Document YGx57m0m2GN2n00bpqoL160wD

file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 1 1 2 BILLY B. SMITH, ET UX : 14TH JUDICIAL DISTRICT COURT 3 VS. NO. 92-2915 : PARISH OF CALCASIEU 4 HARTFORD ACCIDENT AND INDEMNITY COMPANY, ET AL : STATE OF LOUISIANA 5 JOHN W. CRAWFORD, JR., ET UX : 14TH JUDICIAL DISTRICT 6 COURT VS. NO. 92-3098 : PARISH OF CALCASIEU 7 HARTFORD ACCIDENT AND 8 INDEMNITY COMPANY, ET AL : STATE OF LOUISIANA 9 --------------------------------------------------------- 10 DEPOSITION OF OTTO WONG, Sc.D., F.A.C.E. 11 --------------------------------------------------------- 12 13 PURSUANT TO SUBPOENA and the Louisiana Code of Civil Procedure, the above-entitled deposition was taken on 14 behalf of Plaintiffs at 55 Madison Street, Suite 510, Denver, Colorado, on Wednesday, July 28, 1993, at 9:55 15 a.m., before Rebecca L. Semeyn, Registered Professional Reporter and Notary Public within Colorado. 16 17 18 19 file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (1 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 20 21 22 23 24 25 2 1 APPEARANCES: 2 For the Plaintiffs: HERSCHEL L. HOBSON, ESQ. 3 Law Offices of Herschel L. Hobson 2190 Harrison 4 Beaumont, Texas 77701 5 For the Defendants: KENNETH R. SPEARS, ESQ. 6 Jones, Tete, Nolen, Hanchey, Swift, Spears & Massey 7 P.O. Box 910 Lake Charles, Louisiana 70602-0910 8 Also Present: Daniel Teitlebaum 9 --------------------------------------------------------- 10 INDEX 11 EXAMINATION PAGE 12 July 28, 1993 file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (2 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 13 By Mr. Hobson 14 By Mr. Spears 15 3 -- 16 17 EXHIBITS INITIAL REFERENCE 18 1 Dr. Wong's curriculum vitae 6 19 2 List of documents regarding John Crawford, 124 20 Billy Smith and depositions from Ken Spears 21 3 Statement submitted to the OSHA benzene hearing by Otto Wong 22 (Attached to original and copy transcripts.) 23 125 24 25 3 1 PROCEEDINGS 2 OTTO WONG, Sc.D., F.A.C.E., 3 having been first duly sworn, was examined and testified 4 as follows: 5 EXAMINATION file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (3 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 6 BY MR. HOBSON: 7 Q. Would you introduce yourself please, sir. 8 A. Otto Wong. 9 Q. And how are you employed now, Dr. Wong? 10 A. I'm the chief epidemiologist at Applied Health 11 Sciences. 12 Q. And what is Applied Health Sciences, please? 13 A. It's my company set up to do research and 14 consulting mainly in epidemiology. 15 Q. And what kind of a company is it? Is it a 16 corporation? 17 A. Yes, it is. 18 Q. And are you the only shareholder? 19 A. Yes. 20 Q. And where is the corporation located? 21 A. It's located at 181 Second Avenue, Suite 628, 22 San Mateo, California 94401. 23 Q. Are there any other professionals who work with 24 you at Applied Health Sciences on a full-time basis, 25 Dr. Wong? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (4 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 4 1 A. Not on a full-time basis. 2 Q. Are there any full-time employees of Applied 3 Health Sciences? 4 A. I'm the only full-time employee. 5 Q. Who are any part-time scientists who are 6 associated with Applied Health Sciences? 7 A. I have a number of people who work with me on 8 an as-needed basis. I have a Ph.D. in biostatistics, his 9 name is David Ragland; and a master level biostatistician, 10 her name is Nancy Gordon. 11 Q. And could you spell her name, please. 12 A. G-o-r-d-o-n. I have a nurse/programmer, her 13 name is Linda Hebbar, H-e-b-b-a-r. 14 Q. Is it H-a-p, as in Paul? 15 A. H-e-b, as in boy. 16 Q. Anyone else who has associated with you at 17 Applied Health Sciences in the past? 18 A. And I have some data entry research assistants. 19 Q. Would it be fair to characterize the data entry 20 people as nonprofessionals? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (5 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 21 A. Technical people. Nonprofessional. 22 Q. How long has Applied Health Sciences been in 23 business? 24 A. Since January, 1991. 25 Q. Did you have any professional activities since 5 1 leaving Environmental Health Associates besides Applied 2 Health Sciences? 3 A. I assume you're talking about "professional" 4 relating to epidemiology or health? 5 Q. Correct. 6 A. No. 7 Q. So basically you left EHA and started Applied 8 Health Sciences and that covers all your professional 9 work? 10 A. That's correct. 11 Q. What kinds of work do you have ongoing at 12 Applied Health Sciences? 13 A. We have a number of long-term projects ongoing. 14 I can give you some examples. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (6 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 15 Q. Yes, I would like those. 16 A. We are finishing up a long-term epidemiologic 17 study of gasoline workers for the American Petroleum 18 Institute. I am doing a cohort mortality study for USG of 19 people exposed to man-made mineral fibers. 20 I'm working on a cohort mortality study for the 21 Styrene Information Research Center for people exposed to 22 styrene. And I also have some ongoing 23 activities--consulting activities with Mobil and Chevron. 24 These are the projects that I can recall right now. 25 Q. Have there been any projects that you have 6 1 completed while you've been at Applied Health Sciences 2 that are not in the group you just described for me? 3 A. We completed a project for HIMA, H-I-M-A, 4 Health Industry Manufacturers Association. It was a 5 project that carried over from the old EHA days. It was a 6 study of workers exposed to ethylene oxide. 7 Q. Was that published? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (7 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 8 A. Yes, it was published last year in the British 9 Journal of Industrial Medicine. 10 Q. I see you have your CV with you. Does that 11 include your publication list? 12 A. Yes. 13 Q. And it would include the ethylene oxide 14 publication, then? 15 A. Yes. 16 Q. Any other projects that you've completed since 17 you got to Applied Health Sciences? 18 A. Well, I'm drawing a blank. I work very hard, 19 but I'm drawing a blank. What else did we finish? 20 Q. Lawyers know you can make more money if you 21 continue to work on projects than if you finish them. 22 A. I guess some epidemiologists think the same 23 too. 24 I didn't have any major projects that we 25 finished during the last couple of years at this point. 7 1 Q. When you do consulting for Mobil and Chevron as file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (8 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 2 an employee of Applied Health Sciences, which you assume 3 it is, what rate do you charge them? 4 A. 230. 5 Q. Dollars per hour? 6 A. Right. 7 Q. And when you give deposition work like you're 8 doing today, how much do you charge for that? 9 A. 320. 10 Q. Why the difference? 11 A. In general, work related to litigation--they 12 are pretty short-term. It doesn't really provide a 13 long-term stability to the business. 14 Q. Any other reason? 15 A. That's the major reason. 16 Q. Do you charge $320 an hour to prepare for a 17 deposition also? 18 A. Yes. 19 Q. Can you give me some idea of what percentage of 20 your time you spend working on litigation-related 21 activities? And by that I mean not including your 22 epidemiological studies that you work on. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (9 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 23 A. I would say over the last few years it would 24 average to about 30 percent. 25 Q. By "the last few years," is that since you've 8 1 been at Applied Health Sciences? 2 A. Yes. 3 Q. Has the trend been more or less in the last few 4 months--say the last six months? 5 A. I would say it stays about the same. 6 Q. So if you charge more per hour for litigation 7 work than you do for your other kinds of activities, the 8 amount of income you would generate from litigation would 9 be more than 30 percent of your total income? 10 A. No. Let me clarify this. I think you asked me 11 how much I charge for consulting work for Mobil and 12 Chevron. I think that was the specific question. 13 Q. Yes, sir. 14 A. When we do a major study, sometimes we work on 15 a fixed fee basis--a fixed budget. So if we do an 16 efficient job, we can naturally come out ahead. So it's file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (10 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 17 very difficult to just look at the hourly wage to 18 determine how much we make. 19 Q. To make sure I understand what you're saying, 20 if you work on an epidemiology project that has a fixed 21 payment for that project, you can actually make more than 22 $320 an hour? 23 A. If we are efficient. 24 Q. Now, you say that USG was the sponsor of one of 25 the studies. Is that United States Gypsum or what used to 9 1 be United States Gypsum? 2 A. I really don't know. Since I have known them 3 it has always been USG. 4 Q. Where are they located? 5 A. In Chicago. 6 Q. And the American Petroleum Institute, you've 7 done work for them in the past besides the gasoline 8 workers project, correct? 9 A. Yes. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (11 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 10 Q. And who is the styrene project being sponsored 11 by? 12 A. The Styrene Information Research Center. 13 Q. And who is the Styrene Information Research 14 Center, as you know them? 15 A. It's an organization, I guess, sponsored by the 16 industry. 17 Q. And do you know who the member companies are in 18 that organization? 19 A. I know some. 20 Q. Who do you know? 21 A. Dow Chemical, Chevron, some smaller companies, 22 boat manufacturers and so on. 23 Q. Does your study look primarily at styrene 24 manufacturing or does it include styrene consumption? 25 A. Can you clarify "consumption." 10 1 Q. Yes. Converting styrene into another product 2 away from the manufacturing side. 3 A. We look at mainly industrial user's styrene. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (12 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 4 Q. For instance, would it include a styrene 5 butadiene rubber plant? 6 A. No. 7 Q. Or a polystyrene manufacturing facility at a 8 different location than the styrene manufacturing 9 facility? 10 A. No. 11 Q. And what kinds of users come to mind that 12 you've been looking at for styrene? 13 A. Boat makers--manufacturers of boats. Any 14 reenforced plastic manufacturers. 15 Q. You say that is a cohort mortality study? 16 A. Yes. 17 Q. What makes up the cohort? 18 A. Workers in the reenforced plastic industry who 19 are exposed for a minimum of three months during a 20 specific period of time, and I just don't recall the time. 21 Q. Where are you in that project? How close to 22 completion? 23 A. Toward the end of this year, knock on wood. 24 Q. And the cohort mortality for USG, how far along file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (13 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 25 are you on that? 11 1 A. We just started. 2 Q. And the gasoline workers study for the API? 3 A. In fact, I'm writing the report now. Let me 4 back up a little bit. We issued a report on the cohort 5 analysis part two years ago, and subsequent to that we did 6 a case control study and now I'm working on the case 7 control study reports. 8 Q. What are the cases that are going into the case 9 control study? 10 A. All the leukemia cases, kidney cancer cases and 11 multiple myeloma cases. 12 Q. And you say the final report of the case 13 control study is being written now? 14 A. Yes, sir. 15 Q. And when do you expect to have that final 16 report generated? 17 A. I would say within two or three months. 18 Q. Who's the API coordinator for that project? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (14 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 19 A. Dave Monogillo, M-o--you put me in a very 20 difficult situation. I better know how to spell his name. 21 M-o-n-o-g-i-l-l-o. 22 Q. Where is he located? Is he an API employee? 23 A. He's at API in Washington, D.C. 24 Q. Do you have a company coordinator for one of 25 the member companies? 12 1 A. The chairman of the task force is Will Bailey 2 at Chevron in San Francisco. 3 Q. How do you as an epidemiologist go about 4 estimating exposures for cohorts that you study? 5 A. Well, take the API study--the gasoline study. 6 We--"we" meaning myself and also the EHA, because that was 7 started at EHA about nine years ago. We have the 8 epidemiologic portion of the study and another group of 9 investigators handled the exposure assessment part, 10 although we worked together with them. 11 The person who was in charge of the components file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (15 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 12 is Dr. Tom Smith. He used to be at the University of 13 Massachusetts. Now he's at Harvard University. 14 Q. That's a little more specific than I wanted to 15 know. 16 I'm interested in knowing how you, as an 17 epidemiologist, can go about doing these mortality studies 18 or morbidity studies and assessing exposure as a part of 19 them as a general proposition. Can you tell me how that's 20 done. 21 If you were going to teach a class--a bunch of 22 college students how to assess exposures and look at those 23 responses for a mortality or morbidity study, how would 24 you begin that lecture? 25 A. What we do is we collect all the employment 13 1 records of the cohort members in the study. We 2 computerise the records. We print out a list of 3 departments, job titles, and then we would go to people 4 who are knowledgeable about historical exposure, and that 5 would be company industrial hygienists, engineers and also file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (16 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 6 personnel managers--personnel employees, and work with 7 them to determine what exposures were entailed in each 8 department for each job title. 9 They would, in turn, rely on whatever 10 industrial hygiene data that are available, and the 11 historical records of engineering changes, production 12 records and so forth. 13 Q. Can you tell me, from your experience, if you 14 know of any instance where in a project that you've been 15 involved with, the historical exposure information has 16 been validated by talking to the workers who were actually 17 exposed? 18 A. I don't know whether we actually have gone back 19 and talked to some workers and validated that, but in some 20 of our own projects we have input from employees, 21 especially long-term employees, who remember the physical 22 setup of the process and the job description, and so on. 23 Q. I'm trying to ask a very narrow question and 24 I'd appreciate your best recollection. 25 The hourly workers who are the subject of the file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (17 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 14 1 study who would have had the historical exposure you're 2 trying to evaluate, can you give me any instance where you 3 have attempted to validate historical exposures with those 4 workers who actually received the exposures? 5 A. We have input from those employees. 6 Q. How? 7 A. By talking to them in terms of--for example, we 8 see a job title that's no longer used anymore. Let's say 9 the job title was used in the '50s or '40s even. Nobody 10 remembers what the job title was. We would talk to some 11 old, long-time employees. We even pay some retirees to 12 come back and work with us. 13 Q. And so I'm clear, are these hourly workers who 14 have remained hourly workers or are they supervisors who 15 at one time might have been hourly workers or might have 16 some knowledge? 17 A. Both. 18 Q. And the "we" you use, is that Dr. Otto Wong or 19 is that somebody else? 20 A. "We" meaning myself, plus the company file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (18 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 21 industrial hygienist. 22 Q. And so you yourself have actually sat down with 23 hourly workers and validated historical exposures? 24 A. I hate to use the word "validate" because we 25 have input from them in the process. As I said, I don't 15 1 remember--after we complete our assessment, I don't 2 remember that we have actually gone back to some other 3 employees to validate it independently. I don't think we 4 have done that. 5 Q. That's the question I'm trying to ask. I do 6 mean to use the word "validate." 7 "Validate," I think, has the conception with it 8 of something you do after the fact to make sure that your 9 assumptions were correct. Would that be how you would use 10 "validate," or that your information was correct? 11 A. I guess in a sense we have not gone back to the 12 employees for validation. What I was trying to do is 13 point out that we do have input from them to start with. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (19 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 14 Q. Why don't you do that? Why don't you go to 15 employees who are the actual subjects of the exposure 16 categorizations and validate that the exposure category 17 you placed their work in measures up with what they tell 18 you? 19 A. I think it's a practical consideration. The 20 area that we have concerns about would be exposures a long 21 time ago that we don't have industrial hygiene records, 22 and most people--active people don't know anything about 23 those jobs. So we have to identify some long-term 24 employees who actually were exposed. 25 There aren't that many people that would fall 16 1 into that category, and they're very difficult to find. 2 So whoever we could contact, we make use of the 3 information in an input process. There aren't that many 4 people that we can go to, that's all. I think that's the 5 major reason why we don't have the luxury of validating it 6 again. 7 Q. Have you looked? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (20 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 8 A. Oh, yes, we did. 9 Q. You've looked actually and considered 10 validating the historical exposures? 11 A. Right. 12 Q. Have you published that? 13 A. For the API study--the gasoline study 14 publication is coming out under Tom Smith, and I'm one of 15 the coauthors in that publication. 16 Q. And that publication will address the 17 validation of historical exposures or the inability to do 18 it? 19 A. As I said, we did not have validation. We have 20 input from long-term employees. 21 Q. That's what I'm trying to get to, Dr. Wong, is 22 the validation of these historical exposures. 23 A. Well, I told you we don't have validation and I 24 gave you the reason why we could not have the luxury of 25 validating those data. 17 file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (21 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 1 Q. And then I asked you, Have you ever documented 2 your inability to validate historical exposures, and you 3 mentioned Dr. Smith's study, which then I think you said 4 does not address that. 5 So are there any written documents, published 6 or otherwise, that you can direct me to where you have 7 considered the validation of the historical exposures? 8 A. I don't know whether we have documentation of 9 that or not. I can't give you an answer. 10 Q. In considering cause and effect relationships, 11 the response which comes from the historical exposure data 12 is very important, is it not? 13 A. Yes. 14 Q. Would you agree with me that in exposure 15 assessments historically that is the weakest part of the 16 epidemiological study? 17 MR. SPEARS: That being what, Herschel? 18 MR. HOBSON: The analysis of the historical 19 exposures. 20 A. It's one of the potential limitations. I don't 21 know if I would label it as the weakest part. Weak as 22 compared to what? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (22 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 23 Q. (BY MR. HOBSON) Well, death certificates, for 24 instance. They're hard and fast pieces of paper that you 25 can get and they say what they say and that's what you 18 1 have to go by, right? 2 A. Well, yes and no. If you get the right death 3 certificate, yes. It's a piece of paper and you have to 4 go by that. But a lot of times the information on the 5 death certificate may be slightly different from what you 6 have on the employee. 7 The social security number may not be a perfect 8 match, the birthday may not be a perfect match. So 9 sometimes you have to use your judgment to determine 10 whether that is the right death certificate or not. 11 So when you get to the subtleties of every 12 study, there are always some limitations. 13 Q. But the limitations on historical exposures are 14 much greater than deciding whether or not you're going to 15 use a particular death certificate, aren't they? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (23 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 16 A. I would say they are equally important. I 17 don't want to give the impression that historical exposure 18 assessment is not important, but at the same time, death 19 certificates are very important too. 20 Sometimes the difference of two or three deaths 21 due to a specific cause can change the finding of a study. 22 So I would say they're both very important. 23 Q. When you design an epidemiological study, you 24 come up with a protocol and study design that you're going 25 to follow; is that right? 19 1 A. Yes. 2 Q. Is it important not to go back and amend the 3 protocol after you begin the study and begin to actually 4 analyze data? 5 A. Unless there are some major unanticipated 6 changes in the data that requires some modification. 7 Q. What would be some considerations that you have 8 used in the past to modify protocol under those 9 circumstances? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (24 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 10 A. If the sample size turned out to be quite 11 different, then you have to modify analyses. I can't 12 think of any instances that I modified a protocol in a 13 major way; not right at this moment. 14 Q. If exposure categorization is part of the 15 original protocol, do you also define what exposure 16 categorizations are going to be in the protocol initially? 17 A. I would think so. 18 Q. Are there times when you change how that 19 protocol for exposure standpoint will be dealt with after 20 you begin to collect data? 21 A. I don't know how to answer these kind of 22 general questions. If you give me a specific sample, I 23 can probably give you a more specific answer. 24 Q. I'm trying to find out about your experiences. 25 How important is it to determine what protocol 20 1 will be utilized for analyzing exposure categories in the 2 beginning before data is collected as far as how it file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (25 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 3 affects the outcome of the study? 4 A. I think every investigator would try to follow 5 the protocol as closely as possible. 6 In your protocol you plan certain analyses, but 7 if the numbers turn out to be very small in certain 8 categories, then you may have to lump things together and 9 that may be a deviation from the protocol. It may not 10 address certain questions, but still that may be the only 11 way you can do some analysis based on the data that you 12 actually have. 13 Q. Will a carefully thought-out protocol consider 14 those possibilities in the beginning to know what lumping 15 together might be necessary if the numbers are small? 16 A. Sometimes we can anticipate some, I guess, 17 unforeseen changes, but I guess nobody can predict the 18 future. 19 Q. You can't get it always right the first time, 20 but will you agree with me you can anticipate many of 21 these data collection problems, and those that can be 22 reasonably anticipated should be addressed in the initial 23 protocol? 24 A. Sometimes, yes. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (26 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 25 Q. Are there times when you can reasonably foresee 21 1 a problem and you would not address it in the protocol? 2 A. You mean--you're asking me very general 3 questions. I just don't know how to answer them. Again, 4 if you give me some specific samples, I would be happy to 5 come up with a specific answer. 6 Q. So that's a question that's too general for you 7 to give me an answer to? 8 A. It's very broad. 9 Of course everybody would follow the protocol. 10 Everybody is trying to anticipate any unforeseen changes 11 and try to deal with that, but it's not a black and white 12 kind of decision sometimes. 13 Q. As a person who's skilled in statistics, what 14 kinds of problems can you get into by changing protocol 15 after you start collecting data? 16 A. Depends on what kind of changes we are talking 17 about. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (27 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 18 Q. Give me some examples that you have in your 19 mind. 20 A. Well, for example, if you go and do a study and 21 you anticipate a much larger cohort than you actually 22 have, you may have to modify your conclusion because the 23 sample size would not give you as much confidence as you 24 would like. It may not even answer the question that you 25 start with. That's a very common scenario. 22 1 You go into industry, you do a study, you 2 thought you would have a large cohort and you end up with 3 a much smaller cohort. 4 Q. I guess maybe I'm having a difficult time 5 phrasing a question so we can communicate. 6 What I'm trying to find out about is if we talk 7 about data collection, how can changing the protocol in 8 the data collection process after data collection has 9 begun affect the outcome of a study? 10 A. Depends on what the change is. I don't know 11 how to answer your question more specifically than this file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (28 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 12 unless you give me a more specific example. 13 Q. Okay, let's try a couple, then. Let's talk 14 about in evaluating exposure categories and you begin to 15 gather your death certificate information--assuming we are 16 talking about a mortality study--and partway through the 17 study you change your criteria for exposure 18 categorization. 19 A. In terms of what? In terms of duration of 20 exposure? In terms of chemicals? In terms of 21 departments, job titles? 22 Q. All of those? 23 A. All of those. 24 Q. Or any one of them. 25 What kinds of questions does that call to your 23 1 mind, then, about any results that you can gain from such 2 a study? 3 A. Well, I assume that if we modify the definition 4 of exposure--I'm talking about when you go out and file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (29 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 5 assemble a cohort. Regarding exposure, you have to rely 6 on job titles or departments to determine whether someone 7 was exposed or not. If you modify that, then you better 8 have some reason why you're modifying that. 9 The previous definition of exposure must be 10 incorrect or incomplete based on incomplete information. 11 So to me if you modify the definition of exposure in terms 12 of job titles or departments, then I would assume you have 13 better information now as opposed to before. 14 If you modify the duration of exposure--a lot 15 of times we would say we would only study people with at 16 least three months of exposure, six months of exposure or 17 a year of exposure. If you modify that, that may or may 18 not change your study hypothesis at all. 19 Sometimes you have to relax the requirement for 20 minimum exposure so that you have a larger sample size. 21 Sometimes you have to decrease that to make it smaller to 22 make it more manageable. 23 Q. Let's take a hypothetical and see if I can work 24 this through in my mind. 25 Let's say that I'm a large industry and I come file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (30 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 24 1 to Dr. Wong to do a study based on a chemical exposure in 2 my industry and I want there to be--this will be a 3 mortality study--a historical perspective mortality study, 4 and it will have, as a component, exposure categories so 5 that we can get a measure of dose response. Does that 6 make sense so far? 7 A. Yes. 8 Q. You design a protocol for such a study which 9 includes exposure categorization. Data is begun to be 10 collected. You start collecting the death certificates 11 and you begin to do analyses and you do a preliminary 12 analysis of what data you have and you give it to me 13 because I'm the industry, I'm doing the exposure 14 categorization. Is there a problem with that? 15 A. Can I interrupt? 16 Q. Sure. 17 A. When you say, do a preliminary analysis, do you 18 mean mortality analysis actually with results such as 19 standardized mortality ratios? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (31 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 20 Q. Maybe not that sophisticated. Maybe I collect 21 30 percent, 40 percent of the death certificates you 22 expect to find and you start to see what could be a trend. 23 A. Okay. 24 Q. And you come to me, the person who's paying for 25 the study, and say, Well, here's something I see. 25 1 Is that a problem? Can it be a problem? 2 A. I think that's a problem. 3 Q. Why? 4 A. I can speak for myself. 5 Q. Yes. 6 A. I don't think I would ever want to present 7 preliminary analyses based on 30 percent data to the 8 sponsor. 9 Q. Why? 10 A. Number one, I don't think they should be that 11 intimately knowledgeable about some ongoing research in 12 terms of results. They should be aware of the problems 13 and so on, but they should not know preliminary analyses file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (32 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 14 based on 30 percent of the data. 15 And besides, I would never analyze a study 16 based on 30 percent data. It's just not--it's not 17 efficient. It doesn't tell you anything, to start with. 18 Q. I guess there is some level, though, at which 19 you would begin to do some preliminary analysis--some 20 level of data collection. 21 A. The only scenario I can think of is we have 22 about 90 plus percent of the data. When I say that, 23 mostly I mean death certificates; that we would start 24 running some preliminary analysis, make sure the computer 25 programs work and so on. 26 1 Now, that doesn't mean we are not going to 2 modify the data as we get more death certificates coming 3 in during that period of, so-called, preliminary analysis, 4 but I don't think the results would change that much based 5 on some small percentage of additional death certificates. 6 Q. Would you give that kind of an analysis to the file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (33 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 7 sponsor? 8 A. We may present that, yes. 9 Q. What if then the sponsor said, Well, we want to 10 change the definition of the exposure protocol? 11 A. I would ask why. 12 Q. Why would that be of any concern? 13 A. Certainly you would not change your criteria of 14 exposure or any other classification scheme of your 15 cohorts after you look at the results. That doesn't make 16 a whole lot of sense. 17 Q. Why? 18 A. If a study has been going on for six, seven 19 years, the exposure scheme should be worked out completely 20 by then. That really raises some suspicion, why somebody 21 would change the exposure categorization, especially if by 22 doing so you've also changed the result. 23 Q. Let me come at the problem that I'm trying to 24 get your opinion on a little bit differently. 25 Let's assume for a moment that I'm going to 27 file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (34 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 1 hire you to do a study for me--I'm the sponsor--a large 2 industry-specific chemical historical perspective study 3 with exposure categories, and I am interested in whether 4 or not there is a cause and effect relationship between 5 the chemical and a particular disease category. And I 6 know, because I've got a little bit of knowledge about 7 epidemiology, that the more clearcut a dose response is, 8 then the stronger the study results are likely to be. 9 Does that make sense? 10 A. Right. 11 Q. And I want you to go out and collect the death 12 certificate information, give me, then, an idea of where 13 these particular kinds of suspicious diseases lie, and I 14 want to be able then to manipulate the exposure data to 15 soften the dose response curves. 16 How do you guard against that? 17 MR. SPEARS: As I understand the question, 18 you're asking him in a hypothetical situation that this 19 industry wants to commit a fraud, basically. Is that what 20 you're trying to say, Herschel? 21 MR. HOBSON: Sure. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (35 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 22 MR. SPEARS: How would he respond to that if he 23 knew you were? 24 MR. HOBSON: No. I'm asking how he, as a 25 scientist who's going to be dealing with companies who 28 1 might want to commit a fraud-2 Q. (BY MR. HOBSON) --how do you design in your 3 protocols ways to prevent that when the fraud is playing 4 with the exposure data, because now the industries 5 typically do provide you with a lot of the exposure 6 information, correct? 7 A. Yes, but they--let me go back and tell you how 8 we retrieve exposure information from the industry. 9 Normally we don't get exposure data on a 10 specific individual. We get exposure information on 11 departments, jobs, locations, and also calendar years, so 12 it's like a multidimensional matrix. And we ask them for 13 whatever industrial hygiene data they have. Not 14 estimates; real data for those jobs for the specific time 15 period, and so on. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (36 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 16 Then, based on that, you should be the one who 17 knows--we don't, usually. If you go back to the '60s or 18 '50s, you don't have that much data or none at all, so you 19 do have to do some kind of estimates, and that's the part 20 we have to talk to old timers. We talk to the company 21 industrial hygienist and we don't rely on the company 22 alone. We rely on workers as well, long-term employees. 23 And the estimate--historical estimate's got to 24 be consistent with the real data too that we have in the 25 latter part of the study. And we do all this independent 29 1 of the mortality data, and we do all the classification 2 before--even before we analyze the data. 3 Q. How do you know that the exposure 4 categorization is done independently of the mortality 5 data? 6 A. As far as we are concerned, it's done 7 independently. 8 Q. How do you know that? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (37 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 9 A. Because we have the mortality data. We never 10 give the mortality data to the companies until the study 11 is done. 12 Q. But isn't it true, Dr. Wong, that many of these 13 industries that you do work for are the kind we have been 14 describing? 15 They already have a big piece of the mortality 16 data. They may not have as many death certificates as you 17 do, but they have a lot of them through the benefit plans 18 and they have some other epidemiology studies that have 19 been done internally or through other consultants or 20 studies done earlier. They have a piece of that already, 21 don't they, many times? 22 A. They have part of the death certificates. Most 23 likely they would have death certificates for their active 24 employees who die on the job and also death certificates 25 for annuitants who are vested in the retirement plan at 30 1 least for the latter part of the study. 2 Q. So while you keep your mortality data from the file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (38 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 3 sponsor, you know going into it that most of these 4 sponsors, if not--let's say most of the sponsors already 5 have a piece of the mortality information in-house, right? 6 A. Right. 7 Q. So that's what I'm asking you now, is how do 8 you guard against that sponsor who wants to, to use 9 Counsel's term, commit a fraud by manipulating the 10 exposure data to make dose responses mushy? What do you 11 do as a researcher to stop that? 12 A. Again, we would look at the consistency of the 13 data. 14 Q. But there is no data. You said for most of 15 these exposures there is no exposure data. 16 A. We have real exposure data for certain parts of 17 the study period. 18 Q. And usually in the last 15 years? 19 A. Starting in the '70s, yes, we do have that. 20 Q. And even the information for exposure in the 21 '70s is somewhat limited in most instances, is it not? 22 A. We always want to have more data than we 23 actually have, but I don't know whether you'd call it file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (39 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 24 limited or not. 25 Q. If you were going to do a statistical analysis 31 1 of the exposure data that most of the sponsoring companies 2 have, say, up to 1980, isn't it true that they don't have 3 enough data to do really meaningful statistical 4 interpretations by job for exposure? 5 MR. SPEARS: Can you--for the record, I want to 6 object unless you identify what companies we are talking 7 about. I think you asked him specifically there. 8 Q. (BY MR. HOBSON) Pick any of them you would 9 like. 10 What I'm trying to find out, Dr. Wong--take a 11 pipe fitter at a refinery. 12 A. Okay. 13 Q. I will pick Mobil in Beaumont. I don't know if 14 you know anything about that or not. I know you studied 15 that group. But in looking at a pipe fitter at Mobil in 16 Beaumont, Texas for the exposure information for 1977-17 A. Right. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (40 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 18 Q. --do you think that there is adequate benzene 19 exposure data to define any given pipe fitter's exposure, 20 or even a group of pipe fitters at that refinery? 21 A. For that specific study, if we are talking 22 about-23 MR. SPEARS: Counsel-24 Q. (BY MR. HOBSON) Counsel said give you a 25 specific. I got one for you. 32 1 A. In that study we did not have any quantitative 2 exposure estimates, as you are well aware of. We didn't 3 use that in our study for the reason that we just don't 4 have all the data that we need. 5 As opposed to--let's take another example. The 6 API gasoline study that I am doing now. In that study API 7 spent a lot of resources. In fact, they contracted the 8 University of Massachusetts to come up with exposure 9 assessment and that's a major component. 10 Q. How much of that was based on hard data as file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (41 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 11 opposed to estimates in 1975, '76, '77? 12 A. I don't recall. I'd have to go back and look 13 at the report. 14 As I said before, that study was under the 15 direction of Dr. Tom Smith. I work with him. I provide 16 him with employment records and so on, but I'd have to go 17 back and take a look. 18 Q. Let's talk about the CMA benzene study. That 19 had exposure categories, did it not? 20 A. Yes, sir. 21 Q. How much of that information is based on hard 22 data? 23 A. I recall we have industrial hygiene data in the 24 '70s. We are talking about benzene, and benzene is one of 25 the things that people measure for a long time relative to 33 1 other substances. 2 Q. Tell me how you mean that, as your perception? 3 A. Starting in the '70s. 4 Q. That was a long time? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (42 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 5 A. Relative to other substances. 6 And our studies start in the mid '40s. Some 7 company's studies started in the '40s and '50s. So I 8 would say about anywhere between 1/3 to 1/2, depending on 9 which company you're talking about, we have real data. 10 Q. For some part of the time? 11 A. Right. 12 Q. If it came down to a location and a person, how 13 much information on benzene is it your perception existed 14 for the CMA study? 15 A. The way we estimated exposure was based on 16 jobs. 17 Q. Right. 18 A. We divide up the job into a number of what we 19 call uniform tasks. 20 Q. You and I have had a chance to talk about the 21 study many times, but I am trying to get more specific 22 this time. 23 Give my your perception, if you would, sir, of 24 how much information on benzene exposure on an individual 25 cohort member for the exposed group in the CMA benzene file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (43 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 34 1 study there was. 2 A. It depends on the period of employment. 3 For someone who was hired since the '70s, I 4 think we probably have pretty good data on that person, or 5 at least on jobs similar to that person. 6 Q. That's a different question. 7 Jobs similar to that person I will ask you 8 about next, but right now I'm talking about the 9 individuals. What is your perception of how much exposure 10 data you had for the individuals in the cohort? And I 11 realize it varied by time, and that's what I'm trying to 12 find out. 13 A. The way we estimate exposure is based on jobs 14 and not based on specific individuals. 15 It may happen that some of the data on that 16 specific individual went into the exposure estimates for 17 that general category of jobs. So to that extent, we may 18 have some specific information on that individual, but 19 more likely than not we have information on jobs in file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (44 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 20 general categories. 21 Q. Did the information on jobs in the general 22 categories--are those categories across different plants 23 or are they specific for each plant that was studied? 24 We can talk about the CMA benzene study. 25 A. It's a little bit of both. 35 1 What they do is--"they," meaning the industrial 2 hygienists. 3 Q. Of the industry? 4 A. Of the industry--and I would participate in 5 those discussions as well--is to look at how comparable 6 the job titles are across the company and also over time, 7 and we reduce those job titles to some uniform comparable 8 job categories. 9 Q. What happens to the confidence of what you're 10 trying to learn about exposure as you combine and combine 11 again and then extrapolate across, as you've just 12 described? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (45 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 13 A. I mean, in general terms, if we are not 14 confident, we would not use it in our analysis. 15 Q. But your confidence would go down, right, as 16 you continue to combine and average across? 17 A. Well, yes and no. 18 Q. The best of all possible worlds would be to 19 know the exposures of each individual every day they 20 worked, agreed? 21 A. That's the ideal world that the epidemiologist 22 would like to have, but that's not going to happen. 23 Q. Doesn't exist? 24 A. No. 25 Q. And I guess a fallback to that would be if you 36 1 knew one day a week of exposure for every individual that 2 was in the cohort, that would not be as good as knowing 3 every day, but better than what you had, right? 4 A. No. I don't think a day a week is necessarily 5 the second best. That one day a week may or may not be 6 representative. It may be better to look at the overall file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (46 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 7 and come up with an average for that category, for that 8 job title. 9 Q. What about for plant to plant? How does that 10 affect your confidence in exposure categories if you're 11 talking about two different manufacturing facilities where 12 the job exposures are being combined? 13 A. We would combine them only if they're 14 comparable. 15 Sometimes the same job title for different 16 companies may involve different job functions and may not 17 be comparable at all. 18 On the other hand, due to different 19 reasons--employment reasons, union contract reasons--they 20 may call the same job by different names. That's why we 21 need input from the industry, from personnel, from 22 industrial hygienists. 23 Q. You were, I think, personally involved in the 24 exposure category determination in the CMA benzene study, 25 correct--at least part of it? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (47 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 37 1 A. Yes. 2 Q. How did you address a question like this, if 3 you did: One plant has benzene workers who wash their 4 hands and wash their clothes with benzene. 5 What does that do to your consideration of 6 exposure categories in the CMA benzene study? 7 A. Are you still talking about differences between 8 plants? 9 Q. However it figures in. I'm not sure that I 10 know how you did consider it and that's what I'm trying to 11 find out. 12 A. Are we talking about the CMA benzene study? 13 Q. We are. 14 A. In that study--I think I mentioned this 15 before--for each job and for each time period we try to 16 divide a job into different components. 17 In fact, we had a list of over 30, what we 18 call, uniform tasks such as taking samples from the 19 stream, analyzing samples, fixing up a leak, washing tools 20 in solvents, benzene. 21 In fact, I remember using solvents or benzene file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (48 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 22 to wash tools as being a specific task, because that is 23 one event that would entail some exposure, so we had that. 24 And we put that into our equation in considering an 25 eight-hour time-weighted average. 38 1 For certain jobs we would say, How often--based 2 on talking to employees, how often would the employee wash 3 his tools with benzene or solvents and what kind of 4 exposure would be associated with that, and put that into 5 an eight-hour time-weighted average calculation. 6 Q. Where would we go now to find which locations 7 told you that their employees washed their tools in 8 benzene in assessing job exposures for the CMA benzene 9 study? 10 A. That study has been completed several years 11 ago. Many years ago. I don't have the data anymore. I 12 don't know whether CMA still kept the data or not--whether 13 my former employer turned the data back to CMA or not. 14 So I don't know where the raw data are at this file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (49 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 15 point or whether they still exist, but I assume you have a 16 copy of my report to CMA--not the publication, but the 17 report itself to CMA--and attached to that report there is 18 an example of our calculation of our time-weighted 19 average, and we have a list of uniform tasks. I think 20 it's in one of the appendix. 21 Q. Now, let me make sure I understand, Dr. Wong. 22 Did you at one time have the information yourself as to 23 which facility had workers who washed their tools in 24 benzene? 25 A. Yes. 39 1 Q. And where did you get that? 2 A. From the companies. 3 There was a committee formed to come up with 4 exposure assessments. In the CMA study we have estimates 5 for every job and estimates for an eight-hour 6 time-weighted average, and also a time for short-term peak 7 exposure, and that was reported to us on a form. 8 In terms of the 30 or so uniform tasks, for file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (50 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 9 each task we had the percent of time that job title was 10 spent on, the exposure in terms of ppm associated with 11 that task, and then we calculated the eight-hour 12 time-weighted average for that job title. So at one time 13 we had all that data. 14 Q. And you no longer have that data? 15 A. No, I don't. 16 Q. And then you relied again on the companies to 17 tell you whether or not their employees, by job, by year, 18 actually utilized the different tasks that were on the 19 list, and one of those tasks included washing their tools 20 in benzene? 21 A. That information was based on industrial 22 hygienist input as well as input from long-term employees. 23 Q. But it all came from the companies; not from 24 Dr. Wong? 25 A. From companies. I would have no exposure 40 1 information on the employees. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (51 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 2 Q. But what I'm trying to find out, sir, is all 3 the information that you had about these exposure 4 categories in creating the job profiles--the job exposure 5 profiles, came from the companies themselves? 6 A. When you say "companies," you mean company 7 industrial hygienists or you also include employees when 8 you say "company"? 9 Q. I'm including both. 10 A. In that case the answer would be yes. 11 Q. And that was done by year, by job; is that 12 right? 13 A. It would be more correct to say by calendar 14 period. 15 We don't consider single years. Only when 16 there are major changes in terms of exposure we would 17 change to a different time period. So you may 18 be--sometimes it can be a couple years. Sometimes it can 19 be 10, 15 years if there were no major changes during that 20 period of time. 21 Q. So tell me approximately what year these 22 exposure categories were reviewed by the company 23 employees, industrial hygienist employees that were file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (52 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 24 creating this information for CMA? 25 A. It would cover the entire study period. 41 1 Q. So it began about what year? 2 A. The observation period started in the mid or 3 late '40s. 4 Q. I'm sorry. When did the company industrial 5 hygienists and the company employees actually sit down to 6 start working on getting the exposure profiles that were 7 used in the CMA study? 8 A. I don't recall. That's one of the things that 9 we did early on. 10 Q. And about what year was that? 11 A. I just don't recall. I don't remember. Again, 12 it's a long-term study and I just don't recall the exact 13 years. 14 Q. But it was done toward the beginning of that 15 study? 16 A. Absolutely, yes. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (53 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 17 Q. And so whatever time that was, then as of that 18 date there were individuals within the companies that 19 participated in the CMA benzene study who went back and 20 determined what years it was the practice for various 21 employees at various locations to wash their tools in 22 benzene? 23 A. Yes, sir. 24 THE DEPONENT: Can we take a break? Is this a 25 good time? 42 1 MR. HOBSON: Absolutely. Anytime is a good 2 time if you want one. 3 (Break was taken from 11 a.m. to 11:05 a.m.) 4 Q. (BY MR. HOBSON) Dr. Wong, how can you assure 5 yourself, as a scientist who is going to take this 6 information on exposure categories and be the one who's 7 going to publish based on some of this information, how 8 can you take that information and make sure, as best you 9 can, that it's correct? 10 A. Well, to some extent we have to use our file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (54 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 11 judgment. We may not be able to verify on a quantitative 12 basis, but certainly by knowing the job description of 13 certain jobs, certain departments, the process and so on, 14 we have some idea of what kind of exposure you would 15 entail, relatively speaking, to other jobs. 16 So on the relative scale, we should have 17 certain confidence or comfort in exposure assessment. 18 Q. Let's take a specific, and, as best you can, 19 give me the most specific exposure from the CMA study for 20 benzene that you can give me. 21 MR. SPEARS: Are you asking for any particular 22 company or just what he can remember? 23 Q. (BY MR. HOBSON) If you need to rely on the 24 published studies, that's fine, or whatever else you might 25 have, but, in other words, I don't think you can give me 43 1 the exposure of an individual. You might be able to give 2 me the exposure of a craft, maybe you can only give me the 3 exposure of a category of worker, like maintenance, but file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (55 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 4 can you give me, for instance, like a craft at a plant, 5 what exposure--specific exposure categorization? 6 A. We had that information. 7 Q. You had it that specific? 8 A. Certain job title, certain period of time, yes, 9 we have that. 10 MR. SPEARS: I think he said he doesn't have 11 that data. 12 A. We had that information when we did the study, 13 but I don't have the documents themselves now, today. 14 Q. (BY MR. HOBSON) Can you tell me--for instance, 15 in the CMA benzene study, would the exposure category for 16 a pipe fitter at Mobil have been different than a pipe 17 fitter at Monsanto? 18 MR. SPEARS: Dr. Wong, don't guess. He's 19 asking you if you can recall this. 20 MR. HOBSON: No. I'm asking if he had it. 21 A. We had that information. 22 Q. (BY MR. HOBSON) Can you tell me if it would 23 have been different? 24 A. That, I don't know. Whether Mobil's pipe 25 fitter was comparable to Monsanto's pipe fitter? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (56 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 44 1 Q. And even if we are not calling a pipe fitter a 2 pipe fitter at the different plants, I'm talking about the 3 people that did that work. 4 A. That's right. 5 Q. So there was a time when you had in your 6 possession a document that you could have looked at where 7 you could have looked up the job of a pipe fitter--whether 8 it was called a pipe fitter or some other name, but you 9 could have looked up that job that was done at Monsanto 10 and found out if that pipe fitter's job for a particular 11 time period included washing tools with benzene, and you 12 could have also gone to the Mobil category of employee 13 listed as pipe fitter work and found out for another time 14 period, or perhaps the same time period, if that exposure 15 included washing tools with benzene? 16 A. Yes, sir. 17 Q. And all that information came to you from the 18 employees of the sponsoring companies? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (57 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 19 A. Yes, sir. 20 Q. It's correct, is it not, that you never went 21 back to validate that information? 22 MR. SPEARS: Let me object to the word 23 "validation" again, except that I think Dr. Wong did say 24 he did talk to some people himself. Did he not already 25 say that? 45 1 MR. HOBSON: Let's start with your objection. 2 Q. (BY MR. HOBSON) Dr. Wong, I want to make sure 3 we are using words that you and I agree their meaning of, 4 so let's start with "validate." Give me a definition of 5 validate that you and I can use. 6 A. Validate means different things to different 7 people and you may have different levels of validation. 8 The most complete, comprehensive validation 9 would be--I think would be that you would come up with 10 whatever estimates based on certain criteria, certain 11 models. You use those estimates and compare them to some 12 real data, independent data, data that would not include file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (58 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 13 it in the input process--in the estimating process. It 14 has to be independent. 15 That would be the most complete, independent 16 validation. That is the true meaning of the word. But in 17 the real world, sometimes it's very difficult to do that. 18 Another validation may simply mean that you 19 look at your exposure estimates and see whether they're 20 consistent or not. See whether it makes sense in terms 21 of--not on a quantitative basis, but in terms of jobs, the 22 functions, whether it makes common sense. That's another 23 level of validation. It's not as good as having 24 independent data to do that. 25 Q. So if we want to talk about validation, we can 46 1 talk about two levels of validation. One is a validation 2 based on real, independent data, and then a second, I 3 guess, step down in validation is a comparison of 4 nonindependent, but consistent--looking for consistency in 5 data? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (59 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 6 A. Right. And I wouldn't say there are only two 7 ways to do the validation. There are probably many 8 intermediate levels between those two. 9 Q. So we can communicate, let's stick with those 10 two right now, and if we need others, you tell me that we 11 are off base. 12 In looking at the exposures--the exposure 13 categories in the CMA benzene study, did you ever do any 14 validation in the sense that you took real, independent 15 data and validated those exposure categories? 16 A. Not for the entire time period, because that 17 was impossible to do because we simply did not have data 18 for the early part of the study. So there was no 19 data--independent, real data--industrial hygiene 20 measurements, let's say--in the '40s or '50s for us to 21 validate estimates for that time period. 22 Q. Then tell me what you did in later time periods 23 that would be a real independent validation of the data? 24 A. If I remember correctly, we have set aside a 25 small set of data that we did not put into the estimate file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (60 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 47 1 process so that we can verify our estimate process, but 2 then, that period of time really did not need that much 3 validation because we had real data to start with. 4 And that's the irony of the whole validation 5 process, because you can only validate your data if you 6 have real data, and if you have real data, your estimates 7 are pretty good. 8 Q. Let's talk about the real data for a minute. 9 The real data came from the companies too, right? That's 10 the air sampling for employee exposures that the companies 11 did? 12 A. Mostly air sampling data plus some engineering 13 modification data and so on. 14 Q. But all that came from the companies 15 themselves? 16 A. Yes, sir. 17 Q. When we talk about statistics and data 18 collection for statistical purposes, you recognize that in 19 the sense of doing exposure assessments there is a 20 difference between random sampling and nonrandom sampling, file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (61 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 21 do you not? 22 A. Yes. 23 Q. Tell me why that's important, to know whether 24 you're doing random sampling or nonrandom sampling? 25 A. Random sampling is a scientific procedure that 48 1 would give you a, quote, end quote, representative sample. 2 A nonrandom sample may not give you a 3 representative picture of whatever you want to look at. 4 Q. Wouldn't it be true that one of the problems 5 with nonrandom sampling is that you induce bias by the 6 judgment of the person actually collecting the sample? 7 A. Yeah, that's possible. 8 Q. So an example of that would be if you have two 9 manufacturing plants where benzene is being made and you 10 have two different industrial hygienists doing exposure 11 assessments at those plants, neither one of with which is 12 doing random sampling. 13 One industrial hygienist at one plant can say, 14 Well, if I see something that's really obviously wrong, file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (62 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 15 like benzene pouring out of a pump seal and men working in 16 the area, there is no need to take an air sample because I 17 know it's bad. I'm going to stop the flow of benzene, get 18 the area cleaned up and then take a sample to make sure 19 they really got it cleaned up. 20 That would be a nonrandom type of exposure 21 analysis, correct? 22 A. Yes. 23 Q. And I've got another plant where the philosophy 24 is, Anytime I see a problem, I want to go out and document 25 that problem, but I'm not as concerned about what's left 49 1 after I've cleaned up the problem. 2 If the two exposures occurred identically, you 3 would get two different data, would you not, out of those 4 two sampling strategies? 5 A. Two different data for what? 6 Q. You've got two separate plants, two different 7 industrial hygienists deciding what samples are going to file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (63 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 8 be collected. You have the same kind of leaking pump, 9 same circumstances involved with employees being in the 10 area doing their work. 11 One industrial hygienist says, I'm not going to 12 take a sample until I have it cleaned up, and the other 13 one says, The only sample I'm going to take is when I have 14 a leak. 15 That's their sampling strategy. One plant has 16 data that shows low exposure and the other plant has got 17 data that shows high exposure when the same incident 18 occurred, correct? 19 A. Right. 20 Q. How do you, as a scientist, correct for that? 21 A. When we look at data that are distinctly 22 different, then we don't combine them into a group. 23 That's why we need to look at the data, look at the jobs, 24 and see whether they're comparable. Look at the real 25 data, whether we can combine them at all. 50 1 Q. But how would you know what's real if it's not file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (64 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 2 random? 3 A. If they're very different, that tells us. 4 Although by the description of the job title itself it may 5 mean the same thing, in reality they are not and we would 6 not call them the same job. 7 Q. But isn't that just exactly the opposite, 8 Dr. Wong? 9 In other words, if you've got the same pump 10 leaking the same way at two different locations and the 11 only difference is how the person decides to collect their 12 exposure data, you've really got the same job with the 13 same exposure, but yet when you look at the data that goes 14 with them you say, No, these are two different jobs. 15 They're inconsistent because I have high data here and low 16 data here. 17 So you really can't do that, can you, unless 18 you know the sampling strategies being employed at the 19 different plants? 20 A. Okay. I think we have a major confusion here. 21 When you talk about quantitative exposure 22 estimates, all along I have in mind the ultimate goal is file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (65 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 23 to use that quantitative exposure estimate, regardless of 24 job titles now, but quantitative exposure estimates to do 25 a quantitative exposure analysis based on mortality data. 51 1 So it really doesn't matter what job titles we 2 are talking about as long as we rely on the exposure 3 estimates--the numerical exposure estimate itself. 4 Now, your q}es|ion would be appropriate if we 5 look at a certain craft, certain job, and what you're 6 describing to me is maybe at different locations we have 7 different exposure scenarios, and if we lump those 8 jobs--the same jobs, at least in terms of the title in the 9 employment records--lump those jobs together as a group 10 ftually sat down with 23 hourly workers and validated historical exposures? 24 A. I hate to use the word "validate" because we 25 have input from them in the process. As I said, I don't 15 file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (66 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 1 remember--after we complete our assessment, I don't 2 remember that we have actually gone back to some other 3 employees to validate it independently. I don't think we 4 have done that. 5 Q. That's the question I'm trying to ask. I do 6 mean to use the word "validate." 7 "Validate," I think, has the conception with it 8 of something you do after the fact to make sure that your 9 assumptions were correct. Would that be how you would use 10 "validate," or that your information was correct? 11 A. I guess in a sense we have not gone back to the 12 employees for validation. What I was trying to do is 13 point out that we do have input from them to start with. 14 Q. Why don't you do that? Why don't you go to 15 employees who are the actual subjects of the exposure 16 categorizations and validate that the exposure category 17 you placed their work in measures up with what they tell 18 you? 19 A. I think it's a practical consideration. The 20 area that we have concerns about would be exposures a long 21 time ago that we don't have industrial hygiene records, file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (67 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 22 and most people--active people don't know anything about 23 those jobs. So we have to identify some long-term 24 employees who actually were exposed. 25 There aren't that many people that would fall 16 1 into that category, and they're very difficult to find. 2 So whoever we could contact, we make use of the 3 information in an input process. There aren't that many 4 people that we can go to, that's all. I think that's the 5 major reason why we don't have the luxury of validating it 6 again. 7 Q. Have you looked? 8 A. Oh, yes, we did. 9 Q. You've looked actually and considered 10 validating the historical exposures? 11 A. Right. 12 Q. Have you published that? 13 A. For the API study--the gasoline study 14 publication is coming out under Tom Smith, and I'm one of 15 the coauthors in that publication. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (68 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 16 Q. And that publication will address the 17 validation of historical exposures or the inability to do 18 it? 19 A. As I said, we did not have validation. We have 20 input from long-term employees. 21 Q. That's what I'm trying to get to, Dr. Wong, is 22 the validation of these historical exposures. 23 A. Well, I told you we don't have validation and I 24 gave you the reason why we could not have the luxury of 25 validating those data. 17 1 Q. And then I asked you, Have you ever documented 2 your inability to validate historical exposures, and you 3 mentioned Dr. Smith's study, which then I think you said 4 does not address that. 5 So are there any written documents, published 6 or otherwise, that you can direct me to where you have 7 considered the validation of the historical exposures? 8 A. I don't know whether we have documentation of file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (69 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 9 that or not. I can't give you an answer. 10 Q. In considering cause and effect relationships, 11 the response which comes from the historical exposure data 12 is very important, is it not? 13 A. Yes. 14 Q. Would you agree with me that in exposure 15 assessments historically that is the weakest part of the 16 epidemiological study? 17 MR. SPEARS: That being what, Herschel? 18 MR. HOBSON: The analysis of the historical 19 exposures. 20 A. It's one of the potential limitations. I don't 21 know if I would label it as the weakest part. Weak as 22 compared to what? 23 Q. (BY MR. HOBSON) Well, death certificates, for 24 instance. They're hard and fast pieces of paper that you 25 can get and they say what they say and that's what you 18 1 have to go by, right? 2 A. Well, yes and no. If you get the right death file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (70 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 3 certificate, yes. It's a piece of paper and you have to 4 go by that. But a lot of times the information on the 5 death certificate may be slightly different from what you 6 have on the employee. 7 The social security number may not be a perfect 8 match, the birthday may not be a perfect match. So 9 sometimes you have to use your judgment to determine 10 whether that is the right death certificate or not. 11 So when you get to the subtleties of every 12 study, there are always some limitations. 13 Q. But the limitations on historical exposures are 14 much greater than deciding whether or not you're going to 15 use a particular death certificate, aren't they? 16 A. I would say they are equally important. I 17 don't want to give the impression that historical exposure 18 assessment is not important, but at the same time, death 19 certificates are very important too. 20 Sometimes the difference of two or three deaths 21 due to a specific cause can change the finding of a study. 22 So I would say they're both very important. 23 Q. When you design an epidemiological study, you file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (71 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 24 come up with a protocol and study design that you're going 25 to follow; is that right? 19 1 A. Yes. 2 Q. Is it important not to go back and amend the 3 protocol after you begin the study and begin to actually 4 analyze data? 5 A. Unless there are some major unanticipated 6 changes in the data that requires some modification. 7 Q. What would be some considerations that you have 8 used in the past to modify protocol under those 9 circumstances? 10 A. If the sample size turned out to be quite 11 different, then you have to modify analyses. I can't 12 think of any instances that I modified a protocol in a 13 major way; not right at this moment. 14 Q. If exposure categorization is part of the 15 original protocol, do you also define what exposure 16 categorizations are going to be in the protocol initially? 17 A. I would think so. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (72 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 18 Q. Are there times when you change how that 19 protocol for exposure standpoint will be dealt with after 20 you begin to collect data? 21 A. I don't know how to answer these kind of 22 general questions. If you give me a specific sample, I 23 can probably give you a more specific answer. 24 Q. I'm trying to find out about your experiences. 25 How important is it to determine what protocol 20 1 will be utilized for analyzing exposure categories in the 2 beginning before data is collected as far as how it 3 affects the outcome of the study? 4 A. I think every investigator would try to follow 5 the protocol as closely as possible. 6 In your protocol you plan certain analyses, but 7 if the numbers turn out to be very small in certain 8 categories, then you may have to lump things together and 9 that may be a deviation from the protocol. It may not 10 address certain questions, but still that may be the only file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (73 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 11 way you can do some analysis based on the data that you 12 actually have. 13 Q. Will a carefully thought-out protocol consider 14 those possibilities in the beginning to know what lumping 15 together might be necessary if the numbers are small? 16 A. Sometimes we can anticipate some, I guess, 17 unforeseen changes, but I guess nobody can predict the 18 future. 19 Q. You can't get it always right the first time, 20 but will you agree with me you can anticipate many of 21 these data collection problems, and those that can be 22 reasonably anticipated should be addressed in the initial 23 protocol? 24 A. Sometimes, yes. 25 Q. Are there times when you can reasonably foresee 21 1 a problem and you would not address it in the protocol? 2 A. You mean--you're asking me very general 3 questions. I just don't know how to answer them. Again, 4 if you give me some specific samples, I would be happy to file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (74 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 5 come up with a specific answer. 6 Q. So that's a question that's too general for you 7 to give me an answer to? 8 A. It's very broad. 9 Of course everybody would follow the protocol. 10 Everybody is trying to anticipate any unforeseen changes 11 and try to deal with that, but it's not a black and white 12 kind of decision sometimes. 13 Q. As a person who's skilled in statistics, what 14 kinds of problems can you get into by changing protocol 15 after you start collecting data? 16 A. Depends on what kind of changes we are talking 17 about. 18 Q. Give me some examples that you have in your 19 mind. 20 A. Well, for example, if you go and do a study and 21 you anticipate a much larger cohort than you actually 22 have, you may have to modify your conclusion because the 23 sample size would not give you as much confidence as you 24 would like. It may not even answer the question that you 25 start with. That's a very common scenario. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (75 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 22 1 You go into industry, you do a study, you 2 thought you would have a large cohort and you end up with 3 a much smaller cohort. 4 Q. I guess maybe I'm having a difficult time 5 phrasing a question so we can communicate. 6 What I'm trying to find out about is if we talk 7 about data collection, how can changing the protocol in 8 the data collection process after data collection has 9 begun affect the outcome of a study? 10 A. Depends on what the change is. I don't know 11 how to answer your question more specifically than this 12 unless you give me a more specific example. 13 Q. Okay, let's try a couple, then. Let's talk 14 about in evaluating exposure categories and you begin to 15 gather your death certificate information--assuming we are 16 talking about a mortality study--and partway through the 17 study you change your criteria for exposure 18 categorization. 19 A. In terms of what? In terms of duration of file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (76 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 20 exposure? In terms of chemicals? In terms of 21 departments, job titles? 22 Q. All of those? 23 A. All of those. 24 Q. Or any one of them. 25 What kinds of questions does that call to your 23 1 mind, then, about any results that you can gain from such 2 a study? 3 A. Well, I assume that if we modify the definition 4 of exposure--I'm talking about when you go out and 5 assemble a cohort. Regarding exposure, you have to rely 6 on job titles or departments to determine whether someone 7 was exposed or not. If you modify that, then you better 8 have some reason why you're modifying that. 9 The previous definition of exposure must be 10 incorrect or incomplete based on incomplete information. 11 So to me if you modify the definition of exposure in terms 12 of job titles or departments, then I would assume you have file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (77 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 13 better information now as opposed to before. 14 If you modify the duration of exposure--a lot 15 of times we would say we would only study people with at 16 least three months of exposure, six months of exposure or 17 a year of exposure. If you modify that, that may or may 18 not change your study hypothesis at all. 19 Sometimes you have to relax the requirement for 20 minimum exposure so that you have a larger sample size. 21 Sometimes you have to decrease that to make it smaller to 22 make it more manageable. 23 Q. Let's take a hypothetical and see if I can work 24 this through in my mind. 25 Let's say that I'm a large industry and I come 24 1 to Dr. Wong to do a study based on a chemical exposure in 2 my industry and I want there to be--this will be a 3 mortality study--a historical perspective mortality study, 4 and it will have, as a component, exposure categories so 5 that we can get a measure of dose response. Does that 6 make sense so far? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (78 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 7 A. Yes. 8 Q. You design a protocol for such a study which 9 includes exposure categorization. Data is begun to be 10 collected. You start collecting the death certificates 11 and you begin to do analyses and you do a preliminary 12 analysis of what data you have and you give it to me 13 because I'm the industry, I'm doing the exposure 14 categorization. Is there a problem with that? 15 A. Can I interrupt? 16 Q. Sure. 17 A. When you say, do a preliminary analysis, do you 18 mean mortality analysis actually with results such as 19 standardized mortality ratios? 20 Q. Maybe not that sophisticated. Maybe I collect 21 30 percent, 40 percent of the death certificates you 22 expect to find and you start to see what could be a trend. 23 A. Okay. 24 Q. And you come to me, the person who's paying for 25 the study, and say, Well, here's something I see. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (79 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 25 1 Is that a problem? Can it be a problem? 2 A. I think that's a problem. 3 Q. Why? 4 A. I can speak for myself. 5 Q. Yes. 6 A. I don't think I would ever want to present 7 preliminary analyses based on 30 percent data to the 8 sponsor. 9 Q. Why? 10 A. Number one, I don't think they should be that 11 intimately knowledgeable about some ongoing research in 12 terms of results. They should be aware of the problems 13 and so on, but they should not know preliminary analyses 14 based on 30 percent of the data. 15 And besides, I would never analyze a study 16 based on 30 percent data. It's just not--it's not 17 efficient. It doesn't tell you anything, to start with. 18 Q. I guess there is some level, though, at which 19 you would begin to do some preliminary analysis--some 20 level of data collection. 21 A. The only scenario I can think of is we have file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (80 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 22 about 90 plus percent of the data. When I say that, 23 mostly I mean death certificates; that we would start 24 running some preliminary analysis, make sure the computer 25 programs work and so on. 26 1 Now, that doesn't mean we are not going to 2 modify the data as we get more death certificates coming 3 in during that period of, so-called, preliminary analysis, 4 but I don't think the results would change that much based 5 on some small percentage of additional death certificates. 6 Q. Would you give that kind of an analysis to the 7 sponsor? 8 A. We may present that, yes. 9 Q. What if then the sponsor said, Well, we want to 10 change the definition of the exposure protocol? 11 A. I would ask why. 12 Q. Why would that be of any concern? 13 A. Certainly you would not change your criteria of 14 exposure or any other classification scheme of your file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (81 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 15 cohorts after you look at the results. That doesn't make 16 a whole lot of sense. 17 Q. Why? 18 A. If a study has been going on for six, seven 19 years, the exposure scheme should be worked out completely 20 by then. That really raises some suspicion, why somebody 21 would change the exposure categorization, especially if by 22 doing so you've also changed the result. 23 Q. Let me come at the problem that I'm trying to 24 get your opinion on a little bit differently. 25 Let's assume for a moment that I'm going to 27 1 hire you to do a study for me--I'm the sponsor--a large 2 industry-specific chemical historical perspective study 3 with exposure categories, and I am interested in whether 4 or not there is a cause and effect relationship between 5 the chemical and a particular disease category. And I 6 know, because I've got a little bit of knowledge about 7 epidemiology, that the more clearcut a dose response is, 8 then the stronger the study results are likely to be. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (82 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 9 Does that make sense? 10 A. Right. 11 Q. And I want you to go out and collect the death 12 certificate information, give me, then, an idea of where 13 these particular kinds of suspicious diseases lie, and I 14 want to be able then to manipulate the exposure data to 15 soften the dose response curves. 16 How do you guard against that? 17 MR. SPEARS: As I understand the question, 18 you're asking him in a hypothetical situation that this 19 industry wants to commit a fraud, basically. Is that what 20 you're trying to say, Herschel? 21 MR. HOBSON: Sure. 22 MR. SPEARS: How would he respond to that if he 23 knew you were? 24 MR. HOBSON: No. I'm asking how he, as a 25 scientist who's going to be dealing with companies who 28 1 might want to commit a fraud-- file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (83 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 2 Q. (BY MR. HOBSON) --how do you design in your 3 protocols ways to prevent that when the fraud is playing 4 with the exposure data, because now the industries 5 typically do provide you with a lot of the exposure 6 information, correct? 7 A. Yes, but they--let me go back and tell you how 8 we retrieve exposure information from the industry. 9 Normally we don't get exposure data on a 10 specific individual. We get exposure information on 11 departments, jobs, locations, and also calendar years, so 12 it's like a multidimensional matrix. And we ask them for 13 whatever industrial hygiene data they have. Not 14 estimates; real data for those jobs for the specific time 15 period, and so on. 16 Then, based on that, you should be the one who 17 knows--we don't, usually. If you go back to the '60s or 18 '50s, you don't have that much data or none at all, so you 19 do have to do some kind of estimates, and that's the part 20 we have to talk to old timers. We talk to the company 21 industrial hygienist and we don't rely on the company 22 alone. We rely on workers as well, long-term employees. 23 And the estimate--historical estimate's got to file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (84 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 24 be consistent with the real data too that we have in the 25 latter part of the study. And we do all this independent 29 1 of the mortality data, and we do all the classification 2 before--even before we analyze the data. 3 Q. How do you know that the exposure 4 categorization is done independently of the mortality 5 data? 6 A. As far as we are concerned, it's done 7 independently. 8 Q. How do you know that? 9 A. Because we have the mortality data. We never 10 give the mortality data to the companies until the study 11 is done. 12 Q. But isn't it true, Dr. Wong, that many of these 13 industries that you do work for are the kind we have been 14 describing? 15 They already have a big piece of the mortality 16 data. They may not have as many death certificates as you file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (85 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 17 do, but they have a lot of them through the benefit plans 18 and they have some other epidemiology studies that have 19 been done internally or through other consultants or 20 studies done earlier. They have a piece of that already, 21 don't they, many times? 22 A. They have part of the death certificates. Most 23 likely they would have death certificates for their active 24 employees who die on the job and also death certificates 25 for annuitants who are vested in the retirement plan at 30 1 least for the latter part of the study. 2 Q. So while you keep your mortality data from the 3 sponsor, you know going into it that most of these 4 sponsors, if not--let's say most of the sponsors already 5 have a piece of the mortality information in-house, right? 6 A. Right. 7 Q. So that's what I'm asking you now, is how do 8 you guard against that sponsor who wants to, to use 9 Counsel's term, commit a fraud by manipulating the 10 exposure data to make dose responses mushy? What do you file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (86 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 11 do as a researcher to stop that? 12 A. Again, we would look at the consistency of the 13 data. 14 Q. But there is no data. You said for most of 15 these exposures there is no exposure data. 16 A. We have real exposure data for certain parts of 17 the study period. 18 Q. And usually in the last 15 years? 19 A. Starting in the '70s, yes, we do have that. 20 Q. And even the information for exposure in the 21 '70s is somewhat limited in most instances, is it not? 22 A. We always want to have more data than we 23 actually have, but I don't know whether you'd call it 24 limited or not. 25 Q. If you were going to do a statistical analysis 31 1 of the exposure data that most of the sponsoring companies 2 have, say, up to 1980, isn't it true that they don't have 3 enough data to do really meaningful statistical file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (87 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 4 interpretations by job for exposure? 5 MR. SPEARS: Can you--for the record, I want to 6 object unless you identify what companies we are talking 7 about. I think you asked him specifically there. 8 Q. (BY MR. HOBSON) Pick any of them you would 9 like. 10 What I'm trying to find out, Dr. Wong--take a 11 pipe fitter at a refinery. 12 A. Okay. 13 Q. I will pick Mobil in Beaumont. I don't know if 14 you know anything about that or not. I know you studied 15 that group. But in looking at a pipe fitter at Mobil in 16 Beaumont, Texas for the exposure information for 1977-17 A. Right. 18 Q. --do you think that there is adequate benzene 19 exposure data to define any given pipe fitter's exposure, 20 or even a group of pipe fitters at that refinery? 21 A. For that specific study, if we are talking 22 about-23 MR. SPEARS: Counsel-24 Q. (BY MR. HOBSON) Counsel said give you a 25 specific. I got one for you. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (88 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 32 1 A. In that study we did not have any quantitative 2 exposure estimates, as you are well aware of. We didn't 3 use that in our study for the reason that we just don't 4 have all the data that we need. 5 As opposed to--let's take another example. The 6 API gasoline study that I am doing now. In that study API 7 spent a lot of resources. In fact, they contracted the 8 University of Massachusetts to come up with exposure 9 assessment and that's a major component. 10 Q. How much of that was based on hard data as 11 opposed to estimates in 1975, '76, '77? 12 A. I don't recall. I'd have to go back and look 13 at the report. 14 As I said before, that study was under the 15 direction of Dr. Tom Smith. I work with him. I provide 16 him with employment records and so on, but I'd have to go 17 back and take a look. 18 Q. Let's talk about the CMA benzene study. That file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (89 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 19 had exposure categories, did it not? 20 A. Yes, sir. 21 Q. How much of that information is based on hard 22 data? 23 A. I recall we have industrial hygiene data in the 24 '70s. We are talking about benzene, and benzene is one of 25 the things that people measure for a long time relative to 33 1 other substances. 2 Q. Tell me how you mean that, as your perception? 3 A. Starting in the '70s. 4 Q. That was a long time? 5 A. Relative to other substances. 6 And our studies start in the mid '40s. Some 7 company's studies started in the '40s and '50s. So I 8 would say about anywhere between 1/3 to 1/2, depending on 9 which company you're talking about, we have real data. 10 Q. For some part of the time? 11 A. Right. 12 Q. If it came down to a location and a person, how file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (90 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 13 much information on benzene is it your perception existed 14 for the CMA study? 15 A. The way we estimated exposure was based on 16 jobs. 17 Q. Right. 18 A. We divide up the job into a number of what we 19 call uniform tasks. 20 Q. You and I have had a chance to talk about the 21 study many times, but I am trying to get more specific 22 this time. 23 Give my your perception, if you would, sir, of 24 how much information on benzene exposure on an individual 25 cohort member for the exposed group in the CMA benzene 34 1 study there was. 2 A. It depends on the period of employment. 3 For someone who was hired since the '70s, I 4 think we probably have pretty good data on that person, or 5 at least on jobs similar to that person. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (91 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 6 Q. That's a different question. 7 Jobs similar to that person I will ask you 8 about next, but right now I'm talking about the 9 individuals. What is your perception of how much exposure 10 data you had for the individuals in the cohort? And I 11 realize it varied by time, and that's what I'm trying to 12 find out. 13 A. The way we estimate exposure is based on jobs 14 and not based on specific individuals. 15 It may happen that some of the data on that 16 specific individual went into the exposure estimates for 17 that general category of jobs. So to that extent, we may 18 have some specific information on that individual, but 19 more likely than not we have information on jobs in 20 general categories. 21 Q. Did the information on jobs in the general 22 categories--are those categories across different plants 23 or are they specific for each plant that was studied? 24 We can talk about the CMA benzene study. 25 A. It's a little bit of both. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (92 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 35 1 What they do is--"they," meaning the industrial 2 hygienists. 3 Q. Of the industry? 4 A. Of the industry--and I would participate in 5 those discussions as well--is to look at how comparable 6 the job titles are across the company and also over time, 7 and we reduce those job titles to some uniform comparable 8 job categories. 9 Q. What happens to the confidence of what you're 10 trying to learn about exposure as you combine and combine 11 again and then extrapolate across, as you've just 12 described? 13 A. I mean, in general terms, if we are not 14 confident, we would not use it in our analysis. 15 Q. But your confidence would go down, right, as 16 you continue to combine and average across? 17 A. Well, yes and no. 18 Q. The best of all possible worlds would be to 19 know the exposures of each individual every day they 20 worked, agreed? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (93 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 21 A. That's the ideal world that the epidemiologist 22 would like to have, but that's not going to happen. 23 Q. Doesn't exist? 24 A. No. 25 Q. And I guess a fallback to that would be if you 36 1 knew one day a week of exposure for every individual that 2 was in the cohort, that would not be as good as knowing 3 every day, but better than what you had, right? 4 A. No. I don't think a day a week is necessarily 5 the second best. That one day a week may or may not be 6 representative. It may be better to look at the overall 7 and come up with an average for that category, for that 8 job title. 9 Q. What about for plant to plant? How does that 10 affect your confidence in exposure categories if you're 11 talking about two different manufacturing facilities where 12 the job exposures are being combined? 13 A. We would combine them only if they're 14 comparable. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (94 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 15 Sometimes the same job title for different 16 companies may involve different job functions and may not 17 be comparable at all. 18 On the other hand, due to different 19 reasons--employment reasons, union contract reasons--they 20 may call the same job by different names. That's why we 21 need input from the industry, from personnel, from 22 industrial hygienists. 23 Q. You were, I think, personally involved in the 24 exposure category determination in the CMA benzene study, 25 correct--at least part of it? 37 1 A. Yes. 2 Q. How did you address a question like this, if 3 you did: One plant has benzene workers who wash their 4 hands and wash their clothes with benzene. 5 What does that do to your consideration of 6 exposure categories in the CMA benzene study? 7 A. Are you still talking about differences between file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (95 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 8 plants? 9 Q. However it figures in. I'm not sure that I 10 know how you did consider it and that's what I'm trying to 11 find out. 12 A. Are we talking about the CMA benzene study? 13 Q. We are. 14 A. In that study--I think I mentioned this 15 before--for each job and for each time period we try to 16 divide a job into different components. 17 In fact, we had a list of over 30, what we 18 call, uniform tasks such as taking samples from the 19 stream, analyzing samples, fixing up a leak, washing tools 20 in solvents, benzene. 21 In fact, I remember using solvents or benzene 22 to wash tools as being a specific task, because that is 23 one event that would entail some exposure, so we had that. 24 And we put that into our equation in considering an andpoint. 9 A. That's the kind of detail I was not involved 10 in. I don't know whether they regard it the same or not. 11 "They" meaning the industrial hygienists. 12 Q. (BY MR. HOBSON) Coming back to where I started file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (96 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 13 off, I think, at least somewhat, Dr. Wong--how could you 14 as the outside scientist involved in the project assure 15 yourself that these industrial hygiene employees, the 16 employees of the companies sponsoring the studies, didn't 17 obscure or alter exposure patterns by the way they did 18 their work? 19 A. The only thing we can do is to check for 20 internal consistency, see whether there are any big gaps 21 or any changes, jumps in the data, gaps in the data, that 22 are not explainable. We have to start from somewhere in 23 order to do occupational epidemiology. 24 Q. And all of those things that you just told me 25 that you would look at, they still go back to the company 75 1 industrial hygienist or company engineers or the company 2 supervisors, don't they? 3 A. Yes. 4 Q. And I guess I would come back, then, to ask 5 you, to your knowledge, no one who was not an employee of file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (97 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 6 one of the sponsoring companies would take one of these 7 job profiles of the uniform tasks and independently go to 8 the union or go to a union member or an hourly worker and 9 say, Here's what I understand your job to be or to have 10 been. What is your input? 11 You've never attempted, and, to your knowledge, 12 no one else ever attempted that kind of a validation with 13 real independent data for the job tasks? 14 A. I didn't do that in the CMA benzene study, but 15 something similar to that was done in the API gasoline 16 study. I believe Dr. Tom Smith did go back and talk to 17 some long-term employees in terms of his exposure 18 estimates. 19 Q. But in the CMA study, to your knowledge, no one 20 went back and talked to the workers themselves with the 21 uniform task profile for their job and said, Does this 22 match up with your recollection, correct? 23 A. We had input from the long-term employees in 24 the process itself, but I don't remember and I don't know 25 for sure whether the company industrial hygienist went file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (98 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 76 1 back and talked to some other long-term employees again 2 after we ccme up with the estimates. 3 Q. But if I'm a company and I don't want to play 4 fair and I want to cheat on the data, I can do it then 5 that way, can't I? I can shade the uniform task. I can 6 go to people that I either expect to get a certain answer 7 from or I can take their answer and I can alter it and I 8 can give you incorrect information about exposures and no 9 one would be the wiser, would they? 10 A. Theoretically they can try, but remember too 11 they do not have the mortality data. We kept that 12 separate. 13 Q. But didn't you tell me earlier that the 14 companies already have a large portion of the mortality 15 data because they have the death certificates on active 16 employees and annuitants? 17 A. I didn't say "large." I said they have death 18 certificates for active employees and annuitants, but I 19 don't know--I would not think that's a large portion of a 20 cohort study over a period of 40 some years. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (99 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 21 Q. What percentage would you say that would 22 typically make up, the total cohort? 23 A. The death certificates--the proportion of death 24 certificates that we are taking from companies would be 25 very small. 77 1 What we do is when we do our follow-up, so to 2 speak, vital statistic determination--the first thing we 3 do is ask the companies, Whatever death certificates you 4 have on the employees, give them to us so that we don't 5 have to repeat the process. And based on my experience, 6 that percentage is very low. 7 Q. The number? 8 A. A don't have a number. 9 Q. 10 percent, 15 percent? 10 A. Probably in that range. 10, 15 percent. 11 Q. But now I wouldn't have to know the mortality 12 information to be able to confuse exposure, would I? 13 MR. SPEARS: Object to the form of the 14 question. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (100 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 15 You ask an obvious question, Herschel. If 16 someone wants to cheat, they can cheat. That's obvious, I 17 guess. If somebody wants to give you false information, 18 the answer is, yes, they can give you false information, 19 but that's an obvious answer, if that's what you're asking 20 him. Is it possible that someone can cheat on giving 21 information? 22 Q. (BY MR. HOBSON) Well, I guess that's the 23 beginning point, sure, but what I guess I'm asking, 24 Dr. Wong, is isn't it true that you don't really need to 25 know the mortality information to confuse those responses 78 1 if you want to? 2 In other words, if I'm sitting here and I want 3 to create havoc in the interpretation of an 4 epidemiological study and I know that one major component 5 of assessing cause and effect is dose response, really all 6 I have to do is confuse dose. I don't have to know 7 anything about the response to affect the outcome, do I? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (101 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 8 A. No. I think if what you're implying is that 9 you can introduce some random error into the exposure 10 estimate and then the whole thing would disappear--the 11 dose response--that may or may not be so. 12 If you really want to cheat, I think you have 13 to underestimate the exposure of people who die from a 14 certain disease and overestimate people who didn't. So 15 you really have to know some mortality data before you can 16 be successful in confusing the picture. 17 Q. But really, it's not random error I would want 18 to introduce, is it? Isn't it nonrandom error I would 19 want to introduce, because random error would tend to 20 balance out, would it not? 21 A. That's exactly my point. So you do have to 22 know the health outcome of the people you want to change 23 the exposure. 24 Q. Isn't it really true that all you need to know 25 is what the true exposure picture is? 79 1 A. I think you're confusing me now. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (102 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 2 You're talking about how to cheat, and I gave 3 you one scenario. If you really want to cheat, the only 4 way you can do that is by overestimating noncases--people 5 didn't die from a certain disease--and underestimate 6 exposure of those people who die from that disease. That 7 is the only way to get rid of the dose response 8 relationship. 9 Q. Or make it so soft and wishy-washy that it 10 can't be interpreted? 11 A. Well, there is always the scenario in between, 12 but I'm giving you the specific example. 13 Q. You've told me that you don't think it's 14 possible to know the outcome in advance because there is 15 not enough of the mortality information available to the 16 companies, and I'm saying, okay, let's talk about that 17 intermediate step where you just make the exposure data so 18 mushy that dose response can't be interpreted. That can 19 be done, can it not? 20 A. It's a very broad generalization. Anything is 21 possible. 22 Q. But wouldn't it be fairly simple to take some file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (103 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 23 of the exposure profiles and go back and validate those 24 after the fact with the real independent data? Just go 25 talk to the workers themselves and say, We have for this 80 1 job, this job and this job that you didn't wash your tools 2 with benzene, and we have for this job, this job and this 3 job that you did. 4 A. I think we are going back into this question 5 over and over again, and let me make it clear that we have 6 input from long-term employees, and I mentioned to you 7 earlier this morning that it's very difficult to identify 8 those employees. And basically we use all of them that we 9 can identify in the input process and we really have no 10 one else to go to for validation. 11 You're talking about validating exposures in 12 the 30's, '40s and '50s. There aren't that many people 13 around that a company knows of still here for us to talk 14 to. 15 Q. Let me make sure I understand. You say that 16 some of the long-term employees were questioned and I file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (104 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 17 thought you told me earlier that it's your belief that 18 those were supervisory people? 19 A. No. 20 MR. SPEARS: Object to to that. 21 Q. (BY MR. HOBSON) What is your understanding, 22 then? 23 A. He can be a pipe fitter and retires as a pipe 24 fitter. 25 Q. Tell me if you know that was done. 81 1 A. I told you that was done. 2 Q. You're telling me that you know in the CMA 3 benzene study that long-term pipe fitters who were not 4 part of management were requested to validate the exposure 5 of profiles from a uniform task standpoint? 6 A. Herschel, we are getting nowhere with this. 7 I did say that long-term employees, 8 nonsupervisors, nonmanagement, but again, I said we did 9 not validate because we didn't know that many long-term file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (105 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 10 employees to start with. 11 So the small number that we could identify, we 12 used their information in the input process and there was 13 no one left for us to validate. 14 MR. SPEARS: I think he's describing a form of 15 validation, Herschel. I guess it's the question, do you 16 think it's valid in your definition? 17 MR. HOBSON: I think I understand what you're 18 saying now, which is different. My understanding is 19 different than it was earlier. 20 MR. SPEARS: The record will reflect it. 21 MR. HOBSON: Obviously. 22 Q. (BY MR. HOBSON) I'm not saying that you didn't 23 say that earlier, Dr. Wong, I'm saying I didn't appreciate 24 what you said earlier. 25 I take it that there are from time to time 82 1 conflicts in people's memory as to what particular tasks 2 are incorporated into a given job. Would that be fair? 3 A. I'm sure that happened from time to time. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (106 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 4 Q. And so, again, that's where I come back to the 5 needing to go out and after the fact validate this 6 information that if you're--because what I think you told 7 me is at the beginning of the study when you were 8 considering setting up the exposure had classification and 9 finding what uniform tasks were involved in which job, 10 that's when these long-term employees were sought out and 11 questioned and their input sought. Am I right? 12 A. Yes, sir. 13 Q. And so they may have said, Well, yes, these 14 crafts or these jobs or these activities, however the 15 proper wordage was, did or did not involve washing your 16 tools with benzene. And then other people say, Wait a 17 minute. I have a different memory. I don't think it did 18 or I do think it included these tasks. 19 And you can't have it both ways, right? Either 20 that task of washing your tools with benzene is included 21 or it's not and someone has to make a decision. 22 A. That's right. 23 Q. And so after that decision is made, that's what 24 you stick with in categorizing the exposures of the file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (107 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 25 different jobs? 83 1 A. Yes. 2 Q. And you're telling me, then, that because the 3 companies included everyone who was possible to include, 4 the long-term employees that were actually there on the 5 job, there was no way to go out and do a real independent 6 data validation. Have I got your testimony correct now? 7 A. In the CMA benzene study? 8 Q. Yes, sir. We have been talking about the CMA 9 benzene study. 10 MR. SPEARS: And he would say yes, if you are 11 asking him did he go out and validate it to each and every 12 person. I guess you're asking him, As to every person who 13 was studied, subject to that study were they questioned 14 independently by Dr. Wong? 15 MR. HOBSON: No. Not at all. I'm sure a lot 16 of the people in that study were dead. 17 MR. SPEARS: But he had to rely on information 18 given to him by long-term employees who based it on their file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (108 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 19 personal knowledge. So that's validation. 20 MR. HOBSON: I think we have been through 21 what's validation and what's not. I don't want to beat 22 that to death again. 23 Q. (BY MR. HOBSON) But it is correct, Dr. Wong, 24 that the information you had for the exposure categories 25 is all information that came through these 84 1 company-sponsored employees, industrial hygienists that 2 were on the committee? All that information filtered 3 through those individuals on the committee, did it not? 4 A. Yes, sir. 5 Q. And so you, Dr. Otto Wong, or no one who was 6 independent of the corporation took this information from 7 the long-term employees when it was gathered. You relied 8 on the company industrial hygienist and the others to 9 gather that information for you. Is that right or wrong? 10 A. Yes, sir, that's correct. 11 Q. And so trying to move on, because I think we file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (109 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 12 have got it now, there was no one outside the company 13 people, either at your direction or somebody else's 14 direction--some other consultant to the project that was 15 not a company sponsored employee--who validated the 16 exposure tasks for each of the jobs. Would that be 17 correct? 18 A. That's correct. 19 Q. And the reason all this becomes important, 20 Dr. Wong, is that I want to make sure that if I represent 21 someone who was one of of these people who did a job like 22 the jobs that are in your CMA study, and they're alive and 23 they were there in the '40s and '50s and '60s and they 24 come in and say, Nobody ever asked me about any of these 25 uniform tasks and what a pipe fitter did or no one ever 85 1 asked me about these uniform tasks and what a boiler maker 2 did or any other job that comes along, you have no way of 3 telling me that that's so or not? 4 A. That's correct. 5 MR. HOBSON: Off the record. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (110 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 6 (Break was taken from 12:20 p.m. to 12:35 p.m.) 7 Q. (BY MR. HOBSON) Dr. Wong, when you do 8 epidemiological studies for industry, do you consider the 9 fact that there might be some in industries who would want 10 to shade the truth? 11 A. Well, it depends on what kind of data we're 12 talking about. 13 To the extent possible, we always try to have 14 some measure to assess that possibility, either knowingly 15 or unknowingly, whether we have complete records or not. 16 Q. You know that with many of the studies that 17 you're involved with, depending on the outcome, millions 18 of dollars can change hands, whether it's litigation or 19 regulation. Your studies are done for a purpose and those 20 purposes involve a lot of money at risk. You realize 21 that, don't you? 22 A. I'm aware of that. 23 Q. Tell me some of the things that you do in your 24 epidemiological study protocols that are designed into 25 them to make sure that your clients doen't shade the file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (111 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 86 1 truth. 2 A. Well, one of the major considerations we have 3 is how complete the cohort is. If we go out and assemble 4 a cohort based on the employment records made available to 5 us, we want to know that we have all the records. That is 6 something we call a cohort verification, using independent 7 documents to verify whether we have a complete cohort or 8 not. 9 Q. How do you do that? 10 A. There are different ways of doing that. It 11 depends on what kind of independent documents we have. 12 A lot of times we go to the Social Security 13 Administration and get copies of 941 forms, quarterly 14 report earnings, so that we know we have the names of all 15 the individuals who worked there, and compare it to what 16 we have. 17 Q. And how far back in time are 941s kept by the 18 Social Security Administration? 19 A. It depends on how the company is structured. 20 It can go back all the way to the beginning of our study, file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (112 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 21 which sometimes can be the '40s, depending on the employer 22 identification number, which number they use, whether they 23 use one number for all the divisions in a single state or 24 what. 25 If that was the case, it would be very 87 1 difficult to use that. If they use one single number for 2 a specific operation it is much easier to get information 3 from the Social Security Administration. 4 Q. What other things are built into your study 5 protocols to keep companies from shading the truth? 6 A. In addition to social security 941 forms we 7 also request, to the extent possible, union seniority 8 lists as another data source for verification. 9 Q. And this all goes to identifying the cohort and 10 how complete it is? 11 A. Yes, sir. 12 Q. What else is built into your studies to keep 13 companies from shading the truth? What other areas file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (113 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 14 besides just cohort identification? 15 A. Well, there are two types of data that a 16 company provides to us in general. One would be 17 employment records, and that's the one we just talked 18 about, how to verify the completeness of the data they 19 provide to us in terms of employment records. 20 Another set of data or another set of 21 information that we rely on the company for is industrial 22 hygiene or exposure data and we discussed in length this 23 morning how we verify that. 24 Q. Anything else that comes to mind about what you 25 could put in your protocols to keep a company from shading 88 1 their data? 2 A. That would be the two major sources of data 3 that we have to rely on the companies. 4 In terms of mortality data, we don't rely on 5 companies that much. Whatever death certificates they 6 provide to us is fine, but then we would send the rest of 7 the cohort to Social Security Administration, the National file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (114 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 8 Death Index, all those external sources. 9 Q. And can you think of any other ways than what 10 we have talked about this morning already that you can 11 keep the companies from shading the truth about the 12 exposure part of the information you use in your studies? 13 A. The only thing I can think of is going again 14 back to exposure data. After we have exposure estimates 15 we look at internal consistency and at the same time also 16 look at the literature--what has been published for 17 similar populations. And if our estimates are very 18 different from theirs, different from what has been 19 reported in the literature, then again that would raise 20 some questions. 21 Q. And is the comparison with the consistency of 22 the data, is that done by the industrial hygienist on the 23 committee or is that something that you in your research 24 activities do independently of the committee? 25 A. We would do that. When we look at the exposure 89 file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (115 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 1 estimates, let's say for a certain job over the years, we 2 would expect some kind of continuum instead of some gaps 3 or jumps up and down. 4 If we do see some unexpected pattern of 5 exposure, then we would go back and talk to them and ask 6 them to explain to us why that was the case. 7 Q. And the "we" you use is someone other than the 8 company industrial hygienist on this committee? 9 A. "We" meaning myself and--myself. 10 Q. And your team? 11 A. Meaning the people that work for me, not 12 employed by the companies. 13 Q. So then when you do these studies 14 you--independently of the companies that are sponsoring 15 the research, you get all their exposure data; is that 16 right? 17 A. That's the assumption. We get all the data, 18 yes. 19 Q. You did an epidemiological study for Mobil in 20 Beaumont? 21 A. Yes. 22 Q. Did you, independently of the company file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (116 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 23 personnel, get all their exposure data to benzene that 24 they had? 25 A. That's a bad example because we didn't do any 90 1 exposure quantification in that study. 2 Q. Well, you studied the Mobil chemical benzene 3 workers in part of the CMA study, did you not? 4 A. Right. I'm confused. I thought you were 5 referring to the Beaumont refinery study we did for Mobil 6 itself. Are you referring to the Mobil plant in the CMA 7 study? 8 Q. Yes. 9 A. Okay. And the question was? 10 Q. Did you get Mobil's benzene monitoring data for 11 the CMA benzene study independent of the company 12 personnel? 13 A. No. The exposure data were provided to us 14 through that uniform task assessment process. 15 Q. I misunderstood you and I was trying to find file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (117 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 16 out if you, as the independent consultant-17 A. Right. 18 Q. --if you actually got the exposure data so that 19 you could independently look for these consistencies 20 between the periods of time when you had no data to when 21 you had data to see that there were no gaps, that that was 22 the consistency you would expect. 23 A. We checked that mainly through looking at the 24 computerized data we have subsequent to the exposure 25 assessment. 91 1 In other words, based on the committee's 2 assessment we put in the data in our computer data base 3 and we would call up a certain job and look at the 4 exposure pattern over the years. 5 And if--as I said, if there were some gaps or 6 jumps, then we would go back and say, Explain to us why 7 that was the case. 8 Q. But as far as getting industrial hygiene 9 monitoring data, if there was any from Mobil in 1967 or file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (118 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 10 1968, you independently never got that air sampling data 11 and then went to the exposure categories to see if there 12 was consistency? 13 A. No, we did not look at that. We might have the 14 data in terms of data entry, but we did not make that 15 assessment. 16 Q. You're familiar with the term "statistical 17 significance" as used in epidemiological studies? 18 A. I would think so. 19 Q. Would you agree with me that the statistical 20 significance level that is used by epidemiologists is an 21 arbitrary number? 22 A. It's arbitrary, but it's also a convention. 23 Q. Will you agree with me it's a convention that 24 not everyone uses in epidemiology? 25 A. I can't speak for everyone. All I can say is 92 1 it's a common practice. 2 Q. Could the statistical significance be file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (119 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 3 calculated at this 51 percent level? 4 A. Yes, but it's not very meaningful. 5 Q. Well, it certainly would answer the question as 6 to whether or not something was more likely than not to 7 occur, wouldn't it? 8 A. No. These are two different concepts, 9 statistical significance and probability of causation or 10 more likely than not. They are two independent concepts. 11 Q. Maybe we-12 A. I have yet to meet a lawyer who is not confused 13 with this. 14 Q. I don't think I intended to ask that question. 15 Maybe I asked it, but I didn't intend to. 16 A statistical significance level of .49 would 17 tell us that an event was more than 50 percent likely to 18 be real than chance, would it not? 19 A. The 49 percent refers to the probability that 20 we are willing to accept a false negative. 21 Q. Yes. 22 A. But that is not the same as, given that someone 23 has a disease, what is the likelihood of that person 24 getting the disease through certain exposure? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (120 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 25 Q. I agree. Two different questions. 93 1 A. Very good. 2 Q. But you will agree with me that a statistical 3 significance level of .49 does have to do with whether or 4 not the finding, whatever it is that the study is looking 5 at, occurred from chance alone or not? 6 A. Yes. 7 Q. And you could easily calculate the 49 percent 8 level with your statistics, could you not? 9 A. Yes. 10 Q. Have you ever had anyone that you've contracted 11 to do a study with instruct you not to have contact with 12 the study members in follow-up? 13 A. I'm not so sure I understand your question. 14 Q. In doing a study for mortality, for instance, 15 in determining vital status, have you ever had any of your 16 clients instruct you or contract with you not to contact 17 the families direct to determine vital status? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (121 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 18 A. We don't go to the family to determine whether 19 somebody is alive or not. It would be just too expensive 20 to do that on a cohort of tens of thousands of people. 21 We use a centralized data base such as the 22 Social Security Administration, the National Death Index. 23 Q. But even after you go through those 24 organizations you still end up with some percentage that 25 are at that point lost to follow-up? 94 1 A. Right. 2 Q. And what I'm asking you, then, is for that 3 percentage that is lost to follow-up, you want the number 4 to be as low as possible? 5 A. Yes, sir. 6 Q. Has anyone ever contracted with you 7 specifically that you will not go to family members in 8 pursuit of the vital status of these people that have been 9 lost to follow-up? 10 A. We would not propose that either simply because 11 I don't even know where the family is for those cases. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (122 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 12 You don't have an address. You don't have a phone number 13 to call. Those are the difficult ones to locate. 14 In order to reduce that number of lost to 15 follow-up, what we do is we usually go to some kind of 16 commercial credit bureau to determine whether they have 17 information on those people in terms of applications for a 18 loan or some credit history and so on. 19 Sometimes we also go to--if we know the state 20 where they lived last, we go to those states and see 21 whether they still have a driver's license, that kind of 22 thing, but we don't contact families. It's too expensive. 23 It's not very promising anyway. 24 Q. Yes, sir. I understand you're telling me about 25 the different indirect methods you can use to determine 95 1 vital status of people in follow-up with a cohort, but 2 what I'm trying to find out is if you can remember if 3 anyone has contracted with you to prevent you from making 4 direct contacts in determining vital status? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (123 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 5 A. No, because that's never in a proposal. That's 6 never in the protocol to do that. 7 Q. Would you know if anyone who has contracted you 8 to do a study has intended to keep that study secret from 9 the subjects of the study? 10 A. That, I have no idea. 11 Q. As far as you know, there has never been 12 expressed to you an intent by the sponsor to keep the 13 study a secret from the subjects. Would that be accurate? 14 A. That would be correct. 15 Q. Have you ever, that you can recall, been 16 dissatisfied with exposure categories established for 17 cohort in a study that you've been responsible for, but 18 yet gone ahead and used the exposure categories despite 19 your reservations? 20 A. This is not a yes or no kind of question. I 21 guess it's a matter of degree of satisfaction, so to 22 speak. 23 It really depends on what questions you want to 24 ask based on the analysis, based on the data. Certain 25 exposure would be quite adequate for some questions and file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (124 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 96 1 some maybe not. 2 Q. I guess really what I'm trying to find out more 3 about is can you recall an instance where you've had a 4 contract to do a study and you felt like there were 5 improvements in exposure categorization that could 6 reasonably be made and the sponsor said, No, we are not 7 going to do it, and you went ahead and did the study 8 anyway. 9 A. No, I don't recall the situation at all. 10 Q. One of the papers that you brought was the 11 De Coufle study on benzene. Do you remember that one? 12 It's the first one. 13 A. Yes. 14 Q. Mortality Among Chemical Workers Exposed to 15 Benzene and Other Agents. 16 A. Yes. 17 Q. By De Coufle, et al. 18 A. Yes. 19 Q. Do you know whose employees those were that he file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (125 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 20 studied? 21 A. It may be enough--not 100 percent sure. I 22 think it was a Conoco refinery, and then later on they 23 changed it to a chemical plant in the Baltimore area. 24 Q. I think you're right. 25 Q. Have you ever done any work for Conoco? 97 1 A. I assume you mean outside of litigation cases? 2 Q. Even including litigation cases. 3 A. I know I have worked on some cases in which 4 Cononco is a codefendant. 5 Q. Which ones do you recall? 6 A. I don't recall them. I know for sure-7 MR. SPEARS: They are all the ones you have, 8 Herschel. 9 MR. HOBSON: There may be a few I don't have. 10 A. In terms of projects, I don't remember doing 11 any projects for Conoco itself. That doesn't mean we 12 don't have projects through API and Conoco may be a 13 participant in those projects. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (126 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 14 Q. That was going to be my follow-up. Do you 15 know? 16 A. I know you pretty well. Now I can anticipate 17 your questions. 18 Q. There are only a certain number I know to ask, 19 and once I get through those, I'm out. 20 You don't know one way or the other whether 21 Conoco has been a participant in any of the trade 22 association studies, then? 23 A. If I was the principal investigator of that 24 project, I would know. 25 Q. Have there been any that you can recall that 98 1 Conoco has either begun as a participant and didn't finish 2 the study or finished the study? 3 A. Conoco is not a participant in the ongoing API 4 gasoline study. That, I know. 5 I was involved in some specific analysis of an 6 update of the old Tabershaw, T-a-b-e-r-s-h-a-w, Cooper file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (127 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 7 studies, and I don't remember whether Conoco was part of 8 that study or not. That would be the only study that I 9 could think of that Conoco might have participated in that 10 I have some association with. 11 Q. Do you know if Conoco has ever begun to 12 participate in an epidemiologic study and withdrawn? 13 A. No, I don't know that. 14 There was another project that I worked on 15 several years ago when I was at EHA and that was a 16 mortality registry. It's not a study, per se, but a 17 program set up to keep mortality data on a prospective 18 basis. 19 Conoco was one of the companies included in the 20 first phase of the project. In other words, to determine 21 what companies would have adequate data to participate in 22 such a registry if API decided to fund such a registry. 23 Conoco was part of the first phase. 24 Q. Was data actually collected? 25 A. Not mortality data, but in terms of how many 99 file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (128 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 1 records they have, how many years we can go back, yes, 2 there was some data. 3 Q. You're familiar with the Sloan-Kettering 4 work--the registry and perspective study that was done by 5 API Sloan-Kettering? 6 A. I know of the project, yes. 7 Q. Is the Conoco project a different project than 8 that? 9 A. It's almost like the second phase of the 10 Sloan-Kettering project. 11 There were some problems in terms of data in 12 the Sloan-Kettering registry and several years ago that 13 project--I think in the early '80s the Sloan-Kettering 14 project was terminated, but then after a few years there 15 was some interest in the petroleum industry to do 16 something like that again and that's why they sponsored 17 EHA to go out and look at what we can do. 18 Q. But the study was never undertaken, then, 19 following the EHA's work. Is that what you're saying? 20 A. That's correct. 21 Q. Do you recall if there was a final report from file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (129 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 22 EHA on what was found? 23 A. The first part of the study? 24 Q. Yes. 25 A. Yes, there was a report. 100 1 Q. And that was API activity? 2 A. Sponsored by API. 3 Q. Have you done any epidemiological studies on 4 vinyl chloride? 5 A. Yes. 6 Q. Was Conoco a participant in those? 7 A. I don't know for sure. I just don't remember. 8 There were a whole bunch of companies in that study. I 9 just don't recall today at this moment whether Conoco 10 was or was not in the study. 11 Q. What about Vista? 12 A. I don't recall that either. 13 Q. Who do you recall originally was contemplating 14 participating in the CMA benzene study but who ultimately 15 did not complete that project as a participant? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (130 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 16 A. ARCO was one of the participants when we 17 started the project, but later on because of some problems 18 with the data, ARCO withdrew from the study. 19 Q. Anyone else that you can think of? 20 A. That's the only company. 21 Q. Can you recall if there were any companies who 22 had--who appeared to have adequate personnel records, but 23 you who elected not to participate in the CMA benzene 24 study? 25 A. I have no knowledge of that. 101 1 Q. Was there ever a time that the API had any 2 involvement with the CMA benzene project or what became 3 the CMA benzene project, to your knowledge? 4 A. It wasn't exactly what you describe. 5 At one time when we were doing the CMA benzene 6 study there was supposedly a separate but parallel study 7 on the manufacturer's side. 8 The CMA study consists of users of benzene, and file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (131 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 9 supposedly we have an independent study on the 10 manufacturing side that would be member companies of API, 11 and, in fact, I was the principal investigator of that 12 project when I worked for Dr. Tabershaw, but there was 13 very limited company participation on the API side. 14 I remember there were only two companies on the 15 API side: Gulf and also Ashland. By "Gulf" I mean the 16 Philadelphia refinery--the Gulf refinery in Philadelphia. 17 And their contribution to the study was like 90 some 18 percent of the data, and Ashland contributed less than 10 19 percent of the data. 20 The decision was made that based on that 21 limited number of companies in the study, we should not 22 label that as an industrywide study and maybe API 23 shouldn't pay for that study. 24 As far as I know, subsequently Gulf did an 25 independent study of the Philadelphia refinery and I don't 102 1 know what happened to Ashland. 2 Q. Do you know how many companies originally had file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (132 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 3 data sufficient to participate in the API portion the 4 Tabershaw group was involved in? 5 A. I don't know. But on the API side what we are 6 talking about are refineries, and at that time during the 7 '80s there were a lot of cohort mortality studies on 8 refinery workers. So independently the companies are 9 doing studies of their own refineries. 10 Q. But the refinery worker studies, none of those 11 studies done by the API looked at benzene exposures, did 12 they? 13 A. We did. There was a study that I did with Gulf 14 in Port Arthur. We did look at the group of employees at 15 the benzene units. We looked at that subgroup. 16 Q. That was with Dr. Wen at Gulf, right? 17 A. Right. That was a joint project. The Port 18 Arthur refinery study was a joint project between Gulf and 19 Tabershaw. 20 Q. But what I was trying to address was your 21 comment that the API, and, I guess in essence the 22 industry, was already doing studies of refinery workers 23 and what I was trying to find out is whether or not it's file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (133 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 24 true that those API refinery worker historical perspective 25 epidemiological studies did not address benzene exposure? 103 1 A. Well, I just gave you an example. The Port 2 Arthur refinery study was one of those studies conducted 3 in the '80s independent of the CMA or API benzene studies. 4 And one of the analyses that we did in that study was to 5 look at people exposed to benzene. 6 Q. But that, again, would not be an industrywide 7 study? 8 A. No, that's not an industrywide study, that's 9 correct. 10 Q. The Tabershaw-Cooper study of '74, the 11 follow-ups to it that were done by other investigators, 12 those were industrywide refinery worker studies, were they 13 not? 14 A. Yes. 15 Q. And none of those industrywide refinery worker 16 studies dealt with benzene exposure, did they? 17 A. It depends on what you mean by "dealing with file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (134 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 18 benzene exposure." 19 We can safely assume that the majority of 20 refinery workers would have some exposure to benzene--the 21 gasoline which would have benzene content in it, but it's 22 true that none of those studies have sufficient employment 23 history information to identify a separate group as 24 benzene cohorts. 25 Q. So going back to where I think I started down 104 1 this path, can you recall if there were any API member 2 companies who had sufficient record keeping or records in 3 place which could have done a study of benzene exposures 4 in refinery workers similar to what you ended up doing for 5 Gulf in Philadelphia? 6 A. Quite a number of companies have sufficient 7 records to do a mortality study. 8 Q. Of benzene exposures in refineries? 9 A. Right. And in fact, they did do cohort 10 mortality studies of their refinery workers, but the file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (135 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 11 exposure information may not be sufficient for them to 12 identify a separate group out of the entire refinery. 13 Q. Isn't it true that the same historical exposure 14 techniques that were used by the CMA benzene study group 15 are available to be used in refinery worker studies? 16 A. Theroretically, yes, we can apply the same 17 method to estimate historical exposure. 18 Q. So such a study could have been done had one 19 wanted to do it? 20 A. On a theoretical basis I would say yes, but I 21 really don't have much exposure information on the 22 refinery side. 23 Q. Well, at the same time the chemical plants were 24 taking benzene exposure samples in the late '70s and early 25 '80s in response to OSHA regulations, the refineries were 105 1 doing the same thing, weren't they? 2 A. That's why I say on a theoretical basis we 3 should be able to do that. 4 Q. It's more than just theoretical, isn't it? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (136 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 5 Isn't it practical? 6 A. Unless they could not retrieve the data--the 7 historical data. Unless the job description is so 8 different that they cannot come up with some comparable 9 groups. There are a lot of practical considerations in 10 carrying out that process. 11 Q. Well, certainly Gulf in Philadelphia had that 12 ability to do such a study, did they not? 13 A. No. I was referring to the Port Arthur 14 refinery. Subsequently they identified--using the 15 department, the location of work, they identified a group 16 of workers who were employed at the benzene units. I 17 didn't say that about the Philadelphia refinery. 18 Q. Was there a benzene study of the Philadelphia 19 refinery workers or not? 20 A. Not that I'm aware of. There is an overall 21 refinery study at Philadelphia. There is an overall 22 refinery study at Port Arthur, but within that study we 23 also did an analysis for people assigned to the benzene 24 units. 25 Q. Would it be correct to say that as of today file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (137 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 106 1 there is not an epidemiological study of refinery workers 2 that addresses benzene exposure? 3 A. Depends on what you mean by "address the issue 4 of benzene exposure" because when you look at the entire 5 refinery, certainly employees in a refinery are exposed to 6 benzene. And by looking at the job titles, you can 7 identify the groups with potential exposure to benzene 8 relative to some other crafts or some other job titles. 9 Q. Is there an epidemiological study of a 10 petroleum refinery that addresses benzene specifically by 11 breaking down the job, task, however the exposure 12 categories are created, looking for those response 13 relationships? 14 A. Not in terms of quantitative exposure 15 estimates. 16 Q. How about qualitative? Is there even a 17 qualitative estimate of benzene exposure in a refinery 18 population? 19 A. I think at one time in the old API industrywide file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (138 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 20 study they used some kind of ranking classification--high, 21 medium and low. I was not involved in that study so I 22 really don't know how they came up with that 23 classification, but they do report mortality ratios by 24 degree of exposure in terms of high, medium and low. 25 Q. And what do you recall they found? 107 1 A. I don't recall what they found. 2 Q. That's not one of the papers you've brought to 3 the deposition today, is it? 4 A. No. And the reason being, subsequent to that 5 we have individual company studies and they are all 6 components within the old API study. 7 Q. And the individual company refinery workers 8 studies address benzene exposures? 9 A. Not on a quantitative basis. 10 Q. How about qualitative? 11 A. Indirectly through job titles, also maybe 12 through period of exposure. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (139 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 13 Q. Which ones are those? 14 A. For example, the study I did for Chevron, we 15 looked at those people exposed prior to 1947, '48 as 16 opposed to people who were hired subsequent to that. 17 The rationale behind that was the exposure, at 18 least at the recommended standard, dropped from 100 to 35 19 ppm over a period of two years around '47, '48. So that's 20 an indication of level exposure. 21 Q. Is inherent in using that information that you 22 just gave me about the Gulf study the change from 100 to 23 35? 24 A. No Chevron. 25 Q. I'm sorry. Chevron or Gulf? 108 1 A. Chevron. I said Chevron. You need lunch. 2 MR. SPEARS: There is no Gulf any more. 3 MR. HOBSON: I know. 4 Q. (BY MR. HOBSON) Well, let me go back. 5 When there is a change as you described in 6 maximum allowable concentration from 100 to 35 participate file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (140 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 7 per million, do you assume, then, that all exposures are 8 kept at or below those levels in a refinery? 9 A. I can't say all. I can say on an average basis 10 certainly the exposure prior to 1947 or '48 would be much 11 higher than exposure thereafter. 12 Q. What's the basis for being able to say that? 13 A. Because that was the recommended standard 14 during that time. And by talking to company industrial 15 hygienists, that's their impression. 16 Q. You're aware that the old Gulf Oil Company did 17 not have an industrial hygienist before the mid-'60s? 18 A. We are talking about Chevron. 19 Q. But didn't--didn't the study you're talking 20 about occur at the Gulf Port Arthur Refinery? 21 A. No. The Chevron refinery in California. 22 Q. I beg your pardon. I thought you told me 23 earlier about the Gulf refinery at Port Arthur. 24 A. That was a separate study. But when you asked 25 me about analysis by period of exposure, the example I file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (141 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 109 1 gave you was based on two refineries out in California 2 that belonged to Chevron all this time. They are not from 3 Gulf. 4 Q. I apologize. I didn't follow that. 5 Is it reasonable, in your view, to think that 6 the exposures would be the same or similar at other 7 refineries as the two Chevron refineries in California? 8 A. I would say there are some minor differences, 9 but on an average basis it should be similar. 10 Q. Were the refinery workers in California washing 11 their tools in benzene? 12 A. We have examples of that, sure. 13 Q. And were they washing their clothes in benzene? 14 A. I don't know about clothes, but their tools, 15 certainly. 16 Q. They cleaned their hard hats with benzene? 17 A. I don't know specifically about hard hats. 18 Q. For the refineries in California, for what time 19 period was the practice of washing their tools with 20 benzene a practice? 21 A. I don't remember. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (142 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 22 Q. Is this part of the pre-'47 time period or 23 after? 24 A. I'm sure that practice was done prior to 1947, 25 '48, but I don't know whether that was completely gone 110 1 after 1948 or if there were some individuals who were 2 still doing that. I don't recall. 3 Q. Would it be your view that you could have met a 4 35 parts per million maximum allowable concentration and 5 wash your tools with benzene? 6 A. I would think it's possible if the time 7 involved in washing the equipment, or whatever, with 8 benzene is relatively short and the rest of the day you 9 don't have much exposure. Certainly the time--the 10 eight-hour time-weighted average could be below 35 ppm. 11 Q. If you took--tell me, how long do you recall 12 you attributed in a workday to washing your tools with 13 benzene in your tasks for the CMA benzene study? 14 A. I don't have the data now so I really can't file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (143 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 15 answer that question. The time estimates involved in that 16 depends on the location as well. I remember seeing 17 different estimates from different companies. 18 Q. What ballpark are we talking about for those 19 estimates? Are we talking about one minute, ten minutes, 20 an hour? 21 A. Not an hour, for sure. Probably less than an 22 hour. 23 Q. Would ten minutes be a reasonable length of 24 time to wash your tools in benzene? 25 A. I don't recall. 111 1 Q. Well, let's talk about now, then, not the CMA 2 study. 3 As we sit here right now, how long do you think 4 would be reasonable to attribute to a worker who is going 5 to wash his tools in benzene? 6 A. Reasonable as opposed to what I actually had in 7 the study? 8 Q. Yes, sir. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (144 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 9 A. I don't know what's reasonable. Depends on, I 10 guess, how many pieces of equipment you want to wash. 11 If you're talking about personal equipment, 12 maybe no more than 10, 15 minutes. 13 Q. Let's pick 10 minutes. And if we pick 10 14 minutes and we say that the eight-hour time-weighted 15 average that we have of interest is 10 parts per million, 16 what would be the maximum exposure for that 10-minute 17 period that you could have and not exceed the eight-hour 18 time-weighted average of 10 parts per million? 19 A. Assuming there was no exposure for the rest of 20 the day? 21 Q. Yes, sir. 22 A. We can sit down and calculate that. That can 23 be done. 24 Q. It would be 480, wouldn't it? 25 A. You are the industrial hygienist. I take your 1 word for it. 112 file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (145 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 2 Q. Well, it's a matter of parts per million, 3 minutes and working it out, isn't it? 4 A. Right. 5 Q. Would you have any idea what the breathing zone 6 levels of benzene would be if you wash your tools in 7 benzene? 8 MR. SPEARS: Herschel, we are getting 9 theoretical here. Washing in a closed environment? 10 Washing outside in the wind? Washing in a 55 gallon drum, 11 in a sink? We could sit here all day and hypothesize 12 about things. 13 Q. (BY MR. HOBSON) Okay. 14 A. We actually have some specific exposure data on 15 that task in the CMA study. I just don't have the data 16 now. 17 Q. For Mr. Crawford, do you have any estimate for 18 his benzene exposure? 19 A. No, I don't. 20 Q. Have you read his coworker depositions? 21 A. Yes, I have. 22 Q. How about for Mr. Smith? Do you have any 23 benzene estimates for exposure for Mr. Smith? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (146 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 24 A. No. 25 Q. Have you read his deposition? 113 1 A. Yes, I have. 2 Q. How about his coworkers? 3 A. Yes, I have. 4 Q. Can you categorize their exposure 5 qualitatively, either of those gentleman or both? 6 A. I would say they are typical of some of the 7 studies that I have done. 8 Q. Typical for what time period? 9 A. For the same time period from the '50s until 10 now. 11 Q. So it's your impression that these exposures 12 that Mr. Crawford and Mr. Smith received at Conoco would 13 be typical of refinery workers, say, working at Exxon in 14 the same time period? 15 MR. SPEARS: I don't think he said Exxon, 16 Herschel. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (147 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 17 A. I never studied Exxon, come to think of it. I 18 have done studies for Mobil, I have done studies for 19 Chevron, I have done a study for Gulf, but I have never 20 done an Exxon study. 21 So I would say based on my experience, it would 22 be experience derived from Chevron, Mobil, and Gulf. 23 Q. (BY MR. HOBSON) And so you see no difference 24 of any consequence in exposures between Conoco and those 25 other three refining corporations? 114 1 A. Based on the job description that I get for the 2 two plaintiffs from the depositions, I would say their 3 exposure would be comparable to my studies, the studies 4 that I have done. 5 Q. So would it be fair to say, then, you read that 6 it was a practice to wash their hands, wash their tools 7 with benzene at Conoco. You read that, didn't you? 8 A. Right. 9 Q. You accepted that as the truth? 10 A. I assume that's true, and that is not any file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (148 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 11 different from the common practice that I have observed 12 when I did my study. The same practice occurred in my own 13 study. 14 Q. And when you say "in my own study," you're now 15 talking about at Mobil and at Chevron and at Gulf for 16 comparable time periods? 17 A. Right. 18 Q. And, of course, Mobil and Chevron and Gulf 19 employed you in your area of expertise to do those studies 20 where you gained this knowledge; is that right? 21 A. Yes, sir. 22 Q. Will you acknowledge that there were numerous 23 industry publications as early as the 1940s that said 24 washing your tools with benzene was a bad practice? 25 A. I'm not aware of those publications or 115 1 documents that you refer to. 2 Q. You've never read the toxicological review for 3 benzene put out by the American Petroleum Institute? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (149 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 4 A. I don't believe so. 5 Q. Will you agree with me that as of 1948 there 6 were reasonably well-documented cases of leukemia in 7 workers exposed to benzene? 8 A. There are some case reports, if that's what 9 you're referring to, but there weren't any studies. 10 Q. But would you agree that those could be 11 described as reasonably well-documented cases of leukemia 12 from workers exposed to benzene? 13 MR. SPEARS: I object to the form of the 14 question as what you mean "reasonably well-documented." 15 You could drive a truck through that definition. 16 MR. HOBSON: That's one the industry used. 17 It's one Dr. Wong would agree with. 18 A. I guess we can use the word "documented" within 19 the framework of case reports. To that extent, yeah, 20 there are some case reports in the literature, but I still 21 have to say that there weren't any studies--formal 22 epidemiological studies. 23 Q. (BY MR. HOBSON) Could a reasonable person 24 conclude in 1948 that benzene could cause leukemia in 25 humans? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (150 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 116 1 MR. SPEARS: I'm going to object to that 2 question. That is really hypothesizing to the point of 3 infinity now. Could a reasonable person? 4 A. I don't know what a reasonable person would 5 think. All I can tell you, in trying to answer your 6 question, is that as a scientist if I were given the case 7 reports in the '40s or '50s, I would not be able to make 8 the definitive conclusion that benzene had caused 9 leukemia. 10 If that were the case, then I don't think NIOSH 11 would spend the money to do a study on benzene and 12 leukemia and CMA would not fund a study. 13 Q. (BY MR. HOBSON) Of course, NIOSH had to meet 14 certain regulatory requirements before regulations could 15 be passed. You realize that, don't you, Dr. Wong? 16 A. I realize that, but at the same time the 17 regulatory requirements are actually less than the 18 scientific criteria for causation for establishing an file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (151 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 19 association. 20 Q. How certain does someone need to be about 21 benzene causing leukemia before they should warn workers, 22 in your view? 23 MR. SPEARS: Object to the form of the 24 question. That's beyond his expertise. You're asking him 25 for a legal conclusion now. 117 1 MR. HOBSON: No, I'm not. I want to know his 2 personal opinion. 3 A. I am not an expert in that area. 4 I can tell you what criteria were used in 5 determining causation based on scientific data, but I 6 don't know whether you should tell people what you can 7 interpret from those data. 8 Q. (BY MR. HOBSON) If you had information 9 available to you that there were reasonably 10 well-documented cases of leukemia from exposure to 11 benzene, is that something that you would pass on to your 12 workers? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (152 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 13 A. That's really outside my area of expertise. 14 Q. I'm not necessarily asking for an expert 15 opinion. I'm just asking for what you would do as an 16 employer. 17 A. I don't know. I have never been put in that 18 position, so I really cannot answer your question. 19 Q. Do you know what the term "etiology" means? 20 A. As an epidemiologist I would interpret that as 21 the causal agent of that disease. 22 Q. What would the term "etiology established" mean 23 to you? 24 A. For an epidemiologist that would mean there are 25 epidemiological data to support a causal relationship. 118 1 Q. Are you aware that as of 1949 that a 2 representative of the Manufacturing Chemists Association 3 and a representative of the American Petroleum Institute 4 were involved in a publication of a document that said 5 that the etiology had been established that benzene was a file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (153 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 6 human carcinogen? 7 MR. SPEARS: In all fairness to Dr. Wong, 8 what's the name of the article, Herschel? Has he seen it? 9 MR. HOBSON: I don't know. That's what I'm 10 trying to find out. 11 MR. SPEARS: What's the title? Is there a 12 title? 13 MR. HOBSON: It's a real life document listed 14 in our exhibit. 15 MR. SPEARS: Do you know what he's talking 16 about, Dr. Wong? 17 THE DEPONENT: I understand the question, but I 18 don't know of such a document, number one. Number two, 19 whoever wrote that must not be an epidemiologist. 20 Q. (BY MR. HOBSON) Is that something that only an 21 epidemiologist could conclude? 22 A. In terms of disease causation for chronic 23 diseases. I would think epidemiology is the appropriate 24 discipline to address that question. 25 Q. Now, remember my question goes back to 1949. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (154 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 119 1 A. Right. 2 Q. In 1949 would it be so that this would be a 3 questions that only an epidemiologist could address? 4 A. Well, as I said, epidemiology is the 5 appropriate discipline to address causation issues. It 6 doesn't change with time. 7 Q. So if someone were not an epidemiologist and 8 came to such a conclusion about the etiology being 9 established that benzene was a carcinogen--a human 10 carcinogen, they would be acting beyond their abilities? 11 A. I would like to know what data they rely on in 12 arriving at their conclusion. 13 We are talking about a document that I have 14 never seen. I don't know what kind of data they relied 15 on. 16 Q. Wouldn't you expect that if representatives of 17 the American Petroleum Institute and the Manufacturing 18 Chemists Association were going to put their names behind 19 such a statement, that they would have had whatever they 20 felt was sufficient to make those statements? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (155 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 21 MR. SPEARS: I object to the form of the 22 question. You're calling for speculation. I don't even 23 know who the people were. You haven't identified who the 24 people are or the article so how could he possibly know 25 what was in their minds or the basis for their opinion? 120 1 MR. HOBSON: I'm not asking that question. 2 MR. SPEARS: You asked, Wouldn't you expect? 3 He doesn't know who the people are. You haven't told who 4 they are. 5 MR. HOBSON: Representatives of those two trade 6 associations, both of whom he's done trade work for. 7 A. I have no opinion. I don't know those people 8 and I don't know what the basis-9 Q. (BY MR. HOBSON) You know the American 10 Petroleum Institute, correct? 11 A. Yes. 12 Q. And the new Manufacturing Chemists Association 13 which is now the CMA, Chemical Manufacturer's Association? 14 A. Yes. I worked with a very small number of file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (156 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 15 people within those organizations. 16 Q. And based on your work with those organizations 17 do you think they are people that can be looked to to be 18 relied upon for information about chemical hazards? 19 MR. SPEARS: You said--I object to the 20 question. You are saying the organizations themselves are 21 reliable? 22 MR. HOBSON: Yes, based on his experience with 23 them. 24 A. I can't answer that question because the people 25 that I deal with at CMA or API, they are administrative 121 1 people. They are project administrators. They don't 2 participate in the scientific aspects of the studies. 3 I work with the committee formed within API or 4 CMA, and the committee members by and large are company 5 employees, industrial hygienists or epidemiologists or 6 member companies. 7 Q. (BY MR. HOBSON) Can you give my some examples file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (157 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 8 of the CMA or API publications that you have found to be 9 unreliable? 10 A. I don't read their publications. I'm 11 interested in studies. I'm not interested in policy 12 statements. 13 MR. HOBSON: Off the record. 14 (Discussion off the record.) 15 Q. (BY MR. HOBSON) Do you think it's reasonable, 16 Dr. Wong, for people in the general public to rely upon 17 publications of the American Petroleum Institute or the 18 Manufacturing Chemist's Association? 19 MR. SPEARS: I'm going to object to the form of 20 the question. I don't think that's within his expertise 21 and I don't think it's a very well-defined question. The 22 general public? 23 A. I can't answer your question. 24 Q. (BY MR. HOBSON) Can benzene cause CLL, chronic 25 lymphocytic leukemia? 122 1 A. No, not based on epidemiological studies. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (158 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 2 Q. Is that something about which reasonable minds 3 can differ? 4 A. Well, I would assume--by looking at all the 5 important papers, all the pertinent literature I come to 6 the conclusion that exposure to benzene would not increase 7 the risk of CLL, and if somebody offered a different 8 opinion, I certainly would like to see why he or she would 9 arrive at a different opinion, what kind of studies they 10 relied on, and to that extent it may or may not be 11 reasonable. 12 Q. Well, are you so convinced that benzene is not 13 a cause of CLL that you will tell us that anyone who has a 14 contrary opinion is just flat wrong? 15 A. It would be incorrect based on the current 16 knowledge we have. 17 Q. You know that Dr. Morgan has said in his 18 opinion benzene is the cause of CLL, don't you? 19 A. I don't know that. 20 Q. You don't know that? 21 A. I don't know that. I have not read his 22 deposition. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (159 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 23 Q. Well, assume for the purpose of this question 24 that Dr. Morgan has said that in his opinion benzene is 25 the cause of CLL. 123 1 A. I would like to know what kinds of studies he 2 relies on. 3 Q. Are you willing to say he's flat wrong? 4 A. Yeah, I would say he's wrong. 5 Q. Is there any kind of leukemia or other cancer 6 besides acute myelogenous leukemia that you believe 7 benzene is a probable cause of? 8 A. No. 9 Q. Do you have to have aplastic anemia before you 10 can develop acute myelogenous leukemia from benzene 11 exposure? 12 MR. SPEARS: I you object. Now I think you're 13 asking a medical question. He's not a medical doctor; 14 he's an epidemiologist. 15 MR. HOBSON: It may or may not be a medical 16 question. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (160 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 17 MR. SPEARS: It sounds like a medical question 18 to me and I will defer to Dr. Wong. 19 MR. HOBSON: Some people might think medical 20 causation is a medical question. 21 MR. SPEARS: You're asking whether one disease 22 precedes another, and you're asking him can you have 23 something before having another type of disease. That's a 24 medical question. 25 Q. (BY MR. HOBSON) Do you have an opinion or not, 124 1 Dr. Wong? 2 A. I don't have an opinion. I have not looked 3 into that issue and I don't know whether there are data 4 that would tell us one way or another. 5 Q. Have you read--what depositions have you read 6 in this case--or these cases, Dr. Wong? 7 A. Depositions of the plaintiffs and depositions 8 of the coworkers. 9 Q. Which ones? Which coworkers? file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (161 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 10 A. In fact, there is a cover letter that was sent 11 to me together with all the materials. 12 MR. SPEARS: I think it's all the coworkers 13 that can be designated as being coworkers in this case. 14 A. Here's a list of coworkers. 15 Q. (BY MR. HOBSON) And you've read all the 16 depositions listed on this sheet? 17 A. Yes, sir. 18 Q. I notice that most of these have "Summary" 19 after the name. Did you read the deposition or did you 20 read the deposition summary? 21 A. I read both. 22 Q. Okay if I mark this letter? 23 A. Go ahead. 24 Q. Were there any CLLs in the benzene CMA study? 25 A. There were two observed, and we expected about 125 1 1.6 something. 2 Q. And that was in the comparison with the exposed 3 group to the nonexposed or exposed to the general file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (162 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 4 population? 5 A. With the general population. 6 Q. In which groups did the two CLLs appear? 7 A. I don't recall. 8 Q. Would it be specified in the article--the 9 published article? 10 A. Yes. 11 Q. And there were no CLLs in the unexposed 12 workers? 13 A. There was no leukemia at all in the nonexposed 14 group. 15 Q. Let me show you, Dr. Wong, this document which 16 is purported to be a statement submitted to the OSHA 17 benzene hearing by you-18 A. Okay. 19 Q. --dated March of 1986, I believe. Do you 20 recognize that? 21 A. Seems to be my statement, yes. 22 Q. Anything about that statement that you now 23 recognize was incorrect at the time you made it? 24 A. Well, I made that statement in 1986. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (163 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 25 Q. Yes, sir. 126 1 A. Subsequent to that we have a number of studies 2 published after 1986. To that extent, my opinion today 3 might have been modified by those studies published after 4 that date. 5 Q. But really, the first question I was going to 6 ask you ask is, is there anything about that March 1986 7 statement that you now know was incorrect at the time you 8 made that statement? 9 A. I have to go--I have not read this for some 10 time. I would have to go back and take a careful look if 11 you want to be that specific. 12 Q. Nothing comes do mind as you sit here now? 13 A. Not really. 14 Q. Whether your opinions are the same or different 15 now, I guess in order to answer that we would have to go 16 through it line by line and statement by statement? 17 A. I have to go through and try to recall--refresh 18 my memory exactly what I said and whether--based on the file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (164 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 19 studies that I'm aware of published subsequent to that 20 date, whether those studies would have modified my opinion 21 expressed on that date. 22 Q. And is there anything that you know--off the 23 top of your head without going through it line by line or 24 item by item, that you know you've changed your opinion 25 about? 127 1 A. I think in general the issue of cell-type 2 specific leukemia may be somewhat different than the 3 situation we had in 1986 because we have more data now 4 today on that issue than we had before that time. 5 Q. Give me the summary of what your opinion was in 6 1986 and a summary of what it is today on that regard, if 7 you can. 8 A. In 1986 I wasn't aware of some of the writings 9 by hematologists talking about the facts that different 10 cell-types of leukemia represent different diseases. I 11 wasn't aware of that. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (165 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 12 There are also some epidemiologic studies that 13 use--take advantage of some techniques that we only have 14 within the last few years, and that is the identification 15 of specific onco genes associated with specific cell-type 16 leukemia. That wasn't available in the early '80s. 17 On top of that, we also have a lot more 18 epidemiologic studies now on refinery workers that we did 19 not have--at least not published before 1986. So we have 20 a better idea of cell-type leukemia today than back in 21 1986. 22 Q. And so I'm not sure I understand where your 23 opinion has changed. 24 A. Well, my opinion in 1986, if I remember 25 correctly, was that we just don't have enough data on 128 1 cell-type leukemia. There weren't enough studies to allow 2 us to make that assessment. But today what I'm saying is 3 we do have a lot more studies now that we can address that 4 question, cell-type specific leukemia and exposure to 5 benzene. That's on the epidemiologic side. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (166 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 6 At the same time, we also have clinical data to 7 support that indeed different cell-type leukemias are 8 different diseases. 9 Q. And how many industry-sponsored studies on 10 benzene have you done since 1986? 11 A. Certainly the Chevron study was published after 12 1986. The study was done around that time, but we didn't 13 complete that until sometime after 1986. 14 Q. Was that cell specific? Was the Chevron study 15 cell specific for leukemia? 16 A. Maybe I omitted one part in my answer. 17 Based on any single individual study you're not 18 going to be able to come up with a definitive answer on 19 cell-type leukemia simply because the number is too small 20 based on any single study. 21 The assessment has to come from several 22 studies. We have cell-type specific information in the 23 Chevron study for example, but that itself doesn't tell us 24 that there is an association or there isn't because the 25 sample size was not big enough for CLL, for example. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (167 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 129 1 But on top of the Chevron study we also have a 2 number of studies published around that time or subsequent 3 to that. When we combine all the studies together, we 4 have a pretty good picture. 5 Q. And the studies that are combined together have 6 exposure to benzene quantitatively assessed? 7 A. There are some benzene-specific studies such as 8 the Rinsky study, such as the CMA study. On top of that, 9 we also have refinery studies overall. 10 Q. The refinery studies do not address benzene 11 exposure specifically, do they? 12 A. Not specifically, but it's pretty safe to make 13 that assumption that refinery workers are exposed to 14 benzene. 15 Q. Which ones? 16 A. The majority of refinery workers would be 17 exposed to benzene. 18 Q. How about a worker who works in a boiler in a 19 refinery in Port Arthur? Is he exposed to benzene? 20 A. If he walks through the refinery, there would file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (168 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 21 be some exposure to benzene. 22 Q. And so that would be enough to allow you to 23 draw conclusions about cell-type and benzene 24 exposure/causation information? 25 A. Not just based on the example that you gave. 130 1 We are also talking about operators who are constantly 2 exposed to benzene. 3 Q. Which operators are constantly exposed to 4 benzene? 5 A. There is one study that we did with Gulf. We 6 identified a process unit that makes benzene and we 7 studied the operators of those units. 8 Q. Can you accept that operators on a benzene unit 9 can have lower exposure to benzene than employees on a 10 coking unit? 11 A. Sure. 12 Q. How did that happen? 13 A. Depending on the process, the engineering, file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (169 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 14 whether there is any leak and how they sample the products 15 and so on. 16 Q. What about benzene exposure being lower for 17 operators on a benzene unit than operators on a catalytic 18 reforming unit? 19 A. That's also possible. 20 Q. So I guess I go back to the refinery worker 21 studies. Let's take the Mobil Beaumont refinery worker 22 study that you did. Which of those workers would have 23 exposure to benzene and how would you categorize who had 24 higher exposures than lower? 25 A. You're talking about the Mobil study in the CMA 131 1 study. 2 Q. The refinery workers. 3 A. We did not have exposure quanitification in 4 that study. 5 Q. Which refinery worker study are you aware of 6 where there is benzene quantification? 7 A. I'm not aware of any. The only study that has file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (170 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 8 some qualitative benzene exposure data would be the Gulf 9 study that I did with Dr. Wen. 10 Q. And there you did high, medium and low? 11 A. No. We identified a process that involved 12 benzene exposure and studied only those people that worked 13 at those units. 14 Q. So that's one study? 15 A. Right. 16 Q. And was there enough information from that one 17 study of refinery workers that you could get cell-specific 18 information with benzene exposure? 19 A. No, not based on one study. 20 Q. I guess, Dr. Wong, it seems to me like a great 21 leap that you're willing to go from the refinery worker 22 studies where benzene is not qualitatively or 23 quantitatively expressed on the one hand when I question 24 you about that at Mobil and Beaumont to then leap over to 25 while looking at refinery worker studies in general saying 132 file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (171 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 1 everyone is exposed to benzene while working in a 2 refinery. 3 That's not logical and I want you to explain 4 how you can do that. 5 A. It's more than looking at refinery studies in 6 general. I told you we looked at--in forming my opinion, 7 I looked at studies of benzene-exposed workers and that 8 would include my CMA study. That would include the 9 Rinsky--the Rinsky NIOSH study, and that would include the 10 De Coufle study you mentioned earlier, and I also relied 11 on case controlled studies as well. 12 Q. And there is enough information on those that 13 you can do cell-specific analysis and have statistical 14 significance? 15 A. I'm not so sure the question is meaningful in 16 terms of statistical significance. What I can say is 17 based on all the data combined, I think the data that I 18 have--I'm quite confident about the data. 19 Q. Do you find in those same studies you just told 20 me about that there is an increase in AML, acute 21 myelogenous leukemia? 22 A. Yes. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (172 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 23 Q. And there is no increase in the other forms of 24 leukemia or lymphoma? 25 A. It may not be true for every study, but some 133 1 studies report an increase of AML. 2 Q. You said you're looking at them in total. 3 A. Right. 4 Q. And I'm saying, when you look at them in total 5 is the AML level statistically increased and being caused 6 by benzene exposure? 7 A. For some studies, yes. 8 Q. I'm talking about whether you looked at them 9 all. 10 A. I did not combine them and look at the overall 11 result for AMLs. I didn't do that. 12 Q. Did you do that for any of the cell-types? 13 A. Just for CLL. 14 Q. Have you written that up anywhere? 15 A. No. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (173 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 16 Q. Have you reduced that to writing in any way? 17 A. Not yet. 18 MR. HOBSON: Let me get the reporter to mark 19 some of this stuff. 20 (Deposition Exhibits 1, 2 and 3 were marked.) 21 Q. (BY MR. HOBSON) Dr. Wong, we have marked as 22 Deposition Exhibit No. 3 your statement that you gave to 23 OSHA on March the 4th, 1986, correct? 24 A. Yes. 25 Q. And Exhibit No. 1 is your most recent 134 1 curriculum vitae, and I take it that it's relatively up to 2 date? 3 A. Yes. 4 Q. And then Exhibit No. 2 is the letter that sets 5 out the records that you were sent with regard to this 6 case, correct? 7 A. It's two cases, yes. 8 MR. HOBSON: That's all I've got. Thank you 9 very much. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (174 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 10 (The deposition concluded at 2 p.m.) 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 135 1 I, OTTO WONG, do hereby certify that I have 2 read the foregoing transcript and that the same and file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (175 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 3 accompanying correction sheets, if any, constitute a true 4 and complete record of my testimony. 5 6 ____________________________ Deponent 7 8 ( ) No changes ( ) Amendments attached 9 Subscribed and sworn to before me this 10 _________day of __________________, 1993. 11 My commission expires ________________ 12 13 ___________________________ 14 Notary Public 15 Address ___________________________ 16 ___________________________ 17 18 19 BS 20 21 22 23 Smith vs. Hartford Accident and Indemnity file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (176 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 24 25 136 1 STATE OF COLORADO) 2 ) ss. REPORTER'S CERTIFICATE 3 COUNTY OF DENVER ) 4 5 I, Rebecca L. Semeyn, do hereby certify that I 6 am a Registered Professional Reporter and Notary Public 7 within the state of Colorado; that previous to the 8 commencement of the examination, the deponent was duly 9 sworn by me to testify to the truth. 10 I further certify that this deposition was 11 taken in shorthand by me at the time and place herein set 12 forth and was thereafter reduced to typewritten form, and 13 that the foregoing constitutes a true and correct 14 transcript. 15 I further certify that I am not related to, 16 employed by, nor of counsel for any of the parties or 17 attorneys herein, nor otherwise interested in the result file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (177 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 18 of the within action. 19 In witness whereof, I have affixed my signature 20 this ______ day of ____________, 1993. 21 22 23 ______________________________ PATTERSON REPORTING SERVICE 24 Rebecca L. Semeyn Registered Professional Reporter 25 and Notary Public 137 1 PATTERSON REPORTING SERVICE 55 Madison Street, Suite 510 2 Denver, Colorado 80206 3 Otto Wong 4 Applied Health Sciences, Inc. 181 Second Avenue, Suite 628 5 San Mateo, California 94401 6 Re: Smith vs. Hartford Accident and Indemnity VS Nos. 92-2915 and 92-3098 7 Deposition of OTTO WONG 8 The deposition in the above-entitled matter is ready for reading and signing. Please attend to this matter by 9 complying with ALL blanks checked below: 10 ____ arranging with us at 303-320-6628 to read and sign the deposition in our office file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (178 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt 11 ____ having deponent read your copy and sign amendment 12 sheets, if any (orig. signature page enclosed) 13 _XX_ reading enclosed deposition, signing attached signature page and correction sheets, if any 14 _XX_ within 30 days of the date of this letter 15 ____ by ___________ due to a trial date of___________ 16 ____ Due to a trial date of _________ a telephone 17 message was left with ______________________ on _____________ advising the deposition is 18 available for reading and signing until ________ at which time the original will be filed. 19 Please be sure that the signature page and accompanying 20 amendment sheets, if any, are signed before a notary public and returned to our office at the above address. 21 If this matter has not been taken care of within said 22 period of time the deposition will be filed unsigned pursuant to the Rules of Civil Procedure. Thank you. 23 PATTERSON REPORTING SERVICE 24 cc: Herschel L. Hobson, Esq. William B. Baggett, Esq. 25 Kenneth R. Spears, Esq. 138 1 PATTERSON REPORTING SERVICE 55 Madison Street, Suite 510 2 Denver, Colorado 80206 3 4 William B. Baggett, Esq. file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (179 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt Baggett, McCall & Burgess 5 3006 Country Club Road P.O. Drawer 7820 6 Lake Charles, Louisiana 70606 7 Re: Smith vs. Hartford Accident and Indemnity 8 Enclosed is the deposition of: OTTO WONG 9 _____ Previously filed. Forwarding signature page and amendment sheets. 10 _____ Signed, no changes. 11 _____ Signed, with changes, copy of which is enclosed. 12 _____ Unsigned, notice duly given___________ pursuant to 13 the Rules of Civil Procedure. 14 _____ Not signed, notice duly given___________ since trial is set for _________________. 15 _____ No signature required. 16 _____ Signature waived. 17 _____ To be signed in court. 18 _____ Signature pages/amendments to be returned to court 19 on date of trial. 20 _____ Mailed by certified mail # _________________. 21 _____ Hand-delivered _____________________________. 22 Enclosures: (As above-noted) 23 cc: Herschel L. Hobson, Esq. Kenneth R. Spears, Esq. 24 25 file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (180 of 181) [7/1/2010 10:57:08 AM] file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt file:///C|/Users/MadeK42/Desktop/EXPERT%20TRANSCRIPTS/WONG.O.93.07.SMITH.txt (181 of 181) [7/1/2010 10:57:08 AM]