Exploring Quality Control Samples

Quality control (QC) samples are an essential aspect of bioanalysis. When conducting bioanalytical studies, it is critical to ensure the accuracy and precision of the methods used to analyze study samples. One way to achieve this is by using QC samples.

Samples of known concentrations or purities are generally prepared to be similar in composition to the test samples. These QC samples are typically prepared at various concentrations to cover the expected range of the test samples. They are then analyzed alongside the test samples to assess the precision and accuracy of the analytical method.

Imagine you have a dartboard with multiple concentric circles. The bullseye at the center represents the true value or target value of the analyte you’re measuring. The concentric circles around the bullseye represent different levels of measurement variability.

A common approach to measuring the precision of an analytical method is by calculating Coefficient of Variation (CV) values for the QC samples. The CV can be compared to the spread of darts around the bullseye. A low CV indicates that the measurements are tightly clustered around the true value, like a tight grouping of darts on the dartboard. Conversely, a high CV suggests a larger spread of measurements, resembling a scattered pattern of darts on the dartboard. A lower CV indicates higher precision and less variability in the measurements.  It is calculated as the ratio of the standard deviation (a common statistical measure of variability) to the mean of the measured values and is usually expressed as a percentage.

The acceptance criteria for the CV can vary depending on the regulatory agency and the type of analysis being performed, but in general CV values exceeding 15% to 25% are considered unacceptable.

The accuracy of an analytical method is oftentimes assessed with a Relative Error (RE) calculation. The RE measures the closeness of the measurements to the true value. In the dartboard analogy, the RE can be likened to the average distance of the darts from the bullseye, regardless of how tightly or widely they are spread. A low RE indicates that, on average, the measurements are very close to the true value, like the darts being concentrated around the bullseye. On the other hand, a high RE suggests that, on average, the measurements deviate from the true value, akin to the darts being farther away from the bullseye.

A lower RE indicates higher accuracy and less bias in the measurements. The RE is calculated as the difference between the measured value of the QC sample and the known value divided by the known value, expressed as a percentage. The acceptance criteria for the RE also vary depending on the regulatory agency and the type of analysis being performed, but in general are considered unacceptable if they exceed ±15% to ±25%.

The QC samples and their CV and RE values are typically presented in a final report of the test sample results.  Most labs or Sponsors will have an SOP in place to outline which QCs to include in the final report.  The nature of the analysis and the purpose of the study can also play a role, but in general the QCs from each analytical plate of samples, as well as that individual plate’s mean, CV, and RE results (i.e., the intra-plate summary statistics) are listed in descending order. At the bottom of the list of plates, the overall summary statistics for all the plates are calculated (i.e., the inter-plate summary statistics).

The decision to exclude individual QC sample data, or whole plates of data, from the QC table in the final report must be made carefully. If the QCs from a plate fail to meet the acceptance criteria due to an assignable cause, such as a technical error or equipment malfunction, then the data from that plate can be safely excluded from the QC sample table because those data is unreliable. However, if there is no documented assignable cause, many regulators prefer to see the QC data in the table, even if the entire plate failed.  Scientists must then make a careful decision on how, or if, to include failed data in the summary statistics for the final report.

One approach to this problem is to simply provide two sets of inter-plate summary statistics at the bottom of the QC table.  One set contains the statistics from all the passing plates, and the other presents the statistics from all plates.  This level of transparency tends to help temper the concerns of reviewers and regulators who want the confidence that the method was run reliably and as intended for the test samples.

It can be tricky at the intra-plate level.  Sometimes one aberrant QC value can skew the overall results for that plate.  Some labs might deem the value to be an outlier and simply exclude it.  However, in the absence of a documented cause for the aberrant value, it is typically safer to include the intra-plate statistics both with and without the value in the interest of full transparency.  Some labs will choose to exclude the value anyway, so its inclusion doesn’t result in a total plate failure and a need to repeat the analysis of the test samples.  In most cases the aberrant value is a result of a technical error in the lab, or something related to the equipment.  Outlier tests and other justifications have been used to exclude the value. In such cases it behooves the study to address the approach that was taken to exclude the value either as a footnote to the table or in an independent Comments Section in the report.

These types of approaches to data exclusion and inclusion in final report tables should be documented somewhere in an SOP that describes how to handle failed plates.  Having this in place can go a long way to instilling confidence in regulators and other reviewers that the lab has a documented process to handling tricky data situations that arise with QC samples. It can also help regulators understand why the results from test samples can be trusted despite the existence of aberrant individual QC values, decisions to reject whole plates, or decisions to accept test sample results despite these tricky data situations that might otherwise weigh on the reliability of the bioanalytical method.

Led by CEO Sonia Veluchamy, we boast a diverse network of full-time medical writers and regulatory consultants, hailing from various scientific disciplines, with a particular specialization in the regulated bioanalytical domain. This extensive network empowers us to offer comprehensive services tailored to client programs necessitating adeptness in various writing methodologies pertinent to regulated bioanalysis. Among the pivotal subjects we address, the utilization and meticulous reporting of QC samples stand as a cornerstone.

We can be reached at info@celegence.com or contact us online for more information

About the Author
This blog post is written by Jean-François Cazorla, our Scientific Writing Consultant, who has extensive experience and expertise in the field of regulated bioanalytical report writing. Jean’s 20+ years in the field includes overseeing the regulated QC and writing divisions in bioanalytical CRO environments, as well as the individual production of regulated bioanalytical documents for 250+ biopharmaceutical sponsors. During this time, he was involved in small and large molecule bioanalytical programs, with particular emphasis on producing immunogenicity and pharmacokinetic validation and sample analysis reports. He is currently a consultant for Celegence helping a major documentation contract with a Sponsor developing gene therapies.