The ISO 2859 has become with the time an essential tool not only for medical device manufactures but also for all the manufacturing companies which anyhow needs to have a sampling method based on statistical tools applied to their processes.
Moreover, in another post, we have been quite extensively talking about the application of statistical methods to the processes of the Quality Management System, and the AQL sampling methodology can easily be part of the statistical tools that are essential to be applied in the framework of a Quality System of an organization.
The ISO 2859-1 – Sampling procedures for inspection by attributes is the ISO standard of reference for the so-called AQL sampling inspections. Let’s now consider the most important sections of the ISO 2859-1 and move froward for a discussion of the main requirements.
Acceptance Activities and QMS Requirements
If we consider the ISO 13485:2016 there are specific requirements associated to the verification of purchase products, specifically in the section 7.4.3. Basically the requirements are related to the necessity to have a system for the verification that purchased products are aligned with the specifications defined with the supplier. As usual, a risk based approach shall be applied and thus the extent of the verification activities shall be based on the results of the supplier evaluation and proportionate to the risk associated with the purchased products.
If we consider the MDSAP, specific requirements for acceptance activities should also be taken in consideration. For example, Anvisa Regulation requires evidence that sampling plans are based on valid statistical rationale. The organization should also periodically review the sampling methods to ensure they are consistent and suitable for their intended use. The frequency of the review of the sampling plans should be based on the occurrence of nonconforming products, audit results and complaints records.
Very similar requirements are mentioned in the US 21 CFR 820 Code of Federal regulation, thus the FDA regulation for Quality Management System. In fact, in the section in the section 21 CFR 820.250(b), it is reported:
Verify that the manufacturer establishes and maintains procedures to ensure that sampling methods are adequate for their intended use and ensure that when changes occur, the sampling plans are reviewed.
Before starting with the discussion on the requirements mentioned in the ISO 2859-1, it is important to report some definitions, focusing on the words which are not frequently used and thus the meaning might not be clear to everybody in the framework of Acceptance Quality Limits. These definitions are directly taken from the ISO 2859.
- Inspection by attributes
inspection whereby either the item is classified simply as conforming or nonconforming with respect to a specified requirement or set of specified requirements, or the number of nonconformities in the item is counted. - Sampling plan
combination of sample size(s) to be used and associated lot acceptability criteria. - Sampling system
collection of sampling plans, or of sampling schemes, each with its own rules for changing plans, together with sampling procedures including criteria by which appropriate plans or schemes may be chosen. - Acceptance quality limit (AQL)quality level that is the worst tolerable process average when a continuing series of lots is submitted for acceptance sampling.
In the subsequent sections, as mentioned above, we will discuss the requirements associated to ISO 2859-1 with particular attention to the following topics:
- Acceptance Quality Limit
- Lot Acceptance Vs Not Acceptance
- Type of Inspections
- Sampling Plans
The Acceptance Quality Limit (AQL) according to ISO 2859
The Acceptance Quality Limit (AQL) along with the sample size code letter, defines the sampling plans to be used, thus the number of items to be checked and the maximum number of non-conforming items that it is possible to have to accept the whole lot.
In other words, the AQL indicates that the sampling scheme linked to the specific AQL value will accept the lot provided that quality level (percent non-conformities per 100 items) is no greater than the designated value of AQL.
Lot Acceptability
Of course, the acceptability of the lot shall be determined based on the selected sampling plan. When the lot is not accepted, it is basically rejected and it means it cannot be used (for if the sampling is performed in the framework of verification activities for incoming inspection).
When the lot is rejected, disposition of the lot shall be decided. It can be scrapped, reworked or resubmitted for inspection. It is also possible the lot is accepted after having properly evaluated from a risk standpoint the non-conformity.
Similarly, when the lot is accepted, the non-conforming items could be scrapped, reworked or re-submitted for inspection. Obviously, even if the whole lot is accepted, we have the right to do not accept non-conforming items within accepted lots.
When a lot is rejected because of the high number of non-conformities, all the relevant parts shall be informed. As suggested within the ISO 2859-1, rejected lots should be resubmitted for inspection until all items are re-examined, and the supplier is satisfied that all non-conforming items have been removed or reworked. It is possible to decide to use a tighter inspection when the rejected lot is re-inspected.
The type of Inspections
The ISO 2859-1 defines three level of inspections, named:
- Tighten inspection
- Normal Inspection
- Reduced Inspection
Usually, at the beginning normal inspection shall be used. Then, it is possible to change the type of inspection performed based on the results of the inspection previously performed, according to the scheme reported below:
The switching rules between different levels of inspection are reported in the scheme above. When normal inspection is performed, tightened inspection shall be performed when two out of five consecutive lots have been rejected or considered non-acceptable. Of course, it is possible to switch the level of inspection even before reach the level of five consecutive lots.
At the same time, when tightened inspection is performed, it is possible to return to normal inspection level when there are no rejections in 5 consecutive lots.
In some circumstances, it is also possible to decrease the level of inspection from normal to reduced. This is typically feasible when reduced inspection is desirable (for example to avoid cost issue on destructive samples) and the production is at steady rate. To switch, usually there might not have been rejections for 10 consecutive lots.
When we are at the reduced level, it is necessary to return a normal inspection level in case of:
- A lot is not accepted
- Production becomes irregular
How AQL Tables are Used?
There are different tables that need to be taken in considerations to determine the sample size to be inspected and the maximum number of non-conforming items it is possible to have to do not reject the lot under inspection.
Firstly, it is necessary to look at the Sample Size Code Letters, as reported in the table mentioned below.
As it can be seen, it is necessary to select the appropriate inspection level, which basically, designates the level of inspection. Under standard circumstances, the general inspection shall be considered. Under this category, three level of inspections are considered. As it is mentioned within ISO 2859-1, unless otherwise specified, level II shall be used. Level I and III could be used when less or greater discrimination is needed.
Moreover, for additional special levels S1- S4 are defined in the aforementioned table and they can be used with relatively small sample size.
When the sample size code letter is found out based on sample size and inspection levels, it shall be used on other tables to find the related sampling plan. The table to be used depends from the level of inspection (normal, tightened or reduced).
For example, let’s imagine in the framework of incoming inspection control we have a lot of 1500 items. First, in the sample size code letters table we need to find the related code letter. If we consider we are using Type II General Inspection Level, the code letter is K.
Then, if we consider we are using a normal inspection level, we should go to the related table (reported above) and we can find that the sample size code letter K corresponds to a sample size of 125. This number corresponds to the number of items to be checked in the framework of our incoming inspection activity.
Then we need to decide the AQL linked to the specific activity. As a reminder, the AQL corresponds to the number of non-conformities for 100 items. Let’s assume that our AQL for the incoming inspection of our lot sample is 2.5. If we check the table above, this AQL corresponds to a limit of acceptance of 7. It means we cannot have more than 7 non-conformities to accept the lot; thus if we have 8 or more non-conformity within the 125 items to be checked, the lot will have to be rejected.
Conclusions
In conclusions, we have been given a general overview of the ISO 2859 and the related AQL sampling methodology, which can be useful for many different applications, not only for medical device manufacturers.
Specific requirements for verification of purchased products are mentioned within the ISO 13485:2016 or within other regulation such as FDA Quality System Regulation or RDC ANVISA 16/2013 (Brazilian Regulation). Typically, these requirements are associated to the need to have a statistical valid methodology to support the verification activities on purchased products. The classic statistical tool used is the AQL sampling methodology explained within ISO 2859.
Subscribe to QualityMedDev Newsletter
QualityMedDev is an online platform that provides extensive support to medical device manufacturers and consultancy companies in the field of regulatory compliance. We publish blog posts on quality management system topics and provide extensive documentation ready to be downloaded to support the implementation and maintenance of the quality system or product-related certification.
We publish a periodic newsletter aimed at sharing information on the new articles or documents which have been made available through the QualityMedDev website.
If you would like to stay updated with the last news and analysis from the regulatory world for medical device sector, do not forget to subscribe to our newsletter.
QualityMedDev E-book on Computer System Validation
QualityMedDev has published a dedicated e-book with a practical guideline on computer system validation approaches based on GAMP-5 guidance.
Specifically, the e-book covers the following topics:
- Introduction to Computer System Validation, with an overview of the GAMP-5 guideline
- The lifecycle approach to CSV
- A deep dive on risk management requirements linked to CSV
- A discussion on the different GAMP-5 categories for hardware and software and an overview of their impact on the validation approach
- Discussion on the requirements related to Design Review and Traceability
- In depth discussion on the contents of Validation Plan and Validation Report for Computer System validation activities
Do not hesitate to download the e-book!
Coinsmart. Beste Bitcoin-Börse in Europa
Source: https://www.qualitymeddev.com/2021/04/28/iso-2859-aql/
