Quality Assurance and Calibration

Quality Assurance and Calibration Methods

• Section 5-1 discusses the basics of Quality Assurance
• Section 5-2 discusses method validation
• Section 5-3 discusses standard addition
• Section 5-4 discusses internal standards

Case Study: Misconduct at the Gothan Crime Lab?

• A senior technician is accused of falsifying methamphetamine assays
• The accuser is a visiting scientist who have ulterior motives
• An investigation should determine if mistakes are made, and if these mistakes are intentional
• Concluding the data was falsified could force re-examination of 3 000+ drug cases

Basics of Quality Assurance

• Quality assurance results in the right answer for the stated problem when statistics are taken into account
• The results need to have sufficient accuracy and precision to support subsequent decisions
• There is no point in obtaining a more accurate/precise answer than necessary
• Three data quality standards: get the right data, get the data right, and keep the data right

Quality assurance process

• These are the steps for a quality assurance assessment
• State the objectives for the data, what is the main use?
• State the specifications: how good do the numbers have to be?
• Assessment: document procedures and keep records suitable to meet use objectives and verify that use objectives are met

Different types of data

• Raw data: measurements
• Treated data: concentrations derived from raw data by use of calibration methods
• Results: quantities reported after statistical analysis of treated data

Use Objectives

• The goal of quality assurance is making sure that results meet the customer's needs and keep in mind clear and concise use objectives
• Example: drinking water is usually disinfected with chlorine, which kills microorganisms
• Chlorine also reacts with organic compounds present in water to produce chlorinated by-products, many of which are toxic to humans.

Specifications

• Specifications state how good the numbers need to be statistically and what precautions are required in the analytical procedure
• This can include: how shall samples be taken, are special precautions needed to avoid degradation and what level of accuracy will satisfy the use objectives?

Sampling

• First, collect representative samples
• Address heterogeneity of the sample, if applicable
• Trace-metal samples should be stored in plastic or Teflon containers, not glass
• Organic samples should be stored in glass containers, not plastic

False Positive and False Negative Results

• A false positive result has a concentration that exceeds the legal limit, when, in fact, the concentration is below the limit
• A false negative result has a concentration is below the legal limit when it is actually above the limit
• Even well-executed procedures produce some false conclusions due to random error

Choosing a Method

• When choosing a method, consider selectivity, sensitivity, and possible blanks
• Selectivity is the ability to distinguish analyte from other species in the sample and avoid interference
• Sensitivity is the ability to respond reliably and measurably to change in analyte concentration
• Blanks are samples not intended to contain analyte that may inadvertently contain it at trace levels

Types of Blanks

• Method blank: contains all components, including the matrix, except the analyte and undergoes all steps of the analytical procedure
• Reagent blank: similar to a method blank, but it has not been subjected to all sample preparation procedures
• Field blank: similar to a method blank, but it has also been exposed to the site of sampling

Spike Recovery

• Spike recovery is when the response to the analyte can be decreased or increased by something else in the sample
• The matrix is everything in the unknown, other than the analyte
• Spike/fortification: a known quantity of analyte added to a sample

Dealing with Large Numbers of Samples

• Perform periodic calibration checks to prevent method drift
• Method drift happens when changes such as room temperature or spoilage of reagents/standards occur
• Calibration check: periodical recalibration of the method at specified intervals, also known as bracketing the unknowns
• Performance test samples - quality control samples, or blind samples - are samples of known concentration but unknown to the analyst

Assessment

• Assessment is the process of collecting data to show that analytical procedures are operating within specified limits of the Standard Operating Procedure
• Critical documentation needs to be made
• Control charts can be used to monitor performance
• If final results meet the use objectives, the method is fit for purpose

Control Charts

• Control charts are a visual representation of confidence intervals for a Gaussian distribution
• The charts warn when a property strays dangerously far from an intended target value
• For a Gaussian distribution, 95.5% of all observations are within ±2σ, and 99.7% within ±3σ
• ±2σ are warning lines, and ±3σ are action lines

Agencies that Develop Standard Methods

• American Public Health Association (APHA) includes 400 methods for analysis of water, water supplies, and wastewater
• AOAC International (formerly Association of Official Analytical Chemists) has 3 000 standardized chemistry and microbiological methods to ensure safety of food and beverages
• ASTM International (formerly American Society for Testing and Methodology) has 5 400 test methods for 90 industrial sectors
• U.S. Food and Drug Administration (FDA) contains methods for allergens, additives, supplements, pesticides/herbicides, drug residues, toxic elements, bacteria, and microbiologics in food and cosmetics

Method Validation

• Method validation is the process of proving that an analytical method is acceptable for the intended purpose, as defined by the customer
• Method validation includes specificity/selectivity, range, linearity, limit of detection, accuracy, limit of quantitation and precision
• The validation also includes robustness

Specificity (Selectivity)

• Selectivity is the extent to which an analytical method can distinguish analyte from everything else in the sample
• A method is specific if it selects for only one analyte and has no interferences

Linearity

• Linearity measures how well a calibration curve follows a straight line
• The response is proportional to the quantity of analyte
• R2 should be greater than 0.995–0.999 for a good fit
• Absorbance vs Molarity graph should intercept the y axis at 0 within statistics

Residual Plots

• These plots emphasize the difference between calibration data and the least-squares line

Accuracy

• Accuracy can be demonstrated by analyzing a certified reference material in a matrix similar to that of the unknown
• Accuracy also can be shown by comparing results from two or more different analytical methods, or using blank samples spiked with known additions of analyte
• Finally, standard additions of analyte to the unknowns can be used

Types of Precision

• Instrument precision: same quantity of one sample is repeatedly introduced into an instrument
• Intra-assay precision: analysis of aliquots of a homogeneous material several times by one person in one day with the same equipment
• Intermediate precision: assay is performed by different people on different instruments on different days in the same and in Interlaboratory precision is when aliquots of the same sample analyzed by different people in different laboratories

Horwitz Trumpet

• Interlaboratory tests are routinely used to validate new procedures, especially if intended for regulatory use
• These tests often include relative standard deviations of interlaboratory results as a function of sample concentration
• High variations in results are referred to as the Horwitz trumpet

Range and Robustness

• The range is the concentration interval over which linearity, accuracy, and precision are all acceptable
• The linear range is the concentration range over which calibration curve is linear
• The dynamic range is the concentration range over which there is measurable response
• Robustness is the ability of an analytical method to be unaffected by small, deliberate changes in operating parameters

Limits of Detection and Quantitation

• Detection limit is the smallest quantity of analyte that is significantly different from the blank
• Quantitation limit is thesmallest quantity of analyte that can be measured with reasonable accuracy

Detection Limit Equations

• Corrected signal, ysample - blank, is proportional to the sample concentration
• Calibration line:, ysample - yblank = m × sample concentration
• Minimum detectable concentration is obtained by substituting ydl for ysample

Reporting Limit

• Reporting limits is the concentration below which regulations indicate that, the analyte is "not detected".
• This doesn't necessarily mean that analyte is not observed, it just means the analyte is below the prescribed level
• The reporting limit is set at least 5 to 10 times higher than the detection limit

Standard Addition

• Standard addition involves known quantities of the analyte added to the unknown
• Increase in signal indicates how much analyte was in the original unknown
• This method requires a linear response to analyte concentration
• Especially appropriate when sample composition unknown, complex, or affects the analytical signal

Matrix Effect

• Matrix effect represents change in analytical sensitivity caused by something in the sample other than analyte
• Sample composition affects the analytical signal

Standard Addition Equation

• Signal is directly proportional to analyte concentration
• Concentration of analyte in initial solutions / Concentration of analyte + standard in final solution = Signal from initial solution / signal from final solution

Graphical Procedure for Standard Addition

• This graphical procedure is used for successive standard additions to one solution

Internal Standards

• Internal standards has a known amount of a compound that is different from the analyte, which is added to the unknown
• Very useful for analysis when quantity of sample analyzed or instrument response varies from run-to-run
• When sample loss can occur during sample preparation, steps prior to analysis this method is benificial
• It helps if the internal standard being chemically similar to analyte so that effects of the matrix can be minimized

One-Point Calibration

• First, prepare a known mixture of standard and analyt
• Second, measure the relative response of the detector to the two species
• Third, because the detector generally has a different response to each component, measure the response factor

Multpoint Calibration Curve

• A multipoint calibration curve is preferred to average out experimental variability and to verify linearity of response.
• Signal from analyte / signal from standard = F (concentration of analyte / concentration of standard)
• Graphs constructed from Signal Vs Concentration should have near linearity

This lesson covers quality assurance fundamentals, method validation techniques, standard addition methods, and the use of internal standards. It also explores a case study of misconduct at the Gotham Crime Lab, evaluating data integrity and potential consequences of data falsification.