Document Ne68VYzzQqZraJZB80n8vbn5p

From: Sent: To: Cc: Subject: Attach: Bruce Jarnot <jarnotb@api.org> Saturday, February 12,20055:18 PM (GMT) benzene-erp@listserve.api.org; benzene-srp@listserve.api.org; benzconsorttc@listserve.api.org benzconsort-oc@listserve.api.org SHS ... Year-end 2004 CC & EA Progress Reports Fudan Progress Rpt 02 2005.pdf; CC Progress report 02-08-2005.pdf Dear Colleagues - Attached for your review are 2 (of 3) year-end 2004 Shanghai Health Study (SHS) Progress Reports: -- Case Control (CC) Study / Dr. Wong -- Exposure Assessment (EA) / Drs. Liang & Fu The Disease Progression (DP) & Molecular Epidemiology (ME) Progress Report from Dr. Irons should be available shortly, and will be distributed separately. Best Regards - Bruce. ***** Bruce M. Jarnot, PhD., DABT American Petroleum Institute Regulatory and Scientific Affairs 1220 L Street, NW (Suite 900) Washington, DC 20005-4070 phone: (202) 682-8473 fax: -8031 email: jarnotb@api.org SH ELL-MCCLU RG-OS2327 Prog ress Report AML and NHL Case-control Study Activities through December 2005 Prepared by: Otto Wong, ScD. Applied Health Sciences, Inc. 181 Second Avenue, Suite 628 San Mateo, CA 94401 USA Tel: 1 (650) 347-7898 USA Fax: 1 (650) 344-6887 USA ottowong@aol.com February 8, 2005 Patient Enrollment: In the case-control study of acute myeloid leukemia (AML) and non-Hodgkin lymphoma (NHL), patients with preliminary diagnoses of AML or NHL are referred to the Joint Clinical and Molecular Laboratory (JCML) by the 31 participating hospitals for diagnostic confirmation. Only JCML-confirmed cases are accepted. The following numbers were provided by Dr. Richard Irons and represent the most up-to-date figures. As of September 2004, approximately 173 JCML-confirmed AML patients have been enrolled in the case-control study. The trend ofthe accrual rate of AML patients is quite positive and encouraging. Taking the trend into account, we estimate that the number of AML cases by the end of 2006 will easily meet or even exceed the originally targeted number of 500. As of September 2004, there were approximately 150 cases of JCML-confirmed NHL patients enrolled in the case-control study. The initial accrual rate of NHL patients was slightly lower than we originally estimated. The accrual rate in recent months, however, has improved markedly, after we expressed our concerns to the responsible clinical coordinators at the participating hospitals. Taking the recent upward trend into consideration, our revised estimate at this point is that we will have close to 400 NHL cases (maybe even slightly more) by the end of 2006. Our original estimated number of NHL cases was 500, which was derived based on a statistical power analysis. It is encouraging to know that the predicted number of NHL is close to our original target, even though we lost almost a year in time due to the 2003 SARS epidemic in China. 1 SH ELL-MCCLU RG-052328  However, it is important to maintain the accrual rate of NHL patients, as a sample size of 400 to 500 NHL patients will be needed for the planned analyses to be statistically meaningful. A large number of potential risk factors (both occupational and non-occupational) for NHL have been reported in the literature and we have been collecting information on these risk factors in the case-control study. We need to be certain that the sample size will be adequate for analyses stratified by these risk factors. A reduced sample size will greatly compromise the scientific merits of the study. Selection of Controls Previously, we reported some irregularities of control selection at a few hospitals. Some of the NHL outpatient controls at the Tumor Hospital (the participating hospital with the largest number of NHL patients) and Municipal Number 6 Hospital were not "real" patients, but they were relatives, co-workers or neighbors of the clinical coordinators at the hospitals. These individuals were not sick, but registered as outpatients at the hospital solely for the purpose of "participating" in the study. To ensure that all controls were real patients, we modified the matching criteria by requiring all controls be selected from inpatients only (regardless whether the case was an inpatient or outpatient). This modification took effect in mid-2004. We have reviewed all the controls enrolled since then and have not discovered any improper selection. We will continue to monitor the selection of controls in the future. We will continue to impress upon the clinical coordinators that we will monitor the selection of controls closely and irregularities will not be ignored or tolerated. The control selection criteria require that among eligible controls who meet age and gender requirements the two patients with admission dates closest to that of the case be selected. Among the controls recruited in the last six months, we noted a few whose admission dates were substantially later than those of the cases. In one extreme example, there was a difference of 92 days between the admission dates of the case and one of the controls. Although we cannot think of any obvious bias associated with a difference in admission date, this might effectively modify the eligible admission periods of the cases and controls and allow the controls to have a slightly longer opportunity for exposure. I have sought advice from Professor Harvey Checkoway of the Scientific Review Panel and he agreed that no obvious bias would be introduced by the late admission date. Professor Checkoway suggested that in the analysis we should truncate the control's exposure on the case's admission date. This problem in admission date occurred at a small number of hospitals. In the future, we will remind the clinical coordinators at these hospitals on a daily basis if no controls were identified within one week of the case's admission date. For clinical coordinators who do not identify controls in a timely manner, we may have to take more drastic action, including withholding payments. 2 SH ELL-MCCLU RG-052329 QAlQC of Questionnaire Interviews To monitor the questionnaire data quality, I conducted several QNQC checks. I checked completed individual primary and secondary questionnaires. The two most recent QNQC checks were conducted in November 2004. The reports of these two QA/QC checks are attached (Attachments 1 and 2). Overall, I found the quality of the completed questionnaires quite acceptable. The number of errors is very low, considering that there are literally hundreds of questions in each questionnaire. Most of the errors found are minor in nature. Through the QNQC checks, I have identified some problems with the questionnaires, such as open-ended questions on exposure. I have discussed these problems with Tom Armstrong and corrective actions will be taken. Exposure Assessment: Based on preliminary numbers in the entire SHS, we have identified 84 patients exposed to benzene. The most frequent occupations or industries among benzene-exposed patients were: shoe workers, painters, mechanics or machinists, and printers. Among the benzene-exposed patients, the earliest exposure occurred in 1941. The occupations with the highest exposure were brush painters and spray painters. The highest "typical" exposure was 690 mg/m3. The highest maximum exposure was 1300 mg/m 3 , and the highest minimum was 200 mg/m3. The longest duration of exposure was reported in a patient who worked with benzene for 40 years. In addition, several hundred patients were identified to have been exposed to other chemicals. The most frequent occupations or industries among these patients were: farmers, textile workers, teachers, engineers, drivers, cooks, chefs, and military personnel. A recent QA/QC of the completed exposure assignments indicated a significant problem of misclassification (Attachment 3). During my trip to Shanghai in August 2004, I identified several cases of patients (mostly mechanists) who used gasoline to clean machine parts and their hands. These subjects were classified as "unexposed" by the Exposure Assessment Committee in Shanghai. I expressed my doubt and concern regarding the classification. I asked He Yonghua about these subjects during my trip in November, and found out that nothing has been done to reassess their classification. Documentation of the decision process of the Exposure Assessment Committee is inadequate. In my November 2004 trip I tried to go over the decision process 3 SH ELL-MCCLU RG-OS2330 of the EAC in a few patients. Unfortunately, it was not possible to do that based on the JCML database, because certain components are not yet linked. Another problem is that documentation of the EAC decision process is still lacking. A close examination of the current exposure assessment process revealed the following potential problems. The staff in Shanghai is inexperienced and apparently did not understand some ofthe instructions. As a result, the current approach has not been consistent. Furthermore, both record-keeping and documentation have been poor. The value of the database maintained by the Shanghai Municipal Institute of Public Health Supervision (IPHS) has been overstated. The level of cooperation of IPHS has been much lower than we anticipated or expected. It is imperative for us to conduct a comprehensive review to determine the current status of exposure assessment. We need to reassess and revise the procedures and instructions that were developed more than a year ago before any interview of cases and controls took place. For example, based on the feedback from interviews, some questions in both primary and the secondary questionnaires do not appear to be meaningful or productive. A case in point is that some questions in several secondary questionnaires are in English. This would present a major problem as most interviewers and interviewees do not understand any English. Project Management: In a number of discussions, our Chinese colleagues expressed concerns regarding the inadequate communication with the US investigators. We are in the process of setting up routine reports and communications on a periodic basis. We need to encourage the Fudan team to voice their concerns and problems to us so that corrective actions can be taken. We continue to feel the need and demand to devote more time to project management than originally estimated. In particular, the demand for greater physical presence in Shanghai and working with the Fudan team in person continues. 4 SH ELL-MCCLU RG-OS2331 ATTACHMENT 1 QA/QC of Questionnaires by Otto Wong (8 November 2004) QA/QC 1 Included in this QC are 12 questionnaires from September 2004 that were checked by Miao. These 12 were selected randomly (last two ID digits being even) from a total of 42 questionnaires completed in one week in September 2004. 102804 Primary questionnaire: D12 - "Did not bring ID card." Most people remember their ID number. Agree with i'o1iao's overall assessment Secondary questionnaire: (Welder) Q3/4 - Not answered. Agree with rv'liaofs overall assessment 102768 Primary questionnaire: Job 2 OEl - No date for exposure to gasoline. Job 3 OEl - No date for exposure to gasoline. Job 4, EHCl - No month for the beginning date. Beginning year=2000 for Job4, but end date of Job3=2001/07. Agree with rv'liaofs overall assessment Secondary questionnaire: (Carpenter) Q7 - Use of gasoline: no answer. Agree with ivJiao's overall assessment. 5 SH ELL-MCCLU RG-OS2332 Secondary questionnaire: (Rubber) Q2 - No answer for many substances. The same is true for many questions. No indication whether "yes," "no," or "don't know." [Note: The patients simply do not know the technical names of many chemicals, so most answers will be "Don't know." The interviewers get tired of repeating "Don't know.'l Agree v\iith ivliao's overall assessment. 102766 Primary questionnaire: No problems found. Agree IrVittl f\1iao's overall assessment 102764 Primary questionnaire: No problems found. Agree with ~Jliaors overall assessment Secondary questionnaire: (Health care) QG - All answers (mutually exclusive) were selected. Q2 - All answers (mutually exclusive) were selected. Agree v\iith ivliao's overall assessment. 102762 Primary questionnaire: D12 - "Did not bring ID card." Most people remember their ID number. D14 - No month or day for birth date. ivJiao noted problems vvith iVlH5. (She wants a positive indication of "No" for diseases that the patient did not [lave! but there is only a column of "Yes." I would not consider this an error, Howeverf we need to be consistent. Secondary questionnaire: (Farming) Q3A, 3D, 3E - No answer checked. 6 SH ELL-MCCLU RG-OS2333 Agree wittl fv1iao's overall assessment, 102760 Primary questionnaire: No problems found. iVliao noted problems with 1\1HS. (She wants a positive indication of "No" for diseases that the patient did not have; but there is only a column of "Yes," I would not consider this an error. However, we need to be consistent 102728 Primary questionnaire: DAE2 - Changed from "No" to "Yes" without the interviewer's stamp. Agree v\iith i\fliao's overall assessment. 102722 Primary questionnaire: No problems found. Agree IrVittl f\1iao's overall assessment 102682 Primary questionnaire: Job 2, EHB2 - No month for ending date. DAE2 - No answer checked. Agree wittl fv1iao's overall assessment, Secondary questionnaire: (Farming) GF10D - Changed answer without stamp. i\fliao missed the above. 102684 Primary questionnaire: 7 SH ELL-MCCLU RG-OS2334 Job1, OE3 - No answer for "hours/week." f\1iao missed the above, [Note: ~/1iao toid me that in this case it means tile patient did not know, This is fine as long as we all knmv that leaving it blank means "don't knmv,tr] Secondary questionnaire: (Farming) No problems found. Agree with rv'liaofs overall assessment 102680 Primary questionnaire: MH5 - Did not cross out diseases that the patient did not have, as instructed by Miao. i\fliao missed rVlH5, [Note: Miao said this was OK as long as the overall answer was "no,"] 102660 Primary questionnaire: Job 1, OE3 - No answer for "hours/week." f\1iao missed the above, [See note above,] Secondary questionnaire: (Farming?) The copy of 2nd Q was incomplete. Only one page was copied. !'vliao did not note the incomplete copy, QA/QC 2 Included in this QC are 9 questionnaires from October 2004 that were checked by Miao. These 9 were selected randomly (last two ID digits being odd) from a total of 42questionnaires completed in one week in October 2004. 102711 Primary questionnaire: Job2, EHB15 - Frequency not filled out. 8 SH ELL-MCCLU RG-052335 Agree with ~Jliao's overall assessment Secondary questionnaire: (Farming) Ql - No answer selected. Agree with ivJiao's overall assessment. 102659 Primary questionnaire: No problems found. Agree IrVittl f\1iao's overall assessment 102679 Primary questionnaire: No problems found. Agree with rv'liao's overall assessment Secondary questionnaire: (Manager, supervisor) No problems found. Agree with f\1iao's overall assessment 103137 Primary questionnaire: No problems found. Agree wittl fv1iao's overall assessment, Secondary questionnaire: (Food industry) Ql - No answers selected. Agree v\iith ivliao's overall assessment. 103139 Primary questionnaire: 9 SH ELL-MCCLU RG-OS2336 No problems found. Agree wittl fv1iao's overall assessment, Secondary questionnaire: (Printing) Q7- No answers selected. No answers for other questions as well. Agree with rv'liao's overall assessment 103117 Primary questionnaire: Jobl, OEl - No "hours/week." NOEH5 - Change made by Chen. How did he get involved in modifying the Q? Agree with iVliao's overall assessment 103115 Primary questionnaire: No problems found. Agree with [\1iao's overall assessment Secondary questionnaire: (Kitchen worker and dish washer) No problems found. Agree with rv'liao's overall assessment 103075 Primary questionnaire: Jobl, OE3 - No "hours/week." Agree v\iith ivliao's overall assessment. Secondary questionnaire: (Farming) Ql - No answer checked. 10 SH ELL-MCCLU RG-OS2337 GF21 - Some answers unchecked. fv1iao did not indicate that she checked 2nd Q, 103077 Primary questionnaire: No problems found. Agree with rv'liaofs overall assessment Secondary questionnaire: (Military) No problems found, but some multiple choices were not checked (however, "none" was written on the margin. ivJiao noted incompiete answers. Summary Overall, I find the quality of questionnaires quite acceptable. The number of errors is very low, considering that there are literally hundreds of questions in each questionnaire. Most of the errors found are minor omissions. Miao did a good job in checking the questionnaires. Overall, I agree with her assessment. Open-ended questions on exposure may present a problem. The quality and usefulness of the answer depend on the industry, the occupation, the patient and the interviewer. Whether an answer is "acceptable" or not depends on all these factors. We need to remember that the interviewers, for the most part, are not familiar with the industry, the profession, or the tasks involved. We should not be overly critical, nor should we expect too much. If we are overly "aggressive" in getting "details," we may introduce interviewer bias and/or inaccuracy. If we want further improvement in this area, we need to provide very clear and definite guidelines to the interviewers. Actions: (1) Certain parts of some secondary questionnaires are not clear. I will talk to Tom Armstrong about it. (2) Miao will have a meeting with the interviewers on November 9. She will go over the problems with them. She will also let them know that overall they are doing a good job. 11 SH ELL-MCCLU RG-OS2338 ATTACHMENT 2 QA/QC of Questionnaires by Otto Wong (10 November 2004) Included in this QC are 15 questionnaires that have not been checked by Miao. They were selected randomly from the file cabinets where completed questionnaires are stored. The ID numbers of these 15 questionnaires (5 matched sets) are: 2534, 2535, 2536, 2551, 2552, 2553, 2516,2517,2518, 2572, 2573, 2574, 2611, 2612, 2613. Results: There are not major errors or omissions found. Overall, the quality is quite acceptable. The following common problems are noted in some of the questionnaires: (1) Handwriting was difficult to read. This may slow down data entry or create mistakes. (2) When the answer is "don't know," sometimes that box is not checked. (3) When the answers to all the questions in an entire section (particularly secondary questionnaires) are "no,' sometime the boxes are not checked, but the "NONE" in large print is written on the margin in that section. This is acceptable, but we need to be consistent. The number of such errors is very small. Conclusion: The findings are quite encouraging. Miao should provide some positive feedbacks to the interviewers. We need to maintain the level of quality in interviews. 12 SH ELL-MCCLU RG-OS2339 ATTACHMENT 3 OAtOC OF EXPOSURE ASSESSMENT IN THE CASE-CONTROL STUDIES By Otto Wong (12 November 2004) During my trip to Shanghai in November 2004, I conducted a brief QA of exposure assessment in the case-control studies of AML and NHL. I worked mainly with Ye Xibiao and He Yonghua. My QA included the following: 1. Going over the exposure assessment procedure with both Ye and He. 2. Examining a sample of completed primary and secondary questionnaires. 3. Reviewing the exposure classification of the sample of study subjects in (2). He Yonghua provided some statistics of exposed subjects by data sources. As of 1 November 2004, there are 2186 completed questionnaires (for the entire SHS), of which 455 have not yet been entered in the JCML database. Among the questionnaires that have been reviewed by the Exposure Assessment Committee (EAC), 53 subjects were classified as benzene-exposed. The distribution below is based on a table provided to me by He dated 1 November 2004. Exposure data source "Subject's own factory in IPHS database" "Subject's own factory in IPHS database" + "Visit to subject's own factory" "Subject's own factory in IPHS database" + "Surrogate factory in IPHS database" "Surrogate factory in IPHS database" "Visit to subject's own factory" "Visit to surrogate factory" "Visit to surrogate factory" + "Surrogate factory in IPHS database" "Visit to surrogate factory" + "Literature" "Literature" + "Expert" + "Factory visit" All others Number 1 3 2 18 4 5 2 0 1 17 The table provided by He Yonghua is potentially informative. The table in its current form, however, has some problems, such as duplicated categories with different numbers. For example, in the original table, 2 subjects are attached to the row labeled as {"Visit to surrogate factory" + "Surrogate factory in IPHS database"} and 0 attached to another row labeled as {"Surrogate factory in IPHS database" + "Visit to surrogate factory"} . 13 SH ELL-MCCLU RG-052340 Problems aside, the distribution above demonstrates the limited usefulness of IPHS database as the source of exposure information for the subject's own factory. As for "Surrogate factory in IPHS database," I am not sure whether the search is systematic or comprehensive. The search strategy should be closely scrutinized. Also, "Literature" appears to have been underutilized. There are several problems associated with the use of Secondary Questionnaires (2nd Q). 1. It seems that the use of 2nd Q is still not consistent and somewhat subjective. Professor Liang also noted the fluctuating frequency of the use of 2nd Q. I am not sure where the problems lie. In any event, it should be made clear to the interviewers that certain occupations/industries will require 2nd Q. There should be no decision-making or judgment on the part on the interviewers. 2. The original 2nd Q are derived from projects conducted in the US. Some items in the 2nd Q may not apply to the Shanghai area in particular or to China in general, but these items have not been modified to account for locality differences. 3. Some of the questions in 2nd Q are in English. For example, in the 2nd Q for the chemical industry (2nd Q No.8, CHI), some choices of answers to questions are in English, such as "Polyimide," "Cyclodienes," "Carbaryl, carbofuran and aldicarb," "Picloram," "Glyphosate," Barbituates such as phenobarbitol," "Ataractics such as valium or librium," "Dimetapp," and "Elavil" (these terms are copied directly from the 2nd Q). We should not expect an average patient in Shanghai to understand English, let alone technical chemical terms in English. 4. The format of some questions in 2nd Q is confusing. For example, in the 2nd Q for farmers (2nd Q No. 21, GF), the answers to the question "On the average, how often do you apply insecticides, pestiCides, ..... 7" are: u Never u I did, but I don't remember the frequency. U I don't know. 1) times / (day. week. Month. year) 2) hours / (day. week. Month. year) It seems that something is missing and a fourth box labeled as "I did, and the frequency of application was as follows" should be inserted immediately below the third box. Furthermore, my understanding is that the Exposure Assessment Committee (EAC) seldom relied on the 2nd Q for their classification. If the 2nd Q are underutilized, we need to talk to EAC and review and revise the 2nd Q to make them more useful. If the information in the 2nd Q is not helpful, we need to reassess whether we should keep them. 2nd Q are time-consuming and a burden to both the interviewers and patients, because by the time when the 2nd Q is administered, the patient has already been interviewed for 45 minutes to an hour (Primary Q). I tried to go over the decision process of the EAC in a few patients. Unfortunately, it was not possible to do that based on the JCML database, because certain components are not yet linked. Another problem is that documentation of the EAC decision process is still lacking. 14 SH ELL-MCCLU RG-OS2341 During my previous trip to Shanghai in August 2004, I identified several cases of patients (mostly mechanists) who used gasoline to clean machine parts and their hands. These subjects were classified as "unexposed." I expressed my doubt and concern regarding the classification. I asked He Yonghua about these subjects during my trip in November, and found out that nothing has been done to reassess their classification. By now, we have had several months of experience of exposure assessment of actual patients in the study. We should conduct a comprehensive evaluation of the protocols and procedures. We need to identify problems and correct them accordingly. This can only be done by individuals who are experienced in historical exposure assessment in epidemiologic studies and who are meticulous about data collection. We should also reevaluate (possibly downgrade?) our expectation of exposure assessment in the SHS. 15 SH ELL-MCCLU RG-OS2342 Progress Report of Shanghai Health Study January 31, 2005 Prepared by Research group in Fudan University School of Public Health Introduction The Shanghai Health Study (SHS) project represents a comprehensive epidemiologic investigation consisting of three research studies: 1) Disease progression study (DP study) of etiology and pathogenesis of lymphoid and hematopoietic diseases, 2) Case-control study (CC study) of acute myeloid leukemia (AML) and non-Hogkin lymphoma (NHL), and 3) Molecular epidemiology study (ME study) of benzene exposure. Detailed descriptions of these research studies can be found in respective study protocols. To support these three research studies, the Joint Clinical and Molecular Laboratory (JCML) between the University of Colorado and Fudan University was established in 2002. Clinical diagnoses of all patients with lymphatic and hematopoietic diseases recruited for the CC and DP studies are performed at JCML in collaboration with the Department of Pathology at the Shanghai Tumor Hospital and the Department of Hematology at the Huashan Hospital of the Fudan University. Furthermore, analyses of biochemical molecular biomarkers for the ME phase II study are developed at JCML. The creation and development of JCML took place from October 2001 to June 2003. In July 2003 JCML became fully functional and patients were recruited into the research projects shortly thereafter. Another major component of the SHS is exposure assessment, which was designed by ExxonMobil Biomedical Sciences, under subcontract to the University of Colorado, and carried out at Fudan. A number of local government agencies are involved in providing support and information to exposure assessment, including the Shanghai Center for Diseases Control and Prevention (CDCP), Shanghai Municipal Institute of Public Health and Supervision (I PHS) and Shanghai District IPHSs. In particular, one of the major data sources is the exposure database maintained at the Shanghai MunicipallPHS. This report is not designed to account for many of the study successes to date, which are considerable in number. Rather, in this report we will attempt to summarize Fudan's perspectives of the challenges facing the SHS considering the progress to date. We will also discuss Fudan University efforts and suggestions to meet these challenges. SH ELL-MCCLU RG-052343 Part I: CC&DP Studies 1. Executive summary The participating institutions for the CC and DP studies include University of Colorado Health Science Center, Applied Health Sciences, ExxonMobil Biomedical Sciences, Inc., Fudan University School of Public Heath (FUSPH), 31 hospitals from Shanghai, Shanghai Hematology Society, Shanghai Pathology Society, Shanghai Municipal Institute of Public Health and Supervision (I PHS) and Shanghai Municipal Center for Disease Control and Prevention (CDCP). Up to Dec 14, 2004, a total of 2370 subjects have been interviewed. The exposure status of 1811 subjects was assessed. 84 were classified as benzene exposed; of which 57 were assigned quantitative exposure estimates. Quality assurance (QA) and quality control (QC) checks are being conducted on a regular basis. Challenges in the CC and DP studies are summarized as follows: Lack of adequate involvement of IPHS in exposure assessment Overestimated value of IPHS exposure database Inadequate technical resources to support the exposure assessment Unclearly defined or misunderstood terminologies in questionnaires regarding exposure assessment Underestimated the complexity and difficulty of questionnaire interview including the first and second level questionnaires, and factory inspection questionnaire JCML database not fully functional Insufficient support from clinical coordinators at participating hospitals in following control selection criteria Gaps in regular communication between Chinese and US researchers Lack of clearly defined roles and responsibilities of some of US researchers 2. PROGRESS 2.1 Questionnaire interview 2.1.1 Number of participating hospitals A total number of 31 hospitals in Shanghai have engaged in overall SHS that include 16 municipal hospitals, 12 district hospitals, 2 occupational diseases hospitals and 1 teaching hospital affiliated to a Military Medical University. 2.1.2 Number of questionnaires administered Up to Dec 14, 2004, a total of 2370 questionnaires, including 790 cases and 1580 controls, have been administered since mid June 2003 with an average rate of 132 cases per month. 2 SH ELL-MCCLU RG-OS2344 For the 3 person-staffed questionnaire taking team and approximately 22 working days per month, the average workload is about 2 questionnaires per person per day. 2.2 Data entry and errors check The double entry approach by two independent persons was introduced for the purpose of minimizing the errors during data entry process and checking the accuracy of data entered. The major procedures for data entry and errors check are as follows: 1) Creating a report comparing the difference between data entered and the original information 2) Checking entered data along with the original questionnaire 3) Revising check report 4) Accessing database and revising data by the data entry personnel 5) Finalizing check report and compiling it with the original questionnaire by the archiving clerk. 1880 (97.2%) of the totaled 1935 pieces of CC/DP questionnaires have successfully passed the double entry check; 55 were failed (2.8%) which were under further checked .. With further improvement of the JCML database and data entry skill, as well as closer cooperation among team workers, the data entry and error check have been more efficient over time. Observed errors have been greatly reduced; and the time spent for producing a check report dropped from 2 minutes to 1/2 minute. All error check reports produced are now printed out for archiving. 2.3 Exposure assessment The procedure for exposure assessment is based on what US experts have outlined in the protocol with slight modifications: 2.3.1 Procedures of exposure assessment (from EA T to EA) The initial assessment is performed based on the evidence of occupation/job history with potential exposure to benzene obtained from interviewed subjects' questionnaires and literature information. A total number of 1147 literature articles published in the Chinese medical literature related to occupational benzene exposure between 1960 and 2003 were collected and electronically archived by industry and year that covered dozens of industries of relevance. The IPHS database, questionnaire records, literature information and current worksite monitoring data provide a scientific basis for the initial assessment, which will result in a categorized summary identifying "exposure", "non-exposure" and/or "uncertain" . After having completed the initial assessment, the following steps were to have taken place: 1) Questionnaires are doubly checked by the Exposure Assessment Coordinator (EAC), and progressed to data entry if all are considered acceptable; 2) The initially-assessed information such as "benzene exposure", "uncertain benzene 3 SH ELL-MCCLU RG-052345 exposure" and "uncertain exposure to other hazards except benzene" is presented at the weekly Exposure Assessment Committee meeting (EACm) for further evaluation and confirmation. The information of confirmed "non-exposure" and "exposure to other hazards" is entered into the JCML database; 3) In collaboration with IPHS, the EAC should further search for supplementary information from the IPHS database, industry records, and other resources available to make a better assessment; 4) Telephone calls by Exposure Assessment Team (EAT) and/or EAC to the subjects for further information, if possible; worksite survey and monitoring for collecting general information and working conditions of the factory to facilitate a quantitative assessment and to provide evidence in selecting a proper surrogate factory; 5) EAC working along with IPHS staff to search for surrogate factory information from the IPHS database. 6) Preparing all materials relevant to exposure assessment including worksite monitoring data from the IPHS database, current survey data, literature information, etc by the EAC 7) Performing both qualitative and quantitative assessments, and/or recommending a work proposal for further assessment by EAC; 8) EAC working along with EAT to implement the assessment proposal, including simulating exposure assessment, current exposure worksite monitoring, surrogate worksite monitoring; and providing literature information in cooperation with IPHSs, where 10cailPHSs are designated as responsible for providing the worksite; 9) Worksite monitoring and surrogate surveys by EAT; 10) Updating assessment based on the updated information available by EAC; and 11) Data entry and custody. 2.3.2 Key outcomes of exposure assessment Up to Dec 14,2004, 1811 subjects were assessed to clarify their potential exposures. 1181 (65.21 %) of these were identified as "non-exposed" meaning neither exposure to benzene nor to any other occupational hazards; 84 (4.64%) were "exposed to benzene and other chemicals"; 541 (29.88%) were "exposed to other chemicals except benzene"; and 5 (0.28%) were "uncertain" (Table 1.1). Table 1.2 showed sources supporting the quantitative exposure assessment. Table 1.1 Summary of exposure assessment (up to Dec 14, 2004) Status of exposure No. case % Exposure Benzene only 0 I0 Benzene and other 84"( 4.64% chemicals or hazards* Other chemicals except 541 29.88% benzene I Non-exposure 1181 65.21% I Uncertain/pending Total 5 1811 0.28% 100% 4 I SH ELL-MCCLU RG-052346 ? Notes: Of these 84 cases regarded as benzene and other chemicals exposed; quantitative assessments for 57 of the 84 cases have been completely accomplished; while the remaining 27 cases are being further quantitatively assessed. Table 1.2 Proposed evidence-based EA for 84 cases of "Combined exposure to benzene and other chemicals or hazards" Methods Quant. assess't Completed IPHS database relevant factory o No. No yet completed IPHS database surrogate factory IPHS database relevant factory + worksite walk through survey 19 6 2 worksite walk through survey 19 7 Worksite of surrogate factory Worksite of surrogate factory + IPHS surrogate database Literature + surrogate worksite 5 2 8 3 5 Literature + expert + worksite 2 IPHS database relevant factory + surrogate factory PHS database relevant factory+ surrogate worksite worksite walk through survey + surrogate historical information from other channels outside SHA Factory historical information + surrogate worksite 2 2 No. Bn exposed worker 57 27 Total 84 2.4 Newly developed DP study The newly developed phase I study is characterized by the analysis of additional benzene poisoning cases, which constitutes an important component of the DP study although it has not been originally documented in the respective protocol. To keep pace with the newly developed study, a further searching was initiated in collaboration with Dr. Ni Weimin and the Occupational Physician Committee (OPC), as well as the hospitals responsible for the periodical health surveillance. Under a multi-collaborative approach, we acquired the historical data of health monitoring for 194 cases of benzene poisoning (1964-2003) in late 2003. We have started abstracting the medical records of hospitalized patients diagnosed as benzene poisoning (BP), to initiate a study in depth of benzene exposure-effect on peripheral hematopoiesis as well as the bone marrow since September 2003. The addition of BP cases is really a new part of the DP study - since it is not only looking at medical records and peripheral blood counts. It is looking at the progression of disease, and is 5 SH ELL-MCCLU RG-OS2347 accessing clinical records and bone marrow smear. As expected, however, this study has also experienced a series of challenges. 2.4.1 CDCP benzene poisoning database It was thought that the acquisition of the historical data of benzene poisoning and its relevant database would have been available in Shanghai CDCP. Unfortunately, the BP database was missing as a result of the separation of IPHS and CDCP. 2.4.2 Historical data re-collecting from local district CDCPs Shanghai Municipal CDCP has promised to work on re-collecting of the historical database on BP. CDCP has asked for the historical benzene poisoning cases from local district CDCPs. However, the re-collected information provided only a basic "demographic data" of the poisoning cases without acknowledging anything about the occupational exposure level, blood counts or bone marrow features. 2.5 Status of QA/QC 2.5.1 Standard Operation Procedure (SOPs) According to the standard procedure flowchart and SOPs for each study process, QA/QC practical procedures (Version 1) have been established since mid 2004. After several revisions, QA/QC practical procedures (Version 2) have been preliminarily documented. In order to make QA/QC inspection work smoothly, a clearly defined contrast table describing the position title and its corresponding responsibility for each QA/QC coordinator of relevance has also been listed. 2.5.2 QA/QC of CC/DP As SHS is composed of several studies, the QA/QC approach has thus included a number of inspection processes for the three different studies). In order to make the QA/QC inspection approach concrete and to ensure the entire quality of the study project, QA/QC for each study has been divided into several parts (e.g., the quality of the questionnaire, data entry error, control selection, etc.). For each study component, a QA/QC inspection record, i.e., the summary report is required to be available immediately after each QA/QC inspection has been accomplished. QA/QC inspection for the quality of questionnaire, control selection and the data entry error rate are inspected twice every month. The initial EA and final EA are checked once every month according to QA/QC SOPs. Two remediation meetings were held on Nov.9th and Nov.10th 2004, respectively for questionnaire quality and data entry errors. Criteria for questionnaire interviewing and filling-in were emphasized and unified. More emphases were placed on how to properly refer subjects to the second level questionnaire, and filling-in its opened-blanks. Scheduled dates for re-inspection on questionnaires and data entry error were also set. 6 SH ELL-MCCLU RG-052348 Several improvement reports for questionnaire quality and data entry error after the remediation meeting have been accomplished. In order to draw attentions from the study session coordinators and project coordinator immediately if any problem is found during the course of a QA/QC inspection, a written QA/QC report is periodically submitted to the session coordinators of each study. Of which, problems and the corrective actions were noted. By now, QA/QC monthly reports summarizing QA/QC approaches of the past three months have been submitted on a regular basis. 3. Challenges and our efforts 3.1 Cooperation and coordination FUSPH has been facing an unaffordable responsibility shifting that should naturally be borne by other participating institutions of relevance. 3.1.1 Clinical coordinators To fully inform clinical coordinators with the scope and goals of the study and the responsibilities they should take, dozens of clinical coordinators from all 31 hospitals have been gathered at more than 3 meetings held separately since June 2003 as not all coordinators were be able to present them at the same time. A small number of these coordinators were still absent from all meetings; and some of the coordinators were shifted to other departments that are not participants of the SHS. These changes have brought us much more inconvenience than expected and resulted in profoundly negative impacts, such as: 1) many clinical coordinators did not pay enough attention to fully understand what role they should play, how they should explain the Informed Consent in detail to their patients, that have resulted in some sudden withdrawal from the proposed interview; 2) some clinical coordinators did not properly follow the instruction of control selection, which include matching the closest admission date of control with case, randomizing the distribution of control selecting sources, etc. We have held remediation meetings with some coordinators, and have revised some aspects of control selection criteria thus we believe that control selection is now proceeding sufficiently. 3.1.2 Cooperation with IPHS We were confronted with similar difficulties while cooperating with IPHS. For example, the original proposal specified that the district IPHSs should have their own work team responsible for worksite monitoring to collect current benzene exposure data for subjects with potential benzene exposure. Unfortunately, neither the municipal nor the district IPHSs have carried out this work so far. FUSPH has to get everything ready in meeting with IPHSs' requirements, which has brought us an unaffordable burden of both time and manpower. In addition, although the leaders of the participant institutions were designated as persons taking primary responsibility for their participation in the SHS, the major 7 SH ELL-MCCLU RG-052349 professional and managerial transactions are generally handled by FUSPH side only. For example, EA is recognized as one of the most complicated processes which require an active participation of both municipal and district IPHSs conducive to providing adequate information and helping in timely access worksites of relevance to link the 'case' with 'exposure'. However, the responsibility of professional and managerial transactions has not been regularly shared by the relevant institutions, which sometimes put FUSPH in an extremely overloaded and stressful situation. (This shifting burden of responsibility is also an issue for the M.E. study - IPHS wanted to lead the factory selection committee, but they did not. This should be mentioned later - in the M.E. part of the report) 3.2 Questionnaires 3.2.1 Development of questionnaires 3.2.1.1 First level questionnaire The CC/DP questionnaire is composed of the first level and second level texts which were designed in US but being used in Shanghai, China. To make the questionnaire more practical and feasible, FUSPH researchers worked together with US experts making many modifications in terms of its content and format. These innovating measures generated some updated versions of the questionnaire on 2002/04/01, 2002/06/13, 2002/08/20, 2002/11/19,2003/01/27,2003/02/18, etc, based on the initial version of 2001/06. However, some minor gaps still exist between US designed questionnaire and the one commonly used in China. It might be partially derived from cultural difference; but most importantly, it simply reflects the gap between the research desirability and the real practicability. For examples, there are too many "open-ended blanks" in the questionnaire Part 4 "Employment Histories" and Part 5 "Occupational Exposure List" that have been proved too difficult to be completely and accurately answered by interviewees under certain circumstances, such as ill health condition and/or awkward interview environment, since not every hospital was willing to provide a private room for interviews. 3.2.1.2 Second-level questionnaire The second-level questionnaire was designed relying on a study conducted in the USA, which required that 63 types of occupations/jobs should be referred to the second-level questionnaires. FUSPH researchers translated all the 63 types of occupations/jobs into Chinese but some disputes might still exist due to: Conceptual difference in defining and distinguishing occupation and job, work process and job activity; Being bored by the redundancies of too many "open-ended blanks" that often make both interviewer and interviewee confused and lose confidence in fully responding them; and The information obtained from the second-level has not been substantially useful to 8 SH ELL-MCCLU RG-052350 the exposure assessment. Therefore, most questionnaire administers do not appreciate and do not highly value of using second-level questionnaire. Rather, they would preferably concentrate themselves to achieve the requested goals by interviewing the primary questionnaire as perfectly as they could. By the end of 2004, it was found that among a total number of 1924 questionnaires, that were accomplished data entry, 459 of them were referred to the second-level questionnaires accounting for 23.86%. The referring rate did not seem to be as high as expected although 43 of the 63 occupations/jobs listed (68.3%) have been referred for further investigation. Of which, farmer (46.2%), textile worker (6.8%), health care worker (5%), mechanical maintenance worker (2.4%), and cook (2.4%) were ranked as the top 5 occupations/jobs being referred. 3.2.2 Questionnaire interview 3.2.2.1 Pilot tests To investigate the feasibility of the questionnaire, we conducted at least three trials of pilot study in hospitals at various levels prior to the formal study started. The three pilot investigations were initially carried out at the affiliated hospitals and gradually expanded to all the 17 participant hospitals on 02/04/23, 02/05/28 and 02/08/02. 3.2.2.2 Questionnaire interview The experience gained from the pilot studies has built a good foundation for the formal study. However, the rapid growth of the number size of hospitals involved, which has been increased from 17 at the beginning to 31 hospitals up now, has increased difficulties. These constraints include: the mobility of clinical coordinators, disorderly work environment, less personal confidentiality protected (most cases were interviewed by bed-side at a multi-patient ward), and patients at ill health condition or elderly age. All theses constraints were so often frustrating questionnaire takers to get all answers in matching with the requirements of "integrity", "consistency" and "correctness". However, FUSPH questionnaire administrators have been making every effort in allowing for good quality, even if working conditions are difficult. 3.3 Control Selection 3.3.1 Criteria for control selection The original criteria for control selection of CC study were as follows: 1) Same gender; 2) Age 5 years; 3) No disease of the lymph or blood system; 4LCIosest admission date; 5) Same hospital; and 6) If the case is an inpatient, controls will be inpatients as well, if the case is an outpatient, controls will be outpatients as well. In addition, the source of the controls should be randomly selected from the "admission registration center" in each of 9 SHELL-MCCLURG-052351 the hospital involved. 3.3.2 Challenges in practicing the criteria However, the effectiveness of the criteria is very much dependent on how they are followed. Having had some investigations, we found that 1) clinical coordinators could not easily access the "admission registration center" because of its confidentiality; 2) out-patients have always been hasty while they are visiting hospitals, which make them unlikely to be the controls; 3) clinical coordinators are always overloaded and overworked during on their duty that frustrating them to carefully follow control selection criteria; and 4) some of the physicians might not fully understand the importance of the selection criteria. For example, we spoke to a number of directors from the participating hematology institutions to explore the possibility of control selecting through "admission registration center". 14 of 16 directors did not think it is possible to access the registration system frequently. A meeting was held to further discuss on this issue and got one positive response only saying that they have successfully selected controls through "admission registration center". It implied that the gap between scientific rigor and reality of control selection criteria had become an obstacle preventing CC/DP study to move forward smoothly. 3.3.3 Efforts in narrowing the gap Having had an analysis from our own data and an investigation based on more hospitals beyond the project, we found that: 1) there were no significant differences of both demographic characteristics and life style between outpatients and inpatients if two controls were selected rotating from different departments (Annexed table 1); and 2) it showed no substantial difference of disease distribution between in-patients discharged from hospitals in Shanghai in 2002 and the controls selected in the CC/DP study (Annexed table 2). 3.3.4 Revision of the criteria Based on the analysis and investigation, both US experts and FUSPH researchers decided to make a compromise of modifying some of the criteria to make them more practical without introducing significant bias by adding: 1) If a control cannot be found, relax age to 1 0 years; 2) If only one control can be found, we will keep both the case and the one control; and 3) To circumvent the problem of irrelevant people being pretended as "patients" we have limited control selection to inpatient only since early 2004 because anyone can easily become an "outpatient" by going through a simple registration procedure. (It seems that selecting a patient with the 'closest admission date' is still a problem occasionally, but experts believed that it won't bring significant bias to the study.) 3.4 Communications with US experts 10 SH ELL-MCCLU RG-OS23S2 Training courses on different topics with diverse emphases were offered at the early stage that has granted trainees a sound scientific basis of industrial hygiene benefiting to the exposure assessment. However, with the SHS moving forward to a deeper and critical stage, some unsolved problems have become apparent. For example: 1) US experts could not work here as frequent as the project needed, which sometimes leads to a "vacancy" of technical supports; 2) gaps between theory and practice existed due to lack of sufficient practice; 3) unclearly defined role for each US expert at the early stage of the project resulting in team workers confused about to whom they should consult with; and 4) insufficient communication had also brought about barrier and misunderstanding. 3.5 Unstable JCML database JCML database setting has been a complicated issue, which depends on how well the programmer understands the study design and the data per se, and the frequent communication between project designer and software programmer. However, a big gap existed between the two counterparts for quite a while at the early stage. The process of database setting and refining had been taking too long to meet data entry demands. As a result, the database did not run quite well until later 2003, which had hindered data entry and errors check to a certain extent. Having put in greater efforts, the condition has been dramatically improved, which was witnessed by the efficiency of task performance. For example, by double entry check, the complete match rate for a piece of questionnaire with more than 470 variables has been greatly elevated. However, some more problems still exist in terms of data entry processing and exposure assessment report creating. For example, input content might be accidentally missing in variable item; exposure assessment results could only be linked with one job, etc (see details in 4.6 Further improvement of JCML database). 3.6 Challenges facing exposure assessment The exposure assessment has been witnessed as one of the most complicated and difficult parts of CC&DP Studies. Unfortunately, we may have underestimated the challenges of the situation and not been taking adequate action in conquering the difficulties although everyone recognizes that would have become a tough issue. 3.6.1 Overestimating the value of IPHS database Imperfect integrity of IPHS database, inadequate factory data records, job mobility of factory workers involved and un-matched exposure data associated occupation were the key elements affecting the study. For example, a total number of 47 factories were identified as the workplaces that were associated with subjects being investigated. However, there were only 13 of them (28%) with data available at the IPHS database; 6 of the 13 were considered partially useful and none were really meeting the requirements of exposure assessment. In addition, as the relevant information has not been timely updated, it has been hardly sorting out convincible evidence from IPHS database to support the proposed "surrogate exposure assessment". 11 SH ELL-MCCLU RG-052353 3.6.2 Lack of adequate involvement of IPHS It's commonly recognized that the exposure assessment needs close joint-efforts devoted from both sides of FUSPH and IPHS. However, the contract with IPHS was negotiated between University of Colorado (UC) and IPHS. All the activities of IPHS are in accordance with the above subcontract; and FUSPH has no authority of control over IPHS's performance. Therefore, it has always been challenged by: Inadequate information at the IPHS database without effective remediation measure; Conflictive role between governmental health inspectors and scientific researchers borne by a limited number of overloaded IPHS staff preventing municipal IPHS to take a lead towards achieving access to district IPHSs; Lack of an incentive scheme to motivate active participation of district IPHSs. All these issues have badly hindered a further progress of exposure assessment. As being mentioned, exposure assessment will be very time consuming, as it needs a highly multi-institutionally supported and well-coordinated collaboration. The following example presents you a general feature of the entire process for an exposure assessment that took about 8 months to accomplish: Example: Case #867, male, a paint warehouse keeper at the locomotive maintaining workshop of Shanghai Railroad Transportation Bureau during 1988-2003. The hospital-based interview was done on 2003-12-16, which was followed by: 2004-4-10: Searching information from IPHS database: EAT obtained some of the exposure data, but failed to locate the exposure site associated with job-type information; 2004-4-14: Transmitted IPHS a name list of factories that were supposed to be useful for further assessment; 2004-7-23: Made telephone calls to get contact with the subject for further information; 2004-7-30: Conducted field surveys; 2004-8-3: Received a pile of historical data of worksite monitoring from the Health and Anti-epidemic Station, Shanghai Railroad Transportation Bureau; and 2004-8-6: Accomplished the quantitative exposure assessment. From this typical case, it might imply that: 1) significant information resources are needed to simply make one exposure assessment; and 2) to achieve an information transfer with satisfied response usually take quit a while of time period. For the above case, it took about 3.5 months to get a response with respect to the subject of relevance (from 2004-4-14 to 2004-7-30). The retrospective benzene exposure assessment for historical cases of benzene poisoning might be more difficult versus the phase I study as the cases were distributed into more than 150 factories, which urges a closer and more promising collaboration from both municipal and district IPHSs. 12 SH ELL-MCCLU RG-OS23S4 3.6.3 Inadequate scientific supports to exposure assessment team (EA TIEAC) The significant setback with respect with technical support to the exposure assessment issue appeared to be: 1) Different views among researchers We were not surprised by different views sometime existed with respect with either conceptual or methodological aspects of exposure assessment among researchers. However, problems might thus be derived from insufficient communication and lack of prompt response to the disputes, such as not taking a follow-up study or further discussion on the debatable cases soon after the conflict occurs. For example, how should we categorize the investigated subjects with "low" or "very low" benzene exposure that are very much dependent on the pattern of exposure? For ordinary taxi drivers, the "low" or "very low" benzene exposure due to gasoline could basically be neglected but for mechanical workers who were using benzene-containing gasoline as degreaser, the "low" even "very low" exposure could mean much more than those of taxi drivers. Despite the existence of different opinions among researchers is not uncommon what we have been challenged are how to handle these disputes and to reach an agreement timely conducive to the success of exposure assessment. 2) Limited function of temporary EAC experts First, the temporary EAC experts serve on a "part time" basis that might not allow them to spend more time and energy to deal with the problems persistently. Second, as the IPHS database has neither provided proper data for the assessment nor the convincible evidence supporting the option of 'surrogate assessment' that exaggerates the EA issue becoming even tougher. In that case, the temporary EAC experts would have thus confronted more challenges that are unlikely to be solved by them. Third, the proposed systematic approach of 'surrogate and/or simulate exposure' would unable be successfully designed and proceeded without proper historical records and clear definition of the approach that could not be solved by the EAC experts alone either. 4. Suggested solutions 4.1 Human resources More staff is badly needed. For example, the staff who is responsible for immediate questionnaires check after back from the hospitals to avoid mistake occurring while data entering, and the staff working for EAT and/or ME, etc. Several solutions are suggested as 1) The questionnaire administration coordinator should be released from a heavy workload of making questionnaire copies to the immediate Os check after being interviewed from the hospitals; 2) 1-2 more team workers might need to be recruited working for EAT and/or ME surveys; and 3) a multi-disciplinary temporary advisors and IPHS staff members might be invited to participate in the EAC meeting either on a regular 13 SH ELL-MCCLU RG-052355 basis or under certain circumstances whenever it is necessary. 4.2 Communication and technical support More timely technical support from US experts and communications between FUSPH researchers and US experts are crucially important to the further progression of the project. The criteria and requirements in meeting with the study Protocol should be clearly stated and defined for each component of the study. A Quarterly Summary Report focused on work status, QA/QC check and remediation steps taken is suggested to become a routine documented circulation for both sides' researchers to prompt the communication and technical support. 4.3 Second-level questionnaire Some more errors were found in the second-level questionnaire due to unclearly defined job title or work task in the 63 jobs listed, and/or conceptual difference in understanding the definition for job title or work task between China and US. Most of the questions listed in second level questionnaires cannot be clearly answered by interviewees. Further communication on the value of second-level questionnaire and its proper use should be further clarified towards reaching an agreement between US experts and FUSPH researchers. 4.4 Control selection Some of control selection criteria still needed to be verified. It seems that the difference of admission date for the cases and controls has not been definitely regulated. Is "the closer admission date, the better?" For example, the difference of the admission date between Case and Control was found to be as long as 61 days (from RenJi hospital), which has been too hard to judge if it is qualified or not during a QA inspection. Some of the criteria need to be further clarified and documented. FUSPH will pursue regular QA/QC approach to monitor how the revised control selection criteria are followed. 4.5 Enhancement of clinical coordinators' involvement The suggested solutions are:1) pursuing regular QA/QC check on the status of control selection; and 2) enhancing communication between JCML and clinical coordinators via regular meeting and immediate dialogue. 4.6 Further improvement of JCML database More efforts are necessary to make the database function more stable and friendly for users. The following problems have been reported to database developers and should be solved shortly. 4.6.1 Data entry 14 SH ELL-MCCLU RG-052356  Input content might be missing unexpectedly in variable item 'D111S At this address since'; Filled content in the original questionnaires in variable items 'MHD4 and MHH1 "Year Started' and 'Year stopped" ': it was sometimes expressed as 'year', which needs to be converted into 'month'; Some variable items in the screen questionnaire were provided with both choices of 'Yes' and 'No' simultaneously, which needs to be corrected. 4.6.2 Exposure assessment summary report Data couldn't be entered and updated sometimes The first entered character always disappears automatically Exposure assessment results could be input for one job only, which needs to be corrected by linking different documents with the EA decision. Subjects couldn't be linked with his/her factories. Similar corrective procedure needs to be taken. 4.6 Roles of responsibility of US researchers 1) Since there are number of researchers from the States, the role of these researchers should be clearly defined and then FUSPH staff members would know clearly whom they should report to. For examples, FUSPH researchers would like to know who is responsible for making final decision related to each session of these three studies, such as exposure assessment, project management, protocol and/or procedure changing, etc. 2) Greatly appreciate having definitively described roles and responsibilities for each of the US researchers 4.7 A comprehensive re-evaluation of the exposure assessment in terms of expectation and practicality. It is recognized that SHS is approaching into an important and critical stage, while the exposure assessment becoming a crucial component to the success of the CC and DP studies. Therefore, remediation steps towards prompting exposure assessment are urgently needed, for examples:1) re-evaluating exposure assessment in terms of expectation and practicality; 2) listing necessary information resources and the methodological approach for using the resources 4) human capacity building, enhancing and organizing; and 4) setting a clear work process and time schedule towards solving EA problems on a case by case basis. 15 SH ELL-MCCLU RG-OS23S7 Annexed table1 Demographic characteristics of outpatients and inpatients from CC/DP questionnaires accomplished (by Apr 2004) Variables Types of patients Total Outpatients Inpatients SEX Male Female 78f 60.5l 51 39.5 401 320 MARTIAL STATUS DNK DIVORCED MARRIED NEVER MARRIED WIDOWED 0 0 111\ 86.0} 12 9.3} 6( 4.7} 5( 0.7} 7( 1.0} 604( 83.8} 76( 10.5} 29( 4.0} 5( 0.6} 7( 0.8} 715( 84.1} 88( 10.4} 35( 4.1} EDUCATION DNK 0.8} HIGH MIDDLE 37 28.7\ 35 27.1 NONE 5 3.9\ POSTGRADUATE 3 2.3 PRIMARY UNIVERSITY 21 16.3\ 27 20.9 3( 0.4} 196 27.2\ 228 31.6 53 7.4} 4( 0.6} 111\ 15.4\ 126 17.5 4( 0.5} 233 27.4\ 263 30.9 58 6.8} 7( 0.8} 132f 15.5\ 153 18.0 SMOKE* No Yes, but not now Yes, still smoke 72 55.8 16 12.4 41 31.8 447 62.0 120 16.6 154 21.4 519 61.1 136 16.0 195 22.9 DRINK DNK No Yes 0 96f 74.4\ 33 25.6 5( 0.7} 544f 75.5\ 172 23.9 5( 0.6} 640f 75.3\ 205 24.1 CITY Shanghai Out of Shanghai 99f 76.7\ 30 23.3 586f 81.3\ 685f 80.6\ 135 18.7 165 19.4 AGE(yrs)* 1835455565- * P<0.05 24 18.6 25 19.4 32 24.8 28 21.7 20 15.5 131( 18.2} 88( 12.2} 173 24.0 121 16.8 208 28.8 155 18.2 113 13.3 205 24.1 149 17.5 228 26.8 Annexed table 2 Disease distribution of inpatients in 68 hospitals (2002, Shanghai) vs. CC/DP controls Disease Discharged inpatients in 68 hospitals in SHA Number pr(B%)on Controls in CCIDP Number I Proportion (%) Infectious disease and verminosis Cancer 12952 74460 1.90 10.92 b d0 51 16 SH ELL-MCCLU RG-052358 Endocrinopathy, nutrition related disease, metabolic disease, immune disease Blood and hematopoietic sy-stem related disease Mental disease Neurologic system disease and sensory organ related disease Cardiovascular disease Respiratory disease Digestive sy-stem disease Urinary- sy-stem disease Cyesis, baby delivery and puerperal complications Skin and hypodermis disease Musculoskeletal disease, connective tissue disease Inherent abnormality- Perinatal conditions Signs and symptoms with unknown causes Injuries and poisonings I Others Total * including outpatient controls 15859 4448 941 27642 100827 93602 105418 40459 82255 4542 16156 4472 3461 6040 54607 43742 681883 2.33 0.65 0.14 4.05 14.79 13.73 15.46 5.93 12.06 0.67 2.37 0.66 0.51 0.89 8.01 6.41 100.00 R40 0 00 36 8.2 92 20.9 61 13.8 49 11.1 27 6.1 00 2 0.45 ~29 0 W17 0 36 I 8.2 440 Part II: Molecular Epidemiolgy Study 1. Executive summary The molecular epidemiology (ME) study is designed as a two-phase study, phase I and phase II. The ME project is also a study with multi-collaborative institutions, that include 16 units of Shanghai Institute of Public Health and Supervision (IPHSs) at municipal and district levels, Shanghai Municipal Center for Diseases Control and Prevention (CDCP), and factories of relevance. 17 SH ELL-MCCLU RG-052359 The key point of phase I study is to access the historically medical records. Great efforts have been devoted to locating medical records from IPHS and CDCP data system. However, it was witnessed that medical records for phase I study were neither available from IPHS nor CDCP data system. A new Occupational Physician Committee (OPC), which owns a comprehensive network on occupational health surveillance, has taken responsibility for the abstraction of medical records from workers who have been working in factories with benzene exposure. This task has been progressing quite well and should be completed by the end of March 2005. US side experts were also asking for the medical record abstraction although it was not planned at the beginning stage of SHS. After failing to get medical records for benzene poisoning cases from CDCP and IPHS, OPC doctors abstracted the medical records of 195 historical cases of benzene poisoning in 2004. Factory selection is the first big challenge for the phase II study. According to US scientists' suggestions, priority was put on higher exposure factories. We had tried searching for the candidate factories along with four information sources, i.e., the IPHS database, historical benzene poisoning cases, CC/DP provided subjects and local IPHS industrial hygiene monitoring data. Unfortunately, above trials did not work well because of the changing status of exposure, IPHS' conflictive role and its improper management system. Some other information sources such as medical records at occupational health hospitals, and other channels of relevance have been showing an optimistic perspective. The second challenge is to get closer contact with both factory managers and workers by further explaining the real nature of the study because they have been so worried about the socioeconomic and health risks that might stem from their participation in the study. Laboratory analysis of benzene metabolites has been removed from Shanghai CDCP to Fudan since March 2004. With the establishment of the relevant SOPs, the sample analysis has been running quite well. To facilitate phase II study, new institutions, such as occupational health hospitals other than IPHS, should be closely consulted to promote factory selection. IPHS' adequate involvement in the retrospective exposure assessment should also be emphasized. We believe with these new initiatives, phase II is still a highly feasible study. 2. PROGRESS 2.1 The Phase I study The phase I study is based on an abstraction form with 2000 annual health-monitoring records (the relevant factories should have record keeping available for workers potentially exposed to benzene for at least five years). At the same time a retrospective exposure assessment should also be conducted. The study requires that the health monitoring data originate from populations with different gender, age and working 18 SH ELL-MCCLU RG-OS2360 duration, who had been exposed to benzene at a range of exposure levels. 2.1.1 Collaboration with Occupational Physician Committee (OPC) Based on the information obtained from a number of large scale hospitals associated with the Occupational Physician Committee (OPC) and hospitals responsible for periodical occupational health surveillance, we have successfully accessed medical surveillance data for 87 factories covering 5198 workers with potential benzene exposure, as well as detailed information regarding factories' title, location and periodical medical examination records. To get further information upon the exposure and effects data, FUSPH has kept in touch with IPHS towards accessing the exposure data for the 87 factories from the database. The historical worksite benzene monitoring data for 48 of the 87 factories (55.2 %) have been accessed from the IPHS database. However, it witnessed us again that Shanghai MunicipallPHS database does not own worksite monitoring data for all benzene related industries in Shanghai. Having had the successful trials via OPC-related paths, in consulting with Dr. Rob Schnatter, we recruited 30 factories with historical worksite monitoring data from a total number of 48 factories covering 1803 workers for further study. The selection criteria for the 30 factories include:1) number of workers receiving regular medical examination over 20 annually; and 2)having IH monitoring data available. 2.1 .2 Basic work procedures of factories selection The basic work procedures of factories selection are shown in a flow-chart below: 87 factories With 5198 Workers who had receiVed I'lledexal'll 48 factories With historical woJ"ksite rriol1ilori119 data atPHS database Fil1clllys~lE!cte(J 30 factories With 803 Workers Who.had received .med exam Dr. Ni is now coordinating historical abstraction from the 30 selected factories with 1803 workers who had received the periodical medical exam. This should be accomplished by the end of March 2005. Prior to the initiatives with Dr. Ni, we had also accessed the IPHS 19 SHELL-MCCLURG-052361 database in retrieving the worksite monitoring data for another 10 factories associated with benzene exposure, and had selected 4 of the 10 factories as candidates for phase I study, and have abstracted medical data at these sites as well. (Table 2-1). Table 2-1 Summary of the recent progress on phase I study Sources No. of selected No. workers factory related to received med exam benzene exposure Dr. Ni headed ope 30 1803 and relevant hospitals IPHS 4 484 Benzene exposure levels( ppm) 0'" 1367 0'" 263 2.2 The Phase? study 2.2.1 Tasks of Phase II study Phase II study is comprised of Phase lIa and Phase lib. Major tasks of phase lIa study are: 1) Recruit 1000 workers with benzene exposure at various levels 2) Administrator physical examinations, and blood lab analyses (clinical laboratory test + metabolites measurements) 3) Perform industrial hygiene monitoring (area and personal sampling) and chemical analysis Major tasks of phase lib study are: 1) Recruit 200 workers with benzene exposure at various levels (prioritized on workers with higher exposure) 2) Perform blood and bone marrow lab analysis (clinical, polymorphism and metabolism) 3) Perform urine and exhaled breath analysis 4) Perform industrial hygiene monitoring (area and personal sampling) and chemical analysis 2.2.2 Updating of Phase II study Up to December 2004, we have implemented the phase lIa study at 3 factories where we conducted workplace monitoring and workers' health surveillance that resulted in 905 personal samples by 3M badge sampler and 941 area samples by charcoal tubes; and the routine medical examination and blood sample taking of 231 workers. Besides, 3 workers received health surveillance for Phase lib studies (Table 2-2, 2-3). 20 SH ELL-MCCLU RG-OS2362 Table 2-2 Summary of Phase lIa study at three work facilities I Factory Air sampling No. sampled Bn levels (mg/m3) I SHA Rubbery I 3M badge I 273 0- 5217 I (HQ) I Charcoal tube I 379 0- 397 I Xinyuan I 3M badge I 312 0- 258 Rubbery Charcoal tube 177 0- 642 Golden Cock 3M badge 320 0- 1380 Cooperate Charcoal tube 385 0- 1046 56.29 22.64 34.46 75.5 147.08 93.50 Table 2-3 Summary of workers' health surveillance for Phase lIa & lib I Factory I No. Phase lIa I No. Phase lib I I SHA Rubbery (HQ) I 56 I 0 I Xinhui Rubbery 28 3 I Golden Cock Cooperate I 147 I 0 I The protocol for exhaled breath was set successfully and 3 workers have been tested. Meanwhile, the metabolites analysis protocol was developed by FUSPH and more than 150 blood samples were analyzed. The first draft on metabolite analysis is under preparation. 3. Challenges and our efforts 3.1 Phase I study 3.1.1 Medical data retrieving from Shanghai CDCP and IPHS data system Two types of historical data, industrial hygiene (IH) monitoring and annual physical examinations, are necessary for the phase I study. IH data were systematically kept by IPHS. No medical records were available for individual workers in the IPHS database. The key task of phase I is to find workers' medical records from annual physical examinations. Exposure history from medical records will be linked to IPHS IH monitoring data to further retrospectively assess exposed workers. According to the study design, the data should be provided by IPHS database, and CDCP and its selected industries, which was based on the assumption that the Medical Surveillance Center of Shanghai CDCP could have all the annual health monitoring data in Shanghai available. Unfortunately, it is not the case. In fact, the IPHS database owns merely the worksite monitoring rather than the physical examination data. IPHS may even have no information in respect with how many workers were working for the worksite and whether those who have been exposed to benzene were receiving annual physical examinations. On the contrary, the Medical Surveillance Center of Shanghai CDCP own the information regarding the number size of workers with potential exposure to benzene and their health monitoring results but without knowing anything about exposure levels. In that case, we decided to deal with the problems by using a "two-track policy", i.e., from Medical Surveillance Center of Shanghai CDCP to IPHS to locate the exposure worksite 21 SH ELL-MCCLU RG-OS2363 and its exposure level; and from IPHS database to Medical Surveillance Center of Shanghai CDCP to link the exposure with health effects. However, challenges were found from both methods: 3.1.1.1 From CDCP to IPHS There are 35 hospitals involved in the annual physical examinations for workers with benzene exposure. Of which, the Medical Surveillance Center of Shanghai CDCP is merely one of the 35 hospitals accounting for 2.9% of the total, and being responsible for providing the medical surveillance service for only 11 enterprises in total. The Medical Surveillance Center of Shanghai CDCP has kept the health monitoring data for the recent two years only and the previous historical medical data were retained at each of the 11 factories. Having had retrieved the work environment benzene data for the 11 factories in IPHS database, it was found that only one of the 11 factories with worksite benzene level over 40ppm (or 132 mg/m3) covering15 workers who had received medical examination with one year data being recorded. As for the remaining 10 factories, they were either "no benzene exposure data at all", or with "extremely low exposure", that showed limited value for a follow-up study. 3.1.1.2 From IPHS to CDCP We identified factories with benzene exposure from IPHS database. However, except for IH monitoring data, neither data for medical examination, nor the number of workers receiving examinations were available. It was also difficult to get access via IPHS to each hospital where medical records were stored. Similarly, it was difficult to access hospital data progressing from IPHS through CDCP to the hospital. Therefore, it was witnessed that neither IPHS nor the Medical Surveillance Center of Shanghai CDCP database would have provided any supportive evidence to initiate a phase 1 study. 3.1.2 Striving towards success After being defected by the above trials, FUSPH devoted further efforts towards its success. To prompt the pace of the study, we recommended taking an active action to get a closer cooperation with Dr. Ni Weiming and the organization that he heads, the Shanghai Occupational Physician Committee (OPC). The reasons for the action are: 1) Dr Ni is the Chairman of Shanghai Society of Environmental and Occupational Medicine and the Director of the Department of Occupational Disease of Yangpu District Central Hospital. His headed society has built an extensive network of occupational physicians all around China. The Society has been very influential in the field of occupational medicine and has been keeping close contact and academic exchange with most hospitals responsible for periodical occupational health surveillance. The closer cooperation has profoundly accelerated the process of acquiring all necessary information and/or data from the hospitals which is the key to accessing the matched-exposure information from IPHS database. In addition to the phase I study, the newly activated alliance with Dr. Ni has been 22 SH ELL-MCCLU RG-OS2364 expanded to some other aspects beneficial to the SHS project: 1) Identify newly diagnosed benzene poisoning cases from Shanghai and other areas and recruit these subjects to the DP study; 2) Affirm historical benzene poisoning cases from medical records and abstract these records; and 3) Assist FUSPH to search for appropriate candidate factories for phase II study and coordinate studies in these factories. 3.1.3 Next step to phase I study and the expected difficulties 3.1.3.1 Major tasks Completing the abstraction of historical medical records of workers exposed to benzene; Accomplishing data entry for Phase I study Accomplishing the retrospective exposure assessment for phase I workers and the historical cases of benzene poisoning 3.1.3. 2 Further challenges Little value of the historical data because of the default database in terms of limited sample size, incomplete information regarding when, where, and how the data were collected; Worksite monitoring at the 30 factories selected which requires a promising joint-efforts from IPHS that are crucial to the progressive success of Phase I study; More information of the "similar factories/worksites" are urgently needed for the "surrogate exposure assessment" as a great number of factories were shut down, transformed or re-combined during the transitional period of economy reform; Further literature information searching and reviewing are also essential to the retrospective assessment which demands greater intellectual efforts; 3..2 Phase II study 3..2.1 Factory selection Four information sources were identified for selecting candidate factories at early stages of the ME study. These sources are: Shanghai municipal IPHS database; Industrial hygiene(IH) monitoring; Factories with historical benzene poisoning cases Factories with CC/DP subjects exposed to benzene FUSPH has taken measures to keep pace with the development of the study although the team has experienced many hardships. The following text will describe challenges facing in dealing with each source. 23 SH ELL-MCCLU RG-052365 1) IPHS database Based on our own experience integrated with the literature information generate a list of industries or occupations associated with benzene exposure. According to the listed industries/occupations, we worked with IPHS searching for proper candidate factories with relatively higher benzene exposure from the database. So far, we have visited 31 factories retrieved from the database, and found that the majority of the factories were regarded as "low exposure", and most of them no longer existed, either "shutting down" or being merged into some other big industries. There were no systematically historical monitoring data available; and the oral recall information from senior workers or mangers was usually dramatically different from what the database told. The IPHS staff members were not always kept informed on the changing even relayed that they have "no idea at all" about what had happened there. 2) LocallH monitoring To check whether announcing a visit was influencing the exposures, 7 factories with average exposure levels ranging from 0- 263PPM were visited without announcing. 2 factories for Phase lIa study in Shanghai were selected predominantly based on IPHS staff's personal information proceeded by "dropping in". The success of that kind of trial was very much depending on how informative and how qualified the IPHS staff was. It is unlikely to be worked as a routine scheme, unless IPHS can make these qualified staff routinely available to us. 3) Historical benzene poisoning cases In general, we have to have the list of benzene poisoning cases that were reported in the previous year from Shanghai Municipal CDCP; and then forwarded the list to IPHS asking for help in contact with the factories of relevance and making arrangement for the visit and monitoring. We have worked-out for 8 factories and visited there. It was found that 7 of the 8 factories were ranked as "low benzene exposure" ranged from 0- 16.6PPM, which might be due to the data provided have lagged far behind the occurrence of benzene poisoning. The preventive measures had already been taken during the past years; and the work environment had been much improved while we were visiting there. One of the 8 factories, a private-owned plastic manufacturing with 4 employees, remained a relatively "higher exposure level" with an average of 273ppm, but showed no willingness to cooperate with us. Therefore, the path "from case to factory" has not been so successful. 4) CC/DP subjects with benzene exposure Factory names subjected to potential benzene exposure were forwarded to IPHS for data searching. The ME team visited subjects' factories with potential high exposure to get present exposure level. It was found that in most of the factories, the exposure levels were basically low. For example, among the total number of 22 factories we visited, exposures were of "low exposure" ranging from 2-5 ppm. Fewer factories at higher exposure were 24 SH ELL-MCCLU RG-OS2366 found but they no longer existed. In summary, the inefficient factory selection could be attributed to: 1) Exposure levels were changed in the factories with poisoning cases According to the Occupational Diseases Prevention and Control Act, P.R. China (ODPCAct), which has been put into effect since 2002, the working condition causing poisoning, should be effectively controlled by taking immediate preventive measures without any delay. These factories were no longer considered appropriate for the ME (high exposure) study.. 2) The conflicting priorities of IPHS IPHS is a governmental agency; its primary task is enforcement rather than research. IPHS staff members couldn't contribute too much of their working hour to keep pace with the project developing. 3) Management system of IPHS (a) Municipal IPHS works depending on district IPHSs to a certain extent. Generally speaking, municipal IPHS doesn't work with factories directly. To arrange communication with factories and visit there, municipal IPHS has to ask for help from district IPHSs to get things done because municipal IPHS is primarily responsible for professional guidance and technical support to district IPHSs rather than administrative and inspective authorities. The municipal IPHS has no direct authority to control work performance of its affiliated units. In addition, there is not a good incentive for district IPHSs to participate in the study, nor the QA/QC approach is facilitated in inspecting and promoting work efficiency and quality done by IPHS staff members, which resulted in their work status in an inactive and conflicting manner. (b) Embarrassing situation of district IPHSs For ME study, district IPHSs have often been put under an embarrassing situation. On the one hand, we need district IPHSs help in finding out candidate factories with higher exposure; on the other hand, we should sympathize with their embarrassing situation of protecting themselves from being blamed due to "retaining higher exposure" workplace without immediate action taken. Furthermore, IPHSs do not always want to submit all the true data to the municipal IPHS. For example, the higher exposures were mainly reported a few years ago, and the current data usually appear to be much lower than that of previous ones. 3) Communication between FUSPH and district IPHSs Municipal IPHS is unhappy with FUSPH getting contact with district IPHSs directly. 25 SH ELL-MCCLU RG-OS2367 Sometimes, IPHS staff members own more true data than their municipal counterparts, however, the data can only be delivered through the municipal IPHS because the SHS established contract with the municipal institution instead of the district ones. 3..2.2 Cooperation from factories The reasons for those candidate factories who were not always willing to participate in the ME lIa, particularly lib studies, are thought to be attributed to: 1) Economic burden The occurrence of benzene poisoning due to higher benzene exposure at work will be considered a violation of ODPC Act, although the municipallPHS has promised not to put it on that case if the institute are involved in the SHS. Factories with higher benzene exposure will be treated as the workplaces with higher risk, thus leading to more stringent inspection control. As a result, the factories claimed to put much more capital investment to control the exposure levels that would have brought about an unaffordable economic burden. 2) Political concerns At the meantime, workers would have appealed for more monetary compensation and other benefits if they knew the abnormal blood counts and other ill health effects from their health surveillance. All these economically and politically negative impacts on the industries would have created a critical situation of them being faced with a "bank crush" of their enterprises. In addition, workers cannot properly understand how health risks will be found from the bone marrow punching; nobody is really willing to donate his/her bone marrow at any event. Despite all these concerns have been clearly declared in the Informed Consent workers will undoubtedly blame employers based on their cultural and political experience, if anything chaotic did happen. Therefore, the majority of the authorities of the candidate factories are unlikely to be willing to involve in such a conflict and dangerous "game" which might lead to an unaffordable responsibility for the unexpected outcomes. It usually cannot be helped if the authorities of the factories do not want to involve in the business even for the municipal IPHS per se. In conclusion, there is by no means of sorting it out without the cooperation from both employer and employee sides. 3.2.3 FUSPH's efforts on prompting Phase II study 3. 2.3.1 Benzene and metabolites analysis Shanghai CDCP had an agreement to be responsible for benzene and metabolites analysis; and has been trying to set analysis protocols successfully. However, CDCP failed 26 SH ELL-MCCLU RG-OS2368 to set protocol till early 2004. Then, these tasks have been shifted to FUSPH since March 2004. 3. 2.3.2 Successful exploration in collaboration with ope The unsatisfied progress of the Phase II study have prompted FUSPH to consider a more comprehensive and effective way to identifying proper candidate factories for the phase II study. One of the most successful channels is working in collaboration with Dr. Ni, and searching for the proper factories for the phase II study through OPC and its affiliated hospitals, since they are authorized for the diagnosis of occupational disease. We found that this approach has been very beneficial so far. Cases of benzene poisoning being diagnosed by above-mentioned hospitals are most likely to be newly occurred ones so that we could track back to the existing factories where they have recently worked to catch recent information and make workplace monitoring at relatively higher exposure worksites. Hospitals responsible for diagnosis and treatment of occupational disease are usually most reputable and influential to both employers and employees, which creates a better chance of receiving the cooperation without worrying about punishment from IPHS. Accordingly, we have so far successfully accessed a total of 9 factories and one of them has been selected as the field site for the Phase II study. Breaking the impasse, we moved straightly forward toward achieving the goals of the Phase II study. In addition to the 9 factories that were found via the occupational disease hospitals, some 69 factories have also been visited. During the period between March 2003 and December 2004, we found a total number of 78 factories that could be considered as candidates for phase II study, with an average "accessing rate" of 3.5 factories per day. 3 factories have agreed to participate in the phase II study, the success rate was much higher than that of doing it based on the theoretical design only (11.1 % vs. 2.9%) (Table 2-4) Table 2-4 Summary of the worksites seeking Information Types of No. Bn level Selected Success sources information factory (PPM) field for rate (%) Phase II study Theoretical SHACDCP 8 0-273 0 2.9 design BP information approaches CC&DP study 23 0-23.2 0 scheme SHA IPHS 31 0-30.3 0 database District IPHSs 7 0-263 2 Occupational diseases 9 10-432.3 1 11.1 hospitals I Total 178 27 SH ELL-MCCLU RG-OS2369 3..2.4 Work plan for Phase II study and expected challenges 3..2.4.1 Work plan Continuing factory selection primarily via hospitals responsible for occupational disease diagnosis Accomplishing data entry of health surveillance for Phase lIa study Accomplishing data sorting and entering of industrial hygiene monitoring for Phase II study Accomplishing benzene exposure assessment for 231 workers subjected to Phase II study Continuing work on regular QA/QC check for Phase lIa study 3..2.4.2 Expected challenges For the smaller sample size of 231 cases accomplished there existed a big lag behind the targeted goal of 1000 samples; As being stated, there are so many factors negatively affecting the progress of factory selection for Phase II study so that newly developed searching channels and the updated information are really looking forward; As more updated biochemical and molecular markers were introduced into Phase II study, US side put prioritized factories and workers with higher benzene exposures over 40ppm that would have still been a big challenge at the foreseeing future; A huge amount of industrial hygiene data obtained are waiting for a timely sorting out, analyzing and evaluating, " the longer delayed the more trouble and bias created"; There has been almost no way of motivating factories to participate in the Phase lib study, thus frustrating further development of this part study. There is a heavy task of industrial hygiene monitoring prior to the bone marrow punching that would undoubtedly exaggerate the difficulty of Phase lib study. Therefore, an easier and more practical way of selecting the Phase lib subjects is supposed to be based on the results of hematological analysis and mediated by the clinical physicians of relevance. 4. Suggested solutions 4.1 Factory selection method: Efforts should be focused on additional channels of candidate factories searching, such as CDCP, occupational health hospitals, and other paths out of Shanghai 4.2 Adequate involvement of district IPHSs: This has become crucially important to the success of factory selection and exposure assessment: However, FUSPH is unlike to get access its oversights into IPHS because of no direct contract between SHS and district IPHSs. Closer communication and tighter link between JCML and municipallPHS/district 28 SH ELL-MCCLU RG-052370 IPHSs are urgently needed. 1. The value of the study CONCLUSIONS The Shanghai Health Study (SHS) project and its Joint Clinical and Molecular Laboratory (JCML) have provided a good opportunity and proper collaborative path facilitating for conducting a comprehensive epidemiological and clinical investigation with respect to occupational exposure to benzene and its pathogenic effects on lymph and hematopoietic diseases. With the development of the SHS, researchers from both China and US sides have experienced benefits from the collaboration in quit a few fields, for examples: Human capacity building: The extensive training activities shared researchers with the state of arts in industrial hygiene, occupational epidemiology and the pathogenesis of lymph and hematopoietic diseases, which have become the sound scientific foundation of capacity building for the work team. Teamwork spirit developing: A well-designed work procedure, properly trained investigators, adequately supported human and material resources and well-coordinated collaboration among institutions are essential to SHS success. To achieve the goals of such an influential and complicated research project, the Fudan University School of Public Health has been working closely with US experts and researchers of other institutions involved towards creating a harmonized humanistic atmosphere and promising work environment, which is expected to support the projects successful. A far reaching influence for the future scientific research, training and clinical services: The experience that SHS had already gained and would have much more enhanced, as well as the research methodologies, laboratory facilities, and work models that have developed will become a real potential for the scientific research, training and clinical services in China side of relevance in the foreseeing future. It is expected that the Fudan University School of Public Health would have shared more benefits and experiences from the JCML, 2. Significance of collaborating SHS is characterized by its research difficulty and organizational complexity. For example, it is academically integrating a comprehensive epidemiological investigation with experimental and clinical studies; it is administratively involving governmental health agencies, industries, universities, hospitals and research institutes; and it is internationally, co-sponsoring by two countries, the USA and P.R. China. Therefore, "collaborating" has become the key word of the study and the status of the collaborating become crucially important to the success of the study. We have been benefited from the closer collaboration with all the institutions involved and enjoyed working with all the participants even if we had also verified some of the "difficulties" derived from inadequate collaboration that occurred with the development of the study, that were considered as a kinds of 29 SHELL-MCCLURG-052371 "growing pains". People would not be growing up without "pains". It implies that we have to continuously put procedures to rectify the problems; and more challenges are expected to come with the project moving forward. We hope a closer and more promising collaboration would made most of the challenges and difficulties not insurmountable. Respecting each other between China side and US side, effective communication and more transparency in decision making procedure of scientific design, management and co-sharing results, would be key factors benefiting the future collaboration. 3. Team work As mentioned previously, the unique team work spirit of JCML has become a common heritage of the collaboration. Most of the researchers do enjoy working together towards a greater success of the SHS project even if some challenges were often facing. Work coordinating and the QA/QC approach following a virtuous circle of "identification of problems-remediation-re-evaluation-improvement" by the work team's joint-efforts has been developed individually specified to different component of the study project, which will not only ensure the work quality but grant team workers more confidence by maximizing the probability of success for these three important studies. 30 SH ELL-MCCLU RG-OS2372