Statistical validation accuracy

The acceptance criterion for recovery data is 98-102% or 95-105% for drug preparations. In biological samples, the recovery should be 10%, and the range of the investigated concentrations is 20% of the target concentrations. For trace level analysis, the acceptance criteria are 70-120% (for below 1 ppm), 80-120% (for above 100 ppb), and 60-100% (for below 100 ppb) [2]. For impurities, the acceptance criteria are 20% (for impurity levels <0.5%) and 10% (for impurity levels >0.5%) [30], The AOAC (cited in Ref. [11]) described the recovery acceptance criteria at different concentrations, as detailed in Table 2. A statistically valid test, such as a /-test, the Doerffel-test, or the Wilcoxon-test, can be used to prove whether there is no significant difference between the result of accuracy study with the true value [29],... 

Part—I has three chapters that exclusively deal with General Aspects of pharmaceutical analysis. Chapter 1 focuses on the pharmaceutical chemicals and their respective purity and management. Critical information with regard to description of the finished product, sampling procedures, bioavailability, identification tests, physical constants and miscellaneous characteristics, such as ash values, loss on drying, clarity and color of solution, specific tests, limit tests of metallic and non-metallic impurities, limits of moisture content, volatile and non-volatile matter and lastly residue on ignition have also been dealt with. Each section provides adequate procedural details supported by ample typical examples from the Official Compendia. Chapter 2 embraces the theory and technique of quantitative analysis with specific emphasis on volumetric analysis, volumetric apparatus, their specifications, standardization and utility. It also includes biomedical analytical chemistry, colorimetric assays, theory and assay of biochemicals, such as urea, bilirubin, cholesterol and enzymatic assays, such as alkaline phosphatase, lactate dehydrogenase, salient features of radioimmunoassay and automated methods of chemical analysis. Chapter 3 provides special emphasis on errors in pharmaceutical analysis and their statistical validation. The first aspect is related to errors in pharmaceutical analysis and embodies classification of errors, accuracy, precision and makes... 

This section will give examples of how CALPHAD calculations have been used for materials which are in practical use and is concerned with calculations of critical temperatures and the amoimt and composition of phases in duplex and multi-phase types of alloy. These cases provide an excellent opportunity to compare predicted calculations of phase equilibria against an extensive literature of experimental measurements. This can be used to show that the CALPHAD route provides results whose accuracy lies close to what would be expected from experimental measurements. The ability to statistically validate databases is a key factor in seeing the CALPHAD methodology become increasingly used in practical applications. 

Accuracy is a measure of how close to truth a method is in its measurement of a product parameter. In statistical terms, accuracy measures the bias of the method relative to a standard. As accuracy is a relative measurement, we need a definition of true or expected value. Often, there is no gold standard or independent measurement of the product parameter. Then, it may be appropriate to use a historical measurement of the same sample or a within-method control for comparison. This must be accounted for in the design of experiments to be conducted for the validation and spelled out in the protocol. Accuracy is measured by the observed value of the method relative to an expected value for that observation. Accuracy in percent can be calculated as ratio of observed to expected results or as a bias of the ratio of the difference between observed and expected to the expected result. For example, suppose that a standard one-pound brick of gold is measured on a scale 10 times and the average of these 10 weights is 9.99 lbs. Then calculating accuracy as a ratio, the accuracy of the scale can be estimated at (9.99/10) x 100% = 99.90%. Calculating the accuracy as a bias then [(9.99 - 10)/10] X 100% =-0.10% is the estimated bias. In the first approach ideal accuracy is 100%, and in the second calculation ideal bias is 0%. 

Once the time of exposure and laboratory recovery studies are completed, storage stability should be determined. Although a separate experiment Is possible, the simple expedient of storing one or two fortified pads with each worker s pad set will determine storage stability as extraction and analyses proceed. The resulting recoveries also serve as a check on the accuracy of laboratory technical help. The required number of these fortified pads depends on the size of the experiment. The criterion Is to allow for enough measurements to statistically validate the quality of both storage and extraction. He use a minimum of three fortified pads per exposure day. 

The raw data collected during the experiment are then analyzed. Frequently the data must be reduced or transformed to a more readily analyzable form. A statistical treatment of the data is used to evaluate the accuracy and precision of the analysis and to validate the procedure. These results are compared with the criteria established during the design of the experiment, and then the design is reconsidered, additional experimental trials are run, or a solution to the problem is proposed. When a solution is proposed, the results are subject to an external evaluation that may result in a new problem and the beginning of a new analytical cycle. 

Current use of statistical thermodynamics implies that the adsorption system can be effectively separated into the gas phase and the adsorbed phase, which means that the partition function of motions normal to the surface can be represented with sufficient accuracy by that of oscillators confined to the surface. This becomes less valid, the shorter is the mean adsorption time of adatoms, i.e. the higher is the desorption temperature. Thus, near the end of the desorption experiment, especially with high heating rates, another treatment of equilibria should be used, dealing with the whole system as a single phase, the adsorbent being a boundary. This is the approach of the gas-surface virial expansion of adsorption isotherms (51, 53) or of some more general treatment of this kind. 

Validation of the database. This is the final part in producing an assessed database and must be undertaken systematically. There are certain critical features such as melting points which are well documented for complex industrial alloys. In steels, volume fractions of austenite and ferrite in duplex stainless steels are also well documented, as are 7 solvus temperatures (7 ) in Ni-based superalloys. These must be well matched and preferably some form of statistics for the accuracy of calculated results should be given. 

As noted in the last section, the correct answer to an analysis is usually not known in advance. So the key question becomes How can a laboratory be absolutely sure that the result it is reporting is accurate First, the bias, if any, of a method must be determined and the method must be validated as mentioned in the last section (see also Section 5.6). Besides periodically checking to be sure that all instruments and measuring devices are calibrated and functioning properly, and besides assuring that the sample on which the work was performed truly represents the entire bulk system (in other words, besides making certain the work performed is free of avoidable error), the analyst relies on the precision of a series of measurements or analysis results to be the indicator of accuracy. If a series of tests all provide the same or nearly the same result, and that result is free of bias or compensated for bias, it is taken to be an accurate answer. Obviously, what degree of precision is required and how to deal with the data in order to have the confidence that is needed or wanted are important questions. The answer lies in the use of statistics. Statistical methods take a look at the series of measurements that are the data, provide some mathematical indication of the precision, and reject or retain outliers, or suspect data values, based on predetermined limits. 

The validation of a new method is more involved. A method may introduce a totally new system to the analysis, or it may introduce only new components to an established system. In any case, there may be several new techniques, several new pieces of equipment, several new standard materials, all of which need to be validated, both individually and as a unit. All of the method selection parameters mentioned above (detection limit, accuracy, etc.) are part of the validation process. Also, an important part of validation is the establishment of statistical control (see below). 

---