A Model for Statistical Evaluation of Precision Parameters of Microbiological Methods: Application to Dry Rehydratable Film Methods and IDF Reference Methods for Enumeration of Total Aerobic Mesophilic Flora and Coliforms in Raw Milk

1991 ◽  
Vol 74 (1) ◽  
pp. 92-103 ◽  
Author(s):  
Christine Piton ◽  
Rémy Grappin
Keyword(s):  
Statistical Approach ◽  
Collaborative Study ◽  
Raw Milk ◽  
Confirmatory Test ◽  
Plate Count ◽  
Reference Methods ◽  
Microbiological Methods ◽  
Relative Standard ◽  
Standard Deviations ◽  
Total Bacteria

Abstract A new statistical approach for collaborative study data of microbiological methods Is proposed. This Includes a confirmatory test to the Polsson distribution of the number of colonies. In addition, 2 new statistical parameters are used to express precision as a percent of the original unit: the geometric relative standard deviation (GRSD) and the critical relative difference between 2 measurements (RD95). This statistical approach was applied to an Interlaboratory study to assess and compare the precision of both dry rehydratable film (PetrlfUrn® SM and Petrifllm® VRB) methods and International Dairy Federation (IDF) reference methods [total aerobic mesophilic plate count (TAMPC) and violet red bile lactose agar (VRBL) methods] for estimation of total bacteria and collform, respectively, in raw milk. Each of the 14 laboratories In the study analyzed 40 laboratory samples (20 different materials In blind duplicates) for total bacteria and collform counts by both the Petrifllm and standard methods. Repeatability standard deviations (In log10 unit) of TAMPC, Petrifllm SM, VRBL, and Petrifllm VRB were 0.106, 0.089, 0.219, and 0.171, respectively; their reproducibility standard deviations were 0.170,0.167,0.348, and 0.199, respectively.

Comparison of Two Methods for Determination of Total Solids Content of Milk: Collaborative Study

1989 ◽  
Vol 72 (5) ◽  
pp. 712-718 ◽  
Author(s):  
Jenny L Clark ◽  
David M Barbano ◽  
Chapman E Dunham
Keyword(s):  
Collaborative Study ◽  
Raw Milk ◽  
Preliminary Evaluation ◽  
Oven Method ◽  
Total Solids ◽  
Relative Standard ◽  
Aoac Method ◽  
Standard Deviations ◽  
Forced Air ◽  
Solids Content

Abstract Ten laboratories analyzed 9 pairs of blind duplicate raw milk samples for total solids. A direct forced air oven method (4 h at 100°C) and a modification of the AOAC predry method (16.032) were used. Preliminary evaluation of the modified AOAC method indicated that blank determinations were necessary. Total solids content ranged from 12.0 to 14.6%. Average repeatability standard deviations (sr) of the direct forced air oven and modified AOAC methods were 0.019 and 0.017, respectively. Average reproducibility standard deviations (SR) of the direct forced air oven and the modified AOAC methods were 0.042 and 0.047, respectively. Average repeatability relative standard deviations (RSDr) for the direct forced air oven and the modified AOAC methods were 0.149 and 0.136%, respectively; average reproducibility relative standard deviations (RSDR) were 0.327 and 0.370%, respectively. Mean repeatability values (r) and reproducibility values (R) were 0,054 and 0.118 for the direct forced air oven method and 0.049 and 6.133 for the modified AOAC method, respectively. The mean test result of the direct forced air oven method (12.7293%) was comparable to that for the modified AOAC method (12.7273%). The modification of AOAC method 16.032 and the direct forced air oven method have been approved interim official first action.

Dry Rehydratable Film for Enumeration of Total Aerobic Bacteria in Foods: Collaborative Study

1990 ◽  
Vol 73 (2) ◽  
pp. 242-248
Author(s):  
Michael S Curiale ◽  
Therese Sons ◽  
J Sue Mcallister ◽  
Barbara Halsey ◽  
Terrance L Fox
Keyword(s):  
Collaborative Study ◽  
Agar Plate ◽  
Plate Count ◽  
Relative Standard ◽  
Plate Counts ◽  
Agar Method ◽  
Standard Deviations ◽  
Aerobic Plate Count ◽  
Film Method ◽  
Dry Film

Abstract A rehydratable dry-film plating procedure for aerobic plate counts has been compared to the standard agar plate method (966.23B and C, 15th ed.; 46.014-46.015, 14th ed.) in a collaborative study by 12 laboratories. Each laboratory analyzed the normal microflora of 3 samples in duplicate for 6 products. The aerobic plate counts ranged from 1.0 X 103 to 1.0 X 108 cfu/g. The products were flour, nuts, frozen raw shrimp, spice, frozen raw ground turkey, and frozen and refrigerated vegetables. Repeatability standard deviations of the 2 methods did not differ significantly for 13 of 18 test samples. For 1 shrimp and 2 turkey samples, the dry-film method had lower repeatability variances (P < 0.05) and for 1 spice sample the agar method had lower repeatability variances (P < 0.05). Relative standard deviations of repeatability were between 1.7 and 15.5% for the dry-film method and 1.2 and 16.0% for the agar method. Relative standard deviations of reproducibility ranged from 2.4 to 23.4 % for the dry-film method and 2.3 to 18.8% for the agar method. The dry rehydratable film method has been adopted official first action for determination of the aerobic plate count.

Precision and Accuracy of Infrared Milk Analysis

1972 ◽  
Vol 55 (3) ◽  
pp. 488-497
Author(s):  
D A Biggs
Keyword(s):  
Collaborative Study ◽  
Raw Milk ◽  
Milk Analysis ◽  
Calibration Data ◽  
Mean Values ◽  
Total Solids ◽  
Reference Methods ◽  
Precision And Accuracy ◽  
Standard Deviations ◽  
Mean Differences

Abstract Estimates of precision and accuracy of milk analysis by the IR method are reported for comparison studies done at the University of Guelph, the Ontario Central Milk Testing Laboratory in Guelph, the British Columbia Department of Agriculture Laboratory in Vancouver, B.C., and the Dairyman’s Cooperative Creamery Association in Tulare, Calif. The standard reference methods used were: the Mojonnier method for fat, a semimicro Kjeldahl method for protein, a polarimetric method for lactose, the USDA lactometer method for solids-not-fat, and the AOAC method for total solids. Mean differences of 0.01 or less and standard deviations of difference between means of duplicate tests of the order of ±0.03 have been reported for fat, protein, and lactose; whereas, for solids-not-fat or total solids, mean differences have been about 0.015% and standard deviations of difference about ±0.09%. Best results have been obtained when calibrations have been routinely checked with a reference standard method, and have been adjusted when accumulated calibration data showed either the slope or level calibration to be in need of adjustment. Since the IR method essentially reproduces the results which would be obtained by the standard analytical method used for calibration of the IR instrument, significant differences between the results produced by different laboratories with standard methods will cause significant differences between the IR results produced by the same laboratories. This has been substantiated by the results of a collaborative study in which 5 pasteurized homogenized and 5 raw milk samples were analyzed by both standard and reference methods at the above mentioned laboratories and at Ross Laboratories in Columbus, Ohio. Although regression analysis showed that the calibrations at the various laboratories were producing good estimates of the standard values, there nevertheless were statistically significant differences between the mean values produced by these laboratories for both the standard and IR methods. The IR milk analyzer method for the analysis of milk has been adopted as official first action.

Comparison of Babcock and Ether Extraction Methods for Determination of Fat Content of Milk: Collaborative Study

1988 ◽  
Vol 71 (5) ◽  
pp. 898-914 ◽  
Author(s):  
David M Barbano ◽  
Jenny L Clark ◽  
Chapman E Dunham
Keyword(s):  
Extraction Method ◽  
Collaborative Study ◽  
Raw Milk ◽  
Extraction Methods ◽  
Fat Content ◽  
Absolute Level ◽  
Ether Extraction ◽  
Relative Standard ◽  
Standard Deviations ◽  
The Difference

Abstract Eleven collaborating laboratories analyzed 18 blind duplicate pairs of raw milk samples for fat by the Babcock method and by a modified Mojonnier ether extraction method in 7 round robins conducted over a 14 month period. Ten laboratories used the Babcock method and 10 used the modified Mojonnier method. Fat content of samples ranged from approximately 2.7 to 5.6%. Mean test value of samples analyzed was approximately 3.9% fat. Average standard deviations of within-laboratory repeatability (Sr) of the Babcock and ether extraction methods were 0.029 and 0.015% fat, respectively. Average standard deviations of between-laboratory reproducibility (SR) of the Babcock and ether extraction methods were 0.039 and 0.020% fat, respectively. Average repeatability relative standard deviations (RSDr) for the Babcock and ether extraction methods were 0.742 and 0.396%; average reproducibility relative standard deviations (RSDR) were 1.014 and 0.512%, respectively. Mean repeatability values (r) and reproducibility values (R) were 0.081 and 0.111% for Babcock and 0.044 and 0.056% for ether extraction, respectively. The ether extraction method demonstrated consistently better within- and between-laboratory agreement. The overall mean test value for the Babcock method was significantly higher (0.021% fat) than that for ether extraction. The difference between Babcock and ether extraction fat test results was different for different farms. In addition, the mean difference between percent fat determined by the Babcock and ether extraction methods was different in different months. There was no correlation in the difference between Babcock and ether extraction methods with the absolute level of fat in the samples in the range of 2.7 to 5.6% fat. The modifications of the AOAC Babcock method and the modified Mojonnier ether extraction method have been approved interim official first action.

Enumeration of Total Bacteria and Coliforms in Milk by Dry Rehydratable Film Methods: Collaborative Study

1986 ◽  
Vol 69 (3) ◽  
pp. 527-531 ◽  
Author(s):  
Roy E Ginn ◽  
Vernal S Packard ◽  
Terrance L Fox ◽  
◽  
E Arnold ◽  
...  
Keyword(s):  
Collaborative Study ◽  
Plate Count ◽  
Standard Methods ◽  
Standard Plate Count ◽  
Milk Samples ◽  
Standard Plate ◽  
The Mean ◽  
Standard Deviations ◽  
The Difference ◽  
Total Bacteria

Abstract Eleven laboratories participated in a collaborative study to compare the dry rehydratable film (Petrifilm® SM and Petrifilm® VRB) methods, respectively, to the standard plate count (SPC) and violet red bile agar (VRBA) standard methods for estimation of total bacteria and coliform counts in raw and homogenized pasteurized milk. Each laboratory analyzed 16 samples (8 different samples in blind duplicate) for total count by both the SPC and Petrifilm SM methods. A second set of 16 samples was analyzed by the VRBA and Petrifilm VRB methods. The repeatability standard deviations (the square root of the between-replicates variance) of the SPC, Petrifilm SM, VRBA, and Petrifilm VRB methods were 0.0S104, 0.0444, 0.14606, and 0.13806, respectively; the reproducibility standard deviations were 0.7197, C.06380, 0.15326, and 0.13806, respectively. The difference between the mean Iog10 SPC and the mean logio Petrifilm SM results was 0.027. For the VRBA and Petrifilm VRB methods, the mean log10 difference was 0.013. These results generally indicate the suitability of the dry rehydratable film methods as alternatives to the SPC and VRBA methods for milk samples. The methods have been adopted official first action.

scholarly journals Indirect and Direct Determination of the Casein Content of Milk by Kjeldahl Nitrogen Analysis: Collaborative Study

1998 ◽  
Vol 81 (4) ◽  
pp. 763-774 ◽  
Author(s):  
Joanna M Lynch ◽  
David M Barbano ◽  
J Richard Fleming
Keyword(s):  
Standard Deviation ◽  
Nitrogen Content ◽  
Relative Standard Deviation ◽  
Collaborative Study ◽  
Direct Determination ◽  
Raw Milk ◽  
Method Performance ◽  
Relative Standard ◽  
Milk Casein

Abstract The classic method for determination of milk casein is based on precipitation of casein at pH 4.6. Precipitated milk casein is removed by filtration and the nitrogen content of either the precipitate (direct casein method) or filtrate (noncasein nitrogen; NCN) is determined by Kjeldahl analysis. For the indirect casein method, milk total nitrogen (TN; Method 991.20) is also determined and casein is calculated as TN minus NCN. Ten laboratories tested 9 pairs of blind duplicate raw milk materials with a casein range of 2.42- 3.05℅ by both the direct and indirect casein methods. Statistical performance expressed in protein equivalents (nitrogen ⨯ 6.38) with invalid and outlier data removed was as follows: NCN method (wt%), mean = 0.762, sr = 0.010, SR = 0.016, repeatability relative standard deviation (RSDr) = 1.287℅, reproducibility relative standard deviation (RSDR) = 2.146%; indirect casein method (wt℅), mean = 2.585, repeatability = 0.015, reproducibility = 0.022, RSDr = 0.560℅, RSDR = 0.841; direct casein method (wt℅), mean = 2.575, sr = 0.015, sR = 0.025, RSDr = 0.597℅, RSDR = 0.988℅. Method performance was acceptable and comparable to similar Kjeldahl methods for determining nitrogen content of milk (Methods 991.20, 991.21,991.22, 991.23). The direct casein, indirect casein, and noncasein nitrogen methods have been adopted by AOAC INTERNATIONAL.

Microbial Receptor Assay for Rapi d Detection and Identification of Seven Families of Antimicrobial Drugs in Milk: Collaborative Study

1988 ◽  
Vol 71 (2) ◽  
pp. 304-316 ◽  
Author(s):  
Stanley E Charm ◽  
Ruey Chi
Keyword(s):  
Binding Sites ◽  
Specific Binding ◽  
Scintillation Counter ◽  
Collaborative Study ◽  
False Negative ◽  
Penicillin G ◽  
Specific Binding Sites ◽  
Receptor Assay ◽  
Relative Standard ◽  
Standard Deviations

Abstract A microbial competitive receptor assay for detecting residues of antibiotic families in milk was studied collaboratively by 13 laboratories. The drugs and levels (ppb) tested in this study i nclude penicillin G, 4.8; cephapirin, 5.0; cloxacillin, 100; tetracycline, 2000; chlortetracycline, 2000; oxytetracycline, 2000; erythromycin, 200; lincomycin, 400; clindamycin, 400; sulfamethazine, 75; sulfamethoxazole, 50; sulfisoxazole, 50; streptomycin, 1000; novobiocin, 50; and chloramphenicol, 800. In this method, microbial cells added to a milk sample provide specific binding sites for which 14C or 3H libeled drug competes with drug residues in the sample. The UC or H binding to the specific binding sites is measured in a scintillation counter and compared with a zero standard milk. If the sample is statistically different from the zero standard, it is positive. The assay takes about 15 min. The binding reaction occurs between the receptor site and the drug functional group, so all members of a drug family are detected. In this case, beta-lactams, tetracyclines, macrolides, aminoglycosides, novobiocin, chloramphenicol, and sulfonamides, including/^-aminobenzoic acid (PABA) and its other analogs, are detectable. The incidence of false negative determinations among samples is about 1%; the incidence of false positives is about 3%. For negative cases, the relative standard deviations for repeatability ranged from 0 to 5% and for reproducibility from 0 to 6%. For positive cases, relative standard deviations ranged from 0 to 13% for repeatability and from 0 to 14% for reproducibility. The method has been adopted official first action.

scholarly journals Determination of Ochratoxin A in Wine and Beer by Immunoaffinity Column Cleanup and Liquid Chromatographic Analysis with Fluorometric Detection: Collaborative Study

2001 ◽  
Vol 84 (6) ◽  
pp. 1818-1827 ◽  
Author(s):  
Angelo Visconti ◽  
Michelangelo Pascale ◽  
Gianluca Centonze ◽  
E Anklam ◽  
A M Betbeder ◽  
...  
Keyword(s):  
Ochratoxin A ◽  
Red Wine ◽  
Collaborative Study ◽  
White Wine ◽  
Fluorometric Detection ◽  
Immunoaffinity Column ◽  
Relative Standard ◽  
Standard Deviations ◽  
Liquid Chromatographic

Abstract The accuracy, repeatability, and reproducibility characteristics of a liquid chromatographic method for the determination of ochratoxin A (OTA) in white wine, red wine, and beer were established in a collaborative study involving 18 laboratories in 10 countries. Blind duplicates of blank, spiked, and naturally contaminated materials at levels ranging from ≤0.01 to 3.00 ng/mL were analyzed. Wine and beer samples were diluted with a solution containing polyethylene glycol and sodium hydrogen carbonate, and the diluted samples were filtered and cleaned up on an immunoaffinity column. OTA was eluted with methanol and quantified by reversed-phase liquid chromatography with fluorometric detection. Average recoveries from white wine, red wine, and beer ranged from 88.2 to 105.4% (at spiking levels ranging from 0.1 to 2.0 ng/mL), from 84.3 to 93.1% (at spiking levels ranging from 0.2 to 3.0 ng/mL), and from 87.0 to 95.0% (at spiking levels ranging from 0.2 to 1.5 ng/mL), respectively. Relative standard deviations for within-laboratory repeatability (RSDr) ranged from 6.6 to 10.8% for white wine, from 6.5 to 10.8% for red wine, and from 4.7 to 16.5% for beer. Relative standard deviations for between-laboratories reproducibility (RSDR) ranged from 13.1 to 15.9% for white wine, from 11.9 to 13.6% for red wine, and from 15.2 to 26.1% for beer. HORRAT values were ≤0.4 for the 3 matrixes.

Hydrophobic Grid Membrane Filter Method for Aerobic Plate Count in Foods: Collaborative Study

1986 ◽  
Vol 69 (4) ◽  
pp. 671-676 ◽  
Author(s):  
Phyllis Entis ◽  
◽  
J Allen ◽  
A Bhatnagar ◽  
A Brouwer ◽  
...  
Keyword(s):  
Membrane Filter ◽  
Collaborative Study ◽  
Raw Milk ◽  
Plate Count ◽  
Filter Method ◽  
Membrane Filter Method ◽  
Aerobic Plate Count ◽  
Egg Powder ◽  
Raw Poultry ◽  
Whole Egg

Abstract Twenty-one laboratories participated in a collaborative study to validate a hydrophobic grid membrane filter (HGMF) method for aerobic plate count by comparing its performance against the AOAC/APHA pour plate method. Raw milk, raw poultry, whole egg powder, flours, and spices were included in the study. Counts obtained by the HGMF and pour plate methods did not differ significantly, except in the case of whole egg powder, for which the HGMF method produced significantly higher counts. The hydrophobic grid membrane filter method for aerobic plate count in foods has been adopted official first action.

Sulfuric Acid/Hydrogen Peroxide Digestion and Colorimetric Determination of Phosphorus in Meat and Meat Products: Collaborative Study

1991 ◽  
Vol 74 (1) ◽  
pp. 22-26 ◽  
Author(s):  
David K Christians ◽  
Thomas G Aspelund ◽  
Scott V Brayton ◽  
Larry L Roberts
Keyword(s):  
Hydrogen Peroxide ◽  
Sulfuric Acid ◽  
Collaborative Study ◽  
Meat Products ◽  
Colorimetric Determination ◽  
Pork Sausage ◽  
Relative Standard ◽  
Significant Difference ◽  
Standard Deviations

Abstract Seven laboratories participated In a collaborative study of a method for determination of phosphorus in meat and meat products. Samples are digested In sulfuric acid and hydrogen peroxide; digestion Is complete In approximately 10 mln. Phosphorus Is determined by colorimetric analysis of a dilute aliquot of the sample digest. The collaborators analyzed 3 sets of blind duplicate samples from each of 6 classes of meat (U.S. Department of Agriculture classifications): smoked ham, water-added ham, canned ham, pork sausage, cooked sausage, and hamburger. The calibration curve was linear over the range of standard solutions prepared (phosphorus levels from 0.05 to 1.00%); levels in the collaborative study samples ranged from 0.10 to 0.30%. Standard deviations for repeatability (sr) and reproducibility (sR) ranged from 0.004 to 0.012 and 0.007 to 0.014, respectively. Corresponding relative standard deviations (RSDr and RSDR, respectively) ranged from 1.70 to 7.28% and 3.50 to 9.87%. Six laboratories analyzed samples by both the proposed method and AOAC method 24.016 (14th Ed.). One laboratory reported results by the proposed method only. Statistical evaluations Indicated no significant difference between the 2 methods. The method has been adopted official first action by AOAC.

Sign in / Sign up

Export Citation Format

Share Document