on the necessity and frequency of monitoring.
In the control plan, document all KPC/KPP controls that are being performed, whether they are performed in-house, at a supplier site, or anywhere else. Also include inspections performed on the critical items for lot acceptance.
To ensure effective use of KPCs and KPPs, evaluate their appropriateness on an ongoing basis. Base your evaluation on continual review of information gathered from the following sources:
If the reviews determine that the KPCs or KPPs were incorrectly identified, you will need to launch a documented activity to determine the more appropriate KPCs or KPPs and update all applicable product documentation.
Sarah Downing is an ASQ Certified Quality Engineer with Baxter Healthcare Corporation, Medical Products Division. She received a bachelor’s degree in mechanical engineering and a minor in biomedical engineering from the Colorado School of Mines.
(0) Member Reviews