Document M4L9E3E1ka5wg3YNG1K7ao0k

A Retrospective Mortality Study of Workers in Three Major U.S. Refineries and Chemical Plants Part II: Internal Comparisons by Geographic Site, Occupation, and Smoking History Nancy M. Harris, Ph.D.; Leba G. Shalienberger, M.P.A.; Donna L Donaleski, B.S.; Eugene A. Sales, MS A cohort of 21,698 U.S. refinery end chemical plant workers was observed for eight years to determine if there were inteiplant or other variations in causes of mortality. Plant populations in three geographic locations were com bined to develop an internal standard for comparing subgroups within the total cohort. At no one geographic site were consistently different rates for all major causes o; death observed. The adjusted mortality rates for poten tially exposed wnrkers were slightly greater than those for noncxposed workers for most causes examined. Smokers incurred a higher risk of mortality from many causes of death when compared with nonsmokers, regardless of oc cupational category. After controlling for smoking, there remained a slight excess in mortality lor potentially ex posed as compared with nonexposed workers. iVlortality of cohorts of workers from three Exxon refineries .and chemical plants in the United S tates was studied during the period 1970 through 1977.; Compari sons with the mortality experience of the U.S. population have-been reported separately'. This report details the re sults of lire analysis that combined the three plant site populations into a single cohort to address the- following objectives: 1. To develop an internal standard for comparing subgroups within the cohort. 2. To compare directly adjusted mortality rate's among irom the fyxpn Corporation, East Millstone, M). (Dr; Hants, Ms. Donaleski and Mr. Sales); and the Exxon Company, U.S.A., Houston (Ms. ShaUenberger). Address correspondence to Exxon Corporation, P.O. Box ','35, Me rer; Road, East Millstone, hi) 00873 (Dr, HapiS). Journal of Occupational Medici,nc/Vo! 27. No. 5/May 1965 the three plan! sile populations using the internal com parison group as the standard. 3. To compare directly adjusted mortality rates in the total study population by (1) geographic site, (2) occupa tional categories, and (3) smoking history. Materials and Methods The study design, cohort definition, and data handling techniques were discussed in detail in our previous com parisons with the mortality experience of the U.S. popu lation. 1 Work history summaries and smoking histories were used in the present analysis. Smoking histories have been documented in employee medical records since the mid1960s, but consistency of detail varies by plant. The crude measure of "ever-smoked" was indicated by "Y" (yes), "N" (no), or "IT (unknown). Work histories were summarized to derive the occu pation at which each employee had spent the largest por tion of his/her working time. Nine occupational categories were identified for analyses: Offkial/manager, profes sional, technician, office worker/clerk, process operator, mechanical worker, unskilled laborer, service worker, and laboratory technician/Tield professional. Comparisons used the total study population as the standard unless otherwise specified. Occupational cate gories based on employee records were classified into two groups based on their potential for daily exposure to pe troleum, petrochemicals, and other related substances. Mechanical workers, process operators, unskilled labor ers, service workers, and laboratory techniciandieid professionals were categorized as "potentially exposed" while officials/managers, professionals, technicians, and office workers/clerks were considered to be "nonexpoSed." Table 1 shows the occupational categories by job titles and by potential exposure grouping. Directly adjusted death rates were calculated by molti- 361 EXX-MOR-003351 Fig. 1 -- Directly adjusted mortality rates per 10,000 person-years, 1970 through 1977, tor all causes ol death, by occupation (704 employ ees, accounting for 98 deaths, could not be classified by occupation). NDN-EXPOSEO ntteb ef 1C l- POTENT IRLlY EXPOSED U .S. POPULATION s ... 1 - t*i > | .> ! 30 < I ' ?*! // to -- -> 4> u- si ir-'y Prior. --ul J ECU cu*. jom pcMks T0ti*|_ PEHSOHS SOt *2 27S to its 33 JOB ffjjs 2C7 ttc? OS i$t? 70 no 17 sss *nu*BEft5 Xt* PWlfXTHESES ME P*C$ MJ) 10-0&C . Fig. 2 -- Direcliy adjusted mortality rates per 10,000 person-years, 1970 through 1977, for malignant neoplasms, by occupation (704 em ployees, accounting lor TO deaths due to malignant neoplasms, could not be classified by occupation). plying each age-, sex-, race- and calendar year-specific death rate in the subgroup of interest (for example, smok ers) by the number of person-years in the appropriate cell in the standard population (total cohort). The resulting numbers of expected deaths were then summed over at! the celts and divided by the total number of person-years in the standard population. This quotient was multiplied by 10,000 to yield the directly adjusted rates per 10,000 person-years.. Rates among subgroups could then be di rectly compared because the same population distribution was used in the standardization procedure,3-3 The MantelHaenszel procedure, as implemented for life-table data with confounders, was used to test rate ratios for statistical significance.4-4 Results The majority of employees were classified as mechanical workers (33%) or process operators (28%), followed by professionals (11%). Based on occupational category, 75% of the study population had potential for daily exposures to petroleum, petrochemicals, or other factors in the Workplace (potentially exposed), 22% had little potential for exposure (nonexposed), and 3% had unknown poten tial (Table 2). As shown in Table 3, smoking information was available for 70% of the study population. Fifty-one percent were known smokers and 19% had never smoked cigarettes. Those occupational groups more likely to have potential occupational exposures generally had higher proportions of smokers. Professionals had the lowest percentage of smokers; however, they also had the highest proportion of subjects with "unknown" smoking histories (Table 4). Comparison of Directly Adjusted Mortality Rates Among Cohorts at the Three Geographic Sites---Directly adjusted rates for selected causes of death for the three plant co horts and for the U.S. population are shown on Table 5 if there were five or more observed cause-specific deaths at each of the three sites. No one geographic site cohort showed consistently different rates for all major causes of death. Kidney cancer was the only cause of death whose rate was higher than that of the U.S. population for all three plant cohorts. Cancer mortality rates in Bayway/Bayonne were generally higher than those in other plants, although only slightly higher than the U.S- rates, particularly for cancer of the digestive organs (with the exception of pan creatic cancer) and respiratory system. The rate for cancer of the pancreas was higher in the Baton Rouge cohort. Comparisons by Occupational Category-- Mortality rates for potentially exposed workers were slightly greater than those for nonexposed workers for most causes shown, with the exception that there was a twofold greater in crease in cancer of the prostate. For the category of alt causes of death and for diseases of the circulatory system, the differences between the two groups were statistically significant. Mortality due to cancers of the brain/CN'S was elevated considerably for workers classified as non exposed to petroleum, petrochemicals, or other factors in the workplace. Most of the deaths from cancers of the bladder and kidney and from iympbosarcoma/reticulosarcoma and leukemia were in potentially exposed workers (Table 6). The nonexposed technicians had the highest rates for all causes and for all malignant neoplasms (Fig. 1 and 2). Table 7 shows the directly adjusted mortality rates for se lected cancers by specific occupation. No one occupa tional group had consistently higher rates than the others for specific causes of death. Process operators had the highest rates for cancers of the digestive organs, respira tory system, and kidney; these rates were also higher than those of the U.S, population. Mechanical workers had the highest pancreatic cancer rate; this was also higher than the U.S. rate. Comparisons by Smoking History--The following com parisons were based on smaller numbers of deaths, since smoking data were unavailable for 30% of the cohort. Rates were not calculated when the observed number of deaths was less than five. Mortality rates for selected causes of death, shown on Table 8, were compared for smokers (ever smoked) and nonsmokers (never smoked). Smokers 362 Mortality Study/Hanis et al EXX-MOR-003352 No Exposure A--Officials/managers Buyers Chiefs Department beads Directors Managers Superintendents Supervisors : B-~Professionals/nonexposed Accountants Analysts Attorneys Coordinators Editors Engineers . Foremen/supervisors Nurses Physicians Planners Scientific advisors C--Technieians/nonexposed Cameramen Chainmen Computer technicians Data technicians Designers Draftsmen Engineering technicians Hospital/first-aid men Surveyors E--Office/clerical Key-punch operators Office assistant/clerks Office machine mechanics Reproduction operators Table 1 -- Occupational Group Definition Potential Exposure F--Process operators Coke handlers/rollers Dock helpers/masters Gaugers Plant trainees Process operators Process lechnicians Receiving/shipping operators G--Mechanical workers Auto mechanics Boilermakers Carpenters Craftsmen Electricians Instrument men Instrument technicians Insulators Lead burners Linemen Machinists Maintenance foremen Maintenance/mechanical helpers Masons Materials men Measurement men Mechanical asssistant/helpers Mechanical supervisors Motorized equipment operators Oilers Painters Pipe fitters Plumbers Railroadmen Riggers Servicemen Tinners Trainees Transportation operators Warehousemen Welders/burners Wiremen/joumeymen H--Laborers Handymen Laborers TanKermen/deckhands Utilitymen Yardmasfers/cleaners J---Service workers Chauffeurs Chefs/cooks Firemen/Fire equipment . Inspectors Guards Janitors/maids Storekeepers K---Laboratory Technicians/ Field Professionals Blenders/testers Chemists Gas testers Industrial hygienists Laboratory assistants Laboratory foremen Laboratory supervisors Laboratory technicians Oil conservation inspectors Refinery equipment inspectors Research engineers Research technicians Safety inspectors X-ray technicians experienced a higher risk of death due to the majority of causes shown. Of these, the excess numbers of deaths due to smoking were statistically significant for ail causes of death, diseases of the circulatory system, diseases of the respiratory system, diseases of the digestive system and for all malignant neoplasms. Smokers also experi enced a higher rate of digestive organ cancers, but the increase was not statistically significant. Rates for pan creatic cancer, cancers of all genitourinary organs, and prostate cancer were slightly higher for those who had never smoked. Most of the deaths from cancers of the fiver, respiratory system, bladder, kidney, lymphopoietic system and from leukemia occurred in smokers, but there Journal of Occupational Medicine/Vol. 27, No. 5/May 1985 were too few cases in the nonsmoker group to compare rates in these categories. Rate ratios of "ever" to "never smoked" indicated about a twofold difference between smokers and nohsmokers for most causes of death in both exposure categories. In the potentially exposed group the differences between smokers and nonsmokers were statistically significant for all causes of death, for diseases of the circulatory system, and for malignant neoplasms. In the nonexposed group the excesses for smokers were statistically significant for all causes of death and for malignant neoplasms. The ex ception was for cancer of the pancreas, for which, among the potentially exposed workers, the rate was higher among 363 EXX-MOR-003353 Table 2-- Distribution ol Total Study Population by Occupa tional Category Occupational Category No. :' !0 Potentially exposed Process operators Mechanical workers - Laborers Service workers Laboratory technicians/field professionals Nonexposed Officials/managers Professionals Technicians Office/clerical Unknown Total 16,257 4,737 704 21,698 74.9 6,135 28.3 7,162 33.0 1,572 7.2 479 2.2 909 4.2 501 2,476 269 1,491 21.8 3.2 98.9 2.3 11.4 1.2 6.9 Table 3 --Distribution of Total Study Popu lation by Smoking History Ever Smoked No, % Yes No Unknown Total 11,105 4,059 6.534 21,698 51.2 18.7 30.1 100.0 nonsrnokers (Table 9). In this Table, comparisons cannot be made across all four exposure/smoking groups because the two different occupational subgroups, not the total population, were used as the standards. The death rates for all causes among the potentially exposed workers were only slightly higher than or about the same as those among the nonexposed for both smoking groups (Table 10). Lor smokers, the slight excesses remaining among the poten tially exposed, compared with the nonexposed group, were statistically significant for all causes of death and for dis eases of the circulatory system. Discussion The majority of this cohort were mechanical workers or process operators during the major portion of their em ployment periods. Smoking histories were available for 70% of the population. Those groups more likely to have occupational exposures generally had higher proportions of smokers as well. ' Mortality rates for digestive and respiratory cancers were higher in the New jersey cohort. 7 he conspicuously high rate of pancreatic cancer in the Louisiana cohort may ex plain the increased pancreatic cancer risk observed in the total study population.1 Local, state, and county mortality data show similar patterns of cancer excesses for these .geographic sites.7 The kidney cancer rate was the only one found to be higher than that of the U.S. population con sistently across the three plants. However, the magnitude of increase was small and not statistically significant. Ad ditional analyses would be required to determine if these (cancer rate) differences represent true geographic vari ations or differences among plant processes. Detailed comparisons with state and local rates might provide more insight, and further analyses by plant processes at the three sites may be of academic interest; however, considering the low realtive risks observed, additional geographic analyses may not be justified. Occupational groups were examined by comparing po tentially exposed workers with nonexposed workers. The definition of occupational category was based on job titles and not on specific processes or chemical agents at the work sites, Miscfassification may have occurred as em ployees were assigned to an occupational category based Table 4 -- Distribution of Occupational Categories by Smoking History, 1970 Through 1977 Occupational Categories Ever Smoked No, % Never Smoked Nil % Unknown No. % Nonexposed Officials/managers Professionals Technicians Qffice/clerica! Subtotal Potentially exposed Process operators Mechanical workers Laborers Service workers Laboratory technicians/field professionals Subtotal Unknown Grand Total 255 797 144 638 1,834 3,368 4,080 835 241 505 9,029 242 11,105 50.9 32.2 53.5 42.8 38.7 54,9 57.0 53.1 50.3 55.6 55.5 34.4 51.2 too 556 66 409 1,141 1,032 1,313 182 59 : 213 2,799 119 4,859 20.0 22.9 24.5 27.4 24.1 16,8 18.3 T1.6 T2.3 23.4 17.2 16.9 18.7 146 1,113 59 444 1,762 1,735 1,769 555 179 191 4,429 343 6,534 29.1 45.0 21.9 29.8 37.2 28.3 24.7 35.3 37.4 21 0 27.2 48.7 30.1 Total No. 501 2,476 269 1,491 4 737 6,135 7,762 1,572 479 909 16,257 704 21,698 364 Mortality Study/Hanis et a! EXX-MOR-003354 Table 5__ Ditectly ftttjustetl*' Mortality Rates per 1QMQ Person-Years for Deaths Due to Selected Causes in the Baton Rouge, Baytown, anil Bayway/Bayonne Cohorts andfhe U.S. Population, 1970 Through 1977 Cause of death! Baton Rouge Observed No. Rate Death Rate Baytown Observed No. Rate Bayway/Bayonne Observed No. Rate US. Rate Highest Rate Ration All causes of death {000-999) All malignant neoplasms (140-209) All digestive organs (150-159) Stomach (151) Large intestine (153) Pancreas (15?) All respiratory system (160-163) Lung (162) All genitourinary organs (160-189) Kidney (189) All lymphopoietic (200-209) Endocrine/rrutritional/metaboiic diseases (240-279) Diseases of circulatory system (390-458) Diseases of respiratory system (460-519) Diseases of digestive system (520-577) Diseases of genitourinary system (580-629) Symptoms/senility/ill-defined {780-796) ....... All external causes (E8OQ-0399) Total Do. of person-years Total No. of empioyeesii Total Do. of deaths 1,205 249 73 11 24 23 79 78 39 9 25 18 729 86 34 25 8 52 56,012.0 8,662 1,205 238,9 47.5 13,5 1.5 4.5 4.7 15.3 15.2 7.8 1.7 4.9 3.4 784 158 39 6 11 10 54 54 20 S 16 9 147.4 455 13.5 48 6.6 25 5.1 10 1.5 17 9.2 38 46.19? 6 7,322 784 238.7 46.6 14.7 1,8 5.8 3.3 14.6 14.6 4.9 1,2 4.0 7,9 1,209 . 259 84 16 32 :9 84 77 41 7 14' 32 140/6: 762 17.3 50 6.2 34 3.8 : 18 4.7 17 8.9 21 35,525.8 5,776 1,209 239.2 55.6 17,9 2.7 6.7 1.9 1:9.1 17.7 7.3 1.6 3.5 6.9 255.3 51,2 14.2 2.5 5.0 2.8 16.8 16.0 9.0 1.1 4.5 4.6 142.3 143.8 8.4 18.5 7.3 9.5 3.2 3.7 5.1 3.2 5.0 14.3 0,9 1.1 1,3 1.1 1.3 1,7 1.1 1,1 0.9 1.5 1.1 1.7 1.0 0.9 ; 0.8 1.4 1,6 0,6 * Rates for the three plant cohorts and the U.S. population were directly adjusted tor age, sex, race, and calendar year using the total study population as the standard popttfation t Mernationat Glassification of Diseases (IGOAi. eighth revision codes shown in parentheses i Ratio of the highest plant mortality rate to the V S rate The sum of person-years for the Raton Rouge, Baytown, and Bayway/Sayonne cohorts is slightly greater /;'? 8 person-years, or 0.02% of the total person-years] than that for the total study pdptflat/on,-owing to rounding ' (I An employee may have been at more than One plant during this observation period on where they had spent the largest portion of their work ing fibre. it is impossible to determine accurately the mag nitude of misefassification, effects of which tend to dilute any real difference.** *Furthermore, the use of complete work histories for analyses requires more extensive meth odological development. The mortality rates tended to be higher among potentially exposed workers than among nonexposed workers but the differences were minimal. One exception was the nonexposed technicians, who had the highest rates for ail causes and for all malignant neo plasms, This occupational group represented 1.2% of the total population and only 0.9% of the deaths (N = 269; number of deaths, 30). Since these rates are based on a very small number of deaths and since the list of main jobs included in this occupational group does not suggest areas with occupational exposure problems, the: relatively higher rates among the nonexposed technicians cannot Journal of Occupational Medicine/Vo!, 27, No 5/May 1985 be interpreted further, Other exceptions ihetuded the res piratory and digestive system disease rates and the brain/ GN5 cancer rate, which were higher for nonexposed workers. The brain CNS result does not support the re cent finding in the Texas Gulf Coast study, in which the proportion of brain cancer deaths was higher than ex pected among refinery and petrochemical workers who were union members/1 No one occupational subgroup had consistently higher mortality rales tor all causes of death examined. However, rates were found to be generally lower for managers, professionals, and clerical workers. Smokers generally had higher mortality rates than non smokers, with the exception of deaths from cancers of the pancreas, genitourinary organs, and prostate. However, these observations were based on small numbers and interpretations should be made with caution. Smoking history explained most of the differences observed !* 365 EXX-MOR-003355 Table 6 -- Observed Number of Deaths and Directly Adjusted* Mortality Rates per 10,000 Person-Years for Selected Causes of Death, by Occupational Categories, 1970 through 1977 Cause of Death* ______ Occupational Categoryt _____________ _ Potentially Exposed ______ Nonexposed Observed No. Rate Observed No. Rate RR All causes (000-999) All malignant neoplasms (140-209) All digestive organs (150-159) Liver/gallbladder/bile ducts (155-156) Pancreas (157) All respiratory system (160-163) Lung (T62) All genitourinary organs (180-189) Prostate (185) Bladder (188) Kidney (189) Brain/CNS (191-192) All Lymphopoietic (200-209) Lymphosarcoma/reticulosarcoma (200) Leukemia (204-207) Diseases of circulatory system (390-458) Diseases of respiratory system (460-519) Diseases of digestive system (520-577) Diseases of genitourinary system (580-629) Total No. ot person-years Total No. of employees Total No. of deaths 2,657 557 167 14 35 T85 177 87 50 16 18 8 45 .10 15 1,640 124 66 47 236.1 49.8 14.6 1.2 3.0 16.7 16.0 7,7 4.4 1,4 1.6 0.7 4.4 0.9 1.7 145.6 10,8 5.8 4.0 103,360.1 16,257 2,657 443 189.1 99 41.0 25 10.7 0 7 2.8 29 12.3 29 12.3 11 4.8 5 2.3 3 3 , ,/ 7 2.8 9 3.8 3 4 250 106.6 32 14.7 19 7.5 3 30,354.7 4,737 443 1.2** 1.2 1.4 : 1.1 1.4 1.3 1.6 1.9 . .. A .. , 0.3 1.2 ,.. ' 1.4** : 0.7 0.8 * Rales were adjusted for age, sex, race, and calendar year using the total study population as the standard population; rates were not included w the observed number of deaths was less than five t far 704 subjects, accounting for 98 deaths, a category could not be assigned because of missing data t ICDA, eighth revision codes shown in parentheses ; ~ ' . Rate ratio (RR); ratio of rates of potentially exposed subjects to those of nonexposed subjects; statistical significance at p<.05 is indicated by ** tween the occupational groups. Smokers incurred a higher risk of mortality from most causes of death when com pared with nonsmokers, regardless of occupational cate gory. In the nonexposed category, technicians who had the highest death rates for malignant neoplasms also had the highest; percentage of smokers, while the profession als with the lowest death rates also had the lowest per centage of known smokers. In the potentially exposed category, the distribution of smokers was similar and the death rates were a!so similar among the five occupational categories. It appears that differences in death rates across occupational subgroups could be explained to great ex tent by differences in smoking history. Other studies have shown that cigarette consumption is higher among bluecollar workers as compared with white-collar workers.10 Since the majority of the potentially exposed workers in our study could be classified as blue-collar workers, our results are consistent with these findings. There remains a slight excess in mortality for potentially exposed as com pared with nonexposed workers allowing tor smoking dif ferences. However, the data on which this report is based do not allow further conclusions to be drawn because the numbers of deaths are too small in occupation-and smok ing-specific subgroups to permit further analysis. With 366 smoking histories unavailable for 30% of the study pc ulation, it is not possible to separate out completely l effects of smoking from potential effects of a specific! cupation. The exposed and nonexposed groups also rt have varied with respect to other socioeconomic fach that could be related to mortality outcome. Expansion this cohort over time and across other similar worker pc ulations may help us to identify the possible effects such cont'ounders as educational level, family history, efary intake, alcoholic consumption, and other presen unknown factors that may become apparent in the futu Summary A cohort of 21,698 U.S. refinery and chemical pi; workers was observed for eight years to determine if the were inlerplant variations in causes of mortality. Mortal outcomes were analyzed by geographic site, by occuj tional subgroups, and by smoking history. Mortality frt cancer of the pancreas was highest in the Baton Rou cohort, while mortality from digestive and respiratory a cers was highest in the Bayway Bayonne cohort. The interplant variations in mortal ity were consistent with kno' geographic variations. Only kidney cancer rates were cc 5istently higher than those of the U.S. population for Mortality Study/Hams et EXX-MOR-003356 Table 7 -- Observed Number of Deaths and Directly Adjusted* Mortality Rates per 10,000 Person-Years for Selected Malignant Neoplasms by Occupation, 1970 through 1977 ;r. of Dcathl Officials Managers (ti=5U1) rgah cancer satic cancer M5?i; /system (60-163! cancer;162) 1 ;' icer(l89i ; cancer 21 (204-2071 ; ; !li t i5) ;f3l Ta m (tii id lOi .......................... NoiwxvosBd .......... Occupation Group t__________ _______ _ . ..... ...... . . Potentially Exposed Professionals (H ~ 2.476) It .g : fl2) Technicians (N = 269J 12) (2) til (1/9 (t2l (Dl 11.9 (12) m ., Wm (3) m (2) tii Office Clerical (W= 1.491) 6:7 m n> 15.0 H3:| ISO 113) m in tii Process Operators (N --6.135) 1-8,3 (56): 2b 110) 19 A (55); 16 9 (53.1 18 (7> - (2) m Mechanical Workers (M = 7.162) 12.4 (70) 3,2 (18) 15 8 (301 iS.J (86): 1.6 (9) 1,1 (6) 16 . 19) Laborers (K-1,572) ' 16.! on 0 !i : (6) 17 0 (26) 16.4 (25) 1 (2) (01 (ij; Service Workers (11 = 479) 12 ! ('> (0) : 11.2 . (S) (4) (0) 0): m Laboratory Technicians/ Field Professionals (11=969) m (Ti IS.5 (9) IS fi (3) m (si 01 U.S. Rale 14.2 2.8 16 8 16.0 1.1 10 1,9 is were adjusted for age, sex, race, and calendar year using the total study population as the standard population; rates were not calculated when observed number of deaths was less than five e were 704 missing work histories; of these, 98 were for decedents I, eighth revision codes shown in parentheses (her of deaths shown in parentheses tbte 8 -- Observed Number of Deaths and Directly Adjusted* Mortality Rates per 10,000 Person-Years for Selected Causes : of Death, by Smoking History, 1970 through 1977 Cause of Deatht Ever Smoked Observed No. Rate Never Smoked Observed No. Rate Rate* Ratio 1 causes (0004999) All malignant (neoplasms (140-209) All digestive organs (150-159) Liver/gallbladder/bile ducts (155-156) Pancreas (157) All respiratory system (160-163) Lung (162) ` All genitourinary organs (180-189) Prostate; (185) Bladder (188) Kidnev (189) Brain/CNS (191-192) All lymphopoietic (200-209) lymphosarcorna/reticuiosarcorna (200) Leukemia (204-207) Diseases of pirculatory system (390-458) Diseases of respiratory system (460-519) Diseases of digestive system (520-577) Diseases of genitourinary system (580-629) Total No. of person-years Tola! No. of employees Total No. of deaths 1.566 377 96 10 24 146 142 44 22 7 15 9 31 7 12 908 77 52 19 254.9 57,1 14.5 1.3 3.7 22.2 21.6 7.4 4.1 1.2 2.1 1.9 4,2 1.0 1.6 153.1 13.2 7.6 3.8 74,362.7 11,105 1,566 208 43 17 2 7 4 '3 12 10 ; 131.5 24.5 9.2 4.4 ' , .. 8.4; 7.3 1 3 A t 1 1:34 : 7 7 7 26,418.3 4.059 208 92..0 3.8 3.4 4.0 1.9" 2.3** 1.6 0.8 0.9 0.6 .,. 1.7** 3.5** 2.2** 1.0** i?.ter= were adjusted tor age, sex. race, and caiehdar ye using; the total study population as the standard population; rates were not calculated when ie observed number of deaths was less than live . .... 'DA, eighth revision codes shown in parentheses _ aft of rams fot those who ever smoked ;n those who never smoked; steiist,c;!! significance at />.-.05 rs indicated iiy a double asterisk '. a! of Occupational Mcdicine/VoL 27, No. 5/May 198b 367 EXX-MOR-003357 Table 9 -- Directly Adjusted* Mortality Rates per 10,000 Person-Years for Deaths Due to Selected Causes, by Occupational Category and Smoking History Causes of Death** * Occupational Caiegoryf Potentially Exposed Ever Smoked Never Smoked Rafe Ratio Ever Smoked Nonexposed Never Smoked Rate Ratio All causes of death (000-999) , Diseases of circulatory system (390-458) ' Malignant neoplasms (140-209) Digestive organs (150-159) Respiratory system (160-163) Pancreas (157) Kidney (189) Brain (191) Leukemia (204-207) Total No. of Person-years Total No. of employees Total No. of deaths 279.7 169.0 62.1 16.0 24.2 3.9 2.3 1.8 142.6 100.6 26.2 10.0 4.8 1.0 103,360.1 16,257 2,658 2.0** 1.7** 2.4** 1,6 0.3 . . ,, 180 2 104.0 42.6 9.9 16.7 3.0 .; > 99.6 66.5 19.4 v . . 30,354.7 4,737 443 1.8** 1.6 2.2** ; . : - ;. *: * Rates were adjusted for age, sex. race, and calendar year using the occupational subgroup as the standard population; rates were not calculated the observed number of deaths was less than five f For 704 subjects, accounting for 98 deaths, a category could not be assigned because of missing data X IGDA eighth revision codes shown in parentheses Ratio of ever to never smoked; statistical significance at p<.G5 is indicated by & double asterisk Table 10 -- Directly Adjusted* Mortality Rates per 10,000 Person-Years tor Deaths Due to Selected Causes, by Smoking Histos and Occupational Category* Cause of Death* Ever Smoked ........... Rates Potentially Exposed Nonexposed Ratio ........ Never Smoked Potentially Exposed Nonexposed Rate Rati All causes of death (000-999) Diseases of circulatory System (390-458) Malignant neoplasms (140-209) Digestive organs (150-159) Respiratory system (160-163) Pancreas (157) Kidney (189) Brain (191) 256.8 (1,321)11 158.7 (783) 58.0 (321) 14.4 (83) 24.4 (129) 3.4 (19) 2.1 (13) M) 15 (10) 214.2 (210) 115.7 (111) : 48.7 (52) 12.0 (11) 14.0 (16) 3.4 (5) (2) 3.0 (5) (i) 1.2** 1.4*' 1.2 1.2 . 1,7 1.0 .. , . 128.9 (158) 88.2 (103) 25.9 (34) 9.7 (14) (3) 4.8 (6) (1) (2) (i) ' 125.9 (43) 83.6 (26) 21.7 - (8) (2) (i) .... (D (0) (D '(0) 1.0 1.1 1.2 ... * Total No. of person-years Total No. of employees Total No. of deaths 74,362.7 11,105 1,566 26,418.3 4,059 208 Rates were adjusted for age, sex, race, and calendar year using the total study population as the standard population; rates were not accbuntin 98 deaths, calculated when the observed number of deaths was less than five f For 704 subjects, a category could not be assigned because of missing data X IGDA, eighth revision codes shown in parentheses Ratio of rates of potentially exposed workers who ever smoked to those of nonexposed workers who ever smoked; statistical significance at pc.i indicated by a double asterisk ' II Ratio of rates of potentially exposed workers who never smoked to those ol nonexposed workers who never smoked 1) Number of deaths shown in parentheses i8 Mortality Study/Hanis t EXX-MOR-003358 three plant sites. Variations in mortality were observed by occupation and smoking history. This study could not completely discriminate between the effects of smoking and the possible contributory effect of occupation on mortality outcome. However, substantially higher death rates among smokers as compared with nonsmokers per sisted regardless of occupational classification and, in deed, contributed more to observed differences than did occupation. References 1. Hanis NM, Shallenberger LG, Oanafeslci DL, et al: A retrospect five mortality study o( workers in three major U.S. refineries and chemical plants, Part I. Comparisons with U.S. population. J Occup tiled 27:283-292,1985. 2. Colton T: Statistics in Medicine. Boston: Little Brown & Co., 1974, pp 47-49. 3. Monson RR: Analysis of relative survival and proportional mor tality. Computers Btomcd Res 7:325-332,1974. 4. Mantel N, Haenszel W: Statistical aspecLs of the analysis of data from retrospective studies of disease. JNCt 22:719-748,1959. 5. Mantel N: Evaluation of survival data and two new rank order statistics arising from its consideration. Cancer Chemother Rep 50:163 170, 1966. 6. Stagg V, Eng-Hui Can, Kwarciak L: Computer Program for Ap plication of Mantel-Haensze! Testwith Confotrnders, version 3.1. Uni versity of Western Ontario, 19B2. 7. Mason T{, McKay FW, Hoover R, et al: Atlas of Cancer Mortality for U.S. Counties: 1950-1969. U.S, Dept. Health, Education, and Wel fare, 1975. ' 8. MacMahon B, Pugh TF: Epidemiology Principles and Methods. Boston: Little Brown & Co., 263,1970. 9. Thomas TL, Decoufle P, Moure-Eraso R: Mortality among work ers employed in petroleum refining and petrochemical plants, J Oc cup Med 22:97-103,1980. 10. Covey LS, Wynder EL: Smoking habits and occupational status. I Occup Med 23-.537-54Z, 1981. Broker for Medical Services . Those who see occupational medicine in the vanguard of the new national strategy of health promotion and disease prevention worry that industry-sponsored health care is a diversion that may duplicate and undercut the existing health infrastructure. Others doubt that companies can do an adequate job of employee health protection - their unique responsibility -- when they dilute this mission with primary care services gen erally available in the community. Most companies currently stand on a middle ground; their medical departments serve a brokering function. They maintain collegial ties with a far-flung outside medical network, screening employees and referring them for needed care, and following up afterwards. They informally monitor quality of referred care by consultations and by watching the forms that come through from outside practitioners for disability and other insurance claims. As this brokerage function evolves, a company medical department can be more aggressive in managing situations that could compromise health. The company's atti tude is communicated to outside medical providers and the medical department be comes more activist and anticipatory -- thus, more strategic -- in its outlook. y--From "is There a Doctor in-House?" by prana C. Wafsh in Hak'ard Business Review- /uly/August, 1984. Journal of Occupational Madicine/Vol. 27. No, 5/May 1985 369 EXX-MOR-003359