Document rj9peDEop05bYXKrNQ9VL7M7

FILE NAME Firestone FIRE DATE 1978 Sept DOC FIRE038 DOCUMENT DESCRIPTION UNC Mortality study of workers in 42 plants finds SMR for all cancer in retirees Lung Cancer SMR Ex 939 UNC - 1977 Mortality IC4 September 1978 Study Pltf Exh 939 JACKSON 025904 Occupational Health Studies Group SJ THE UNIVERSITY OF NORTH CAROLINA AT CHAPEL HILL The University of North Carolina at Chapel Hill NCNB Plaza Suite 32 322A Chapel Hill N.C. 27514 September 27 1978 Joint Firestone Occupational Health Committee Addressees Dr. Dr. Mr. Mr. Mr. Mr. L. H. Ballou Theodore Georgiadis R.V. Hassmiller Peter Bommarito Louis Beliczky Jake Miller Gentlemen RECEIVED 1978 ner 1978 INDUSTRIAL HYGIENE Enclosed is a copy of the report of the 1977 Mortality Study I.C. This study was begun by the Occupational Health Studies Group in January 1978. This is the first application of surveillance techniques using mortality registry and population enumeration procedures Although it was our stated intent to report in July I feel that this report still demonstrates the timeliness with which surveillance activities can yield health information We were able to get data in time for this report from 42 of the 48 plants and local unions with which we work in the Joint Occupational Health Program We expect to report further for all 48 plants later in the year If you need additional copies of this report we will be pleased to provide them Sincerely yours dcm Robert L. Harris Jr. Ph.D. Director Enclosure Identical letter and enclosure sent to all Joint Committees JACKSON 025905 1977 MORTALITY STUDY RECEIVED OCT 1978 INDUSTRIAL HYGIENA INDUSTRIAL DEPARTUERT Occupational Health Studies Group School of Public Health University of North Carolina Chapel Hill North Carolina IC SEPTEMBER 1978 JACKSON 025906 1977 MORTALITY STUDY by Dale R. Herman M.S.P.H. Dragana Andjelkovich M.D. _ Robert DesJardins Ph.D. 4 Dianne Byrne M.S.P.H. Eileen Cella M.S. 7 Occupational Health Studies Group School of Public Health University of North Carolina . _ Chapel Hill North Carolina Submitted To > The Firestone General Goodyear and Uniroyal Joint Occupational Health Committees SEPTEMBER 1978 JACKSON 025907 ACKNOWLEDGEMENTS We acknowledge with appreciation the professional advice of Dr. Robert Harris Dr. Michael Symons and Dr. Herman A. Tyroler the statistical assistance of Ms. Claire Bourguet and Mr. Donald Doerfler the programming expertise of Mr. Leland Jones and Mr. Chester Hoover and the invaluable assistance in manuscript preparation and corres- pondence Ms. Julie Banhart and Ms. Jackie Brown JACKSON 025908 TABLE OF CONTENTS wane LIST OF TABLES- eeee nese eee nee anne wenn i LIST OF APPENDICES-- ii EXECUTIVE SUMMARY- fanaa SUMMARY- seen - iii .I INTRODUCTION AND STUDY DESIGN-- nent 4] II DATA COLLECTION PROCEDURES- eawennecnne 6 A. The Creation of the Workforce Enumeration File --- 6 Selection of the Study Population-- a B. Creating the Death Certificate File- -13 C. The Creation of the 1977 Mortality Analysis File- File- ------- 18 D. Processing Comparison Population Data- 20 III ANALYSIS AND RESULTSee eee eee eee were 21 IV SUMMARY AND CONCLUSION- _ 30 REFERENCES .32 TABL eee eeE e n S e - e 34 JACKSON 025909 LIST OF TABLES Table 1 Table 2 , Deaths Distribution of Population and a by Age and Sex wae e eset ncceee 34 Observed and Expected Deaths and SMR's for Major Disease Categories among Males for All Plants Combined- saa Doran anon naa= nana anna 35 Table 3 Observed and Expected Deaths and SMR's for Neoplasms among Males for All Plants Combined-- - 36 Table 4 : | Table 5 Table 6 Observed and Expected Deaths and SMR's for Circulatory Disease among Males for All . Plants Combined-- Combined-- nner nnnn nese ene 37 Observed and Expected Deaths and SMR's for Major Disease Categories among Females for All Plants Combined- Combined- ------ - . Observed and Expected Deaths and SMR's for Neoplasms among Females for All Plants Combined 38 39 Table 7 Table 8 . Observed and Expected Deaths and SMR's for ' Circulatory Disease among Females for All ; Plants Combined-- Combined-- errr ennn--- a 40 - SMR's by Plant for Males for Selected Major Disease Categorieas nne- nen- e anne nen . 41 Table 9 Table 10 SMR's by Plant for Males for Selected Causes . of onan Death on nanan wennen ns nanan -~---- Major SMR's SMR's by Plant for Females for Selected ' Disease Categories- Categories- - 1.42 43 Table 11. Table 12 ' Comparison of Mortality Rates Between Rubber| Industry All Plants Combined U.S. Population for All Causes of Death among Males- 44 Comparison of Mortality Rates Between Rubber- Industry All Plants Combined and U.S. Circulatory among Population for All Deaths Males- more seennen- meee eter tn enn r enn ne- mann: 45 Table 13 Comparison of Mortality Rates Between Rubber | Industry All Plants Combined and U.S. Population for All Neoplasms among Males- 46 JACKSON 025910 LIST OF APPENDICES Appendix I. Appendix II List of Individual Plants nnn 47 Employees the Diagram : Types of Employees -Included in Study Population- Pieietatatatatatatatateteteetatate _. 48 ene Appendix III . Data Analysis Program-awen------- neeneesnnee . 49 } - ii JACKSON 025911 Executive Summary Mortality studies have great value in assessing the health experience | Occupational years of occupational groups past Over the several the | Health Studies Group has completed a number of studies of mortality experience among populations of workers from individual plants and local . unions These studies have dealt with deaths which occurred in past years Most have used the 10 year interval 1964-1973 The Occupational Health Studies Group as part of its obligation under the Company Joint Occupational Health Prograhams also been developing a health surveillance system A major part of such a system is surveilolf acunrrcenet mortality experience The study described in this report is the first application of surveillance techniques to current mortality experience The study covers 1977 the most current . year of mortality experience and it includes all of the plants and | local unions inthe Joint Occupational Health Program _ Oneobjectivoef the study was timely reporting Our stated intention | was to report in the first six months of . 1978. Although this report has been delayed beyond that time a number of tasks which have been done' will be useful in in the future subsequent reportansd will | | permit more rapid reporting | The University of North Carolina works with 48 in plants the Joint> Occupational Health Program Data relative to this study for six of these plants were not available during the first half of this year As a consequence this report is limited to analysis of data from 42 plants Further analyses are underway and 48 plants will be included in subsequent report iii JACKSON 025912 A total of 1,277 deaths are known to have occurred in 1977 among the 79,500 active and retired rubber workers from the 42 plants included in this report The relatively few deaths among women 149 deaths do not permit extensive analysis The overall mortality experience for the year among female workers was not markedly different from that of the general population More detailed analysis among male workers was possible The Standardized Mortality Ratio SMR which compares the mortality experience of the study population to that of men of the same : ages in the general population is used in reporting results An SMRof 100 means that the study population and the general population had the same mortality experience A higher value greater than 100 means that ; more deaths occurred among the study population than were expected based general population experience and an SMR lower than 100 means fewer than expected numbers of deaths occurred among the rubber worker group ~ The SMR for all causes of death in 1977 among male rubber workers in the 15-64 year age range was 73 that for retired workers in the 64- _ 84 year age range was 106. The SMR of 73 for the active age range reflects the healthly worker effect which has been identified in our studies and by others with actively employed workers The SMR of 73 for deaths from all causes in the active age range makes SMR values greater - than than 100 for any specific causes of death even more noteworthy | | would be the case with a higher causes SMR The SMR for all forms of cancer among retired rubber workers in the study was 119. This is a statistically significant excess and consistent with findings of several other epidemiologic studies done for past years SMR's for a number of site specific cancers were elevated for iv JACKSON 025913 | this group Particularly noteworthy is the statistically significant excess of lung cancers SMR = 145 this is in contrast to the results of our earlier mortality studies on several individual plant populations This finding merits further serious investigation An excesosf cancers of the lymphataincd hematopoietic systems SMR = 135 among active workers was also found this excess although real was not statistically significant the 95 percent level of confidence A similar excess SMR = 130 was found for leukemia among retired workers These latter two findings are consistent with results of past studies Overall and cause specific mortality experience as was expected varied among the individual plants Some statistically significant excesses were merit further observed for attention some causeosf death in some | plants These | The importance of annual mortality surveys increases as times goes by and as results for successive years become available Such surveys provide data for comparisons and for detectioonf trends and changeisn the health of working populations A major purpose of studies such as this is to identify plant population and disease types which need more detailed investigatiwiotnh the ultimate goal being to recommend actions which remove hazards to the health of workers BATO JACKSON 025914 I. INTRODUCTION AND STUDY DESIGN . The 1977 Mortality Study is an analysiosf the deaths among approximately aproximately 94,000 active and retired hourly workers who were members of the United Rubber Workers Union and who during the year 1977 were forty employed at or retired from one of the plants which are | affiliated with the occupational health research program at the Uni- | | | versity of North Carolina Past cohort studies conducted by the Occupational Health Studies .- Group OHSG at the University of North Carolina have concentrated on single plants and have been retrospective in nature reconstructing a workforce at some time in the past and then following that group of people for long periods to determine their mortality experience The- cost of such studies and the time involved prevented us from conducting- similar studies for all of the plants in the industry Yet observations made at one plant raised questions about other similar plants It was - clearly desirablteo be able to study all of the plants in the industry Other studies done in the past have examined data from a large number of plants but have done so using a proportional analysis which again raised questions makinigt obvious that a more definitive study was necessary Past studies have gone back in time and although they could give much information about the experience of workers some years back we felt that it was important to conduct a study based on the most current data - available which would be more pertinent to workers today - The OHSG has been collecting death certificates on an ongoing basis for most of the plants in this study since 1972. We have not previously | enumerated the entire population of workers for the four companies aan JACKSON 025915 Based on our experience it was obvious that until a weestablished continuous surveillance system which included enumerations of both the entire workforce and all of the deaths an industry mortality study ~ would be difficult time consuming and expensive With the establish- ment of an efficient means of enumerating the population as part of the surveillance system this study has become possible | 7 The initial plan for this study was presented at the Company University scientific meetings at the OHSG in December 1977. With the cooperation of representatives present at the meet froi m tn he g Unios n . and from each of the four companies the study was begun in January - | | | 1978 The purposes of this study are 1 To test and documetnhte feasibility of conducting a cross- - company mortality study on an ongoing basis as part of an | | overall surveillance system in the industry 2 To explore differences morttha e l moi rtat liy ty experience of the rubber industranyd that of a comparison population 3 To identify differences in mortality experience between plants and processes 4 To compare the mortality experience in 1977 with that reported earlier by the OHSG This study is unique our knowledge this is the first time that the annual mortality experiencoef a working population has been studied in a cohort fashion We see as an important contribution to the ~ health of workers in the rubber industry to be able to condausc tut dy ~ of this nature and report on it within six to nine months of the end of a year The value of this study will increase as the study is repeated JACKSON 025916 each year permitting comparisontso be made between years and making it possible to observe trends in mortality in the rubber industry wide The study demonstrates Pay annual mortality study that it is within six feasible to prepare an industry- to nine months after the end of the year provided that most of the information is which required for the study be made available in computer readable form from corporate | headquarters rather than from the individual plants Although few companies maintain files that are specifically designed for occupational health research almost all of the information which is required for a mortality analysis can be obtained from files which are maintained for operational purposes such as the administration of benefit programs and personnel management With the cooperation of the participating com- panies this information can be obt qua itei rapin dly e andd can be processed very easily thus making it economically feasible to prepare an annual mortality study for the active + Since the major objective and retired employees of the 1977 Mortality Study was to develop the data collection procedures and the data analysis programs which are required to enumerate a complete study population and to investigate the causes of death among the deceased members of that population this report 1 summarizes the procedures which were used to collect data for the 1977 Mortality Study 2 discusses some of the difficulties which~ | were encountered in implementing these procedures 3 describes the data analysis programs which were developed and 4 presents some pre- ._ liminary results from an analysis of the data for 79,496 employees who worked at 42 plants second report which will be completed by the -. end of the year all 48 plants will contain more 7 a extensive analyses of . : the data from . ee | JACKSON 025917 A mortality analysis is the first step in a comprehensive program of health surveillance Morbidity information environmental exposure | data and personal health information are also required to monitor the health conditions in an industrial setting The 1977 Mortality Study is the beginning of a comprehensive health surveillance system which is being developed by the OHSG for the workers in the rubber products industry Other OHSG surveillance research projects are presently : developing procedures for the collection and processing of several types : of related information These studarie e als so concerned with the development of procedures which will protect the privacy of the individuals who are included in the study population As the results of these studies become available they will be incorporatinetdo future mortality studies so that the scope of the analyses can be expanded to provide a more accurate assessment of health conditions in the rubber products industry 7 The 1977 Mortality Study is a cohort study of mortality in the rubber industry A cohort study is a study of a group of people the cohort or study population who all have some characteriisn tcoimcmosn This study population is followed for a period of time to observe the occurrence of disease among its members The experience of the studystudy population is compared with the experience of some other population often the total U.S. population to detect what health hazards the . study population may be exposed to For the present study the common characteristics are membershiipn the United Rubber Workers Union URW and employment in one of the plants participating in the Joint Occupational Health Program in the four companies Firestone General Goodyear and Uniroyal in the rubber JACKSON 025918 industry The group people defined by these characteristics can actually be separated into two distinct cohorts The first cohort consists of all URW members who were active employees at one of the 48 plants listed in the appendix at any time during the year 1977. The second cohort consists of all URW members who were retired from one of plants these 48 prior to Januar1y 1977 but who were living on that | date Although the discussion which follows on data collection and preparation deals with these as a single population the analyses were done separately and the results section of this report treats them as ae separate groups The data necessary for this study were a complete enumeration of the study population including age sex racae nd employment status on each person in the study date of death and cause of death for each - deceased person in the study the number of people in the comparison population in this case the U.S. population together with the age sex and race distribution of this population and the number of deaths by cause of death for each age sex and race group in the comparison ; population The information on the study population came from records from the four companies and from death certificates obtained from the companies their insurance carriers and from state health departments The number of people in the comparison population came from a publica- tion of the Bureau of the Census 1 and the information on deaths.in deaths.in the U.S. came from data generated by the National Center for Health : Statistics and provided to us on computer tapes 2 The most recent information available on deaths within the U.S. U.S. was for the year1975.so 1975.so the data for that year were used for comparison- JACKSON 025919 II DATA COLLECTION PROCEDURES The 1977 Mortality Study required the creation of two major da,,>a files the Workforce Enumeration File and the Death Certificate File These files were created separately and are maintained separately to provide a cross checokn the completaned n ace cursacsy of the informa- - tion which was collected for the study Section A below describes the 7 procedures which were usetdo create the Workforce Enumeration File and . Section B the describes procedures which were used to collect informa- tion for the Death Certificate File After both files been.com- been.com- pleted the information from the Death Certificate File was merged with the information on the Workforce Enumeration File to creat thee 1977 - Mortality Study File which was used to prepare the mortality analyses The procesosf merging the two files is a critical step in the study procedure because any unexplained deaths which are identified at that point must be explained before the mortality analysis can be undertaken The procedures which were used for merging the two working files to create 1977 Mortality Study File are described in Section C. Sec- tion D describes the processing of data for the comparison population . A. The Creation of the Workforce Enumeration File The first step in the creation of the workforce enumeration file active was the collection of information concerning all and retired | employees from each of the participating companies We requested this information from corporate headquarters rather than from the local- plants for the following reasons JACKSON 025920 1 If the data we receive for each company were processed at one central location the methods used to collect code and store the information would be more likely to be consistent 2. Computer readable data which is much easier for us to work with are more likely to be available from corporate head- quarters | 3 The data for all the plants for each company would be sent to us in common format with the same information being avail- able 4 We would work with a smaller number of computing centers and all the data for one company would arrive at OHSG at the same - time and could be copied and editeidn one batch . The strategoyf obtaining the required information from a single source for each company worked quite well. We were able to obtain a single file of retired employees from each of the four participating companies three of these files were sent to us in a computer readable form We were also able to obtain most of the information concerning - active employees from a single source for each of three companies a few | plant files were required to supplement the data received from one . | company but the fourth company could only provide us with files from each of the participating plants All information on active employees was received ina computer readable form except for the information one plant which was provided on computer prepared listing - Selection of the Study Population In January 1978a letter was sent each company requesting the following information for all hourly employees who worked at one of the BATO JACKSON 025921 participating plants during 1977 and retired employees who had previously worked at one of the participating plants and were living as of January 1 1977 Identifying Information Name .Social Security Number and Company Identification Number Demographic Information Date of birth Sex Race Employment Information Planatt which the employee is currently working if active or the last plant at which the employee worked if retired - Date of initial employment and - Union Membership Current employment status and the the continuous service date SS date this status became effective If avail theafb olll owie ng information was also requested Department in which the employee currently works Current job assignment code or job title Date and type of retirement if retired Date of death if deceased ; The responses to this request varied from company to company generally included information on all hourly employees The first step in the processing of this data involved the selection of those employees who were to be included in the study population Three basic constrainwterse critical to the formulation of the criteria / which were used to select the study population 1 Members of the study population had to be members of the URW Union and had to work one of the 48 plants which | | | included in the OHSG contract . 2 A person included in the study population only if the person was entitled to life insurance benefits so wecould be . JACKSON 025922 fairly sure that we would learn of each death through the company insurance office if the person died during 1977 3 Data for all categories of people included had to be consistently recorded in the data files from each of the four ; companies which were included in the study . The first constraint is probably not too critical to the resultosf this study but is does prohibit the inclusion of salaried employees , such as supervisors who frequently work for many years as hourly em- . ployees and who still have environmental exposures which are similar to | | those of the hourly employees The second constraint is more important because it eliminates all employees who left the companies prior to 1977 who did not receive a termination allowance or a pension This constraint not critical to the 1977 Mortality Study because we made no efforts to obtain information on former employees as would be done in retrospective study However this constraiwniltl become more important for the 1978 Mortality Study because we must decide if we will keep short term employien ethse workforce enumeration file and attempt to obtain death certificatferosm | sources other than the company insurance offices are presently at plant analyzing the data which was collected one pilot surveillance . | a part of the experimental surveillance system to guidues in the development of appropriate selection rules for the 1978 Mortality Study The third constraint had an important effect on the 1977 Mortality Study Since some employees who received termination awards or deferred pensions do not qualify for life insurance and since we were not able to obtain information concerning these people from some of the companies we decided to eliminate all of the employees in this group from the BATO JACKSON 025923 10 study population even though we received a large number of death certifi- cates for these employees We are investigating alternative procedures | which might be used so this group of employees can be included in future | | | mortality studies | Using these constraints the following rules were developed to oa select employees for inclusion in the study population These illustrated in Diagram 1 in the appendix | 1 The employee had to be an hourly worker - who was a member of the URW Union and who worked at one of the 48 plants which were included in the study- 2 The employee had to work at least one day during 1977 if she had not retired before Januar1y 1977 and was not on | leave for the entire year 3 The employee had to have received a pension during 1977. if she had retired before January , 1977 4 Employees who were on leave during 1977 were includeidn the - population provided that they had life insurance benefits so i that the company would have received a death certificate if they had died during 1977 It should be noted that the use of these rules eliminated the | | following following employees from the study population | . 1. employees who were on Jan priou r toa Janr uary y 1 1977 and did not return to work before December 31 1977. Because of the Union members' rights there are probably very few if any employees in this category who will return to work at a later date If such employees do exist they will be picked- - up in future mortality studies. JACKSON 025924 11 2 All employees who went on educational leave or military leave priotro January , 1977 and did not return to work before December 31 1977. There is a very small number of employees in this group These employees will be picked in subse- quent mortality studies if and when they return to work .3 All employees who terminated with a lump sum payment award or | who left with a deferred vested pension prior to January 1 1977. Employees in this category who retired during 1977 were : included in the 1977 Mortality Study and will be included in all subsequent studies We hope to be able to add all of the employees in this group retroactively to the workforce enumeration file for future mortality studies 4. All salaried employees unless they transferred from hourly to salaried status during 1977. | We will not be able to include these employees in future studies unless arrange to collect the death certificates for all salaried employees and obtain agreement to extend the study to include selected employees who are not currently members of the URW Union Since the data for retirees and active employees were received from - different sources for each company , we initially created two separate files for each company file for active employees and one file for retirees These files were edited to identify missing data - duplicate ~ records and data discrepancies - Whenever data for an employee were missing or apparent discrepancies were discovered record was checked and corrected if necessary For example employees employees who records which indicated that they began work before they were 14 years old oo sometimes before they were born or were still actively employed BATO JACKSON 025925 12 although they were more than 65 years old were identified and the accuracy of the recorded information was checked Since the corporate headquarters do not have the original data files for the hourly em- ployees error lists had to be prepared for each plant These lists were sent to the personnel offices at each of the plants with requests to check the original records to obtain the correct or missing informa- tion At the time of this report we have received the missing informa- tion or corrections from all but 12 plants eight plants have not yet > responded to our requests and four plants have not yet been contacted because we receive thde initial information concerning the workforce at these plants very recently Only six of the twelve plants have a signi- ficant amount of missing information thus we were able to include 42 plants in the preliminary analyses which are included in this report Race was information not recorded on most of the files which were obtained from the corporate headquarters Since race is known to be an important variable in mortality analysis we made a special effort to obtain race information for the 1977 Mortaltiy Study Race informationinformation which had been obtained for several plants from company for earlier OHSG studies was transferred to the Workforce Enumeration Files for both the active and retired employees Lists of employees for whom we did /. each the not have race information were prepared for plant because only | source of race information is the original employee records which are a stored at the plants These lists were sent either directly to the local planot r s to the corporate headquarters for forwardtionthge . local plants At the thoi f ts his report race information has been obtained for all active employees of three companies and all re- tireeosf one company Since race is still missing for a large number -- JACKSON 025926 of employees the preliminary mortality analyses which are included in this report do not use race as a variable However race information will be used in the mortality analyses which will be run at a later | | | | date After the active and retiree files for each company had been edited . and all available correction entered the two files for each company were merged to create a single Workforce Enumeration File for each company Since the employees who retired in 1977 were entered in both the active and retiree files the duplicate records for these employees | had to be removed The resulting file was later merged with the Deat|h- Certificate File to create a 1977 Mortality Study File for each company B. Creating the Death Certificate File Death certificate processing is an important part of any mortality study , especially if the study involves comparison among companieosr comparisons with national vital statistics . Comparability is best achieved by following clearly defined processing procedures which in- | clude the coding of the cause of death by certified nosologist an expert in determining the underlying cause of death from the information provided by the attending physician on the death certificate and in coding this information in accordance with International Classifi- | cation of Diseases . | Although the size of the 1977 Mortality Study required few modi- fications and careful implementation the procedures which were used to process the death certificates for this study are essentially the same . as the procedures which have been used by OHSG personnel to process death certificates for retrospective epidemiologic studies The major - - steps in this procedure are described below JACKSON 025927 14 In January 1978 a letter was sent to the medical director or his counterpart at each company requesting a copy of the death certificate . for each hourly active employee or retiree who died in 1977. The medical directors forwarded this request either to the company insurance department or to the insurance company which underwrites the company's life insurance program Since OHSG receives many death certificates for other research projects some of the requested death certificates were already on file at OHSG before the request was sent However most of the death cer- tificates had to be duplicated and processed after the request was - received in January Most of these death certificates were received by the end of March but a large number of certificates were delayed for several months We received more than 110 death certificates almost 10 of the total number received after July % the last large group | was received about August 15. Although some death certificataerse still outstanding the time of this report we had received complete death certificate information concerning about 99 of the known 1977 deaths In addition to death certificates we also requested proof forms and a year claims listing as they were available a Although claims listings do not include information on the cause of death and may not always be complete they are very useful for checking our records to be sure that we have received all certi deaf thicecrta ift ice ats es When a requested death certificate had not been received we checked with the company to determine if they had received it If they had not yet receiveidt we requested supplementary information such as the social. security number and the state in which the death occurred so could request a death certificate from the vital statistics of office that JACKSON 025928 .15 state Because of the lengthy delays involved in obtaining information from some states we attempted to obtain as much information as possible from the companies However if the company did not have the death certificate the state vital statistics office is usually the only other | source of information _ Death certificates were usually received and processed in groups | Therefore the processing procedure was designed so we could account for each certificate in a group after the processing for that group of death certificates had been completed The steps in the processing procedure are as follows 1 Since some companies send OHSG copy of the death certificate for all deceased employees the first step in processing each group of death certificates was to remove all certificates for | employees who were obviously not included in the 1977 Mortality oo Study All certificates for deaths which did not occur in 1977 were removed first other certificates were removed if information provided by the company indicated that the employee was not an hourly worker nota member of the URW Union or did not work at one of the 48 plants which were included in the study The number of death certificates which were removed.. - durti hisnstg ep was noted and the death certificates were filed for future reference or use in other OHSG studies 2 All county city and local death certificates were removed and replacement certificates were requested from the state office. _ of vital statistics It has been found that the forms issued ~ by countoyr local registrars of vital events may omit some of the causes of death or list them in a different - sequence than _ JACKSON 025929 16 the official state certificate does Since these discrepancies can affect the coding of the causoef death we use only those death certificates which are issued by the state All of the informatioonn the remaining death certificates was checked for legibility and completeness The information was other : also checked to see if it was consistent with any infor- . mation which was available from proof forms or claims listings Minor errors omissions or inconsistencies were corrected by using informatito hant was available at OHSG or by contacting the corporate insurance office or the local- | planatt which the employee workeIdf much of the information was illegible a copy of the death certificate was requested from the state vital statistics office If the causoef death was illegibloer listed as pending the state vital statistics office was contacted to obtain a supplementary medical certi| fication of the cause of death Each death certificate which was complete and legible was copied and a nosology form was attached The certificates were counted and the group of certificates was sent to the nosologist for coding | When the coded certificates were returned to OHSG they were checked to be sure that they were all returned and that the coding was complete If the nosologist had not been able to read the cause of death information a new copy of the death~ certificate was requested from the state vital statistics office and the certificate was recycled through the processing -- ae procedure JACKSON 025930 17 6 The coded death certificates were then sent to data processing to be converted into a computer readable form and entered into the death certificate file The information on each certificate was typed on an OCR Optical Character Recognition typewriter and was entered into the computer file via an optical character reader Each certificate was typed by two different operators | and both copies were entered into the computer which compared - ~ 7 them to determine if they were exactly the same If any ' differences were noted the records were printed the differ- . ences were rectified and both records were again entered into the computer When no differences were detected the information was added to a working file When the processing for the entire batch of death certificates was completed it was added to the Death Certificate File 7 After most of the death certificates had been received and processed the information on the Death Certificate File was - merged with the information on the Workforce Enumeration File Since death certificate information is available from a restricted number of sources and is not available in a computer readable form the processing of death certificatweis ll always be time consuming However the time required to process the death certificates for an annual _ mortality study could be substantially reduced if all of the companies would followa few guidelines when they prepare death certificates for OHSG We would like to suggest that the following guidelines be adopted for future studies 1. We are aware of the fact that the copies of the death cer- tificates which are sent to the companies are sometimes barely _ JACKSON 025931 greatly legible Our work would be expedited if the copies of | death certificates which are sent to OHSG are as clear as 7 possible | | 2. Checking for the completeness of death certificates is greatly 7 expedited by the availability of a claims listing Thus we et would appreciate receiving a copy of the year claims | co listing if it is available - . 3 It is useful to have the employee's employe's pay status hourly or salary and plant entered on the death certificate especially especially if this information is not on an accompanying insurance claims listing or if a proofdeath insurance form containing this information is not attached to the death certificate : C. The Creation of 1977 Mortality Analysis File After the Workforce | Enumeration File for each company was completed it was compared with the Death Certificate File and death information te was added to all matching records on the workforce file During this process lists were made of all of the records either file for which matching records should have existed but were not found All of these unmatched records were then checked to determine why matching records did not exist . . . All employees on the workforce file who were coded as deceased in 1977 should also have had a record on the Death Certificate File Some - did not To determine why these death certificate records did not . employees exist we contacted the local plants to find out if these had | indeed died in 1977 and were URW hourly workers In the majority of cases the date which was recorded as the death date was actually a JACKSON 025932 .19 benefit payment date and the death itself occurred prior to 1977. Since this study did not include deaths which occurred before 1977 these records were removed from the workforce enumeration file At the time of this report all but five deceased records on the workforce file which did not match the death certificate file have been explained and | 7 appropriate corrections made Since each record on the Death Certificate File represents a 1977 | death there should also have been a matching record on the Workforce Enumeration File. All death certificate records without a matching Work- force Enumeration record were listed and the cause of the discrepancy : was investigated The majoritoyf these matches were for employees- who left the company witha termination award priotro 1977 and who - ; were not included in the study population for the 1977 Mortality Study In a few cases we discovered that an employee for whom we received a death certificate did meet the criteria for inclusion in the 1977 Mor- - tality Study but the employee's record was not included in the workforce ' information which we had received from the company Apparently these records had been deleted from the company's file when pension benefits . were discontinued thus we did not receive the record when they copied . the file for us in the Spring of 1978. By using some information which - one company had provided for another study in 1976 we were able to identifalyl such deleted records for that company and to add every| record which met our selection criteria to the 1977 Mortality File For another company letters had bebe written to the local plants to determine if the unmatched records should be addetdo the Workforc~e~ Enumeration File At the time of this report answers had not been received from all plants so there are still 23 records of 1977 deaths' _ ae JACKSON 025933 20 for which we do not have a record the Workforce Enumeration File We ~ cannot add these 23 records to the 1977 Mortality Study until we have oe received information from the local plants Because of the problems we have had in obtaining some information and removing all known errors from the information which we have received the mortality analyses which are presented in this report are preliminary analyses which are based on approximately 85 percent of the total study population By the time the final report is written before the end of 1978 we hope to be able to determine whether these 23 deaths should be included and we hopeto have adequate data for the six plants which are excluded from this report D. Processing Comparison Population Data As mentioned in the Introducdt atia ofonr the U.S. population came froma publication of the Bureau of the Census This data was key oo . punched and double checked for accuracy by OHSG staff The information on deaths in the U.S. however came to us on computer tape The file we received from the National Center for Health Statistics had about 1,900,000 records on it or onerecor fodr each death in the U.S. during the year 1975 identification information had been removed This massive file had to be copied and summarized by the OHSG data processing staff The summary data from this file was then compared with the ; published mortality figures for 1975. The 1975 mortality data has not yet been published in the detail needed for this study so the procedure used provided us with early access to recent data' JACKSON 025934 III ANALYSIS AND RESULTS The mortality of the rubber workers studied was analyzed and is reported here in terms of age standardized mortality ratios SMR's SMR's The SMR is the ratio of the number of deaths in the study population to the number of deaths which would be expected based on the mortality mortality - - experience of the comparison population To adjust for the age distri- bution differences between the study population and the comparison population the | year age group numbers of expected and these are added deaths are calculated for together to get the total each five | number expected Although the final formula for an SMR includes only the number of deaths observed and expected - oo observed number of deaths = SMR number of deaths expected X 100 the earlier steps in this calculation involve the computation of death greater rates The SMR is uaually multiplied by 100 so that an SMR than | 100 means that more deaths were observed in the study population than would be the case if the same death rates appltioetd he study popula- tion as were present the comparison population Conversely an SMR less than 100 means that the study population had fewer deaths than it present would have had the same death rates applied to it as were in > | | the comparison population . Mortality rates for most diseases change over time It is there-- fore important that the comparison population data be as close time to the study population dataas possible As was mentioned above the . most recent U.S. data available were for the year 1975. These were obtained from the Bureau of the Census and the National Center for Health . Statistics The death rates used for comparison in this study JACKSON 025935 22 were then calculated the OHSG Since the data from the National ; Center for Health Statistics NCHS were provided on computer tape they were checked for accuracy against preliminary figures publishebdy the | | | | NCHS Because of the uniqueness of this study in looking at mortaliftoyr . many plants during only one year a special computer program was written to analyze the data This program is discussed briefly in the appendix - The mortality experience of the rubber workers in this study was analyzed separatefloyr males and females for 42 of the 48 plants in the study The remaining six plants will be in included the final report to be released befo theree nd of the year These six plahn avetbes en , omitted from this report because of inadequate daIt t a was mentioned earlier that this studiys compriosf etwdo separate populations namely active employees and retirees These are separated in this report by age those under 65 being considered as active employees and those over 65 being consiads e rer tie red es Because there are retirees under the age of 65 a further refinement of this separatiwoilnl occur in future | | reports Death certificates have been located for 99 percent of the known~ deaths All causes of death stateodn the death certificates were coded by a qualified nosologist according to the Eighth Revi ofs thi e Io nten r- 7 national Classification of Diseases Adapted for Use in the United States (3) ICDA i | Included in this study are 48 plants ranging in size from less than 100 to over 5500 hourly workers During the year 1977 over 65,000 people were active employees employees in plants These plants are scattered throughout 22 states The major concentratioonf the plants in the . JACKSON 025936 23 East North Central geographic region including Ohio Indiana Illinois Michigan and Wisconsin where 20 plants are located with a totaolf over 23,000 actively employed hourly workers In Ohio alone there are ~~ seven plants from all four companies employing approximately 21,000 hourly _ workers In addition to the active employees there are over 25,000 retirees from these 48 plants As mentioned mentioned above this report covers 42 7 of the 48 plants including 79,496 active and retired rubber workers As can be seen from the age and sex distribution in Table 1 79 percent of the population was in the active age range while the remaining 21 per- - cent was at least 65 years of age Approximately 87 percent of the acti werv e e mals es and 78 percenotf the retireweesre males to occurred During 1977 a total of 1277 deaths are known have _ within the study population of 42 plants Most of these occurred among retirees with 331 deaths among those under age 65. Of these 1277 deaths| 1128 were among males and 149 among females . Tables 2 thr 7 o of tu hisg reph ort present SMR's for various disease categories for all 42 plants combined Because all plants are combined ~ for these tables many very different processes and chemical exposures are included and it should not be assumed that because an SMR is low : there are no problems in the industry relative to that cause of death : Looking at large heterogeneous groups such as the rubber industry as a whole may mask specific conditions which exisitn subgroups of the population 4,5 On the other hand a high SMR does not necessarily - mean that every plant in the industry has a problem with that particular disease At the same time high SMR in these tables should be taken seriously Because it is based on the experience of thousands of workers it is not likely to be a chance happening It may either represent JACKSON 025937 24 a serious problem at a few plantosr it may be a rather widespread _ problem in the industry _ The next step the analysis is to look at individual plants to see if there are particular problems at some plants Tables & through 10 present SMR's for the individual plants In order to prevent iden- tification of specific plants each plant was assigned a numbeart random. and only that number appears as identification in the tables In addition the numbers of observe andd expected deaths do not appear in Tables8 throug1h0. The URW will given a list of the plant codes and each company will be given a list identifying only its own plants Tables 2 through 7 show the numbers ofobserveandd expected deaths for each cause listed SMR's however have been calculated for only - those causes which have at least five observed deaths becaoufsthee statistical instability of rates which are based on small numbers While not reporting some high but probably meaningless SMR's some posbust ipobssliby ly meaningful SMR's will remain unreported well Although Tables 8 through 10 do not report numbers of observed and expected deaths the same rule applies for calculating SMR's Dashes in the tables represent plants or diseases with fewer than five deaths as well as the six plants not included in this report The test of statistical significance used in this report is the .2 test expected x observed deaths - x _ expected deaths 2 deaths ) > .. Although SMR's are not reported for fewer than five deaths each cause of _ death was examined and a test for statistical significance based on the Poisson distribution was usetdo check the probabilitthayt what was ~~ observed was a chance occurrence These cases are discussed later BATO JACKSON 025938 26 do not currently have death certificates were included in this category When death certificates are obtained these deaths will be reclassified | | to other disease groups Diseases of the circulatory system account for about 54 percenotf all deaths Similarly in the U.S. general population about 55 percent of all deaths are attributed to circulatory diseases The next most frequent cause of death was neoplasms which accounted for 23 percent of the total deaths Because these two groups of diseases consist of categories - several smaller disease of special interest they are pre- | sented in more detail in Tables 3 and 4 As seen in Table 3 several site specific neoplasms have elevated SMR's in the study population In the active group stomach cancer~ and neoplasms of the lymphatic and hematopoietic elesv howaet leveatd ed SMR's Within this latter group the category other neoplasms |. lymphoid tissue polycythemia vera and myelofibrosis had a statistically i elevation significant threefold in the SMR . Among retired men several site specific cancershad elevated mortality ratios Interestingly while our earlier mortality 7 study showed fewer deaths than expected due to cancer of the respiratory system for both active employees and retirees this study a shows s,,>a- tistically significant elevation for lung canceramong retirees Table 4 presents mortality ratios for certain selected diseases of the circulatory system for males from all plants combined In all cases fewer deaths were observed than expected among those in the active age range Among retirees the number of observed deaths was quite close to what was expected JACKSON 025939 .25 In this report the results of the cause specific mortality for all plants combined and for individual plants were presented primarily for large disease groups and for a limited number of smaller groups or individual causes of death Additional analyses for other plant combi- combi- combi- nations and for other diseases may be included in future reports Tables 11 through 13 present a comparison of combined plant death : rates with U.S. death rates for each five year age group for total aan deaths total neoplasms and for total circulatory disease These rates | are expressed as the number of deaths per 1000 population The discussion which follows is by no means intendetdo beexhaustive but highlights soom f te he more noteworthy findings in these tables . Table 2 presents the observed and expected numbers of deaths and the corresponding SMR's SMR's for males for major disease groups The SMR - for all causes combined for men age 15-64 is 73 which represents a statistically significant deficit of deaths This is usually usually explained by what is known the healthy worker effect 6 In other words the fact . that the study population is composed of people who are healthy enough to be working while the comparison population in additt o i heao ltn hy workers includes the disabled the institutionalized and others in ill health means that the study population should compare favorably to the comparison population A high SMR then becomes all the more noteworthy Statistically significant deficits also appear in the same age group for , circulatory diseases respiratory diseases accidents poisonings and violence and digestive diseases Two disease groups have statistically . significant elevations in their SMR's among the retirees These are for . neoplasms and for symptoms and defined conditions The SMR for this last group is artificially infsl inca e st omee knd own deaths for which we ce JACKSON 025940 | 27 Because of the small number of deaths among females the tables for 7 females show much less detail than those for males Table 5 which is for all plants combined has one SMR which is statistically significant That one is for accidents poisonings and violence among the active aged women This is particularly noteworthy since five of the seven deaths in this group were suicides These deaths were not restricted to a any specific plant It is interesting that the SMR for total deaths ICDA 000-999 does not follow the same pattern among women as it did among men that is with a low SMR for those aged 15-64 and an elevated . . SMR for those aged 65-84 65-84 _ Instead for women the two SMR's are quite similar The SMR's for the same two age groups for circulatory disease are likewise quite similar to each other Neoplasms however , show what is a more typical pattern among men The elevated SMR for neoplasms among retirees appears to be consistent for several specific neoplasms - as shown in Table 6 specifically digestive cancer breast cancer and | | genitourinary cancer Although the SMR of 221 for cerebrovascular disease in Table 7 / among those people over 85 years of age is statistically significant it . may not be very meaningful because of the unreliability of diagnoses for people of that age Yet this is something which should be kept in mind when future studies are done Consistent results over several years are | Te likelyto be meaningful wg : As mentioned above Tab8laends9 give plant specific SMR's for males for select causes of death To illustrate the earlier mentioned 7 reference to the masking effect when large heterogeneous populations are studied notice that although the SMR of 106 for all causes among retirees - is not particularly noteworthy some individual plants do show SMR's that are quite high BATO JACKSON 025941 '. 28 Two plants show statistically significant deficits of deaths in the age span 15-64 for total deaths ICDA 000-999 Plant 8 has an SMR of 46 while plant 18 has an SMR of 57. For plant 18 this deficit also . shows up for circulatory disease in the same age group with an SMR of 46. Although the SMR for neoplasms in age group 65-84 at this plant is only 112 the SMR for lung cancer ICDA 162 as seen in Table is oan . 7 186 whicihs statistically significant a , Plant 10 shows a statistically significant elevation among those age 65-84 for deaths due to neoplasms with an SMR of 222. Plant 33 had three times number of total SMR of 302 was not attributable deaths to any expected specific among the retired This | cause however The elevation due to circulatory disease in plant 36 SMR of 208 is largely due to cerebrovascular disease notice the SMR of 452 in Table 9 al- though deaths due to ischemic heart disease were also elevated This particular plant is being investigated further because 82 percent of all deaths in the plant were due to circulatory disease The SMR of 186 for neoplasms among retirees at plant 41 is statistically significant Similarly Table 9 shows an elevateSMdR of 226 for lung cancer in the . same age group for this plant Plant 47 has consistently high SMR's for myocardial infarction for both age groups 15-64 and 65-84 Boot f th he SMR's 283 and 231 are statistically significant Plant 48 has three | SMR's for the age group 65-84 which are statistically significant These are for total deaths SMR = 132 neoplasms = SMR 172 and myocardial infarctSiMoRn= 198 Although the SMR of 240 for cir- culatory disease among those people over 85 yearosf age is statistically a e 7 significant as mentioned above this is something to be watched in future studiebust it may not be important because of diagnostic unre- a JACKSON 025942 liability Only five plants had enough female deaths to report plant specific SMR's These are shown in Table 10. Although not statisti- cally significant the elevated SMR's for total deaths in the 15-64 year old age group in plants 18 and 39 the SMR of 132 for total deaths among retirees at plant 26 and the SMR of 153 for neoplasms among retirees in plant 32 should be noted and checked for consistenciyn later studies Although SMR's were not reported when there were fewer than five observed deaths yet each SMR was examined and tested for statistical significance Statistically significant elevations in SMR's with fewer than five deaths appear in Tables 8 through 10 as *** In each of these cases there appear in Table 9 was for more lung than one death but less cancer among actives at than five These plant 33 and in Table 10 for all causes and circulatory diseases among retir atepe las nt 8 and | for neoplasms among retirees at plant 26. Although these are based on few deaths they should be noted for further observation As can be seen in Table 11 the death rates for all causes combined is lower in the rubber industry than in the U.S. population for all ages up through age 64 although the differences are not statistically sig- nificant through age 49. After age 65 the rubber industry rates are higher than the U.S. rates for some age groups and lower for others The comparison of the rateisn Table for circulatory diseases does not show any pattern with the rubber industry having higher rates than the U.S. for some age groups and lower rates for others Table 13 which is for neoplasms shows a statistically significant difference in the rates for age group 70-74 In this case the rubber industry has a higher death rate than the U.S. population JACKSON 025943 IV SUMMARY AND CONCLUSION The principal purpose of the 1977 Mortality Study has been to demonstrate the feasibility of conducting a mortality study on an annual basis as part of a health surveillance system and at the same time to report on that study within a reasonable amount of time This has been accomplished Some of the problems encountered in this first study of _ this type such as obtaining race data will be avoided in subsequent studies because the data will not need to be collected again for the vast majority of the population As mentioned before the importance of an annual mortality study increases as time goes by because it provides valuable data for comparison and for observing trendosr changes in the health of working populations over time The function of studies such as thiiss to indicate what plants and disease groups particularly need further investigation with the ultimate goal being to recommend actions which remove hazards to the | health of workers . A final report on which will include all this study will be prepared by the end of 1978 48 plants as well as further analyses of the 42 plants included here Additional checks will be made on the quality of the data presented here for example checking for additional death certificates checking on the correct identification of people for whom we have death certificates but who are not part of this report and checking the accuracy of the nosology coding although samples of records checked in the past have indicated a high degree of accuracy As men- tioned earlier this study is in reality made up of two separate popula- tions actives and retirees All 1977 actives regardless of how long JACKSON 025944 31 they worked in a plant were included in the study Because a minimum ' length of service is required for retirement the retirees all had to have at least ten years of service in a plant This gives the two groups quite different exposure characteristics For this reason we hope to continue the analysis of this data by looking at actives and retirees separately This report separated on age prior to 65 and 65+ which approximates the retiree separatiobunt because many people retire prior to reaching age 65 this is not as precise as we would like Other analyses planned for the next report will include length of service and race neither of these variabebinl g reeas dy for use at this time We also plan to group plants according to products to see if there are problems which common to plants of one type Individual company analyses will be done for the next report as well The final report on this study will also include analyses using other methods of standardizing for age Instead of using only the U.S. population for comparison populations based on smaller geographical areas mabye used and the total rubber industrmayy be use asda comparison for specific companies or plants JACKSON 025945 32 REFERENCES U.S. Bureau of the Census Current Population Reports Series 25 No. 614 Estimates of the Population of the United States by Age Sex and Race 1970 to 1975 U.S. Government Printing Office Washington D.C. 1975 A description of the data available on computer tapes is available | from the U.S. Government Printing Office in a publicatieonn- titled Standardized Micro Tape Transcripts | National Center for Health Statistics Eighth Revision Inter- national Classification of Diseases Adapted for Use in the - United States U.S. Government Printing Office Washington D.C. Mancuso TF Ciocco A Attar AA An epidemiological approach to the rubber industry JOM 213-232 1968 . Andjelkovich D Taulbee J Symons M Williams T Rubber Workers with Reference to Work Experience 405 1977 Mortality of| JOM 397- McMichael AJ Standardized Worker Effect Scratching 168 1976 Mortality Ratios and the Healthy Beneath the Sur.fJOa M 1c 65e - McMichael AJ Andjelkovich DA Tyroler HA Cancer Mortality Among Rubber Workers An Epidemiologic Study Ann New York Acad Sci 125-137 1976 . JACKSON 025946 ' TABLES NOTE In Tables 2-7 which follow the number of observed and expected deaths for all causes all neoplasms and all circulatory which appears at the bottom of each | table may not equal the sum of the observed and expected deaths for the individual diseases listed in the table This is because the disease groups are not always exclusive and because the bottom line may include diseases than are listed in the tables mutually more JACKSON 025947 34 wy TABLE 1 Distribution of Population and Deaths by Age and Sex Age Males Population Females . Total Males Deaths Females - Total 15-19 510 87 1597 0 20-24 5433 640 6073 10 25-29 8148 725 8873 30-34 7989 8734 8 35-39 5844 | 684 6528 8 1 1 1 11 0 9 1 9 1 40-44 4440 581 5021 11 1 12 45-49 4414 710 5124 24 1 25 50-54 5765 1006 6771 33 5 38 55-59 6182 1479 7661 74 10 84 60-64 5871 1544 7415 119 14 133 65-69 4697 1367 6064 180 23 203 70-74 | 75-79 80-84 | 85+ 3674 2482 1487 704 1108 669 375 136 4782 3151 1862 840 217 183 158 94 29 246 26 209 19 177 17 111 Total | 67640 11856 79496 1128 1277 JACKSON 025948 TABLE TABLE 2 OBSERVED AND EXPECTED DEATHASND SMR's FOR MAJOR DISEASE CATEGORIES AMONG MALES FOR ALL PLANTS COMBINED Cause of Death Circulatory Diseases Neoplasms Respiratory Diseases Accidents Poisonings Violence Digestive Diseases Endocrine Nutritional Diseases~ Symptoms defined Conditions Nervous System Diseases Genitourinary Diseases Infective Diseases Blood forming Organ Diseases Congenital Anomalies Mental Disorders Skeletal Diseases All Causes Combined ICDA Codes 390-458 140-239 460-519 800-999 520-577 240-279 780-796 320-389 580-629 000-139 280-289 740-759 290-315 710-738 Observed 15-64 65-84 Expected 85+ | 15-64 65-84 85+ 137 74 8 42 13 5 6 5 3 0 1 1 1 0 411 58 175 57 842NOH120 18 842NOHO120 26 =. 842NOH120 17 842NOH120 15 842NOH120 5652OH 842NOH120 5652OH 842NOH120 | 5652OH 842NOH120 5652OH 842NOH120 5652OH . 8422NOHO1220 5652OH 842NOHO120 5652OH . 842NOH120 174.80 89.27 19.65 65.49 24.23 6.88 7.82 3.80 3.53 3.28 0.86 0.93 4.59 0.78 401.65 147.56 57.88 20.15 20.15 13.05 8.05 4.63 10.59 4.69 .1.56 0.31 1.99 1.10 67.10 11.23 8.38 2.44 2.07 1.59 1.27 0.32 2:28 0.58 0.26 0.03 0.25 0.13 SMR'S 15-64 65-84 85+ ** 83 41 ** 54 73 77 132 -- --- -- -- -- 102 119 98 89 129 130 186 108 57 107 -~ 86 116 95 -- -Sate - ~ -- -- -- -- -- -- -- -- -- 000-999 296 738 94 406.15 693.97 98.01 98.01 ** 106 96 | Note . The SMR's for Symptoms and defined cates for known deaths are included in and the deaths reclassified to another Conditions are artificially inflated because missing death certifithis category temporarily death certificates can be obtained disease category ; = Statistically significant at the 0.05 level of significance ** Statistically significant at the 0.01 level of significance | JACKSON 025949 TABLE 3 OBSERVED AND EXPECTED DEATHS AND SMR's FOR NEOPLASMS AMONG MALES FOR ALL PLANTS COMBINED ; ; Cause of Death ICDA Codes Observed 15-64 65-84 Expected 85+ | 15-64 65-84 - 85+ Buccal Cavity Pharynx Digestive Organs Peritoneum ' Esophagus Stomach Large Intestine . Rectum Liver Pancreas Peritoneum Unspecified Digestive Respiratory System Larynx Trachea Bronchus Lung Other Unspecified Respiratory Bone Skin Breast Genitourinary Organs Prostate Bladder Other Unspecified Urinary Other Unspecified Sites Brain Nervous System Lymphatic Hematopoietic Lymphosarcoma Multiple Myeloma , Leukemia Lymphatic Leukemia Myeloid Leukemia Monocytic Leukemia Other Lymphoid Polycythemia Vera Myelofibrosis Benign Neoplasms Unspecified Neoplasms ' All Neoplasms 140-149 150-159 150 151 153 154 155 157 158 159 160-163 161 162 163 170-174 180-189 185 188 189 190-199 | 191-192 200-209 200 ; 203 204-207 204 205 206 ! | 202,8,9 ~ 210-228 230-239 140-239 1 5 1 3.27 3.41 0.23 18 48 5 21.52 41.14 3.44 3 4 1 2.88 . 3.34 0.19 5 8 ONHOOH 3.51 7.03 0.68 0.68 6 19 ONHOOH 6.79 15.11 1.35 2 OH724610-2 ONHO OH 1.97 4.31 0.45 0 OH7246210-2 ONHOOH 0.79 1.10 0.05 2 OH724610-2 ONHOOH 4.61 8.21 0.53 0 OH724610-2 ONHO H 0.20 0.25 0.01 0 OH724610-2 ONHOOOH 0.11 0.29 0.04 27 OH7246210-22 NONO 35.39 46.65 1.62 . 0 NONO 1.48 1.84 0.08 25 OH7246210-22 NONO 33.41 44.23 1.50 2 OH7246210-2 NONO 0.28 0.34 0.01 2 OH7246210-2 OH 2.72 2.53 0.29 6 OH7246210-22 OH 7.60 28.54 3.64 4 OH7246210-22 HO 2.94 19.37 2.73 1 9 HO 1.71 6.01 0.72 1 4 ONQOHOH 2.31 2.88 0.16 8 11 ONQOHOH 9.17 11.61 0.82 2 1 .0 3.05 1.82 0.03 11 9 ONQOHOH 8.42 12.28 1.05 2 -0 ONQOHOH 1.63 1.92 0.12 2 1 ONQNOHHOHO 1.19 2.22 0.15 2 7 ONQOHOH 3.11 5.37 0.60 1 3 ONQOHOH 0.71 1.72 0.25 ane) 4 ONQOHOH 1.70 2.27 0.21 1 0 ONQOHOH 0.08 0.19 0.02 a 4 ; 5 1125 0 1.48 2.21 0.16 1125 0 1 1 25 . 0 74 175 13 0.40 0.79 | 89.27 0.41 0.04 0.97 - 0.11 147.56 11.23 * u Statistically significant at the 0.05 level of significance ***== Statistically significant at the 0.01 level of significance 1 wa ow i woot SMR's 15-64 65-84 85+ 12128 12128 12128 142 88 ----- 76 -- 75 --- 79 --= - 87 -- 131 ------- 147 -- 145 -- -- 114 ~~ 126 -- 139 -- ee 85 -- -~ -~ = -- 141 -- -- 145 -- --- -~ -- 112 98 150 -- -- -- =~ - - 95 -- 73 -- -ae G -- -- 130 } -- | -- | ~- 338 --- 83 ---- 119 1 12 1 116 36 JACKSON 025950 TABLE 4 OBSERVED AND EXPECTED DEATHS AND SMR's FOR CIRCULATORY DISEASE AMONG MALES FOR ALL PLANTS COMBINED ; Cause of Death Chronic Rheumatic Heart Disease Hypertensive Disease Ischemic Heart Disease Acute Myocardial Infarction Cerebrovascular Disease All Circulatory Diseases ICDA Codes Observed Expected 15-64 65-84 85+ | 15-64 65-84 85+ SMR's 15-64 65-84 85+ 393-398 400-404 3 20 1 5 1 3.10 3.46 0.20 -- Tose -- 2.29 5.77 1.25 87 -- 410-413 410 430-438 103 294 41 131.38 274.36 41.64 78 107 98 62 152 12 14 70 10 87.98 148.48 14.34 ** 102 84 18.17 71.15 14.79 77 98 68 390-458 137 411 58 174.80 401.65 67.10 ** 102 86 * = Statistically significant at the 0.05 level of significance ** = Statistically significant at the 0.01 level of significance 37 37 JACKSON 025951 TABLE 5 OBSERVED AND EXPECTED DEATHS AND SMR's FOR MAJOR DISEASE CATEGORIES AMONG FEMALES FOR ALL PLANTS COMBINED Cause of Death Circulatory Diseases Neoplasms Respiratory Diseases Accidents Violence etc. Digestive Diseases Endocrine Nutritional Diseases Symptoms defined Conditions Nervous System Diseases Genitourinary Diseases Infective Diseases Blood forming Organs Congenital Anomalies Mental Disorders Musculoskeletal Diseases Complications of Pregnancy All Causes ICDA Codes Observed Expected 15-64 65-84 85+ | 15-64 65-84 85+ . SMR's 15-64 65-84 390-458 140-239 460-519 800-999 520-577 240-279 780-796 320-389 580-629 000-139 280-289 740-759 290-315 710-738 630-678 11 53 12 15.60 67.68 10.85 71 78 10 27 2 15.08 22,42 | 1.17 66 120 | 3 3 1 1.91 5.05 0.82 -- - 7 3 Qa 3.02 2.52 0.30 232 -- 2 2 0 2.55 3.21 0.31 - -- 1 2 0 1.39 3.95 0.29 - ~- 1 3 0 0.68 1.17 0.18 = - e) 0 Q - _ -~- ~- ~~ 0 NHOHOO 1 0.57 1.64 0.20 | -- ~~ @) NHOHOO 0 0.41 0.72 0.07 -- =~ 0 NHOHOO 0) -~- -~ -- -- 6 NHOHOO 0 0.12 0.08 0.00 - -- 0 NHOHOO 0 -- -- -- -- 0 NHOHO O 1 0.25 0.42 0.02 -- ~- 0 NHOHOO 0 -- -- -- -- -~ . 85+ 111 -_ ---a --_ -- _ --- - 000-999 35 97 17 82 88 118 * = Statistically significant at the 0.05 leveolf significance 38 ro + JACKSON 025952 TABLE 6 OBSERVED AND EXPECTED DEATHS AND SMR's FOR NEOPLASMS ~ : AMONG FEMALES FOR ALL PLANTS COMBINED Cause of Death Digestive Organs Peritoneum Esophagus Stomach | Large Intestine Rectum Gallbladder Pancreas Respiratory System Trachea Bronchus Lung Bone Skin Breast Breast Genitourinary Organs Genital Organs Cervic Uteri Ovary Fallopian Tube Other Unspecified Urinary Other Unspecified Sites| ; Brain Nervous System Lymphatic Hematopoietic Lymphosarcoma Multiple Myeloma . Other Lymphoid Polycythemia Vera " Myelofibrosis Benign Neoplasms ICDA Codes Observed Expected ' 15-64 64-84 85+ | 15-64 65-84 85+ SMR's 15-64 15-64 65-84 85+ 150-159 HO OHMA 9 0 3.20 7.39 0.46 -- 122 -- - 150 151 153 154 156 HO OHMA 1 0) 0.18 0.24 0.02 -- -- -- HO OHMA 1 0 0.37 0.93 0.07 -- -~ -~ HO OHMA 4 O 1.39 3.35 0.22 -- =~ - ~ HOOHMA 1 - 0 0.30 0.67 0.05 -- ~ -- ~- HO OHMA 1 0 0.16 0.47 0.02 -~ _ a | 157 HO OHMA 1 0 0.65 1.45 0.07 -- -~ -~ 160-163 162 HO OHMA 0 1 2.35 2.29 0.07 -- -- -~ HO OHMA 0 H HO 2.25 2.18 0.07 -- -- -- 170-174 HO OHMA 6 HHHO 3.84 3.96 0.20 -- 152 -- | 174 HO OHMA 6 HHHO 3.53 3.56 0.16 -- 169 ~- 180-189 180-184 0 Q5A HHHO 2.63 3.69 0.17 136 ated DOONHO 4 O 2.29 2.83 0.12 - -- -- 180 DOONHO 1 O 0.62 0.53 0.02 -- ~~ -- 183 DOONHO IMHOMAL O 1.15 1.28 0.05 -- -- -- 189 DOONHO IMHOOMAL O 0.21 0.41 0.02 -~ -- -- 190-199 | 191-192 DOONHO 3 O 1.47 2.39 0.13 -- -- -- DOONHO 0 O 0.43 0.36 0.00 -- -- -- 200-209 DOONHO 3 O 1.12 2.13 0.10 -- -- -~ : 200 203 DOONHO IMHOMAL O 0.23 0.38 0.02 ~~ -- Seed O 1 O 0.20 0.41 0.02 -- -- < : : 202,8-9 202,8-9 210-228 Q 1 0 .0.19 0.46 0.02 | | || | 0 1 0 0.08 0.10 0.00 | | | All Neoplasms 140-239 10 27 2 15.08 22.42 1.17 66 120 -~ 39 JACKSON 025953 TABLE 7 OBSERVED AND EXPECTED DEATHS AND SMR's FOR CIRCULATORY DISEASE AMONG FEMALES FOR ALL PLANTS COMBINED Cause of Death Chronic Rheumatic Heart Disease Hypertensive Disease Ischemic Heart Disease Acute Myocardial Infarction : Cerebrovascular Disease All Circulatory Diseases ICDA Codes Observed Expected SMR's 15-64 65-84 85+ 15-64 65-84 85+ | 15-64 65-84 85+ 393-398 400-404 410-413 .410 430-438 106 02320 0 0.62 1.00 0.05 11211 -- 106 02320 0 0.36 1.38 0.22 11211 -- Sted 106 35 5 9.31 42.28 6.40 64 83 78 3 02320 1 5.64 20.23 1.86 11211 109 -- 3 10 6 3.10 15.61 2.72 11211 64 221 390-458 12 53 12 15.60 67.68 10.85 77 78 111 * = Statistically significant at the 0.05 level of significance 40 ao 1 vr JACKSON 025954 41 TABLE8 SMR'S BY PLANT FOR MALES FOR SELECTED MAJOR DISEASE CATEGORIES All Causes Circulatory Neoplasms Accidents : 000-999 390-458 140-239 800-999 . | | | | Plant 15-64 65-84 85+ 15-64 65-84 85+ 15-64 65-84 85+ 15-64 65-84 85+ , 1 72 109 -- 95 93 0 -- - 122 -- Joo - 2 -- -- -- - = -~ - -- - = 3 101 -- oe -- soo -- oo -- 4 5 6 3.7 8 9 10 11 12 76 ~~ 46 --. 139 - 100 -- = wwe wl 118 -- 79 0 = 91 190 -=-+ oe -~~ a a 77 -- 151 ~a .98 == ania ~ a 104 = mee - - = a oe -- mae ~ Pore == a ~~ _ wee ~-- 222 -- oo a = == == ~~ oe --- _ -~ -a ~ -_- oe - _ --- _ oe ae = = -- * : 13 14 15 . _ ~~ - ~e = _ ~- 1112111 ~ -~ - _ = -- _ _ ~ - - 16 - -- ~- oo -- 1112111 -- - -- mats - . 17 - aan 1112111 - a -~ - a 18 19 ** a 101 ~- 109 -- ** -- 1112111 98 1112111 97 ae- 112 . 109 -- --- = a 20 -~ = -- 1112111 -- -- = -- aoe 21 94 -- =n -- 1112111 = -- -- =e -- -- = 22 ~~ -Q = -- - oa -- - - - . 23 79 71 178 -- 70 -- - -- = _ - - = 24 73 135 = -- 114 == 166 130 -~ -- - = . 25 -- wee - eee -~ - = -- _ 26 27 28 108 _ _~ ~~ -- == = 130 __ -- 145 _ = - = -PTO a- oe _ _ - = -- a -- moe -- -_ - F 29 -- - = -- -- = -- -- -- = 30 ~~ ~~ = ~ -- = -- -- = -~ - is : 31 -- = -- - =- - - = ~~ = 32 107 91 145 95 - = -- - _~ -- + 33 -- 302 - -~ - - = -- - = -- - - = 34 -- = -- ~~ - o- -- - = 35 -- -- = -- - = 36 37 2168121 2168121 139 - ~ = -_ 208 ~ - = --~ -- = -- +e -- = Oo 38 2168121 - = -~ - = -- -- = - - = 39 40 2168121 2168121 81 - -= --- 69 = = --- 96 -- --- --- - = = 41 42 2168121 2168121 117 - as 56 82 -- -- 213 213 186 - -- 181 -~ _- - a 43 44 45 2168121 2168121 -- 107 86 ~ Oo 64 -- -- 102 -- 77 = = 213 -- 0 -- 106 ~- ee -- ose : 46 ~- - a= - - - Bt n All 47 48 Plants 100 86 ** 130 132 106 -- 165 96 | 159 90 78 129 138 102 == 240 | 86 -== 83 =~ = 172 -- 119 116 * = Statistically significant at the 0.05 level of significance ** Statistically significant at the 0.01 level of significance -~ 188 -- --oo -- -- 211 ** - os nae a~ a = a 89 = JACKSON 025955 TABLE 9 SMR'S BY PLANT FOR MALES FOR SELECTED Causes of dEATH 42 + Plant Acute Myocardial | 15-64 105 410 65-84 73 Infarc 85+ Cerebrovascular Disease 430-438) 15-64 65-84 85+ Cancer Trachea Bronchus 15-64 162 65-84 85+ Lung 113 ~~ -- 102 -- 96 -- 87 -- 186 ~ -- 452 73 87 -- 127 --- 226 -- 127 283 231 ** _- 198 -- 112 -- 145 -- at) All Plants 70 102 84 77 98. 68| 75 145 * Statistically significant at the 0.05 level of significance. ** ** = Statistically significant at the 0.01 level of significance *** *** *** Statistically significant at the 0.05 level of significance SMR not reported because of less than five deaths JACKSON 025956 TABLE 10 | - . SMR'S BY PLANT FOR FEMALES FOR SELECTED MAJOR DISEASE CATEGORIES 43 All Causes 000-999 Plant | 15-64 165-84 85+ Circulatory 390-458 15-64 65-84 85+ Neoplasms 140-239 15-64 65-84 85+ Accidents 800-999 15-64 64-85 85+ ey toa All Plants 82 88 118 71 78 111 *** level Statistically significant at the 0.05 of . of less than five deaths * Statistically significant at the 0.05 level of 66 120 232 significance SMR not reported significance -- oe because - JACKSON 025957 TABLE 11 COMPARISON OF MORTALITY RATES BETWEEN RUBBER INDUSTRY ALL PLANTS COMBINED AND U.S. POPULATION FOR ALL CAUSES OF DEATH AMONG MALES Age Study Population Person No. of . Years Deaths 15-19 20-24 426.1 4849.9 25-29 7608.7 30-34 7429.6 35-39 5357.8 40-44 | .4278.9 45-49 4398.5 50-54 5824.1 55-59 6065.4 60-64 5712.8 65-69 4507.3 70-74 3433.5 75-79 2299.4 80-84 1287.0 85+ 557.7 .0 10 9 8 8 11 24 33 74 119 180 217 183 158 94 Deaths Per 1000 Population . ' Rubber Mortality U.S. Mortality Kate Rate 0 2.062 1.183 1.077 1.493 2.571 5.456 5.666 12.200 20,830 39.935 63.201 79.586 122.766 168.549 1.473 2.094 1.999 2.058 2.765 4.191 6.670 10.439 ** 16.151 * 25.216* 36.373 55.602 82.589 115.933 175.726 44 Note Calculation of person is explaineidn the appendix * Statistically significant difference at the 0.05 level of significance- ** ** Statistically significant at difference the 0.01 level of significance- JACKSON 025958 45 TABLE 12 COMPARISON OF MORTALITY RATES BETWEEN RUBBER INDUSTRY ALL PLANTS COMBINED AND U.S. POPULATION FOR ALL CIRCULATORY DEATHS AMONG MALES Age 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 | 85+ Study Population Person No. of Years Deaths 426.1 4849.9 7608.7 7429.6 5357.8 4278.9 4398.5 5824.1 6065.4 5712.8 4507.3 3433.5 2299.4 1287.0 557.7 0 1 1 2 2 15 14 40 61 90 115 108 98 58 Deaths Per 1000 Rubber Mortality ~ Rate Population U.S. Mortality Rate 0 0.206 0.131 . 0.269 . 0.187 0.467 - 3.410 2.403 6.595 10.678 19.968 33.494 46.969 76.146 - 103.999 0.043 0.066 0.117 0.244 0.623 1.422 2.746 4.830 ** . 7.971 12.916 19.432 31.223 ... 48.987 73.204 120.295 Note Calculation of person years is explaineidn the appendix ** = Statistically significant difference at the 0.01 level of significance JACKSON 025959 46 TABLE 13 COMPARISON OF MORTALITY RATES BETWEEN RUBBER INDUSTRY ; ALL PLANTS COMBINED AND U.S. POPULATION FOR ALL NEOPLASMS AMONG MALES Age Study Population Person No. of Years Deaths 15-19 - 426.1 20-24 | 25-29 4849.9 7608.7 - 30-34 7429.6 35-39 5357.8 40-44 4278.9 45-49 4398.5 50-54 |. 5824.1 55-59 6065.4 60-64 5712.8 65-69 4507.3 70-74 3433.5 75-79 2299.4 80-84 85+ 1287.0 557.7 0 1 1 0 1 4 6 10 18 33 50 60 36 29 13 Deaths Per 1000 Rubber Mortality Rate Population U.S. Mortality Rate 0 0.206 0.131 0 0.187 0.935 1.364 1.717 2.968 5.777 8.752 17.475 15.656 22.533 23.310 0.074 0.096 0.130 0.188' 0.323 0.655 1.356 2.437: 4.055 6.493 9.112 12.834 * 16.408 19.183 20.124. Note Calculation of person is explained in the appendix * = Statistically significant difference at the 0.05 level of significance - JACKSON 025960 FIRESTONE Akron Albany Bloomington Decatur Des Moines Los Angeles ete Memphis Noblesville Pottstown Russellville Ruselvile Russellville Russellville SaSlianalsinas GENERAL Akron Mayfield Waco . GOODYEAR . Akron Conshohocken Cumberland Danville Freeport Gadsden Jackson Lincoln Los Angeles ' Madisonville Marysville New Bedford North Chicago GS St. Mary's 2 Sun Prairie . Topeka 47 APPENDIX I LIST OF INDIVIDUAL PLANTS GOODYEAR continued Tyler Union City Windsor UNIROYAL . Chicago =-. : Chicope _ Chicopee ...... Detroit Detroit ; Claire Eau Claire Los Angeles Masland 7 Mishawaka Naugatuck Chem & Syn Naugatuck Footwear ' . Opelika. Painesville Port Clinton Coated Fabrics Port Clinton Industrial Products ; Scotts Bluff Washington JACKSON 025961 APPENDIX II DIAGRAM 1 TYPES OF EMPLOYEES INCLUDED IN THE STUDY POPULATION Employment Status as of December 31 1977 + Jan. 1 1977 | | | | | |. | | | | + Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec. 31 1977 Active hired before Jan. 1 1977 Active hired during 1977 Sick leave Military or Educational leave Other leaves off Transfer to salary Quit discharged Retired prior to 1977 normal early disability Retired during 1977 early normal disability deferred vested Termination Award Deceased during 1977 | ; ~ date of hire | | | z | | | 4 date of exit dateof off date of off + 90 days date of transfer | date of exit | ; . ve date of - retirement : | date of exit i date of exit | date of death Employees who 1 began a military or educational leave before 1/1/77 2 retired before 1/1/77 with deferred vested benefits but received no benefits in 1977 3 received a termination award before 1/1/77 population 1/1/77 before were not included in the 1977 Mortality Study or 4 died 7? Zoe : a era 48 . are c i 7 JACKSON 025962 -49 an APPENDIX III Data Analysis Program 2 Past mortality studies done at the OHSG have identified study populations at given points of time and then have followed these pop- ulations for a number of years usually ten This study has some unique featuir n e ths at it is for only one year and people can enter the study at different times and can leave the study as well Diagram 1 7 illustrates the categories of people who enter or leave during the year 1977. Because of these differences from past studies a new computer | | program was written to analyze the data Basic to a cohort mortality study is the population from which the deaths occur The death rates in the comparison populatioarne applied to this study population to determine the number of expected deaths Yet when people enter and leave the study population it is difficult to determine exactly how many people are in the study population at any one time or even durinagn entire year Including people fora whole year who either entered or left during the year results in overestimating the population and consequently overestimating the number of expected deaths and underestimating the SMR An illustration may be helpful Suppose that a person is hired on February 1 1977 and quits on June 1 1977 If he died after June 1 theries a good chance that we will never learn of his death his death cannot be included in the study then neither ' should he be included in the underlying study population Yet if he 2 had died on May 25 1977 his death would be recorded on company files we would learn of it and we would get a copy of the death certificate It becomes clear then that he be should included in the study population JACKSON 025963 . 50 for that portion of the year during which he worked and for which his death will be recorded in company records or in other words for that portion for which he qualifies for life insurance The concept of the person year was used to account for the number of people in the study : population Person years are the number of years in this case a fraction of a year which a person contributes to the study population The task of the computer program is to calculate for each person that fraction of a year that he contributes to the total person years of the population This is made more complex by virt ofute he fact that - expected deaths are calculated separately for each race group so that the person years must be accumulated separately for each of these groups. The program that was written at looks each individual record If a person was hired before January 1 1977 and was still working on December 31 1977 the program adds one person year to the appropriate ce race category If however a person was hired during the year . or left other than for retirement or leave of absence the program cal- culates the fraction of the year he was eligible for life insuranacned . that fraction of a person year is added to the appropriate group This entire process is complicated further by the fact that each person a _ birthday during the year so that the person years of any individual may be divided between two different age groups The end result is a popu- lation which has been broken down into 60 different race groups with the appropriate number of person years in each group After the study population has been processed the expected number of deaths can be calculated The number of deaths in each race disease group in the comparison population are read into the computer These numbers are divided by the corresponding race population . ro | ? oa 7g JACKSON 025964 51 figures to get the comparison population death rates These rates for each disease group are multiplied by the study population figures to get the number of expected deaths for each race disease group There were 78 different disease groups studied for this report resulting in a total of 4680 possible disease groups Finally the deaths in the study population are tabulated according to the same race disease groups The SMR's can then be calculated for any combination of these groups by summing the observed and expected deaths across groups and then dividing the sum of the observed by the sum of the expected JACKSON 025965