Mark G. Klang, MS, RPh, BCNSP, PhD
Core Manager, Research Pharmacy
Memorial Sloan Kettering Cancer Center
New York, New York
To properly determine the accuracy of a pharmaceutical product or preparation, a test must be designed specifically for that evaluation. The procedure that is selected must be verified through a process referred to as Method Validation.
Method Validation consists of several mathematical and procedural evaluations to determine if the test is adequate for the specific entity being reviewed. United States Pharmacopeia <1225> addresses the compendia procedures that include cGMP practices [21 CFR211.194 (a)] and the Tripartite International Conference on Harmonization (ICH) documents.
Method validation determinants include the following parameters: Accuracy, Precision, Specificity, Detection Limit, Quantitation Limit, Range, Linearity, and Robustness.
Accuracy: The accuracy of an analytical procedure is the closeness of the test result to the true value. The accuracy also relates to the purity of the sample and the confidence interval (or regression (R2) for the slope of linearity test (5 samples) to be close to 1 (>0.995 is ideal).
Precision: The precision of a particular test is the degree of reproducibility of a result. This requires multiple samplings that are compared on the basis of relative standard deviation (RSD). A good RSD is calculated by dividing standard deviation by the average. An acceptable RSD is <0.05. An ideal RSD is <0.02. This factor of repeatability is also referred to as ruggedness when applied to the inter-day variation associated with a set of analysis.
Specificity: Specificity applies to tests to ensure the identity of the drug being tested. This criterion is similar to accuracy, as the purity and lack of degradation must be determined through this type of testing. Quantitative analysis such as ultraviolet spectroscopy has no value for this type of testing as opposed to qualitative analysis as Raman or infrared would.
Detection Limit/Quantitation Limit: The limit of detection (or LOD) is the lowest amount of sample that can be assessed. The detection limit exceeds the Level of Quantification (LOQ) which relates to a specifically measured value. LOD is useful for measuring chemotherapy exposure or drug leakage, whereas LOQ is best for stability analysis when evaluating degradation of a sample.
Range: Range is a function of LOQ as both the minimum and maximum values are identified.
Linearity: Linearity involves selecting a group of samples that approximate the value of the test from 80% to 120% and can be compared by a mathematical function of their relationship. It is not crucial that the curve-fitting function yields a linear result, but that a regression analysis approached 1.0. Five values are minimal to ensure the best evaluation of the test sample.
Robustness: Robustness is the test's capacity to remain unaffected during the operation. Day-to-day variance could be considered a measure of a procedure's ruggedness or robustness.
The above criteria mark the processes conducted to determine if a particular test is suitable for the analysis undertaken. It is not sufficient to take a measurement and assign a value based on that outcome. It is necessary to first go through the process to validate if that procedure meets the above tests.