EDITOR’S NOTE: This is the first in a series of highly popular blog posts that we are republishing in order to share their important subject matter with a wider audience and those who may be new to GxP Lifeline.
For most companies in the life science and manufacturing sectors, the function of quality assurance occupies a space independent of that of actual manufacturing and operations. The same could be said for quality control. But what is the difference between these two essential components of quality manufacturing and why do they matter?
In theory, the difference is that quality assurance activities are proactive and intended to prevent the production of nonconforming products. Quality control activities are reactive and intended to detect and set aside nonconforming products using inspection and testing mechanisms.
In practice, the difference is that quality assurance sets the rules and standards to achieve product quality while quality control is directed at inspecting and testing the product against those pre-set rules and standards.
Using the analogy of a college campus representing a medical device/pharma company, quality assurance would be the city’s government and administration while quality control would be the assigned police officers that patrol the college campus. The assigned police officers (i.e., quality control) are tasked with enforcing the laws and ordinances that the city’s government and administration (quality assurance) set within the jurisdiction of the college campus (i.e., company manufacturing operations). If the city’s government and administration (quality assurance) was a missing piece and the laws and ordinances were set and enforced by the police officers (quality control) assigned to each college campus, consistency in the lawfulness and justice at all college campuses in the city of Chicago would, at the least, not be able to be guaranteed, and at most, differ considerably.
Enjoying this article? You may also enjoy this White Paper:
Six Ways to Optimize Your Quality Management System and Ensure FDA and ISO ComplianceDownload Free White Paper
There is no such thing as a non-independent quality assurance group. Independence from manufacturing operations and adequate authority is what validates quality assurance as quality assurance, according to the U.S. Food and Drug Administration (FDA) and ISO 13485 regulations. Organizational freedom or independence does not necessarily require quality assurance to be a stand-alone group; however, there must not be any conflict of interest. Typically, it is very challenging to remove all conflict of interest for quality assurance personnel without separating these individuals into a stand-alone group or organizational position.
There are many conflicts of interest that come to mind when quality assurance is not its own group apart from manufacturing/operations or development. The following lists some of these conflicts of interest and illustrates the conflict using an example situation:
Example A: Reporting structures, department metrics and performance goals/indicators
Quality assurance personnel report to a chief operations officer. A key performance indicator (KPI) for operations is the quantity of product meeting first-pass acceptance.
Quality assurance may be involved with this goal by providing visual inspection training to materials receiving personnel so that there are fewer raw materials or components accepted that have defects causing the assembled device to be rejected during in-process or final testing. This involvement is not a conflict – if quality assurance does well at the task of providing visual inspection training to materials receiving personnel by making it effective and robust, it will help support the goal of increasing first-pass acceptance rates.
However, quality assurance is also responsible for choosing acceptance limit quality (AQL) inspection levels for product attributes. Quality assurance has an interest in reducing the number of nonconforming products that end up in consumers’ hands in order to prevent customer complaints, required reporting to regulatory authorities, and adverse regulatory actions, such as recalls. If quality assurance does well at the task of choosing AQL inspection levels that give higher confidence that nonconforming product is not released from manufacturing (which requires inspection of a larger sample size versus a smaller sample size), it will likely hurt operations’ chance of meeting the goal of increasing first-pass acceptance rates. This is a clear conflict of interest. Quality assurance would have to fail at the task of choosing AQL inspection levels that give higher confidence that nonconforming product is not released from manufacturing in order to help support the goal of increasing first-pass acceptance rates.
Example B: Maintaining an unbiased approach when making quality decisions
Quality assurance personnel are routinely involved in meetings that manufacturing holds with supplier A. Individuals from Supplier A that participate in these meetings often informally discuss concerns and doubts about their processes, personnel or resources. Because of this, the quality assurance personnel potentially have cause or concern to impose stricter quality standards on Supplier A due to informal gossip or opinions that may neither be substantiated and nor validated by data relevant to the supplier’s products. Imposing stricter quality standards on this supplier amounts to dedication of resources, including choosing to conduct an on-site audit of this supplier as opposed to having them fill out the routine supplier survey.
Supplier B is currently meeting performance goals for supplier incoming inspections, and Supplier C is currently exceeding performance goals for supplier incoming inspections. Quality assurance chose to send these suppliers the routine supplier survey due to the biased determination to audit Supplier A on-site. If quality assurance didn’t have bias in choosing Supplier A, the chances and decision of the on-site audit being given to Supplier B or Supplier C would have been fair.
As it turns out, an on-site audit of Supplier B would have exposed the fact that equipment being used at its manufacturing facility is past its calibration expiration date and in fact, out of spec. This resulted in the material Supplier B provides to the company to be out of specification for an attribute that is neither inspected nor tested during incoming inspection, but rather later on in the manufacturing process during in-process device testing. As it turns out, an on-site audit of Supplier C would have exposed the fact that this supplier is exceeding performance goals because they are falsifying records and don’t have adequate identification and traceability procedures. The material received by Supplier C has biocompatibility requirements and the incoming inspection includes review of the applicable ISO 10993 testing records and certificates, which were falsified. The material they supply is a variation of the material specified in the design and was not tested per ISO 10993. Because of their inadequate identification and traceability procedures, Supplier C assigned it the same part number as the original material and didn’t identify the design change in the product’s name or description.
The table below is a deeper dive into the conflicting interests relating to a quality assurance unit’s responsibility of selection of statistical technique used for design verification testing (DVT). As you can see, development and manufacturing/ops have different interests when the listed considerations are analyzed. If quality assurance staff were to be incorporated into either the development organization or the manufacturing/operations organization, they would not have the required independence to make the appropriate decision.
|Considerations||Interest||Corresponding Statistical Technique||Interest||Corresponding Statistical Technique|
|Amount of Data Required||Choose statistical technique with minimal amount of data required to ensure adequate resources and due dates are met for test report writing||Descriptive Statistics (mean, SD, Range, etc.)
|Use of design verification testing to gain characterization data on product attributes to use for automated testing systems / equipment used within manufacturing||Graphical Methods (Scatter/Line Plots, Pareto Analysis, Frequency Over Time Plots, etc.)|
|Sample Size||Minimal sample size or alignment of sample size to development budget for prototypes and engineering builds||Hypothesis Testing
Acceptance Sampling Plans (for minor/ negligible defect severities with higher AQLs)
Design of Experiments
|Larger sample size to achieve greater predictability of manufacturing acceptance / rejection rates||Confidence Levels or Reliability Levels (high)
Acceptance Sampling Plans with lower AQLs or Analysis of Variance (ANOVA)
|Cost of correcting a nonconforming product||Cost to reputation - state of the art / design excellence||Design of Experiments
Acceptance Sampling Plans with lower AQLs
Confidence Levels or Reliability Levels (high)
|Avoidance of production shutdowns or recalls||Acceptance Sampling Plans with lower AQLs
Confidence Levels or Reliability Levels (high)
|Specification Changes||Developing precise specifications according to materials and technology||Graphical Methods (Scatter/Line Plots, Pareto Analysis, Frequency Over Time Plots, etc.)
Design of Experiments or Analysis of Variance (ANOVA)
|Manufacturing equipment needed to test updated specifications (abilities of current DMR) and impact to process validations||Descriptive Statistics (mean, SD, Range, etc.) – can be used to evaluate Descriptive Statistics of manufacturing equipment|
|Supplier Performance / Supplier Evaluation – qualification of supplier for updated specifications or updated requirements (potential impacts to environmental monitoring, equipment needs, etc.)||Descriptive Statistics (mean, SD, Range, etc.) – can be used to define requirements for suppliers or requirements for incoming materials|
Now that you understand that independence is a must in order for quality assurance to assure product quality, let’s clarify the responsibilities of both quality assurance and quality control, and how both these functions are intended to successfully interact with the organization.
Quality assurance is traditionally tasked with the following responsibilities, which are intended to be proactive and prevent the production of non-conforming products:
Kim Washburn is a regulatory affairs professional who has covered the U.S., European Union, Canada and over 70 global countries. She brings more than 15 years of experience in the medical device, in-vitro diagnostics, pharmaceutical and biologics industries. Her areas of expertise include regulatory approval and new product development support, regulatory labeling, medical device software, FDA-compliant design control and risk management, and technical writing. Prior to consulting, Washburn held roles at Abbott Laboratories, OrthoSensor, Nipro Diabetes Systems, and University of Miami Tissue Bank. She received a bachelor’s degree in biology/computer science from Illinois State University, after transferring from West Point (United States Military Academy).