In the Lab - Quality Assurance and Quality Control

Quality assurance (QA) and quality control (QC) are integral parts of any analysis program. The QA process consists of management review and oversight at the planning, implementation, and completion stages of the analytical activity to obtain the required data. The QC process includes activities required during sample preparation and analysis to produce the desired data quality and to document the quality of the analytical data.

During the planning of a chemical analysis program, QA activities focus on defining data quality objectives and designing a QC system to measure the quality of data being generated. During the implementation of the data analysis effort, QA activities ensure that the QC system is functioning effectively and that the deficiencies uncovered by the QC system are corrected. After the analytical data are produced, QA activities focus on assessing the quality of data obtained to determine its suitability to support decisions for further monitoring, risk assessments, or issuance of advisories.

It is the responsibility of each fish monitoring program manager to ensure that the analytical QC program is adequate to meet program data quality objectives for method detection limits, accuracy, precision, and comparability.

Quality Assurance and Quality Control Considerations

Quality Assurance Program

Each laboratory performing chemical analyses in fish and shellfish contaminant monitoring programs must have an effective QA program to ensure the quality, defensibility, and comparability of the data. The laboratory should have quality system documentation that includes tissue analysis.

Quality Assurance Project Plans

The Quality Assurance Project Plan (QAPP) describes the overall program objectives and the data quality requirements and documents specific QA/QC procedures. The project manager should provide the laboratory with a copy of the QAPP and identify the portion of the QAPP that applies to the laboratory.

Sample Analysis Recommendations

QA/QC recommendations are intended to provide a uniform performance standard for all analytical protocols used in fish and shellfish contaminant monitoring programs and to enable an assessment of the comparability of results generated by different laboratories and different analytical procedures. These recommendations are intended to represent minimum QA and QC procedures for any given analytical method. Additional method-specific QC procedures should always be followed to ensure overall data quality.

For sample analyses, minimum QA/QC requirements consist of (1) initial demonstration of laboratory capability and (2) routine analyses of appropriate QA/QC samples to demonstrate continued acceptable performance and to document data quality in a tissue matrix.

The QC requirements for the analyses of target analytes in tissues should be based on specific performance criteria (i.e., warning and control limits) for data quality indicators such as accuracy and precision. Warning limits are numerical criteria that serve to alert data reviewers and data users that data quality may be questionable. A laboratory is not required to terminate analyses when a warning limit is exceeded, but the reported data may be qualified during subsequent QA review. Control limits are numerical data criteria that, when exceeded, require suspension of analyses and specific corrective action by the laboratory before the analyses may resume. The laboratory must detail the procedures for corrective action in the event of excursion outside warning and control limits in its quality system documentation.

Qualified laboratory personnel should review the results for the various QC samples analyzed with each batch of samples immediately after the analysis of each sample batch to determine when warning or control limits have been exceeded. When established control limits are exceeded, the laboratory should take appropriate corrective action and, if possible, reanalyze all suspect samples before resuming routine analyses.

Reference Material

A reference material is a material or substance of which one or more properties have been sufficiently well established to allow its use for method evaluation or characterization of other materials. Reference materials may provide information on method accuracy or interlaboratory comparability. Unfortunately, there are limited reference materials that would be suitable for fish and shellfish contaminant monitoring programs. Most reference materials are shelf-stable materials that have been dried and homogenized into powders. As a result, their relevance as a routine QC operation for wet tissue samples may be limited for some types of analyses. For example, most of the methods for organic contaminants involve some form of extraction with an organic solvent. Many of those solvent extraction procedures behave very differently in wet tissue samples compared to dry materials. In contrast, the strong-acid digestion procedures applied for metals analyses are much less affected by the water content of the samples. Therefore, project planners are cautioned to avoid relying solely on the use of reference materials as a demonstration of laboratory capabilities or routine performance. However, if appropriate reference materials are available at a reasonable cost and a clearly stated purpose is detailed in the monitoring program, then reference materials can be part of the QA/QC program, but they should not be a requirement.

Calibration and Calibration Checks

Calibration and calibration checks should be included in each analytical method. The laboratory staff should follow the method as written. Proper maintenance and calibration of equipment will ensure optimum operating conditions. Calibration materials and standard solutions must be stored in accordance with the method and vendor specifications.

Assessment of Detection and Quantitation Limits

It is the responsibility of each laboratory to determine detection and quantitation limits for each analytical method for each target analyte in a fish or shellfish tissue matrix.

The U.S. Environmental Protection Agency defines the method detection limit (MDL) as the minimum measured concentration of a substance that can be reported with 99% confidence that the measured concentration is distinguishable from method blank results. The MDL is determined as described in 40 CFR Part 136 Appendix B as revised by the 2017 Methods Update Rule.

A method quantitation limit (MQL), the minimum concentration allowed to be reported at a specified level of confidence without qualifications, is derived by the laboratory for each analyte. Analysts must use their expertise and professional judgment to determine the best estimate of the MQL for each target analyte and be able to document that derivation.

Assessment of Method Accuracy

Method accuracy shows the closeness of agreement between what is measured in the analyte by the method and equipment and the actual amount. The laboratory should be required to prepare and analyze QC samples at the frequencies described in the methods (often at least once for every batch of up to 20 field samples that are prepared together). Method accuracy may be assessed by analysis of appropriate reference materials, laboratory control samples, or matrix spike samples, whichever are specified in the analytical method. The unspiked sample should be analyzed first and then spiked at 3-5 times the background level.

Accuracy is calculated as percent recovery from the analysis of reference materials, or laboratory control samples, as follows:

% Recovery = 100 (M/T)

Where:

M = Measured value of the concentration of target analyte
T = "True" value of the concentration of target analyte.

Accuracy is calculated as percent recovery from the analysis of matrix spike samples as follows:

% Recovery = [(Ms - Mu)/Ts] x 100

Where:

Ms = Measured concentration of target analyte in the spiked sample
Mu = Measured concentration of target analyte in the unspiked sample
Ts = "True" concentration of target analyte added to the spiked sample.

Percent recovery values for spiked samples must fall within established control limits as detailed in the analyte method. If the percent recovery falls outside of the control limit, the analyses should be discontinued, and appropriate corrective action taken. The samples associated with the spike should be reanalyzed, but if this is not possible, all suspect data should be clearly identified. If corrective action is not possible, the laboratory may label this as “matrix interference.”

Poor performance on the analysis of reference materials or poor spike recovery may be caused by inadequate mixing of the sample, inconsistent digestion or extraction procedures, matrix interferences, instrumentation problems, or the fact that the method is not well suited for tissue analyses. The first step in separating out some of these causes of poor performance is to examine the laboratory control sample results. The laboratory control sample is prepared in a “clean” reference matrix. If the laboratory control sample results fall within the project or method acceptance limits, that indicates that the laboratory procedures and instrumentation can produce high quality data, that the overall procedure is appropriate, and points to a sample-specific source of the issue.

Matrix spike samples are not needed for methods that employ isotope dilution quantitation. Isotope dilution methods add isotopically labeled compounds to every sample, resulting in labeled compound recovery data to assess accuracy for every sample, not just the matrix spike sample. Therefore, matrix spike analyses do not need to be run when employing an isotope dilution method. However, the isotope dilution method to be used must include acceptance criteria for labeled compound recoveries specific to tissue samples.

Assessment of Method Precision

Precision is defined as the agreement among a set of replicate measurements without assumption of knowledge of the true value. To measure method precision, there should be one duplicate sample per sample preparation batch of up to 20 samples. Based on the observed precision compared to all data reported, a determination should be made as to whether the analysis should be rerun.

The most common estimates of precision are the relative percent difference (duplicate samples) and the relative standard deviation (three or more samples).

A matrix spike and matrix spike duplicate pair may be run to assess accuracy and precision.

Precision estimates from analysis of unspiked laboratory duplicates, matrix spike duplicates, and repeated laboratory control samples analyses must fall within specified control limits stipulated in the method. If these values fall outside control limits, the analyses should be discontinued, appropriate corrective action taken, and the samples associated with the duplicates reanalyzed.

The choice of which QC sample type to use may be method specific and may differ across the methods used for different analytes. For example, for those metals that are usually found in most environmental matrices, the methods may rely on a laboratory duplicate analysis to assess precision. Conversely, for other metals and for many organic contaminants that are not routinely found in every sample, a matrix spike duplicate sample may be used in conjunction with the matrix spike sample described above to overcome the fact that “non-detect” results in unspiked samples cannot be used to evaluate precision with a formal metric.

Estimates of the overall variability of contaminant concentrations in fish or shellfish populations and the consistency of the sampling and analysis procedures can be obtained by collecting and analyzing field replicates. It is recommended that duplicate samples be collected at ten percent of the screening sites as a QC check. When used in conjunction with the estimates of laboratory precision, field duplicates can be useful in separating out sources of variability.

Routine Monitoring of Contamination

Contamination can be a limiting factor in the reliable quantitation of target contaminants in tissue samples. Field and laboratory personnel should follow proper handling procedures and decontamination procedures [LINK to page in Lab section] to avoid contamination of samples in the field and laboratory.

If the laboratory is responsible for tissue sample preparation, equipment rinsates from the sample prep equipment should be analyzed at the beginning of the study and preferably once with each batch of 20 or fewer samples that are prepared. Method blanks are aliquots of the clean reference matrix carried through the complete analytical procedures. Method blanks should be analyzed at the frequency (e.g. once with every 20 samples) described in each analytical method.

Reagent blanks are prepared from all reagents used in the analytical procedure. Each lot of analytical reagents should be analyzed for target analytes of interest prior to use.

Contamination in an equipment rinsate or method blank sample may not always indicate contamination of the tissue samples; therefore, analysts and program managers must use their best professional judgment when interpreting blank analysis data. In the 1980s, as part of Superfund’s Contract Laboratory Program, the EPA developed a series of “National Functional Guidelines” for data review to promote consistency in assessments of the data the EPA was collecting about hazardous waste sites. Those guidelines included what are now commonly known as the 5-10x rules for method blank contamination. That is, if a sample result is greater than 10x the method blank, the sample result is acceptable. If the sample result is less than 5x the method blank, it may be considered a non-detect. Blank concentrations should always be reported with each associated sample value, but the laboratory should not employ “blank subtraction” when reporting results unless that approach is explicitly called out in the method.