Document RagyyJ4zZxmNRMgLnZE9pGXqk

A O \V LABORATOIRE D'HYGIENE ET DE CONTROLE DES FIBRES MINeRALES 10, rue de la P6pini&re, 75008 PARIS T6I. (1) 45 22 12 34 - Fax 42 94 98 86 Paris, April 4th 1991 Mrs JANIS WOODSON REYNOLDS CERTAINTEED CORPORATION 750 E.SWEDESFORD ROAD VALLEY FORGE 05 PA 19482 U.S.A. Ref.: AIA Fibre Counting Trial - 1990 Dear Mrs JANIS WOODSON REYNOLDS, You will find enclosed a copy of the final report of the 1990 - A.I.A.- Fibre Counting Trial in which counting results of each participant are taken into account. In this report, an individual position data sheet for each counter of your laboratory is included giving his classification among 3 groups according to his performances. Yours sincerely. Alain Baujon Director Association rgie par la loi de 1901 PLAINTIFF'S EXHIBIT CT-67I CTD026238 A.I.A. FIBRE COUNTING TRIAL - 1990 INDIVIDUAL POSITION DATA SHEET Name : Mrs LAURA SHEMANSKI Company : CERTAINTEED Your countings and corresponding normalised results are reported below. Slide W 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 Results F/mm' 87.3 269.8 128.7 231.5 123.6 191.1 82.1 66.2 259.7 66.9 404.1 49.7 200.8 338.6 175.4 Norm. Results 0.96 0.84 0.84 0.99 0.89 1.06 1.06 0.73 0.87 0.57 0.90 0.82 1.26 1.19 0.98 Normalised Mean : Coefficient of Variation : Group 0.93 18.4 1 Your individual position is plotted on the following diagram: Group 1 : Counters with normalised mean in the range 0,75 - 1,25 and coefficient of variation below 30 % Group 2 : Counters with normalised mean in the range 0,5 - 1,5 and coefficient of variation below 50 % Group 3 Counters with normalised mean outside the range 0,5 - 1,5 or coefficient of variation above 50 % CTD026239 A.I.A. FIBRE COUNTING TRIAL - 1990 INDIVIDUAL POSITION DATA SHEET Name : Mrs JANIS WOODSON REYNOLDS Company : CERTAINTEED Your countings and corresponding normalised results are reported below. Slide 1 2 3 A 5 6 7 8 9 10 11 12 13 14 15 Results F/mm* 60.3 447.9 161.6 359.3 113.2 185.1 32.1 78.2 412.2 132.7 866.2 90.5 200.7 388.0 242.8 Norm. Results 0.66 1.40 1.06 1.54 0.82 1.03 0.42 0.87 1.39 1.14 1.93 1.49 1.26 1.36 1.35 Normalised Mean : Coefficient of Variation : Group 1.18 32.6 2 Your individual position is plotted on the following diagram: GROUP 2 1----- T----r 0.4 **i ---- 1-------1" 1.2 1.6 NORMALISED MEAN Group 1 : Counters with normalised mean in the range 0,75 - 1,25 and coefficient of variation below 30 % Group 2 : Counters with normalised mean in the range 0,5 - 1,5 and coefficient of variation below 50 % Group 3 : Counters with normalised mean outside the range 0,5 - 1,5 or coefficient of variation above 50 % CTD026240 A.I.A. FIBRE COUNTING TRIAL 1990 INDIVIDUAL POSITION DATA SHEET Name : Mr SCOTT HOWARD Company : CERTAINTEED Your countings and corresponding normalised results are reported below. Slide N- 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Results F/mm* 126.1 745.2 374.9 141.5 338.6 93.0 119.7 228.0 276.9 673.3 77.7 289.5 378.6 250.4 127.4 Norm. Results 1.38 2.33 2.46 0.61 2.44 0.51 1.55 2.53 0.93 5.78 0.17 4.76 2.37 0.88 0.71 Normalised Mean Coefficient of Variation : Group 1.96 80.6 3 Your individual position is plotted on the following diagram: isfh ~r T 0.4 0.8 1.2 1.6 NORMALISED MEAN Group 1 Counters with normalised mean in the range 0,75 - 1,25 and coefficient of variation below 30 % Group 2 : Counters with normalised mean in the range 0,5 - 1,5 and coefficient of variation below 50 % Group 3 : Counters with normalised mean outside the range 0,5 - 1,5 or coefficient of variation above 50 % CTD026241 LABORATOIRE D'HYGIENE ET DE CONTROLE DES FIBRES MIN6RALES 10, rue de la P6pini6re, 75008 PARIS T<l. (1) 45 22 12 34 - Fax 42 94 98 86 RESULTS OF THE A.I.A. FIBRE COUNTING TRIAL - 1990 - MARCH 1991 Director: Alain BAUJON Report prepared for: ASBESTOS INTERNATIONAL ASSOCIATION 10 rue de la P6piniere 75008 PARIS FRANCE Association rgie par la loi de 1901 CTD026242 LABORATOIRE D'HYGIENE ET DE CONTROLE DES FIBRES MINeRALES 10, rue de la P6pini6re, 75008 PARIS T6I. (1) 45 22 12 34 - Fax 42 94 98 86 March 1991 RESULTS OF THE A.I.A. ASBESTOS COUNTING TRIAL - 1990 - 1/ - INTRODUCTION : The 1990 slide exchange for asbestos fibres counting is the third one organized by the Asbestos International Association (A.I.A.) among its company members where microscopists have to evaluate asbestos samples taken at the workplace in factories or at production sites. The main objective of this slide exchange is to improve the reliability and the repeatability of the fibre countings, as it is recognised that large variations in the results can arise if the method is not correctly applied. The same procedure as for the previous A.I.A. exchanges was applied (as approved by the A.I.A. Dust Advisory Panel, and participants have been classified categories of performance. slide into 3 Association r4gie par la loi da 1901 /. CTD026243 2.- 2/ - PROCEDURE : Asbestos dust samples were collected in various manufacturing plants (abestos-cement, friction materials, textiles and mines) to cover the main applications of this raw material, and 4 laboratory samples were prepared from non-asbestos fibres. Samples were taken on 47mm membranes cut after sampling into 4 equal parts and clarified according to the recommended acetone-triacetin method (see RTM1). The method allows to prepare 4 equivalent sets of 15 slides, assuming that each 1/4 sector of a filter has the same fibre density. This criteria was first checked by the L.H.C.F. prior to circulation of the slides. Furthermore, the validity of this procedure was further checked using the counting results found by the group 1 counters. The 4 sets of slides were then circulated among the A.I.A. association members which in turn were responsible for organizing their own slide distribution among the participants. All the counting results were returned to the L.H.C.F. according to a standard format. 3/ - PARTICIPANTS : 17 countries took part in the exchange involving counters from 40 laboratories whose experience in fibre counting was evaluated through a questionnaire. 1035 results from 70 individual counters have been collected . Participants were asked to count the slides according to the method which they normally use or which is required by their national regulation. 60 counters reported to apply the A.I.A. RTM.1 method or a similar home one while 7 counters reported to use the NIOSH method as a standard. The surface area of the graticules used for counting are distributed as follows: Graticule Area mm2 < 7500 7500 - 8200 8200 (0=1OO2 mm) 10000 > 10000 Number of counters 5 56 72 CTD026244 3 4 / - FIBRE LOADINGS 11 out of the 15 samples vrere taken from asbestos factories or mines, and 4 asbestos-free samples were mounted with rockwool, ceramic fibres and wollastonite. Product group categories related to the samples and their corresponding mean fibre density (fibres per mm2), are given in the following table: SAMPLE N 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 PRODUCT GROUP MEAN DENSITY FIBRES / mm2 Friction Material Wollastonite Asbestos Mine Asbestos Board Rockwool Asbestos Mine Ceramic Fibres Friction Material Asbestos Cement Friction Material Asbestos Cement Asbestos Board Asbestos Cement Ceramic Fibres Asbestos Cement 91 321 152 234 139 181 77 90 297 116 448 61 159 285 179 CTD026245 5 / - RESULTS : 4.- 5.1. Data Processing : The following calculations are used to classify the participants into 3 different groups: 1) A provisional mean of the counting results for each sample is calculated. 2) Results are converted to normalized values (1) by reference to the mean of the counts. 3) Counters are then provisionaly classified into 2 groups : A - Those whose normalized mean (1) is outside the range 0.5-1.5 OR whose coefficient of variation (1) is greater than 50 %. B - All the others. 4) Steps 1 and 2 are repeated without the A counters results. 5) Finally, all the counters (including category A) are classified into 3 groups as shown on diagram 1: CLASSIFICATION OF COUNTERS ACCORDING TO THEIR NORMALIZED MEAN AND COEFFICIENT OF VARIATION oz p h 6Ho b. NORMALISED MEAN Diagram n 1 (1) See Chap.8 for definitions CTD026246 .../... 5.- 5.2. Interpretation of the results : 5.2.1. - This two steps procedure gives a better evaluation of the mean counts, excluding those counters whose performance appears to be largely outside the range accepted by the D.A.P. 5.2.2. - GROUP 1 defines counters whose performance appears to be quite good in view of the experience accumulated in this field; 5.2.3. - GROUP 2 is intermediate either because the coefficients of variation (CV's) are too high or the counts are too far away from the mean and in some cases for both reasons. 5.3. Final results : The position of each individual counter is indicated on Diagram n 2 below showing the X.Y. plotting of their normalized means and coefficients of variation (CV's). - 31 counters (44.9 %) have been classified in GROUP 1 - 26 counters (37.7 %) have been classified in GROUP 2 - 13 counters (17.4 %) have been classified in GROUP 3 POSITION OF THE 70 COUNTERS ACCORDING TO THEIR NORMALIZED MEAN AND COEFFICIENT OF VARIATION NORMALISED MEAN Diagram n 2 .../... CTD026247 6.- 6/ - STATISTICAL ANALYSIS OF THE DATA : Some further informations are presented in this chapter to allow a better understanding of the evaluation procedure. 6.1. Distribution of the 1035 normalized values : Diagram n 3 shows that 76.6 % of the results are within 50 % of the expected mean and 46.9 % within 30 %. DISTRIBUTION OF 1035 NORMALIZED RESULTS NORMALIZED RESULTS Diagram n 3 A few up to 6 times and therefore variations of counts are much too high or too low, raising or falling down to 1/10 of the expected mean cannot be considered within the acceptable the method. CTD026248 .../... 7.- 6.2. Normalized mean of a counter : This parameter shows the trend for a counter either to count higher or lower than the mean of his colleagues. However it is of limited interest if isolated from any information on the variability of the same counter (coefficient of variation - see chap.6.3) With regard to the mean, diagram n 4 shows that 59 out the of 70 (84.3 %) counters' performance are within the range 0.50-1.50 and 47 (67.1 %) are within the range 0,75 - 1.25 . DISTRIBUTION OF NORMALIZED MEANS Diagram n 4 .../... CTD026249 8 .- 6.3. Coefficient of variation (C.V.) : As reported under chapter 6.2, the individual coefficients of variation are very important to classify the counters because even high unsteady countings may sometimes result in a normalized mean figure well within an acceptable range. As shown on diagram n 5, 58 out of the 70 values (82.9 %) are below 50 %, but the proportion of values below 30% drops down to 54.3 % (38 values). DISTRIBUTION OF COEFFICIENTS OF VARIATION Diagram n 5 .../ OTD026250 9 .- 7/ - NON-ASBESTOS FIBRES SAMPLES 4 samples in each set of slides are non-asbestos. They were prepared articially in the laboratory from rockwool, ceramic fibres and wollastonite. As their number is small,a cautious interpretation of the results is required. It must be reminded that optical microscopy does not allow to discriminate asbestos from non-asbestos fibres and that every fibrous particle should be counted as a "fibre". However, some counters obviously ignore the fibrous particles which morphology does not meet the typical asbestos one. Despite the small number of non-asbestos samples, it was possible to classify the counters according to the WHO-EURO-MMMF criteria which applies to the mean only: GROUP 1 : GROUP 2 : GROUP 3 : Counters having 75% of their counts in the range 0.70 - 1.70 Counters having 75% of their counts in the range 0.55 - 2.20 Counters having more than 25% of their counts outside of the range 0.55 - 2.20 Correspondingly, the counters fall into the following groups : - 50 counters (71 . 5%) have been classified in GROUP 1 - 5 counters (7 . 1%) have been classified in GROUP 2 - 15 counters (21 . 4%) have been classified in GROUP 3 CTD026251 t 10 Diagram n6 shows the distribution of 280 normalized results (all non asbestos counts): 79.1% of the results are within 50% of the expected mean, and 64.4% are within 30%. The distribution is significantly better than that of all fibre type countings (diagram n3). Such a finding is coherent with the fact that it is often easier to count non asbestos fibres than asbestos fibres because their diameters are generaly greater, and countings fields containing agglomerates or split fibres are very rare. DISTRIBUTION OF 280 NORMALIZED RESULTS (NON ASBESTOS SAMPLES ONLY) NORMALIZED RESULTS Diagram n 6 8/ - DEFINITIONS : (1) Normalized result = ratio of a counting result to the mean of counts on the samples. (2) Normalized mean (inter-lab index) = average of the counter's normalized results. It is an indication of systematic overestimation or underestimation. (3) Coefficient of variation (intra-lab index) = the standard deviation (of a counter's normalized results) divided by the normalized mean. It is an index of dispersion of counter's results. CTD026252