Document qwREzxnd95O1vR4m0dGKeO2G

A STATISTICAL EVALUATION OF RESPIRATORY SYSTEM INDICATORS AMONG PIPE COVERERS_____________ 10-21-70 MGOtt ST0006508 INTRODUCTION This report presents statistical findings developed from an epidemiologic std pipe coverers. The report does not include a review of literature on the subject or a medical interpretation of the results. In the location under study, the job of pipe coverer has existed since 1910. A list of 173 men who h2ve held this job was prepared from company personnel records. Virtually all men who worked on the job for at least a year are included in the list, as well as most of the men who worked less than a year. As of January 1970, the status of the 17 3 men was as follows: Still Working For Company 130 Active Pipe Coverers No Longer Pipe Coverers 94 36 Quit or Released FromCompany 18 Retired Deceased ~h2- | 1 *-4-3)4 Within the identified exposed-to-risk population, the relationship between time-cn- job and health indicators reflective of the well-being of respiratory function was investigated. The health indices examined in this study include (1) the mortality rate for cancer of the respiratory system (based on company insurance records, 1940- 1970; (2) the incidence of morbidity episodes involving diseases of the respir atory system (based on company insurance records, September 1969-August 1970); and, (3) vitalometer measurements (derived from a multiphasic health screening program, 1967- 1970). METHODOLOGY Selection Of Exnosed-To-Risk Population And Controls Pipe coverers were identified through personnel lists prepared on a monthly basis by the company. The personnel lists are ordered by department and job classifi cation within department and are retained as far back as the 1920's. However, only the January listings are retained beyond ten years. Since pipe coverers tend to remain on the job for a long period of time, this method is adequate for identifying the exposed-to-risk population. Two sets of controls were utilized in this study. The first set of controls consisted of the total plant population and the second consisted of pipe fitters. While the total ST0006509 c 2- - c plant population wii selected- for jconyenipncej-a _proba_bLlity sample of the pipe fitter population'was Trele'cle'd;'6'tf'2airse''the3'e'~meli'had jobs closely related to the pipe coverers except for exposure to^ asbe_sto_s. The pipe fitters were identified by the same method as was used to identify the"pipe coverers. Because there are many more pipe fitters than coverers, a thirty percent sample was drawn by terminal digit sampling of company identification numbers. Method of Morbidity Analysis For the purpose of this study] a morbidity episode is an absence characterized by the filing-of a proper accident and sickness benefits claim covering a continuous period of absence from work""'ThVciaim includes a physician's certification of i . tii'. -i ' . vr ; .- n. ' .. the medical cause of the absence. In most cases, the morbidity episode must extend beyond seven calendar days before a record of the event is created. The program, as such, is available to hourly employees of the company. The expected number of disability days shown in Tables 1 and 2 are based on a five year age grouping of data on over 5, 000 hourly employees. Method of Vitalometer Analysis The company's multiphasic screening program, offered on a voluntary basis, calls for F.Z. V. Total and F.E.V. 1 Sec measurements to be recorded for each partici pant. A Collins Vitalometer produces the tracings which are digitized by a tech nician. For analysis purposes, the most positive of three repeated tracings for each individual is computerized. A regression equation was developed from the total screened population to adjust for height and age. Monthly comparison of measured results has shown that, over the three years of the program, consistent readings have been obtained. Using data generated through a multiphasic screening program has the advantage that neither the subject nor the technician is aware of uses to which the results might be put. Participation in the health inventory program was about bT'vc during the first cycle of the .program.__Hy.combining first and second cycles, %'italometer readings were obtained for 101 men in the'pipe coverer population. Several selection factors which could bias somewhat the findings of this study should be noted. First, it was not feasible to follow up on the 18 employees who left the company, two of th-ese men-were known to have left because of poor health. Secondly, a morbidity study covering a period two years following the first cycle of the multiphasic screening program revealed that participants in the program experience fewer illnes s - absence episodes than non-participants. The study did not enable us to determine the reasons for this difference. K, ~ ,-r. While the same selection factors apply to the control populations, caution must be axerctsed in-interpreting-results because of ihese unknowns. . - Method'of Mortality Analysis. -i = ... The mortality analysis is. based on copies of death certificates received by the com pany Croup Insurance Department- proof.of death. Asrmentioned previously, individuals releaj-ed from thevcumpany. -excluding retirees, will be lost to a mor- tality study based-cn.insur&ncc records. The two control groups referred to in Table 6.were studied over a more, recent time period than the pipe coyerers, even though the first, twenty .year s. folio wing, initial. place mention, the. jcrb was not.accumu lated as part sof. the-, expo-sure-toT-.risk..years.foxupipe coverers- . HoweY.tr* .since the mortality rate.due to-caaceroof the .respiratory-system-has-been increasing.over the years, this comparison should error slightly on the conservative side. The mortality' rates dor the. three groups were calculated on an age-adjusted basis. Smoking History : -.: A smoking history collected as part of the multiphasic screening program was not j_udged as satisfactory for. use in this study. The set of questions concerning smoking history has- been restructured and appears to be producing results that can be useful in.the future. The .lack-of.smoking history does impose considerable limitations on the conclusions that can be drawn- from .the study.. - - - ST0006590 RESULTS Table 1 shows that while no pipe fitter employment groups differ from the total plant in disability days due to respiratory disease, there are differences among the pipe coverers. The largest difference occurs among pipe coverers with 10 years or more of continuous employment on the job. These differences, however, are not fully reflected in total disability days. In Table 2, the number of episodes involving respiratory disease are compared. Based on a poisson approximation test, significant differences between observed and expected episodes were found for three employment groups among the pipe coverers and for no pipe fitter groups. However, the significant differences in the continuously employed group might be expected to carry over to the total group. Vitalometcr Analysis The readings shown in Table 3 are height and age corrected. Only after ten years of employment do significant differences between corresponding groups of pipe coverers and fitters appear. The prior-to-employment group consists of men who worked at other jobs during the first cycle of the multiphasic screening pro gram and later became coverers or fitters. A non-linear regression equation STOOOG59J c -4- c was developed for F.E.V. Total and F.E.V. 1 Sec using years since first exposure as a pipe coverer for the independent variable. The regression lines, along with the'scatter "ploT of individual readings, are shown in Figure I through Figure IV for both fitters and coverers, . .. Mortality Analysis ... ..... All causes of death among.pipe coverers up to. 1.9.70 ax.e. listed in Table 5. These deaths are categorized by interval between .initial employment as a pipe coverer and death, and by length.of.employment as a pipe coverer. An excess of total deaths.and respiratory cancer deaths among pipe coverers, as compared to two other groups, is indicated in Table 6v..Tbe numb.er._of observations is quite small. In addition, two of the men_ dying from respiratory cancer also worked with lead arsenate. SUMMARY The data presented in this study were derived only from ongoing routine functions of the company Medical Department. The investigation of vitalometer readings, morbidity episodes due to respiratory disease, and mortality due to respiratory cancer, indicated that pipe coverers performed less well than did the comparison groups. There are two important limitations of the study aside from the sine of the exposedto-risk population. First, no smoking history.was obtained. Second, follow-back of individuals no longer with the company was not undertaken. c> , V*.- 1'.. c c 9-10-71 y.coct STATISTICAL EVALUATION OF RESPLRAIORY SYSTEM INDICATORS AMONG INSULATORS ST 0006599 Introduction This report presents statistical findings developed froa an epideniclogic study c: insulators enployed by the Texas Division of The Dow Chemical Company. The report does not include a literature review or a medical interpretation of the results. These crucial elements of a study are properly left to the industrial physician responsible for the health of these workers. The objectives of the statistical evaluation were: (1) to identify through per sonnel records all employees who worked as insulators at the Texas Division; (2) to determine the current status of these workers and the dates of their em ployment as insulators; and, (3) to examine the relationship between exposure duration and several respiratory system indicators, namely the number of morbidity episodes due to diseases of the respiratory system, the number of deaths due to respiratory cancer and two respiratory function measures, the forced midexpiratory flow C25--752^ a"<f the forced expiratory volume at three seconds (tEVj). The insulator craft has existed since 19H at the Texas Division. Between 1511 and 1571, seventy-two men who worked as insulators were identified through per sonnel lists. As of August 1971, the status of these workers was as follows: Total------------------------------------------------------ 72 Still Working For Texas Division Active Insulators ----- --------- 29 Not Active As Insulators - 1 30 Released From Company-------------------- 29 Retired ------------------------------------------------ 9 Transferred From Texas Div. -------- 2 Deceased ---------------------------------------------- 2 Two factors which pose limitations on the statistical analysis are evident from the above tabulation. These factors are: (1) the rather small site of the exposed- to risk population (71 men), and, (2) the large percent cf these employees released fr~n urnolo--lent (10*). ST0006G00 c c-2- Methodology Seniority lists for the insulator craft which were available through the Personnel Department were reviewed to identify the population-at-risk. Since insulators participated in the periodic health exam program, a double check of the population was possible. The verification procedure revealed that the seniority lists had provided complete and accurate identification of the population. The work history card for each in-scope employee was examined to determine the date of birth, the current status of the employee, and the dates during which the employee worked as an insulator. Years-cn-job was subsequently used as an indicator of relative expo sure levels. Morbidity Data Morbidity episodes are recorded in the enployee's nedical record. The date of the episode and the physician's diagnosis are listed for each event. Extracting this information was cost tine consuming because of the need to pull individual records and the unedited nature of the data. Since no prior use of these data for statistical purposes is known to the author, it was necessary to assume that the completeness of reporting and quality of the information was adequate. Morbidity data for the exposed-co-risk population were abstracted beginning in I960. The number of episodes by cause and the frequency of episodes per 100 vork-years-at-risk were tabulated for the total insulator population. A matched pairs design was used to compare the mor bidity experience of the insulators with that of a suitable control group. In this case, the control group was selected from Power Department employees. Resources were conserved by restricting the comparison to the 24 insulators who worked con tinuously from 1960 through July 1971 versus 24 Power Department employees of the same age. These employees also worked continuously during the observation period. The pairing procedure eliminated the need to take into account the age of the worker or the period of observation in comparisons between the two groups. Until such time that routine data processing of morbidity data is initiated, che matched pairs type :f analysis ;-.H remain the most efficient way to statistically treat morbidity data. rC -3- ST000G60 I Scircaetric Measureser.es Spiroaetric neasurenents for insulators were provided through the periodic health exaa prograa. The cost recent testing for each insulacor was selected for analysis. Readings were available for 42 insulators. Two ceasures, the Fir25-75" c^e FEVj, were selected for analysis. The age and height, as well as scokir.g history, were obtained for each insulator at the tiae of his exaa. Because of several limitations, three categories--nonsaoker, current seekers of at least one pack per day, and all others,--were used to describe the saoking history for each worker. The objective of this phase of the study was to deteraine if the two respiratory function ceasures, corrected for age, height and saoking history', were related to years-en-jeb. Age and height corrections for the FEVj were based on "Guides To .. The Evaluation of Feraar.ent Inpaimenr - The Respiratory Systea" (1965). A factor c .95 was used to adjust FVC readings to FEV-j readings. Age and height corrections for the FEFjs.pj- were obtained froa "Spircaetric Standards For Healthy Ncnsaokir.g Adults" by Morris, et al (1971). The FEF25_75 readings were dichctcaized according to the sign or the deviations froa predicted corns to siaplify the analysis. Mortality Data The corraliry analysis was based on copies of death certificates received by the Conpany Group Insurance Departaent as proof of death. Individuals released free the Cocpany, excluding retirees, are lost to any cortality study based strictly cr. insurance records. The potential biases introduced by this loss of ir.foraation have been discussed in the literature. The control group for the cortality cccparisens consisted of a cohort of 6500 Midland location hourly workers followed between 1954 and 1969. The expected nuabers of deaths were calculated based on nortaiiry rates for the cohort population, adjusting for age by the indirect method. c c-4- STC006G02 Results Morbidity Analysis The overall morbidity experience for insulators from I960 to August 1971 is sum marized in Table 1. As expected, episodes related to diseases of the respiratory system account for over forty percent of the total absences due to sickness. Table.2 shows a comparison of morbidity experience between insulators and Power Department employees. The between individual range in the number of total episodes (0 to 20) indicates the vide variability in utilization of group insurance benefits. There are no statistically significant differences at the .05 level between the two groups. This is true both with respect to total episodes and vith respect to episodes due to diseases of the respiratory system. Srircnetric Analysis Table 3 shows the average deviations of FZV3 readings from predicted norms by years of exposure. N'o significant association at the .05 level was found between years of exposure and FIVj. Adjustments for smoking history had little effect on the re sults. The scatter plot of individual FEV^ deviations from predicted versus yearsor.-job is shown in Figure 1. In contrast, Table 5 indicates that nonsmokers perform near predicted norms and about .4 liters better than the average of either of the two smoking groups. Similar results are obtained for the percent of FEF25--7571 readings below predicted norms. Thus, smoking history was found to be related to respiratory function performance while years on the job was not. The possible effects on these readings of selection biases, such as population attrition, are unknown. Mortality Analysis Table 6 summarizes the mortality findings of the study. The observed deaths among insulators is less than expected in each category of comparison. However, since the expcsea-to-risk population is relatively small and the period of observation is short, the favorable experience to date does not, in itself, insure similar results the future. cC -5- Surmarv The statistical evaluation was carried out along the lines set forth in the Introduction of this report. While the necessary data were indeed available, they were soaetiaes a burden to gather. This was especially true when they had been aicrofilaed. With regard to the insulators, no evidence was found to indicate an association between years on the job and the respiratory systea indicators which were investigated. The nany qualification stateaents which appeared in the report attest to the fact that the data were collected under real world and not laboratory conditions. ST000G603 V cc Table 1. Episodes Of Morbidity Observed Since 1960 Aaor.s 46 Insulators Actively Eaployed As Of 1960. ST000G60l Cause All Causes Diseases of the Respiratory Systea Accidental Injuries Diseases of the Digestive Systea Diseases of the Circulatory Systea Diseases of the Ger.ito-Urinary Systea Diseases of the Nervous Systea and Sense Organs Infective & Parasitic Diseases Diseases of the Musculoskeletal Systea & Connective Tissue Necplasas - Total Endocrine, Nutritional & Metabolic Diseases Mental Disorders Diseases of the Skin & Subcutaneous Tissue Ail Other Causes Nuaber Of Itisodes 166 70 20 18 14 14 8 7 7 4 -- -- -- 4 Nuaber Of Episodes Per 100 Person Years At Risk 39.1 16.5 4.7 4.2 3.3 3.3 1.9 1.7 1.7 #0 -- -- -- ,9 cc Table 2. Comparison Of Morbidity Episodes Occurring Between January 1960 And July 1971 For 24 Insulators Paired With 24 Power Department Eaployees Based On Age. ST0006605 Pair Number 2 ' `3 4 c 6 t 8 0 .Total Episodes____________________ Resoiratorv Episodes Insulators Power Deot.Entl. Insulators Power Deot.EnDl. ...2. ___ 17___ _____ o__ 10 '0 i i8 7 0 04 50 15 I3 6 4 01 00 3 20 28 62 31 25 5 3 53 !0 L 0 .4 4 312 6 20 30 20 13 2 5 10 14 2 0 15 1 1 1c 3 4 10 00 1l 17 6 16 10 19 11 20 !i 6 -L, 8 22 | 3 23 1 6 l1 24 U Total Nunoer o: Epi- 110 40.0 12 7 5 1 i2 5 2 6 i 123 1 uu .1 Pish 18 7i 2 2 3 5; i 1 * 1 i \ 3i 1| 1 3 1 i 49 1/. b lc5. 1** cc Table 3. Mean FEY, Readings Expressed As Deviations Iron Predicted Norns By Exposure Group, Adjusted And Unadjusted For Snoking History. Exoosure Grouo < 10 Years 1 10 - 19 Years 20 - 29 Years Average ___ Exposure Years _Nuaber of Men 3.1 8 16.1 17 24.5 17 'FEV3 In Adjusted For Snokine Historv Liters Unadjustec For Snokine Histcrv -.13 -.13 -.34 CO 1 -.19 -.14 9 0 9 (J 0 0 0 1 S' Table 4. Percent Of FZT25--7SZ R^a^l^gs Belov Predicted Nom By Exposure Group, Adjusted And Unadjusted For Snoking Eistory. Exoosure Grouo < 10 Years 10 - 19 Years 20 - 29 Years Average Exposure Years 3.1 16.1 24.5 Nunber of Ms n 8 17 17 I Of FE?2t_75T Belov Predicted Ncm Adjusted For Snokine Histcrv Unadjusted For Srokir.c Histcrv 38 38 42 47 38 35 Table 5. Mean FEV3 Readings Expressed As Deviations Fron BelovPredicted Korns And Percent Of 1^^25-75% Readings Predicted Norn By Snoking Category. Snokine Cateeorv Nonsnokers Current Snokers Of At Least One Pack Per Day All Other Nunber of Men 10 13 19 Deviations Of FEV^ In Liters Fron Pre dicted Noms + .09 I Of rEF^s_7 52 Belov Preoicted Notts 10 -.32 -.31 46 53 k. cc Table 6. Observed And Expected Deaths Anong Insulators By Interval Since First Exposure . Interval Since First Exposure To cal Less Than 20 Years More Thar. 20 Years Years Exposed' -- --------- To I a"I "Deaths ......... To Risk Observed Expected Deaths Due To Respiratory Cancer Observed Expected 942.1 2 4.4 806.2 2 2.7 135.9 0 1.7 0 .3 0 .2 0 ,! 09'J0001S ST0006608 cc FIGURE I. SCAT"IT. PLOT OF DEVIATIONS IN FEY3, FROM PREDICTED VERSUS YEARS ON JOB FOR 42 INSULATORS (ADJUSTED FOR SMOKING HISTORY) T t > i 0 1 G_ I'+ 1 I"! YEARS ON JOB OLVIATIONTi OF FEV, FROM PREDICTF.D cc FIGURE I. SCATTER PLOT OF DEVIATIONS IN FEV3 FROK PREDICTED VERSUS YEARS ON JOB FOR 42 INSULATORS (ADJUSTED FOR SHOEING HISTORY) ST0006609 D E V IA TIO N S OF FF.V, FROH PREDICTED b S 10 1C- 14- lb tfii.-pt-iMrt--:! YEARS ON JOB -S * fc- tta llt 1+