Mahzad Pakzad The difference is in the level and depth of precision, reliability, "Control over Variation, Quality, Sterilization", class I is a lot more forgiving than class III!

Mahzad Pakzad The class III are usually products that have DIRECT IMPACT on health of a patient and can cause serious injury, and normally the process steps, the equipment, the materials purchased, the tooling,... everything is deeply analyzed and controlled.

Class one may have some effect in limited way on patient, therefore only limited steps of the process or materials are analyzed and/or controlled.

For example an instrument used to check sample of urine or blood are not necessary designed and developed with materials and processes that WOULD NEVER ALLOW CONTAMINATION INSIDE OUT (processed in vacuum and/or clean environment), they just are anti-rust/ resist contamination to a level, probably coated and have Exterior sterilization; versus an implant (Class III) has to be manufactured using materials and processes that maintain its purity, inside out!

Julie Omohundro I think the only significant difference is in the level of the regulatory scrutiny.

Mahzad, I think you are mixing up the process with the validation and the device with the process. I think validation is more about the process than the device, although the two are not unrelated. There are Class II devices that require sterilization; there are Class III devices that don't require sterilization. There are implants that are not Class III devices, but of course they still require sterilization. If a device requires sterilization, then the sterilization validation should be the same, regardless of whether the device is Class II or III. If it doesn't require sterilization, then there is no sterilization process to validate, but that isn't dependent on classification.

Mahzad Pakzad The title in discussion is not asking about sterilization only, it is asking about VALIDATION; Validation involves process design and control as well product design control;

You may not get a successful validation result, if you cannot "Provide Evidence That Your Process Can deliver The Expected Quality, Consistently", and that takes Providing Evidence that Process Is Controlled Every step that impacts reliability and/ or Consistency of quality, Companies design TESTS TO PROVE the product robustness regardless of variation between its highest and lowest possible limits.

But again this type of validation is important for products that their failure to comply with design spec would harm patients, mostly Class III, some class II, Class I process/product quality validations are a lot more forgiving!

Aura Sa The difference is the accuracy of the equipment.

Julie Omohundro Mahzad, I think sterilization was offered as just one example of a manufacturing process that requires validation, and used to show why validation is not determined by the device classification, not with the thought that the question was limited to sterilization only.

I am not sure what you mean by "more forgiving," other than patients are less likely to get hurt if the process fails. But again, that really depends more on the nature of the manufacturing process and the consequence of its failure, than the classification of the device. They are highly related, since device classification also reflects the consequences of device failure, but I will still go with it is the nature of the process that drives the validation, not the device classification.

It is also useful to consider that validation is not done solely with a mind to patient harm, but also to business harm. Customers aren't going to be happy with a poorly manufactured Class I device any more than they would be happy with a poorly manufactured Class II or III, just because it has a lower regulatory classification. The MAIN purpose of manufacturing process validation is to assure that your products are manufactured with competitive reliability. If your product is the "low-cost leader," then maybe your customers will tolerate a bit less reliability from it than competitor devices, but only a bit less, and at the same price, not at all.

Julie Omohundro Aura, I would think the accuracy of the equipment would reflect the specified tolerances, not the device classification? And that tolerances are not based on device classification, but on the amount of variability in the process that testing has shown can be tolerated before the process fails, i.e., does not produce a device that meets performance standards.

Dan O'Leary When people tell me the class of the device, my next question is where? The classes in the US, EU, and Canada don’t align.

I infer that you are in the US and therefore the question refers to QSR requirements. The answer would be similar for ISO 13485:2003.

In QSR, there are many validations, but none of them are called “manufacturing validation”. I infer you mean process validation under 820.75. There are some very specific cases which require process validation (cannot fully verify) and very specific requirements for the result of the process validation (high degree of assurance). Notice however, that the class of the device is not one of the factors. You must perform the same type of process validation with the same rigor regardless of the device class.

In contrast, it is worth noting that design validation in 820.30(g) does not apply to most class 1 devices.

Karen Boyd, ASQ CQA I would also assume that "manufacturing validation" refers to processes undertaken during manufacturing of devices. Pay particular attention to processes that affect the integrity / FFF of the finished products and special processes (such as welding, heat treatment, passivation, etc.). Regardless of device classification, aim for consistency, reliability, and safety in design / development validation.

Jakob Nielsen Hi everybody its process validation for the injection moulding process...

Michael Chellson, RAC Jacob, there is no difference in the process validation requirements for different classes of devices. However, with that said, the potential for risk of the process must be considered when creating a process validation plan.

During Risk Management activities, the process hazards, harms, and risk to the functionality and safety of the finished device should be identified FIRST, (PFMEA) and the validation plan & protocols created to ensure that the identified risks have process controls in place to mitigate those risk elements that could have significant impact on the safety or effectiveness of the device.

Aura Sa Sorry, but when I talk about accuracy, I mean class I, II, III and IV in medical equipment.

Mahzad Pakzad Class III injection molded parts normally have higher control over Materials' quality specifications and their reliability / consistency, Mold design/maintenance, Process Control, and Environment that process is at, and in some cases they require a special coating.

Each specific part may have its own specifications and criteria, depending on design and application!

Dan O'Leary In the US, the risk classes primarily determine the path to bring them to market. With an exception for most Class I devices the required QMS is the same. (The exception is in design controls.) The manufacturing portion does not change by device class.

Process validation is the required method when the process output, the product, is not fully verified. The idea is that a process should produce only conforming product. If you verify each product, then you sort out the nonconforming. If you don’t fully verify, then you must control the process input parameters to provide a high degree of assurance of only conforming product.

These production requirements apply equally to all device classes.

Mahzad Pakzad QUOTE from FDA: "(g) Design validation. Each manufacturer shall establish and maintain procedures for validating the device design. Design validation shall be performed under defined operating conditions on initial production units, lots, or batches, or their equivalents. Design validation shall ensure that devices conform to defined user needs and intended uses and shall include testing of production units under actual or simulated use conditions. Design validation shall include software validation and risk analysis, where appropriate. The results of the design validation, including identification of the design, method(s), the date, and the individual(s) performing the validation, shall be documented in the DHF. " Source: http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/cfrsearch.cfm?fr=820.30

Design validation of injection molded parts depend on their processes, and vise versa process control of injection molded parts depend on the design!

If your part design quality expectation and higher/lower risk expectations vary, then you will have a variation in validation, and that should be obvious, because your product and process design specifications and requirement would be different, and validation expects you to check and validate all influencing factors,from raw materials to process steps, to tools, environment, and controls in place!

--> The only thing that is the same for validation of class I, II and class III, is the Philosophy that all critical criteria must be identified, tested, validated, , ... but ACTUAL product design criteria (effectiveness, quality, reliability,..), process design (effectiveness, quality, reliability,..), and procedures' thoroughness vary!

Mahzad Pakzad Class three usually Have higher different Impact on Patient's health and Various Critical To Quality items for both products and processes that must be monitored and controlled during the manufacturing, higher risk for class three is the difference and that call for more thorough validation design/plan and acceptance. Quote:
* "What does it mean for FDA to "classify" a medical device?

FDA classifies medical devices based on the risks associated with the device. Devices are classified into one of three categories—Class I, Class II, and Class III.

Class I devices are deemed to be low risk and are therefore subject to the least regulatory controls. For example, dental floss is classified as Class I device.

Class II devices are higher risk devices than Class I and require greater regulatory controls to provide reasonable assurance of the device’s safety and effectiveness. For example, condoms are classified as Class II devices.

Class III devices are generally the highest risk devices and are therefore subject to the highest level of regulatory control. Class III devices must typically be approved by FDA before they are marketed. For example, replacement heart valves are classified as Class III devices. " Source: http://www.fda.gov/AboutFDA/Transparency/Basics/ucm194438.htm

Dan O'Leary It is really important to remember that the risk class of the device is not important in QSR, with one exception. Design control does not apply for most Class 1 devices. The risk class determines the requirement to bring the product to market, (510(k), PMA, PDP, etc.) not the requirements for design and production. When it does apply (Class 2, class 3, device with software, etc.) the requirements do not differ by risk class.

The requirement in 820.75(a) is to validate processes, when required, with a high degree of assurance. To change the definition of high degree of assurance based on the device class is, in my opinion, an invitation for a warning letter.

The device class is not relevant to ensure that validated processes produce only conforming product.

marc henn Agree with all of the previous comments, however, have a question.. For some Class 1 devices, when looking at the FDA codes, some Class 1's are GMP (part 820) exempt with the exception of vigilance and record keeping as stated per the classification. During FDA audits, in my experience, due to this type of code/stated requirements, validation was not an area audited. Thoughts?

Dan O'Leary The classes are defined by Act of Congress in the Food, Drug, and Cosmetics Act. In some sense, the use of 3 classes is a coarse tool. Consequently, FDA has made distinctions in some of the classes. Some Class 1 devices are exempt from nearly all portions of QSR. An example is the manual toothbrush that is not sold sterile. Consequently these devices do have to implement process validation in 820.75. There are also exemptions in the UDI rule for devices that are exempt from (most of QSR).

The issue is which controls are necessary to assure the manufacturer markets only conforming, safe, and effective devices. A control either applies or doesn’t, as in the case of process validation or design validation.

marc henn Is 820.75 still applicable in a classificaiton code like the below? If 820.75 was applicable wouldn't it be listed? See below, copied directly from a classification code.

The device is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to the limitations in 872.9. If the device is not labeled or otherwise represented as sterile, it is exempt from the current good manufacturing practice requirements of the quality system regulation in part 820 of this chapter, with the exception of 820.180, with respect to general requirements concerning records, and 820.198, with respect to complaint files.

Dan O'Leary 820.75 is not applicable for this device. The only applicable sections of Part 820 are 820.180 and 820.198. Other parts and sections, such as Part 803 or Part 806, continue to apply.

Julie Omohundro Mahzad, the FDA definition you quoted is for design validation. Design validation and manufacturing process validation are two completely different things.

Dan O'Leary Joanne – My point is that there is no difference in the level of control related to the class of the device. Either process control is required or it is not. When process validation is required it must provide a high degree of assurance that the process produces only conforming material. The degree of assurance does not change by device class.

There is no such thing as a Class I, Class II, or Class III process validation. The classes are defined by law to prescribe the pre-market submission requirements (PMA, 510(k), PDP, etc.) and other controls. Device class is not an element of process validation.

Auke Poutsma Dear Dan et al,
The classification of Medical Devices is based on 1) Rislk classification and 2) on the clinical application and human risk exposure (temporary, interventional etc.). Based on the risk and complexity of the devices, the proof of safety, performance and efficacy will increase. The processes it self needs to comply with the regulations as defined by the Authorities. It is evident that Class 1 with low risk doens/t need the same package of proof as a class 3 product. It is obvious that highn risk devices require more extensive testing, proof and longer timelines for market approval. The validation of porocesses is required for all products.

Dan O'Leary You say the validation of processes is required for all products. I trust you mean that when a process meets the requirement for process validation then validation is required regardless of the device class.

Bringing a product through the pre-market submission process depends on the product class, but once successful the control in QSR applies equally regardless of class.

Dan O'Leary Joanne -- Just to make the issue clear, General controls apply to all device classes. Class 2 and Class 3 device require general controls and other controls. In some sense the controls are cumulative, with Class 1 having the least controls.

General controls include QSR, which is part of my point that device class is not relevant to process validation.

Martin Lunn Simple answer is no.

process/manufacturing validation "provides a high degree of assurance that a specfic process, method or system will consistently produce a result meeting pre-determined acceptance criteria" MHRA orange guide 2014

z) Validation means confirmation by examination and provision of objective evidence that the particular requirements for a specific intended use can be consistently fulfilled. 21 CFR 820.3

Don't have the ISo definition to hand but it is along similar lines.

depending on classification you may have differing levels of control e.g. Cp=1.67 or Cp=2 confidence/ reliability 95%/95% or 99%/99%

Martin Lunn I can't find the specific text at the moment but there is something along the lines of "only processes where the output cannot be fully verified must be validated" otherwise known as spcial processes such as sterilisation , welding etc where to determin if it is good or bad you destroy your product.

The other place validation comes into its own is to prevent you as the manufacturer doing 100% inspection of every feature. Inspection is a non value added cost so you wouldf want to minimise it. If you validate and show you have a process in control then you reduce your inspection and therefore costs. Why wouldn't you want to do this?

Julie Omohundro Martin, yes, I was taught that, when it comes to validating processes (rather than design), "validation is the alternative to verification." I think in practice validation is so often logistically and financially preferable to verification, people tend to start thinking of it as being required when it isn't. Just like people sometimes think ISO 13485 is required for CE Marking, when it isn't, but it virtually never makes any sense to pursue any other alternative.

In my experience you will have a hard time defending to any regulatory body why your process is not validated. Just inspecting/testing product is no longer sufficient. You need to prove that your process is capable of ensuring the specification criteria is consistently met. Unless you have a validated test method with a low error rate and a solid Cp/CpK on the critical to quality specifications, good luck defending a non-validated process no matter the device classification.

Martin Lunn I beg to differ here. A regulatory body is there to ensure regulations are followed and validation is not mandatory. it is the manufacturers responsibility to ensure that the product meets the requirement and how they do it is up to them. It is easy to defend a non-validated process if the controls and verification are there and in most manual or operator dependent operations verification is all that can be done

you do not need to prove your process is capable. you do not need a good Cp or Cpk if you can verify that what leaves the factory meets specification. You may have a lot of scrap and waste but that is of no concern apart from your bottomm line profit. if its a lot then investing in a validated process may save you money.

As an example look at making a cup of coffee. how do you validate that you have turned the kettle on? We can validate that the kettle will heat water to a set temperature before turning off but yu cannot validate someone turning it on. it can be verified by by putting thermometer into the coffee when its made (or tasting it). similarly putting milk or cream into it.. Now you could go out and buy yourself a vending machine with alarms and nbells and whistles that will make you a consistent cup of coffee or buy a kettle and rely on yourself puttin water in it, turning it on and letting it boil. Slightly cheaper to buy the kettle. I coud put an extra verification step in by testing the heat of the water by a thermometer which I could calibrate and then validate that it always takes the temperature in the same place or i cantouch the side of he kettle and if its hot its probably hot enough for my drink.

Julie Omohundro Martin, I am not sure what you mean by validating someone turning it on, nor why you couldn't, nor why you would want to.

To me, manufacturing process validation confirms that the manufacturing process will reliably produce specified results. By "the manufacturing process," I mean the entire process followed to produce the product, not an individual step in the process. In your example, the process specifies that the kettle will be turned on. Whether or not someone actually turns the kettle on is conceptually irrelevant. If they don't turn it on in spite of the fact that this step was specified in your process (i.e., written procedures), but you reliably get the specified results, you are still good to go. In this example, you probably won't get the specified results, so you can't validate the process, but the point is the results are what counts, not the individual steps. On the other hand, if you write good instructions and hire someone to turn the kettle on, and train them to turn the kettle on, then you should get the specified results, and you have validated the process that includes turning the kettle on.

Another example would be a visual inspection of the product for manufacturing flaws. If this step is skipped, the overall manufacturing process can still produce the specified results if the processes upstream are sufficiently robust that there are never any manufacturing flaws to be found upon visual inspection.

You hire people, you train them, you purchase equipment. You give the people you hired and trained a set of written procedures, you turn 'em loose on the floor, and then....what happens, happens, and product comes out the other side.

As long as you can demonstrate that the product that comes out the other side will reliably meet specifications once you turn 'em loose on the floor, it doesn't matter if no one you hired had the qualifications you said they should have, nor if your training didn't teach them anything you said they needed to be taught, nor if a single person follows the process in your written procedures, nor whether the equipment you purchased actually met your purchasing criteria. All that matters is whether the product will reliably meet specifications when the process is implemented, regardless of whether it is being implemented the way you wanted it to be implemented. As long as the product reliably meets specifications, there is no need to take any action to "fix" the process by making them do it the way you wanted it done.

In practice, for a number of reasons, usually good ones, manufacturers will choose to validate many individual steps in the manufacturing process, in addition to the manufacturing process itself, but the fact that you can't validate an individual step does not mean you can't validate the manufacturing process that includes the step, and, if the overall manufacturing process can be validated, it is not necessary to validate the individual steps.

IMO.

All that said, you can definitely validate the step of turning the kettle on. You write a procedure that includes the step, "turn the kettle on." You train someone on the procedure. You determine acceptance criteria (e.g., "turns the kettle on 99.9% of the time") and then you run the process a whole bunch of times. If the results show that the kettle will be reliably turned on at least 99.9% of the time, the process is validated. However, the process that is validated is not the process of turning on of the kettle, but the hiring, training, and giving of instructions to the person who is supposed to turn on the kettle.

Now everybody tell me where I'm wrong, please. I'm so not the validation expert.

Martin Lunn Julie,

IMO

a pocess that is operator dependent cannot be validated

Where you say your training process and wrioting of instruction is validated isn't IMO strictly true. you have verified that the instruction works you have verified the training on that intructuiion has worked but to validate you would need to train lots of employees on lots of instructions to get to a validated process. but even then would you train a whole bunch of guys and let them loose without verifying their output first?

I was trying to think of an everyday process as an example but that seems to have added more uncertainty. Trying to think of a generic example is difficult. but will work on it :)

Dan O'Leary Martin,

I’m intrigued by your statement, “a process that is operator dependent cannot be validated”. I’m skeptical. To explain, let’s explore the process validation requirements in 820.75.

In 820.75(a) processes are validated with a high degree of assurance when the process output, the product, is not verified. Process validation determines the combinations of process inputs that produce satisfactory output, i.e., conforming product. This provides an assurance that product conforms without verification.

In 820.75(b), the process must be operated by a qualified individual. This goes beyond the job training requirement in 820.25(b).

Notice that neither of these subsections mentions operator dependence, However, one could make the case, and I do, that operator qualification recognizes that some process steps could be dependant on special skills of the operator.

There are many processes in which process input parameters, operator skills, and lack of product verification come together. One example is welding. Correct parameter setting for the welding equipment is necessary, but not sufficient. You also need a welder who knows how to make the weld, i.e. a qualified operator. However, in an operation that didn’t verify every weld, the process would require validation.

The determining factor for required process validation is the associated verification step. If the product is not 100% verified (perhaps it is destructive, verification uses a sampling plan instead of 100%, etc.) then it must be verified. To come back to the initial question, device class is not a factor.

Martin Lunn ok.

this is going off track a little from the OP regarding device classification and validation.

if we look at this welding process I will try to explain whta i meant.

its is operator dependent.

the operator is a process input.

To validate you would challenge the inputs by changing them within a prescribed window. changing the operator in this case does not mean any operator, even of a similar skill set, would make an acceptabl;e product. To me you can only show that one operator can manufacture this product so the process cannot be varied by operator.

I believe you would validate several processes - each process being defined by an individual operator - that are very similar and give the required output but this would not allow me to have any operator making this product.

so each operator requires validation as part of his/her process and the process is not validated for any operator.

I think this explains my statement.

Not sure I am

Dan O'Leary Martin – I think this gets to the crux of the problem.

You have convolved having a process validated with one of the common techniques in the OQ model.

First, neither FDA QSR nor ISO 13485:2003 requires use of the traditional IQ/OQ/PQ model. It is a useful approach, but not the only method. In the OQ phase, one explores the parameter space to find the combinations of process input inputs that produce conforming product. Often, this is an application of DOE and often the technique is two-level fractional factorial designs. This method produces the worst case input parameter settings and allows setting of the input parameters to achieve the desired result.

As I say above, this method is not required. What is required is a definition of the parameter space that provides satisfactory output. This could have been determined empirically over time. These empirical limits may not be the optimum settings, but as long as they work they are fine.

Similarly, there is no requirement to test every potential person who might work in the process. It is sufficient to say that a “welding certificate” or other qualification is sufficient. If you want to introduce variability in an OQ setting, then the input parameter is a two-variable attribute – qualified operator or not.

Martin Lunn I wold disagree with your two level - qualified or not.

lets look at a test method validation. Gauge R& R. this takes into account the operator and if the variation from the operator is too great then this highlights that the test is not validated. you are removing the dependence of the method or process off the operator so it is generally acceptable. If it is only one operatot r that can use it then yes it may be validated for him/her but you would need to validate each and every operator.

Julie Omohundro How is "operator dependent" different than "equipment dependent"? If the variation from the gauge is too great then the process won't be validated either, right?

Martin Lunn they are interchangeable IMO. and I agree if the variation caused by gauge is too great then you do not have a validated process.

Julie Omohundro Do you think this means that a process that is gauge dependent cannot be validated, either?

Martin Lunn Not entirely sure what you mean by this. If you mean the process is verified by a gauge but through studies you have determind that you cannot rely on the gauge to give you the accuracy you require then I can't see how you could validate it with that gauge. You would not be able to rely on the results obtaned so how can you determine the performance of he process?

Julie Omohundro I meant that you said that you don't think a process that is operator dependent can be validated. If you also think "operator dependent" and "gauge dependent" are interchangeable, it follows that you don't think a "gauge dependent" process can be validated either, right? Just trying to better understand your perspective here. It seems to me that all processes are dependent on one thing or another, and yet many can be validated.

Martin Lunn I said that if it were operator dependent then you would need to validate for each and every operator so a control of certified or not is not acceptable.

so changing the words if it is gauge dependent then you need to validate for each and every gauge so a control of certified (calibrated) is not acceptable.

just becasue Gauge R and r typically uses one measuring stick and 3 operators doen't mean you can't interchange and use one operator with 3 measuring sticks to determine if it is indeed gauge dependent.

these 3 measuring sticks cold be the same type from different manufacturers or differnt measuring sticks e.g. steel rule, vernier caliper micrometer touch probe CMM optical CMM or whatever.