Regulation and process planning: a package deal30 October 2015
Validation of the packaging process is an important step in ensuring that the machinery used to form, seal and assemble a sterile barrier system is suitable for the task at hand. Putting a robust test plan in place is therefore essential, as this guarantees that the end user receives a sterile device while mitigating the risk of discovering design failures late in the development process. Dan Burgess, principal packaging engineer at Boston Scientific, explains how to most effectively carry such a plan out.
When it comes to validating the processes associated with sterile barrier packaging, there are a few key things to keep in mind:
- What regulatory requirements do I need to meet?
- What business requirements do I need to meet?
- Are there any aspects of my process that are unique?
Understanding each of these key items will ensure that a robust process is developed and in place for the manufacture of your medical device package.
Regulatory requirements related to the process validation of sterile medical device packaging are primarily going to come from the US FDA code of federal regulations and ISO 11607-2: 'Packaging for terminally sterilised medical devices', part two: 'Validation requirements for forming, sealing and assembly processes'. The standard is applicable to industry and healthcare facilities, and wherever medical devices are packaged and sterilised.
Following is a breakdown of some of the key items in this regulation with which companies must comply.
Performing an installation qualification (IQ) is essentially confirming that equipment does what it is supposed to do. It is similar in nature to what many industries would refer to as a factory acceptance test (FAT). In the case of a heat sealer, for example, if the item is supposed to heat up to 300°F, it needs to be tested and documented to confirm its capability to do so.
The process of performing this testing would include a test protocol, approved prior to test execution, and a corresponding report documenting the objective evidence from testing.
Operational qualifications (OQ) involve defining the operating window for a process, which consists, in this example, of a piece of equipment and the components that interact with it. (Components can be assemblies that are modified or consumed during the process.) Typical inputs for a heat sealer study may include temperature, pressure, time, component, operator and equipment; these should be evaluated in more detail. There are a number of reasons why inputs need deeper evaluation, from a component-focused view and an operator-focused view.
For a component-focused standpoint, it is in many cases desirable to seal multiple sterile barriers on the same piece of sealing equipment without changing the sealer input settings (such as time, temperature and pressure). This is sometimes possible, but all designs to be sealed should be included as inputs in the OQ process study, to capture the relationship between the process inputs and the outputs measured. If materials of the components are the same, using the components with the shortest and longest seals is a good approach.
If the materials of the components are different, each unique material combination should be considered as an input to the OQ study.
Operator variation is often caught during the last step in the validation process, but if it is suspected that operators may have a significant impact on the output of the process, it's good practice to include multiple operators in OQ work, to characterise their levels of influence on the process before optimal process settings are chosen. The key benefit of this is the opportunity to identify early in the process steps that can be taken to minimise their effects.
A good example of this is the effect operators can have on the sealing process when loading flexible barriers into a sealer. Samples can incur misaligned seals or wrinkling in the seal area due to pressure applied to the centre of the component when held in place by the operator.
Similar to the effect that multiple components can have on a process, equipment variations can also result in inconsistencies and mistakes. If there is a desire to seal the same part on multiple sealers, then, all the sealers should be included as inputs to the development process. This is because each piece of equipment, no matter how well machined, will have subtle differences that may affect the output.
Performance qualification (PQ) is typically the last step in the process and is often paid the least attention to, as many think the work performed during IQ and OQ has flushed out all the major issues that could arise during normal manufacturing. While it may be true that the first two major steps in the process address many possible sources of variation, the two that are generally not well understood going into PQ are the effects of lot-to-lot material and operator variability.
In order to understand these process inputs, it is best practice to run multiple lots of products through the line. Three lots is a common number to use, and is called out as such in ISO 11607-2, but if a lot consists of ten units, relatively little process variation related to the packaging is likely to be captured during a run of three lots of ten - 30 total units. A better rule of thumb, therefore, might be to use the number of units that are built in a shift, or an hour, and to choose the quantities to represent the three groups of evaluated units during PQ.
The practice should be applied to the packaging from internal and external perspectives, as the vendor-applied seal and internal manufacture seal are applied using two distinctly different sets of process inputs. It is also important to note that if the right work has been done during previous steps (IQ and OQ), parts produced during this process should represent production grade units. This means that if they meet specification, they could be used for production. This would be especially true for samples made by a converter for a medical device manufacturer.
Section 4.3 of 11607-2 outlines the requirement that test methods must be validated, but it is purposefully vague on the details of such validation - perhaps in order to preserve some flexibility for the many different manufacturers and processes that exist. What these requirements are saying is that companies must do their homework to understand which test methods are appropriate for the specifications being assessed, and must know how much error can be attributed to the method as it relates to the results.
For the validation and sampling plan portions of the requirement, a risk-based approach can be an ideal method. Key challenges for test method validation are destructive tests and attribute outputs. Destructive tests, like peel testing, are harder to validate, as every time the test is executed the test sample is destroyed, making it difficult to isolate the method's repeatability.
The challenges associated with attribute test methods are driven by two things:
- the number of samples required, due to the attribute test method outputs consisting of pass/fail results
- the difficulty an operator might have in determining whether a test sample passes or fails to meet requirements.
The first item can be addressed by finding a way to turn attribute results into variable results. The second item can be addressed through training, using a machine to perform the inspection, or making visual aids to help an inspector with their decision-making process.
What business requirements do I need to meet?
Business needs should be considered when developing a process validation plan, to ensure that the inputs and outputs of the process - like package design and customer usability - are well understood. Best practice is to develop a list of questions related to the inputs and outputs of the process, such as:
- What type of package design am I working with?
- How is the package assembled?
- Is there a certain line speed that needs to be maintained?
This will enable the identification of the items needed to complete validation and potential roadblocks to the completion of projects.
Are there any aspects of my process that are unique?
This question can be challenging to answer, and may depend somewhat on answers derived from the business needs above. Normal packaging processes, working from the inside out, include: sealing of a flexible (such as a pouch or bag) or rigid (such as a thermoformed tray with lid) barrier; applying a label to the barrier; placing the sterile assembly, along with literature such as directions for use (DFU), in a sales carton; labelling the sales carton; placing multiple sales cartons in a shipper; labelling the shipper, and stacking multiple shippers on a pallet.
During this process, there are several points in which unique challenges can occur. Take, for example, the sealing of the sterile barrier. Several quality control questions must be addressed.
First: is there enough room to fit the end of the pouch in the sealer? When sealing a pouch-style barrier, there will typically be a guard in front of the sealer's seal bars to prevent injury to operators. This can mean there is not enough room for the seal area to be in a smooth and flat position prior to sealing. Options for dealing with this are:
- applying an initial seal to make the material lay flat, then a second seal that acts as the sterile seal
- redesigning the component to add appropriate length
- developing and incorporating a tensioning device into the sealing equipment, to make the seal lay flat.
Then: can the operator adjust inputs to the sealing process during normal manufacturing - and how does this affect the validation work I need to do?
This is a favourite question of mine, and in the case of flexible barriers, it affects the processes of creating seals applied by the converter and the end user. Since the end user is the owner of the medical device design, including the packaging, it is ultimately their responsibility to ensure that the entirety of the sterile barrier meets the related regulatory requirements. This includes seals applied by suppliers. It is therefore key to work with suppliers on their process validation activities during the development of sterile packages to ensure they meet these requirements.
Next: back to the question of operator adjustments for sealing process inputs (such as time, temperature and dwell). During the validation process, with the supplier and in house, it should be defined whether operators are allowed to make adjustments. If they are, this must be taken into account during the PQ part of process validation activities.
Typically during PQ, three lots of components are manufactured at the optimal process settings to capture variations in manufacturing. This is an acceptable practice if the process inputs cannot be changed; however, if they can be changed, OEMs should consider running multiple lots at the edge of the allowed settings to capture process variation at those settings, too. The importance of running multiple lots is illustrated by the graph opposite. If process inputs established during the OQ part of process validation are used during manufacturing without OEMs first understanding the effects of lot-to-lot variation at those settings, there is a chance that parts will be created outside the specification limits.
Developing an understanding of ISO 11607 and doing the appropriate work to demonstrate compliance will certainly functionally address the regulatory aspects of process validation for sterile medical device packaging. However, this might not be enough. Careful planning and working knowledge of an OEM's and converter's packaging processes are needed to make sure the right questions are asked and answered regarding business and unique processing needs.
Dan Burgess is a principal packaging engineer at Boston Scientific in St Paul, Minnesota, where he manages the cardiac rhythm management (CRM) packaging engineering team. Burgess has ten years of medical device industry experience, with a focus on technology and test methods, prior to which he worked as an injection mould maker and machinist for eight years.