Document rw8pK46a1MaZ6je33j2ov65e

American Journal of Industrial Medicine 28:49-70 (1995) Assessment of Mortality in the Construction industry in the United States, 1984--1986 C. Robinson, PhD, F. Stem, ms, W. Halperin, mo, mph, H. Venable, ms, M. Petersen, PhD, T. Frazier, ScM, C. Burnett, pis, N. Lailich, msph, J. Salg, PhD, J. Sestito, JD, and M. Fingerhut, PhD Construction, one of the larger industries in die United States, employs 7.6 million workers, many in skilled trades occupations. Previously published data about potential, worksite exposures and mortality of construction ate workers are limited. We analyzed occupation and industry codes on death certificates from 19 U.S. states to evaluate . .. mortality risks among men and women usually employed in construction occupations. Proportionate mortality ratios (PMRs) for cancer and several other chronic diseases were . significantly elevated among 61,682 white male construction workers wbo died between 1984 and 1986. Men younger than age 65, who were probabjy still employed immedi ately prior to death, bad significantly elevated PMRs for cancer, asbestos-related dis eases, mental disorders, alcohol-related disease, digestive diseases, foils, poisonings, traumatic fatalities that are usually work-related, and homicides. Elevated PMRs for . many of the same causes- were observed to a lesser degree for blade men and white . women whose usual industry was construction. In addition, women experienced excess ' cancer of the connective tissue and suicide mortality. Various skilled construction trades had elevated PMRs for specific sites, such.as bone cancer and melanoma in brickma- ' sons, stomach cancer in roofers and brickmasQos, kidney and bone cancer in concrete/ tenazzo finishers, nasal cancer in plumbers,-pulmonary tuberculosis in laborers, scrotal cancer and aplastic anemia in electricians, acute myeloid leukemia in boilermakers, rectal cancer and multiple sclerosis in electrical power installers, and lung cancer in structural metal workers. Using a standard population of bine collar workers did not result in fewer elevated PMRs for construction workers. Despite lifestyle differences and other limitations of the study, the large numbers of excess deaths observed in this study indicate the need for preventive action for construction workers, o I995waey-lis*, Inc.* Key words: mortality survdDance, occupational health, occupational <n--construction indns- try, skffied trades occupations, carpenters, electrical workers, sflkotb, asbestods, ' bine collar workers . ! f i Division of Surveillance, Hazard Evaluations and Held Studies Surveillance Branch (C.R., W.H..H.V., T.F., C.B., N.L., J.S.. J.S.). Industry-Wide Studies (F.S.. M.F.), and Support Services (M.P.), National Institute for Occupational Safety and Health, Cincinnati,.OH. - Address reprint requests to C. Robinson, Division of Surveillance, Hazard Evaluations and-Field Studies, MSR-18 NIOSH, 4676 Columbia Parkway, Cincinnati, OH 45226. Accepted for publication March 27, 1994. 1995 WUey-Liss, Inc. *This article is a US Government work and, as such, is in the public domain in the United States of America. SC-ELEC-10900 50 Robinson et al. . INTRODUCTION ' Construction is one of the larger industries in the United States, employing 7.6 million workers in 1992 [U.S. Census, 1992], yet published studies about potential worksite exposures and health of construction workers are limited. This may be due partially to the difficulty of studying construction workers, because of the lack of - centralized personnel and medical records. Also, construction workers change jobs frequently because worksites are temporary and change constantly as the work moves ahead. Both skilled trades people and unskilled-workers may be employed on con struction worksites, involved in the construction, renovation, repair, or demolition of ' residential or industrial buildings, bridges, toads, ships, or tunnels. Each occupation or special trade may have unique health ami safety risks specific to its tasks and work practices. In addition, several trades people and unskilled workers may share a work area, exposing each other as "bystanders" to toxic substances. Thus construction worksites present special problems in the study, identification, and control of health hazards. .- In the tradition of William Farr, founder of modern surveillance. [Langmuir 1976], we have characterized the modality risks of construction workers using death certificates coded for usual occupation. We undertook the. present analysis because ffeath certificates coded by occupation and industry for the last decade from many U.S. states recently became available. This permitted evaluation of rare causes of death for recent years for skilled trades and construction laborers. We reviewed previously published reports in the context of our study results. Our goals were to provide estimates of the mortality experience of construction workers, to form hy potheses for further research, and to identify and prevent hazardous occupational exposures. '- PREVIOUS REPORTS Construction workers in building, highway, and special trades have been re ported previously to experience excess mortality from cancer and from circulatory, respiratory, nervous, and digestive system disease and fatalities (Registrar 1978, 1986; Milham 1983; COMS, 1987). The Registrar General's Decennial Supplement on occupational mortality 1970-72 for England and Wales [Registrar, 1978] reported statistically significant excess deaths due to cancer of the esophagus, rectum, larynx, and lung for all construction workers. In 1986, significant excess risk was repotted fin- cancer of the esophagus, rectum, and lung and for falls [Registrar, 1986]. In the United States, a proportionate mortality ratio (PMR) study of deaths occuring during the years 1930-1979 in Washington State [Milham, 1983] found that men working as construction laborers had significant excess mortality due to respiratory tuberculosis; cancer of larynx, lung, Madder, and connective tissue; disorders of character, be havior, and intelligence; other, respiratory disease; accidental falls; cirrhosis of. the liver; and homicide. California construction industry workers who died from 1979 through 1981 were reported to have excess deaths due to cancer of lip,, oral cavity, pharynx, lung/ urinary organs, and other digestive Organs, ischemic heart.idisease, stroke, cirrhosis, chronic obstructive lung disease, falls, and suicide [COMS, 1987]. Swedish construction workers who participated in the Bygghalsan program of regular voluntary medical exams were found to have significant excess cancer of the lung, larynx, and gastrointestinal tract [England, 1981; Fletcher, 1990]. The fatality rate * Mortality Among Construction Trades Workers 51 for the U.S. construction industry has been reported from the National Traumatic Occupational Fatality surveillance project to be 25.61 per 100,000 workers far 1980 1989 [NIOSH, 1993], ranking construction as the second most hazardous industry in terms of traumatic fatalities. . ' In addition to the raised PMRs reported above for all construction industry workers, differentials in mortality risk have been reported for various skilled trades within the construction industry. Table I. compares PMRs observed for selected skilled trades by two of the largo' previous reports, England and Wales 1979-1983 [Registrar, 1986] and Washington State 1970-1979 [Milham, 1983). Only PMRs that were significantly elevated in at least one of the two reports are shown. Most striking were the similarly elevated PMRs across many trades for excess mortality due to all cancer, respiratory cancer, mental disorders, liver disease and cirrhosis, respiratory disease mortality, and traumatic fatalities. . .. MATERIALS AND METHODS ~ Data Sources We analyzed occupation and industry-coded U.S. death certificate data for the years 1984-1986. The data were, pooled from 19 states in the United States, which together with the National Institute for Occupational Safety and Health (NIOSH), National Centerfor Health Statistics, and National Cancer Institute have shared the added costs of. coding occupation and industry on their death certificates. The 19 states are Colorado, Georgia,.Indiana, Kansas, Kentucky, Maine, Missouri, Ne braska, Nevada, New Hampshire, New Mexico, Ohio, Oklahoma, Rhode Island, South Carolina, Tennessee, Utah, Vermont, and Wisconsin. About 6% of die U.S. employed population worked in the construction industry, and most of the 19 states in our study were representative of the overall United States. Only one state differed by more than 2%; this was New Mexico, where 9% were employed in the construc tion industry [U.S. Census, 1982]. . Identification of Construction Industry Workers and the Trades iDeatb certificates for decedents age 20 and over who both resided and died in . one of the 19 states were the source of age, sex, race, usual occupation and industry, and underlying cause^of-death. Occupation and industry entires on 876,731 death certificates (560,370 white men, 69,564 black men, 214,647 white women, and 32,150 blade women) were provided by the states to NIOSH. Housewives, the unemployed,.part-time workers, and unknown occupations were excluded from the analysis. Coding of usual occupation and industry stated on the.death certificate by . next-of-kin was done by clerks trained in procedures developed by the Bureau of the Census for NIOSH and NCHS [NCHS, 1987]. Table II shows the occupation codes used to select the skilled trades, the laborers, and the code for the construction industry [U.S. Census, 1980]. - .' The skilled construction trades included painters, plumbers, insulators, brick- masons, carpenters, operating engineers, elevator installers, electricians, roofers, * structural metal workers, concrete and terrazzo finishers, electrical power installers and repairers, operating engineers, tile setters, carpet installers, plasterers, boiler makers, and drywall installers. The unskilled trades included construction laborers (Table II). . The mortality experience of construction workers was compared first with all 52 Robinson et al. TABLE I. Comparison of Construction Trade Mortality Reports for England and Wales, 1979-1983, and Washington State, 1970-1979___________ '___________________ England and Wales Washington State Causes of death* '- - No. deaths PMR* No. deaths PMRb Carpenters (554, 561, 569*) (411*) ' Malignant neoplasms (MN) (140-208) MN nose, middle ear, sinuses (160) MN pleura (163) MN prostate (185) MN lymphatic and hematopoietic tissue - (200-208) Mental disorders (290-319) Chronic obstructive pulmonary disease and. - allied conditions (490-496) . . Diseases of esophagus (530) Fatalities due to falls (880-888, 9293) . 3771 7 47 219 250 59 457 14 96 105f 136 361* 121r 115* 137* 85 186* 165r 1070 .6 2 139 91 4 280 NR* 73 101 296* 85 108 93 241 116* -- 149* Electricians (555. 575, 576?) (421") Malignant neoplasms (MN) (140-208) MN stomach (151) : .. MN pleura (163) Acute myeloid leukemia (2050) 2385 229 38 . 34 105* 115* 406f 149" 428 17 2 .4 104 99 154 Painters, paperfaangets (556, 579, 583*) (694*) Malignant neoplasms (MN) (140-208) MN trachea, bronchus, and lung (162) Mental disorders (290-319) Chronic obstructive pulmonary disease and . allied conditions (490--496) Chronic Gver disease and cirrhosis (571,5721-5728) . External causes of injury and poisoning (800-999) ' Railway accidents (800-807) - '' ' Unintentional poisonings (850-869,9292) Unintentional poisonings by drugs, medicaments and biological* (850-858) .1 . Fatalities due to falls (880-888, 9293) Suicide and self inflicted injury (950-^959) 3706 . 1754 60 606 58 664 8 . 46 35 94 218 106f 123r 139* 115f 73* 111* 50* 198f 273f . 163r 127* 405 150 9 116 53 140 1 NR NR 19 34 106 118* 113 138* 152* 91 - 117 93 Plasterer* (584) (505*) . Malignant nepplasma (MN) (140-208) 744 115* MN tradjea, bronchus, auditing (162) . 364 138* Chronic obstructive pulmonary disease and allied conditions (490-496) 155 158* 58 120 24 161' 18 173* Plumbers, pipefitters (557, 585. 587) (51(7*) Malignant neoplasms (MN) (140-208) MN esophagus. (150) MN trachea, bronchus, and lung (162) ' MN pleura (163) ' ' Chronic obstructive pulmonary disease and allied conditions (490-496) ' ' 2079 28 908 32 298 112* 127 121* 428* 111 383 16 134 T- 118* 235* - 122* " -* -- 93 132* (continued) Mortality Among Construction Trades Workers 53 TABLE L (Continued) Causes of death* Glorias, roofers (589, 595*) (514*) Malignant neoplasms (MN) (140-208) . MN trachea, bronchus, and lung (162) Chronic liver disease and cirrhosis (571, 5721-5728) External causes of injury and poisonings (800-999) ' fatalities due to fans (880-888,9293) England and Wales Washington State ' No. deaths PMRb No. deaths PMRb 456 111* 221 139* 10 . 83 217 145/ 62 611' 66 35 19 57 4. 102 157' 225' 103 127 Laborers, construction (869) (982*) . Pulmonary tuberculosis (011) 22 MN trachea, bronchus, and lung (162) ' 958 MN penis and other male genital .organs (187) 8 Mental disorders (290-319) .. , 46 Cerebrovascular disease (430-438) 604 Pneumonia and influenza (480-487) 346 Chronic obstructive pulmonary disease and allied conditiocs (490--496) 346 Peptic ulcer (531-534) . . - 69 Chronic liver disease and cirrhosis (571, 5721--5728) ' . . 48 External causes ofinjury and poisoning ' . (800-999) 487 Motor vehicle fatalities (810-825, 9290) 112 Fatalities due to falls (880-888, 9293) 66 Injury undetermined whether accidentally or 1 purposely (980-989) 65 152 122* 259*. 1651 US' 133f 2 . 151 NR 16 NR 41 127 116 -- 164* .-- 92 125f 102 I43f . 6 128* 123 92 121. 148' 105 309 . 118' 78 116 122 173' .32 173' 151' NR -- 'International Classification ofDiseases, 9th rev. Geneva; WHO, 1978. - ''Proportionate mortality ratio. ' ' cOcntpation codes for l/nited Stater and England and Wales based on occupation codes from the US Bureau of Census, 1980 Census of Population, Alphabetical Index of.Industries and Occupation, Final ed., US Dept of Commerce, Bureau of the Census, 1981, publication PHC-80-R3. - ^Occupation codes for Washington State based on a modified version of that published by US Bureau of die Census, Classified Index of Occupations aid Industries, I960 Census of Population, Washington, DC, USPO, I960. "Statistlcally sigmficant, p < .OS; ' 'Statistically significant, p < .01; . NR not reported. workers. The analysis was repealed using a blue collar comparison population con sisting ofall workers Whose occupations were classified in the categories ofprecision crafts workers, operator, or laborer. The blue collar analysis was run to control for the known association of many of the outcomes with risk factors such as smoking and/or lifestyle of bine collar workers. - Proportionate Mortality Ratio Analysis .. ' Proportionate mortality ratio analysis based on the underlying cause of ddath was used to evaluate the mortality patterns of construction workers. The underlying causes of death were coded by state nosologists according to the International Clas- 54 Robinson et aL TABLE H. Index to Table and Codts Used to Select Construction Occupations for Analysis Occupation census code 060(industrycode) - 553.563-4 554, 567-69 556,579 557, 585.587 593 . 844 543 . 595 597 588 555, 575-6 577 869 565 566 573 596 . 643 . 584 . ' Table number - - Occupation tide - .' m Construction industry '' IV . Brictanasons, stonemasons, apprentices IV - Carpenters, apprentices ' - IV Painters, construction and maintenance IV Plumbers, pipefitters, steamfitten, apprentices IV Insulation workers IV Operating engineers ' IV - Elevator installer and repairers IV. Roofers IV Structural metal workers IV Concrete and terrazzo finishers . IV Electricians, apprentices IV Electrical power installers and repairers IV Construction laborers IV . THe setters, hard and soft - IV Carpet installers IV Drywall installers '' IV ' . Shcctmctal duct installers IV ' Boilermakers IV ' Plasterers . . sificadon of Disease, ninth revision [WHO, 1978]. PMRs were calculated using a computer program developed at NIOSH [Dubrow, 1986] that was designed to cal culate PMRs feu occupations or industries for population-based data. This program ralmiates proportionate mortality ratios (PMRs) by comparing the proportion of deaths from a specific cause within a specified occupation or industry group wifii the proportion ofdeaths due to that cause among all decedents. PMRs woe computed after stratification for age (ages 20-64; ages 6S+), race (black, white), and gender. PMRs were computed for all ages and for white males who died before age 65, since usual occupation would be reported more accuratelybefore retirement. A PMR above 100 is considered elevated over the average for all occupations. Exact 95% confi dence intervals for the true mean observed deaths were calculated by assuming that the observed deaths were distributed as a Poisson random variable. The endpoints of these confidence intervals were divided by the expected deaths to yield 95% confi dence intervals (shown in the tables) on. the true PMR [Bailar, 1964]. Statistical .significance (p<.05 for a two-sided'test) and 95% two-sided confidence intervals were determined, but should be evaluated in the context of hypothesis generation because multiple comparisons were made [Rothman, 1986]. PMRs are usually computed when data for the population at risk are not avail able and rates of death or SMRs cannot be calculated. The population at risk for this study includes all persons who ever worked in the construction industry in these states and were at risk of dying January 1,1984 through December 31, 1986. Because data were not available for the population of construction workers at risk-,' we evaluated proportionate mortality based on cumulative deaths over a 3-year period. Thus our PMRs indicate whether the proportion of deaths due to a specific cause appears to be high or low for a particular occupation. Mortality Among Construction Trades Workers 55 RESULTS The proportionate mortality for the years-1984 through 1986 of construction industry workers and skilled trades is summarized in Tables ID--IV. Because the . purpose of this analysis was to identify increased mortality, only statistically signif icant elevated PMRs, PMRs for ischemic heart disease, and the all-cancer PMRs are shown in most tables. However, for a few specific causes of interest, elevated PMRs. were listed. PMRs for all causes of death could not be listed due to space limitations.* Because our data did not indicate tire circumstances of the traumatic fatality, we . excluded most fatality PMRs from our report, except for falls, electrocutions; unin tentional poisonings, homicide, suicide, and fatalities classified as usually work- related (WHO, 1978]. . PMRs for Construction Industry--All Ages . Elevated PMRs.are shown in Table DI for 61,682 whitemen (WM), 9,379 black men (BM), 1,225 white women (WW), and 68 black women (BW), all of whom bad -been usualfy employed in the construction industry. .. - Construction workers experienced mortality deficits for ischemic heart disease . and infective and parasitic disease. Exceptions were blade male construction workers Who bad a significant excess ofpulmonary tuberculosis (Table ID) and Hade women whose deficit of heart disease was not statistically significant. Statistically dgnificant excesses occurred across a broad panorama of causes for all race^-gehder groups. They were cancer of the esophagus (WM, BM), cancer of tbclarynx (WM), cancer of the lung (WM, WW. BHiOi cancer of connective tissue (WW), cancer of foe bladder . (WW), nonmalignant respiratory disease (WM, WW), chronic liver disease and cirrhosis (WM), accidental poisonings (WM), falls (WM, BM), usually work-related . fatalities (WM), homicide (WM, BM), and a grouped category of alcohol-associated [Anon, 1992a] disorders (WM, BM).. . ' PMRs for Construction Industry--White Male3 Under Ago 65 / We evaluated PMRs for white construction industry workers who died before . age 65 (Table ID). These men experienced a deficit of ischemic heart disease deaths, . similar to that observed for men of all ages. Except for a new agnificant excess for cancer of the nasopharynx (PMR= 158, p<.QS, 95% 0(100--237), 23 deaths ob- . - served), most site-specific cancer PMRs remained about foe same-magnitude. How ever, tire PMRs far respiratory diseaseincreased substantially, particularly for asbestosis,. silicosis, and foe overall category of all pneumoconioses. White male construction workers who died before age 65 experienced a more than threefold increased risk for deathdue to asbestosis (PMR--393, p<.,01,95% 0(196--703), 11 rifatha observed) and for silicosis (PMR=327, p<.01, 95% 0(149--620),.9 deaths - observed). All PMRs for traumatic fatalities increased slightly. - PMRs by Skilled and Unskilled Trades--White Males . Table IV lists significantly elevated PMRs that occurred among skilled and unskilled trades. For brickmasons elevated risks occurred for cancer of the stomach,. larynx, and lung, as well as for mental disorders, nonmalignant respiratory disease, skin disease, homicide, ami alcohol-associated disease. Carpenters were observed to Date available on request to the authors. 56 Robinson et a!. TABLE m. Proportionate Mortality, Construction Industry, 19 P-S. States. 1984-1986* Cause of death* . - PMRC OW ' 95% CT WHITE MEN . '' AH deaths . Cancer of esophagus (ISO) . Cancer of larynx (161) Cancer of hmg (162) . Cancer ofpleura or peritoneum* (1588, 1599, 163) Cancer of penis (187) ' Mental disorders (290-319) Ischemic heart disease (410-414) NonmaHgnant respiratory disease (490-519) Asbestosis (SOI) - ' ' Chronic liver disease and dnhcrsis.(57i. 5721, . 5728) .V . Unintentional poisonings* (850-869,9292) ' FaDs (880-888,9293) . Fatalities usually work-related (846, 916-921, 923-927) Homicide (960-969) . . . Alcohol-associated diseases* (291, 303, 3050, 3575, 4225, 5353.7903,5710-5713, 8600, 8601) *. . . ' . ' 100 115* 127* I14r 141 150 121* 94* 119* 295* 112* 120* 137* 125* 142* 133* 61682 365 201 5944 33 23 639 15937 4436 24 003-127) (108-144) 011-116) (97-198) (95-226) 012-131) (93-95) (116-122) (189-439) 1078 . (106-118) 252 (106-136) 507 026-150) 362 . 720 (112-138) (132-153) 861 (124-142) WHITE MEN WHO WED UNDER AGE 65 All cancer (140-208) Cancer of nasopharynx (147). Cancer of larynx (161) Caocer of lung (162) Cancer of pleura, peritoneum? (1588, 1599, 163) Cancer of penis and other male genital organs (187) . Mental disorders (290-319) . ' . ' Ischemic heart disease (41&-414) Nonmalignant respiratory disease (490-519) Chrome obstructive lung disease (490-496) Pneumoconioses (500-505) , Asbestosis (501) 1 Silicosis (502) Digestive diseases (320-579) ~ Chronic Uver disease and cfarhods (571,5721, 5728) Unintentional poisonings (850-869, 9292) Falls (880-88, 9293) Fatalities usually work-related (846, 916-921, 923-927) Homicide (960-969) Alcohol imprinted diseases1 (291,303, 3050, 3575,4255, 5353,7903. 5710-13, 8600.8601) ' 98* 158* 136* 110* 131 . 129 131* 91* 117* 122* 182* 393* 327* 116* 118* 120? 177* 130* 144* 135* 5357 .23 94 2401 12 (96-100) (100-237) (110-166) (106-114) (68-229) 7 308 4757 889 733 22 11 9 1091 - (52-266) (116-146) (88-93) 009-125) 013-131) (114-275) (196-703) (149-62Q) (110-122) 738 (110-127) 226 (105-136) 254 (156-200) 325 (116-145) 690 (134-156) ^ - - 660 (125-145) BLACK MEN . All deaths Pulmonary tuberculosis (002) Cancer of esophagus (150) Cancer of lung (162) ' ' -r ' 100 9379 160* 32 (109-225) 131* 158 (112-153) 105 782 (98-112) (continued) Mortality Among Construction Trades Workers 57 TABLE IH. (Continued) Cause of death* . . . PMR' OW 95% cr BLACK MEN (continued) . Cancer of penis (187) . Secondary, ill-defined cancer (195-199) Ischemic heart disease (410--414) Mental distxdm (290-319) Fallt (880-888, 9293) Honridde (960-969) Alcohol-associated disrates1 (291,309, 3050. 3575,4225, 5353, 7903, 5710-5713, 8600. 8601) . 1 ' .. 130 iw' Vf 119* I74r 115' * U8f 5 194 1476 191 85 532 265 (42-304) (103-137) (86-94) (103-137) (139-215) (IQS--125) (105-134) WHITE WOMEN All deaths - " Cancer ofesophagus (150) '. Cancer of lung (162) Cancer of connective tissue (171) . Cancer of bladder (188) . Secondary, in-defined (195-199) ' Ischemic heart disease (410-414) Nonnulignant respiratory disease (490-519) Suicide (950-959) . .. . . 100 165 126* 374f . 248* 128 84> I63r . 124 . 1232 4 81 . 10 9 29 233 67 . 24 " _- (45-423) (100-157) (180-688) (113-471) . (86-IS4) (74-96) (126-207) - (79-185) BLACK WOMEN . ; : . All deaths ' . All cancer (140-208) Cancer of lung (162) Ischemic heart disease (410-^414) - Namnallgnanl respiratory disease (490-519) Homicide (960-969) . 100 130 . 272 . 70 183 144 68 20 (79-200) 6 . (100-593) 8 (30-137) .3 (38-536) 3 (30-421) 'Analyses based on occupatioa-codcd death certificate data for 1984-1986 provided to N105H by 19 U.S. states for evaluation and follow-up.Occupatinnal coding was supported by NIOSH, National Cancer Institute, and National Center for Health Statistics. . *Code of the International Classification of Diseases, 9th rev., 1978, inpsrentbeses. 'Proportionate mortality ratio (observed/expeettd mortality X 100). ^Observed. .' 'Exact 95% confidence intervals for the true mean observed deaths woe calculated by assuming that the observed dwnht were distributed as a Poisson random variable. The endpoints of these confidents intervals were divided by the expected deaths to yield 95% confidence intervals (shown in die table) on the trite PMR fBaflar JC, 1964], fp<0.01. *May indude mesotheliomas. . 'Way include lead potsonblg. ' '. - 'Carnes of death attributed by the .National Center far Health Statistics to alcohol-induced mortality include alcoholic psychoses, alodial dependence syndrome, sondependent abase of alcohol, alcoholic polyneuropathy, alcoholic cardiomyopathy, alcoholic gastritis, chronic liver disease, and cirrhosis spec ified as alcoholic, excessive blood level of alcohol and accidental poisoning by alcohol, not elsewhere classified. These causes exdude accidents, homicides, and other causes indirectly related to alcohol use _ (Anon, 1992a). V<0.03. . ' . ' ' -' . . ; 58 Robinson *8 al. ... . TABLE iy. Canse-Spedflc Proportionate Mortality by Construction Occupation, 19 HA 1984-1986, White Men*_____________ ._______________________________________ ' ' Cause af death6___________ _________________________ PMRC . Obs* 95% CP BHdanasons . AD deaths .. AD cancerCancer of stomach (111) Cancer of larynx (161) . Cancer of lung (162) . ' ' Cancer of bone (170) Melanoma (akin) (172) Cancer of bladder (188) - Mental rfisardeo'(290-319) ' Ischemic heart disease (410-414) ` Nonmalignant respiratory disease (490-319) Pneumoconioses (500-505) . Skin disease (680-709) Musculoskeletal disease . Homicide (960-969) Alcohol-associated diseases'1 (219. 303.3030.3373, 4255, 5353. 7903, 5710-13, 8600, 8601) 100 108 . 208' 213* vxf 249 130 ' 132' 187f 89f 133r 264 273*. . 165 17Sr 181* 2,351 594 32 13 240 3 11 19 38 590 191 6 7. 7 28 44 (100-118) (142-293) (113-365) (106-137) (51-727) (65-233) (80-207) (133-257) (83-97) (115-154) (97-574) (110-563) (66-339) (116-252) (132-244) Carpenters . . ,. .' -. All deaths. ' ' AB cancer Cancer of long (162) Cancer ofpleura and peritoneum* (1588, 1589, 163) Ischemic heart disease (410-414) NonmaBgnant respiratory disease (490-519) pans (800-888,9293) Homicide (960-969) Alcohol-associated diseases* (219, 303, 3050, 3575, 4255,5353.7903.5716-13.8600,8601) 100 102 3,731 (100-105) 116' 1,489 (111-121) 163 14 (89-274) 94f 4,302 (92-97) 124f 1,278 (118-130) 133f 135 (112-158) 145f 151 (123-17Q) 126f .176 (110-176) Construction and maintenance painters All deaths Viral hepatitis B . All cancer (140-208) Cancer of larynx (161) Cdnccr of long (162) ~' Cancer of bladder (188) Mental disorders (290.-319) . . . Ischemic heart disease (4I()-4l4) NbmnaSgnant respiratory disease (490--519) SIScosis (502) . Diseases of stomach, duodenum (531-537) ' Chronic Uver dbease and cinhosis (571,5721, 5728) Ails (880-888, 9293) Unintentional poisonings' (850-869, 9292) Homicide (960-969) .Alcohol-associated diseases* (291, 303, 3050, 3575. 4255, 5353,7903, 5710-13, 8600, 8601) 100 139* 103 165* 124f 137* 182* 88r 127* 449* 151* 123* 149* 153* 144* 175r 5,902 64 1J413 25 615 50 91 1,450 468 4. 33 . 109 52 26 58 102 (107-178) (98-107) (107-144) (114-134) (101-180) (146-223) (84-92) (116-139) (123-1151) (104-213) (101-143) (111-175) (100-225) (110-187) (143-212) (continued) Mortality Among Construction Trades Workers 59 TABLE IV. (Continued) Cause of death6 PMR* Obs" 95% CT Plumbers : All deaths' . All cancer (140-208) . Nasal cancer (1600, 1601, 1603-9) Cancer of lung .(162) . Cancer of pfean' (163) Ischemic bent disease (410--414) - - Nonmalignantrespiratory disease (490-319) Asbcstosis (301) : Unintentional poisonings* (850-869,9292) . 100 106* 287 - 110* 327* 98 113* 1097* ' 140 5,214 1324 2 508 5 1,462 378 8 . 17 (101-111) (35-1037) (101-120) (106-763) (94-102) (102-125) (474-2162) (82-225) Insulation workers ' .' All deaths AD cancer (140-208). Cancerof tong (162) Cancer of peritoneum1 (158) Cancer of pleura1 (163) . Ischemic heart disease (410-414) Nomnalignant respiratory disease (490-519) ; AsbestOsis (SOI) . ' ' . . . ' 100 130' 193' 13486' 2467' 67f 152* 23197' Operating engineers All deaths - . All cancer (140-208) Nasal cancer (1600, 1601, 1603-1609) Cancer of tang (162) Ischemic heart disease (410-414) . Nohmalignant respiratory disease (490--519) Fatalities usually work-related (846,916-921, 923-927) 100. ' 104 271 120* . 97 112 _ 257' 267 101 5i .4 2 48 22 .9 . (122-182) (157-229) (1989.-) (298-8913) (50-89) (102-202) *)-. 3731 981 -3 434 1022 255 41 " '` (97-110) . (56-792) (109-131) (92-102) (99-127) (184-348) Elevator installers and repairers ' All deaths ' All cancer (140-208) Cancer of long (162) - Ischemic heart disease (410-414) Pneumonia and influenza (480-487) . ' Non-malignanl respiratory disease (490-519) ' 100 122 203* 40' 248 189 . 63 19 (73-190) 12 (105-355) 7 (15-81) 4 - (68-635) 7 (76-390) Roofers AH deaths . Mental disorders (290-319) Ischemic heart disease.(410-414) Nonmalignant respiratory disease (490-519) Asbestosis (501) Chronic liver disease and cirrhosis (571, 5721-5728) Falls (880-88, 9293) Homicide (960-969) Alcohol associated diseases6 (291, 303, 3050, 3375, 4255, 5353, 7903, 5710-13, 8600, 8601) - 100 195' 74' 149' 1873* 168' 287' 190^ 272' 967 18 ' 165 67 2 30 19 36 (116-308) (63-86) <115-189) (227-6766) (114-240) (173-499) (133-263) 37 (191-374) (continued) 60 Robinson.et al. TABLE IV. (Continued) Cause of death* . -- Structural metalworkers . All deaths ' - . An cancer (140-208) Cancer of lung (162) Ischemic heart disease (410-414) Falls (880-88, 9293) . Fatalities usually work-related (846,. 916-921, 923-927) .. Homicide (960-969) . Coucrete/terrazzo finishers AH deaths . Cancer, of stomach (151) . Cancer of bone, cartilage (170) Kidney caacer(I89) . Ischemic heart disease (410-414) . . Electricians . .. AH deaths '- An cancer (140-208) Cancer of long (162) . Cancer of peritoneum1 (158) '\ . Cancer of pleura (mesothelioma) (163) Cancer of penis (187) Cancer of scrotum (1877,-1879) . Aplastic anemia (284) Ischemic heart disease (410^-414) Nonmalignant respiratory disease (490-319) Asbestesb (SOI) . Disease of pancreas (377) - Fatalities usually work-related (846, 916-921, 923-927) Electrocutions (925) . PMRF W 96 129* . 97 344* 310* 167* 100 . - .179 959f . 303* 94 100 104* 114* . 223 331* . 230 1056* . 235* 94* 110* 349 178 . 215* 653*. Ohs* ' 95* CP 1064 245 122 281 21 17 16 (85-109) (107-154) (86-109) (213-526) (181-497) (105-228) 516 6 (66-389) 3 (198-2801) 9 (139-576) 130 . (79-111) 6345 1570 626 8; 6 . 4. 2 8 1678 436 3 15 (100-109) (105-123) (97-443) (122-719) (68-641) (128-3813) (101-463) (90-98) (101-119) (72-1019) (100-293) - 55 (162-280) 33 (450-917) Electrical power installers and repairers , Alldeaths AU cancer (140-208) . Cancer of rectum (154) ' Cancer of gall Madder (156) - Cancer of brain, nervous system (191-192) Multiple sclerosis and disease of central nervous system (340-341) Ischemic heart disease (410-414) Nonmalignant respiratory disease (490-519) Fatafitres usualiy work-related (846, 916-921, 923-927) Electrocutions (994) " . 100 108 195 343 172 539* 96 .120 626* 2720* . 788 200 6 4 .8 3 214 59 19 16 (94-124) (72-425) (93-878) (74-339) . . : (111-1575) (84-110) (92-115) . (377-978) (1555-4417) . Construction laborers AU deaths . Pulmonary tuberculosis All cancers (140-208) -. . . 100 213* 95* 9639 10 2072 .: r #' (102-392) (92-99) (continued) . Mortality Among Construction Trades Workers 61 TABLElV.(Co>tdatuJ) ____________ . . . Cause of death1' __________ _________________ ' PMR5 , Q6s<95 CI* - Construction laborers (continued) . . Cancer of lip, oral cavity, pharynx (140-149) - Cancer of lung (162) MenialfiswdHs (290-319) Ischemic Mart disease (410-414).. . ' Cttonic pulmonary bcart.discase (416) . Nanmafignanttespbatory disease (490^519) . Pucuunmia, influenza (480-467)' ' Aabotosl* (501) Silicosis (502) . . Chrome fiver (titra.tr and cirrhosis (571. 5721-5728) Sudden death (canse unknown) (798-799) Falls (880-888,9293) '' V Homicide $60-969) ' . ' AJdtol associated diseases11 (291 *30*,' 3050, 3575, 4355,5353,7903; 5710-13,8606;.8601) 130 113?. 128*. 9lr 16*: 128*: 114* '250. 153 133* 124* 138? 1751 61 .894. 1OS. 2?50 .' 12' 685' 276 3 2 211 . 139 84 216 . (99-167) (J06-I21) (105-155) (68-94) (85-286) (Il8-i37)\ (jOl-128) (52-731) (18-552) (116-153) (104-147) (110-171) (152-200) 163* 182 (140-188) Tile setters ' All deaths . ' AD cancer (140-208)' Cancer of lip, ond, pharynx (140-149) Cancerof lung (162) . Cancer of brain and nervous systcm(191-192) Ischemic heart disease (410-414) Nonmafignant respiratory disease (490-519) ' . 100 115 186 169* 221 77 133 184 54 (86-150) 2 (22-672) 31 (115-240) 3 (46-646) 39 (55-106) 14 (73-224) Cupet installers ' ' All ilmlw AD cancer (140-208) . Cancer of lip. oral, pharynx (140-149) Cancer of lung (162) Ischemic heart disease (416-414) .. . Nonmalignant respiratory di$ease-(490-5!9) Homicide (960-969) 100 87 191 132 88 161* 217* . 317 60 3 33 64 23 14 (66-112) (39-558) (91-185) (69-113) (102-242) (119-364) DrywaB installers AD deaths AD cancer (140-208) Cancer of Bp, oral, pharynx (140-149) Cancer of lute (162) ' Mental disorders (290-319) Ischemic heart disease (410^414) , Alcohol associated diseases'* (291,303, 3050, 3575, 4355, 5353, 7903.5710-13, 8600. 8601) 100 105 232 121 25ff 87 226* 355 77 4 31 10 . 66 15 (83-131) (63-595) (82-171) (123-471) (68-111) (126-372). Sbeetraetal duct installers AD deaths AD cancer (140-208) Cancer of larynx (161) Secondary and iU-defmed cancer (195-199) . Diseases of the nervous system and sense organs (320-389) . 100 196 109 SO (81-143) 388 2 (47-1400) 188 6 (69r-409) 138 '4 (38-353) (continued) 62 Robinson et al. TABLE IV. (Continued) Cause of death* . . . . Sbeetmetd duct Installers (continued) Iscbendcheart disease (4ICH414) Other ischemic heart tfiscase (411-414) . Diseases of arteries arid capillaries (440-448) Alcohol associated diseases* (291.303, 3050,3575, - 4255, 5353, 7903, 5710-13, 8600, 8601) . PMR* . Q6s? ' ' 95*CP: 110 60 139 31 188 8. 299* ' (84-141) (94-197) (81-370). (110-650) Boilermakers -- Afideaths . All cancer (140-208) Cancerof larynx (161) Acute myeloid leukemia (2050) - Parkinson's disease (332) - . Ischemic heart disease (410-414) Other forms of heart disease (420-429) Asbestosis (501) - . ' '. 100 106 172 2S9 174 85* 119 3238* 679 . 168 3 ' '4 . list1 75 .',3 (91-123) (36-503) (71-663) (48-447) (72-98) (93-149) (668-9462): Plasterers . All deaths All cancer (140-208) ' -' Cancer of lip, oral cavity, pharynx (140-9) . Cancerof larynx (161) . Cancerof lung (162) .. Cancer of brain and nervous system (191--192) Ischemic beast disease (410--414) ' ' Chronic liver disease and cirrhosis (571, 5721-8) 100 ICS 206 646* . 127 181 87 189. 320 74 (80-129) 3 (43-603) 5 (210-1507). 32 (87-180) 3. (37-528) 79 . (69-109) 8 (81-372) `Analyses based on occupation-coded death certificate data for 1984-1986 provided to NIOSH by 49 U.S. states for evaluation and follow-up. Occupational coding was supported by NIOSH, National Cancer Institute, and National Center for Health Statistics. . ''Code of the International Classification of Diseases, 9th rev., 1978, in parentheses following, `proportionate mortality ratio (observed/cxpectcd proportionate mortality x 100). ' 4Observed. ' . cExacf95% confidence intervals for the true mean observed deaths were calculated by assuming that the observed Ariiihe were distributed as a Poisson random variable. The endpoints of there confidence - intervals were divided by the expected deaths to yield 95% confidenceintervals (shown in tbe table) oh the true PMR (Bailar JC, 1964). .' V < 0.01. . p < 0.05. ''Causes of death attributed by- the National Center for Health Statistics to alcohol-induced mortality include akotoCc psychoses; alcohol dependence syndrome, noadepeadent abuse of alcohol, alcoholic polyneuropathy, alcoholic cardiomyopathy, alcoholic gastritis, chronic Over disease and cirrhosis spec ified as alcoholic, excessive blood level of alcohol and accidental poisoning byalcobcl, not elsewhere classified. These causes exclude accidents, homicides, and other causes indirectly related to'alcohol use [Anon, 1992a]. *May include mesotheliomas. ' . _ _ JMay include lead poisoning. . have significantly elevated PMRs for lung cancer, nonmalignant respiratory disease, falls, homicide, and alcohol-associated-disease. Construction and maintenance paint ers experienced statistically significant elevated PMRs for viral hepatitis B, cancers of the larynx, lung, and bladder, mental disorders, nonmalignant respiratory disease, silicosis, diseases of the stomach and duodenum, chronic liver disease and cirrhosis. * Mortality Among Construction Tirades Workers 63 fells, unintentional poisonings, homicide, and alcohol-assOciated disease. Plumbers' were observed to have statistically significant elevated PMRs for all cancer-, cancer of die lung and plana (mesdthidioitiaj; nonmalignant respiratotydisease, add asbestosis. fpgnlatnrs had statistically significant elevated PMRs for aU cancer, cancerof the lbng,. peritoneum, and pleura.(mesotbelioina), nonmalignant; respiratory disease, andaS^ bestosis. Table IV. also shows that significantly elevated PNnis were oteerved amoog operating engineers.for luhg.cancet.and injuries:ufeifly. work-related. Only (jjtdtal deaths werereported far etevatnrinstallers andiepairas.during thestadyyfeais. Small numbers for cause-specific deaths resulted in elevated PMRs for lung canceir. pneu monia and influenza, add nonmalignant respiratory disease. Only die PMR for lung cancer was statistically signiScam. . ' ;" . ' Roofers are reported in Table IV to have significantly elevated PMRs for mortal disorders, nonmalignant respiratory disease,- asbestosis, chronic liver disease and cirrhosis; falls, homicide, and alcohol-associated diseases. Structural metal workers bad elevated PMRs for lung cancer, falls, injuries usually work-related, and homi. eides. Significantly .elevated PMRs were observed atdohg'cOnaete and temizo fin^ idiers for bbne and kidney cancer, although many PMRs were elevated. . Table IV shows significantly elevated PMRs among electricians-for all cancer, cancer of the lung, pleura (mesothelioma), and scrotum, aplastic anemia, norimalig- nant respiratory disease, injuries usually work-related, and electrocutions. PMRs were elevated for different causes among the much-smaller group:of electrical power installers and repairers, but only the PMRs for multiple sclerosis; injuries usually work-related, and electrocutions were statistically significant PMRs among construc tion laborers were significantly elevated.'for pulmonary , tuberculosis, Jiing cancer, mental disorders, nonmalignant respiratory disease, pneumonia and influenza, : chronic liver disease .and cirrhosis, sudden death, falls, homicides, and alcohol- associated disease. Statistically significant elevated PMRs for cancer of the lung were observed among tile setters; nonmalignahtrespiratorydiseaseand homicide among carpet installers; mental disorders and alcobol-assodated diseases among drywali installers; and elevated but not significant PMRs for cancer of the larynx, nervous system disease, and cardiovascular disease among sbeetmetal duct installers, except for alcohol-associated disorders, which were significantly elevated. Table. IV shows sig nificantly elevated asbestosis' mortality for boilermakers.. A significantly elevated PMR was found For cancer of the larynx for plasterers. Blue Collar Analysis - . The results of die analysis using the blite collar comparison group were very similar to. those reported above. Overall, fnany PMRs had decreased slightly from those obtained Using all workers as a referent group (Table IV), but remained statis tically significant. However, six construction trades had PMRs that decreased and became statistically non-significanL They were: cancer of the larynx (PMR --143) and silicosis (PMR=293) among construction and maintenance painters; lung cancer (PMR = 109) among brickmasoos; lung cancer (PMR= 103) and nonmalignant'res piratory disease (PMR=98) among electricians; lung.cancer (PMR = 100) and nonmalignant respiratory disease (PMR=98) among plumbers; nonmalignant respiratory disease. (PMR = 145) among carpet installers; and lung cancer (PMR= 117) and ho micide (PMR= 153) among structural metal workers. 64 Robinson et aL ^ - Additionally, six construction trades had PMRs that Increased and became statistically significant when the blue collar group was used for comparison. .They' were cancer of the kidney (PMR--323, p<.05) among insolation workers; diseases of the arteries (PMR=206, p<.05) among plasterets;strpke(PMR-- 130, p<.05j, of the thymiisfPMR = 1586, p<.05),and diseases ofthe. arteries (PMR= 161, p<.05) among electrical power installers and- repairers; diseases of die' pancreas (PMR= 183, p<.Q5) among painters; and unintentional poisoning (PMR = 1&3, p<.05) among drywall installers. . ' DISCUSSION Previous studies of male Swedish and British construction industry workers, as well as those in the states of California arid .Washington, have reported similar patterns of excess mortality from cancer and from respiratory,;nervous, hnd digestive system disease, as has. our study. These patterns' were,observed among*white men.in this Study and forthe other .race-sex groups toaleSSer degree. . . ' The most striking excesses we observed Were for respiratory disease related to asbestos and other hazards, the most well documented construction .worksite hazard.- In our study, the highest PMRs for white male construction workers under 65 are for asbestosis (PMR=393) and -silicosis (PMR=327), both increased .more than three fold (Table ID)- Three priorreports [Englund, 1982; Enghofcn, 1987; Fletcher, 1990] noted that the-excess hing cancer that was observed among Swedish construction . workers was attributable to occupational exposure to asbestos. This excess, persisted even after adjustment for smoking. Although we could not adjust for smoking or alcohol, the presort report observed excess lung cancer and mesothelioma mortality among black and white female constniction workers, and white male$ under age 65 at death, and for men eriiployed in many of the stalled trades arid for laborers (Tables ID-TV)..This.may suggest that more emphasis should be placed on the control of silica-and asbestos-related diseases. Feasible arid safe substitutes for silica in abrasive blasting should be identified. Significant excess mortality was noted for mental disorders .for white and black men usually employed as construction workers, for white construction workers under age 65,;.and for painten, brickmasoris, roofers, construction laborers, and drywall installers. Neurologic impairment and neurobehavioral deficits have been repented for workers exposed to persutentlyelevated levels of organic solvents or to lead in the work environment, particularly constniction painters [Baker, 1986, 1988; Fuller, 1987]'.. These agents may enhance or cause cancer, other chronic disease, and/or mental disorders. . ' ' ' - Daring health hazard evaluations conducted between 1981-1983 by the Na tional Institute for Occupational Safety and Health (NIOSH), elevated ambient as bestos levels were reported for several constniction occupation sites involving skilled -trades workers. Thse included electrical work, drywalling, insulation, carpentry, and roofing. However, asbestosis PMRs are probably underestimated (using underlying causes of death) on death certificates due to a change in the nosology decision table for occupational lung diseases tbat was implemented by the National Center for Health Statistics in 1982 [U.S. Health, 1991]. The change, i.e., deletion of the specificity clause has resulted in some occupational long disease deaths being coded to the more general category of obstructive pulmonary disease. Constniction site Mortality Among ConstnictionTrades Workers 65 woikers may be unaware of exposures to asbestos, lead, and noise, and materials specific to particular trades [England, 1981; Schneider, 1993]. . New findings of cartcer and other health outcomes woe observed for specific, construction trades as follows: brickmasons--bone cancer and. melanoma (skin); shetfenetal workers-1--cardiovascular disease'; plumbers--uasal cancer; construction laborers--pulmonary tuberculosis; electricians^--sabtal cancer and aplastic! anemia; electrical power installers and repairers--rectal cancer and multiple sclerosis; con crete andtenazzo workers--bone and kidney'cartcer; boilermakers--acute myeloid. leukemia; aqd-white women construction workers--cancer of the connective tissue. Although these, art new findings for construction workers, most of these caiises have been previously linked by epidemiologic or toxicologic study to specific occupational exposures. The following briefsummary suggests hypotheses far further investigation in construction workers. Prevention of cartcer should become a focus for future research strategies. ,. Bone cancer, seen for brickmasons and Concrete andtenazzo workers, has been linked with radiatron exposure pARC, 1971]. Additionally, osteogenic sarebina has - been produced in rabbits by intravenous. administration of beryllium compounds . PARC, 197}]! In Sweden, a fivefold excess of .bone sarcomas wasreported in communities with air pollution resulting'.from ferrous metallurgy plants [Lindahl, 1972]. Construction work at nuclear facilities may involve radiation exposures. /. Melanoma has been linked wjtir exposure to ultraviolet radiation in sunlight, early exposure in childhood arid adolescence beingpredictive in-some studies (Os- terlind, 1988]. Paffenbarger found that paid outdoor work before college was asso ciated wife increased incidence of melanoma [Paffenbarger, 1978]. Sunlight exposure to construction workers might be evaluated. Nasal cancer has been linked with nickel refining and chrome pigment manu facturing (Schottenfield, 1982], but riot reported in previous studies of construction plumbers and pipefitters [Kaminski, 1980; Cantor, 1986], Nickel and chrome expo sures could.occor in welding, but mortality is not the bestraeasure of nasal cancer occurrence because of its high survival rate. ' Although cases of silico-tuberculosis have been reported in sandblasters [Bailey, 1974], excess mortality from pulmonary tuberculosis has been associated recently wife gold miners exposed to silica [Steenland, 1995] and it has been linked to lower socioeconomic status.and other occupations [MMWR, 1995]. Sir Percival Pott's observation of an unusually, high frequency of scrotal (skin) cancer among London chimney sweeps in 1775 was later shown to be caused by contact with coal tar products [Lueke, 1907; Passey, 1922]. A later association was wife cutting oils [Cruickshank,1950], Both agents were also linked with lung cancer. The potential far skin contact with coal tar products, benzo(a)pyrenes, or ingestion of cutting oils for electricians during construction work might be evaluated and has been implicated for roofers. . '. Radiologists who practiced before radiation exposure guidelines were imple mented have been observed to have excess deaths from aplastic anemia and skin cancer [Matanoski, 1975]. Electricians may have exposures to electro-magnetic fields or radiation, depending on their work setting. . Multiple sclerosis has not been reported previously in connection wife occupa tion, although there have been several reports of astrocytomas in multiple sclerosis * . 66 Robinson et al. '. . ' '' i j [Reagan, 1973], implying a link with .elevated brain cancer previously re- j ported for men in electrical oanipatioris[Thoma?, 1986]. " ' ' ' 'f An elevated kidney cancer rate, observed jo construction concreteiand terrazzo : wprkere.wasptevicrasly reported for coke oven workers (Redmond, 1972] and Was 1 linlmri with exposure to asbestos [Selikof]M979] and dinette smoking. Acute meloid 1 leukemia has been reported in connection with-exposure to nonionizing radiation : exposure [Savilz, .1987] and with tteWpatiortal exposure to benzene [McMichael, 1975], but not- for a previously studied dAbrt of boilomakas [Beatiihiont, 1980], ! . Cancer of the connective tissue: hasbeen associated with. dioxin and arsenic . exposures, but not construction occupations* and occurs less often in women than in , men [Schottenfeld, 1982}. However; dioxin and arsenic bavebeen used in herbicides and mayoccur on construction sites in treated lumber. .flie dangers of construction sites are well known, for catastrophic outcomes. High rate* for traumatic fatalities thstdccurred at woik for thecbnstfuction industry have been previously documented [Stoirt-Wiegand, 1988; Bell, 1990; NIOSH, 1993], Exposure to Various chemicals, duSts,solvots, andotheragents with toxic : properties has been recently reported:, for U.S: construction industry worksites (NIOSH, 1988, 1990a,b; Schneider*1993] and earlier for Swedish construction in dustry worksites[Englund, 1981]. Construction site workers may be at risk for health ' problems associated not only with toxic substances used for their own job tasks, but with the. "bystander" and."downstream''.exposures present in shared work spaces. Construction industry workers are protected from hazardous exposures under the current OSHA permissible exposure limits (PELs). However, unlike PELs for general industry, PELs for the construction industry-are currently based on the 1970 Tluesbold Limit Values (TLV) recommended by the A1HC [Anon, 1992b], many of which exceed NIOSH Recommended Exposure Limits (RELS). . . Lifestyle! difference? in more frequent use of alcohol ami tobacco have been reported for construction workers [Harford, 1992], particularly laborers [Mandell, 1992; Burkhart, 1993]. Although it did not.explore all differences, our blue collar analysis had adjusted for lifestyle differences-related excess mortality for construction workers. Some PMRs that were significantly elevated for lung cancer woe reduced and became nonsignificant; other lung cancer PMRs remained or became significantly elevated. Tins finding is similar to that of a previous US. study [Sinjgleton, 1989]. Singleton studied death certificate data for California and adjusted for lifestyle, using National Health Interview Survey and U.S. Census data to describe the smoking, alcohol, and socioeconomic status of occupations. Singleton [1989] also found excess mortality that persisted even after adjustment. For example, of nine construction occupations.that had significant excess' lung cancer, five excesses (Standardized. Mortality Ratios--SMRsj remained significant after adjustment for lifestyle-related factors. Our mortality study results may underestimate the . magnitude of nonfatal dis eases and disorders. Opportunities for morbidity surveillance should be pursued in tandem with obtaining exposure data. . .- ADVANTAGES AND LIMITATIONS Analyses ofstatements of usual occupation and industry on death certificates are advantageous for the construction industry because it is a widely dispersed workforce ' Mortality Among Contraction Trades Workers 07 for which epidemiologic studies axe difficult and sparsely ideated, due to the lack of organized records. Additionally, the large number of deaths available far the present analysis made-possible the evaluation of mortality risks for rare causes of death . anywig white men usually employed in construction occupations. The large number of deaths alsomade it posable to evaluate mortality risks for black mod nod women and white women construction industry workers' but not rare causesof'death. - The interpretation of the FMRs is limited by misclassificatioii of causeof-death and tisual bciftipatian as repotted on deatbcertificates [Rothman, 1986]. The accuracy of occupational information on dcath certificates compared to interviews has been reported to bc ~ 56.8% fin- construction [Schade, 1988]. Four other studies have restated. higher rates of agreement for long-term white wbikers acrpss all occupa- tions.Thestudies reported. 73.5% [Steenland, 1984], 80% [Peterson, 1974], 75% [Mflham, 1976], and 75% [Wegman, 1978] agreement between occupation as listed on tbe desdi certificate and as detenxuhed.by interviewing next-of-iin in case control studies. Die,degree of misclassification of cause of death varies by disease [Kircher, 1985]'. Whehiompared to autopsy findings, death certificate diagnoses was foohd to be most accurate for cancer, with lesser degrees of accuracy for diagnoses of circu- . latory, digestive, respiratory diseases, and; mental disorders. Both"ofthese misclas-. sificationbiases would tend to lower risks toward the null. Radier than adjust for die large hutnber Of statistical comparisons, we chose to draw on other criteria, such as previously published reports, besides statistical precision in emphasizing associa- . dons'. . - . limitations in the PMR method may also bias, risk estimates. For example, . when the overall mortality rate of the working population is higher than that for the comparison population', as it has been reputed to be for many California construction trades [COMS, 1990], the PMR (as an estimate of die SMR) will be artificially decreased [Decoufle,1980]. Proportionate cancer mortality analysis has been sug gested as one-approach to the reduction of artificial changes in the PMR icaased by - deficits in mortality. However, PMR analysis, when used for population-based studies Of workers, has been proposed to be less biased than SMR analysis due to the decreased healthy worker effect, i.e., overall lower death rate among working pop ulations relative to the general population. CONCLUSIONS Tire patterns of excess mortality among white male Construction industry work ers were, similar to those observed in other studies, although we observed them to be occurring before age 65, that may have been employed at death. Significant excess mortality was observed for asbestosis and silicosis in l).S. construction, industry workers who died before'age 65. Deficits occurred for cardiovascular diseases for all trades except sheet metal workers, although adjustment for blue collar status appeared to elevate cardiovascular mortality for some trades. New findings were excesses for bone cancer and melanoma in brickmasons,sinonasal cancer in plumbers, cardiovascular disease in sheetmetal workers, pulmo nary tuberculosis and silicosis in constroction laborers, scrotal cancer and aplastic anemia in electricians, rectal cancer and multiple sclerosis in electrical power install ers and repairers, bone and kidney cancer in concrete and terrazzo workers, and acute myeloid leukemia in boilermakers. Sex differences in risk occurred iu that women 68 Robinson et aL . % * ... construction workers experienced excess mortality from cancer of the connective tissue and- of the bladder. '.. .. - -' ' This report presents a wealth ofinformation on the mortality patterns of workers across; a. wide spectrum pfconstruction trades. A national research strategy should be developed, to identify die most plausible and worrisome oif these associations for ; nomination for moreintensiYe researcb.Highpriority should be given fo findings that are of Wgfc relative and absolute risk and are consistent between independent studies. as'weiras Consistent over tfroe: Sustained .mortality surveillance over long periods of timW will probably be one.of the most effective means of identifying construction industry problems necessitating close attention, in comparison with the inevitable fluctuations in rates due to chance or bias. Further, a national intervention strategy should be begun to develop and implement effective means of intervention among: construction worksites. ; . j ACKNOWLEDGMENTS : .. We are .indebted to the National Institute for Occupational Safety and Health clerical staff, particularly Lisa Thomas. We are grateful for the' technical contribu tions of Steve Adams, Mary Carroll,. Lynette Harde, Terri Schnorr, Randy. Young, arid others. We thank tire State bealth departmeots fbf supplying the death^certificate; dataand Lairy Fine, TerfMeinhaidt, David Brown, Dick Lemen, Sandy Tilfetti; and j others, for helpful advice and encouragement. We thank Knut Ringen. Pete Stafford, i and many, individuals inthe AFL^QO construction trade unions, the Center to Protect! Workers' Rights, and the Occupational Health Foundation. REFERENCES . Anon (1992a): Alcohol-tAduced deaths. Monthly Vital Statistics Report 40(8)(S):23,49. . Anon (1992b): Proposed Rule: Department, of Labor. Federal Register. 57(114)26002-26041. Baflar 1C DL Edesor F (1964): Significance factors for the ratio of a potsson variable to its expectation. Biometrics 20A39--643... ' .' Bailey WC, Brown' M. Buechner HA (1974): Silico-mycobacteria) disease in sandblasters. Ain Rev RespirDis 110:115. '. ' . . ' BakerEL, Fine L) (1986): Solvent neurotoxicity: The caveat evidence. JOM 28:126-129. Baker EL, Lett RE,Eisen EA, Poflucx LI, Ptantamura DL, Larson M, Wolford R (1988): Neurobehav. kml effects of solvents in ctidistructinn painters. JOM 30(2):116-123. Beaumont JJ, Weiss NS (1980): Mortality of welders, shipfinets and other metal trades workers hr , Boilermakers Local No.l04,AFL-CIO. Ami Epidemiol 112:775-786. ' . Beil CA, Stout NA, Bender TR, Conroy C, Crouse W, Mycn JR (1990): Fatal occupational injuries in rite United Slates, 1980 through 1985. JAMA 263(22)3047-3060. Burkhart O, Schulte P, Robinson C, Sieber W, Vosstnas P, Ringen K (1993): Job tasks, potential exposures, and health risks of laboien employed in the construction industry. .Am J Ind.Med 24:413-425. . "; Cantor KP, Sontag JM, Held MF (1986): Patterns of mortality among plumbers and pipefitters. Am J Ind . Med 10:73-89. ' . COMS--California Occupational Mortality 1979-1981 (1987): Figures 2B, P.20 and 4A, P.tU.and5A. ' - P. 145, Sacramento. - . '` Crakkshank CND, Squire JR (1950): Skin cancer in the engineering industry from the use of mineral oiL BrJ Ind Med 7:1-11. Decoufle P, Thomas T, Pickle L (1980): Comparison of die proportionate mortality ratio and standardized ' mortality ratio risk measures. Am J Epidemiol 111(3)263-269. Mortality Among Construction Trades Workers - 69 Dubrow R. Spaeth S (1986): Proportionate Modality Ratio Analysis System--Version IV (draft). Eaghobn G, Enghrod A. Fletcher ACi Haffin N (1987):.Respuatoiycsncer-incidence in Swedish coa- s&vctitM workers exposed to matwnade mineral fibres atd asbestos. Ann Occup Hyg 31(4B): , 663.-675. . . England A (1981): Otemica] Health Hh&^.foHelandb'Mfed): "Human JtetorsTsrganoi&ics for Civil Ehgheeand-Con$tnieri<w'Maiia(ge." Nv York: Wiiqr &Sons,.Chap. 4. .' England A, Eng***n G (1982): AshtotoMdatod cancer in Swedish construction waiters. Archivum ' litunonotogiaeetTlierigriaeExpeiii^^ . . fidSpr AT, Baker EL,'Lets RB(1987): Neurobebaviondeffecti of occupational exposure to organic :.v Sofvents among eonstrodiori pabsteis. Br J Ind Med 44(5)292-308. Heitbert (1990): Lung cancer attrilwfoicfooccupatirm in the Swedish construction industry. Presented - on.September 26, 1990 at the 23ld brternadonal Congress of Occupational Health,. Mootieal, . '' '. . . Harford T, Brooks S (1992): Cuitteisihartaiity aid occupation. I Stud Alcohol 53:463-468. International Agency fot Research on Cancel (tARC) (1971): IARC Mooogtapha on the evaluation of carcinogenic risk of chemicals toman, Vol. I. Lyont lARC, pp 17-18. KarafoskiR, Geissest MP, Dacey E (1980): Morality analysis ofphanBen and pipefitters. JOM 22(3): iGrcher f.NebooJ, Bnrdo H (1983): The autopsy'as a measure of accuracy of the death certificate. N ' ' &igl | Med 313:1263-1269. j ' :.; ' . !. . . A (1976): WflltatO Farr.ffooiuteofinodtTBcPucepta ofsurveilbace. In! J Epidemil 5(l)cl3-- ' ... '.>*. '. '' ' ' ' . Linidahl.Q0972): The gcographical yariatharof cases of bone sarcoma. Lakaitidniagen 69:2945-2948; Lueke AW (1907): Epithefioma in carton workers. Cleveland Med J 6:199-202. - ' MaadcH W. Eaton W. Anthony J. Gamsoh.R(1992): Alcoholjam and occupations: A review and analysis of l04ocorpatior4. AlcoholCIin Exp Res 16(4*734-746. . MatanoslpGM, SettrcrR.'SariwellFB.ctal. (1975): The current mortality rate* ofradloiogists and other ' physician specialists: Specific<3hsesof deatb.AmJ Epidemiol 101:199-210; . ' . McMidute) Al. Spirtas R, Knpper LL, et al. (1975): SoNent exposnrr: and leukemia among rubber workers: An epidemiologic study. JOccup Med 17234-239; . - ' M3ham.SJ 0976): Occupational Mortality in Washington State 1950-1971. NIOSH, Cincinnati. . MiBiam SJ (1983): Occupational Mortality, in Washington State 1950-79. DHHS (NIOSH) Publication No. 83-116. NIOSH, 1983, Crncinnati. NCHS (1987): Instruction Manual Partl9:' Industry and Occupation Coding for Death Certificates. Hyattsvilfe,'MD,NCHS. ' ' . ... NlOSH (1988, i990a,b): National Occupational. Exposure Survey. DHHS (NIOSH) Publication No. 88*106, 89-102,89-103. Washington, DC: US Department of Health and Hitman Services.. NIOSH (1993): Fatal injuries to workers in tbe United States, 1980-1989: A decade of saiveillaiice. NIOSH Report 93-(08. Osteriipd A, Kicker M. Stboe B, Jensen 0 (1988): The Danish case-control study ofcutaneous malignant melanoma. JL Importance of UV-light exposute. Ini J Cancer 42319-324. 'Paffenbarger RS, Wing AL, Hyde RT (1978): Characteristics in youth predictive of adult-onset malignant lymphomas, melanbmas amfleukemias: brief coaununicahon. JNCT60:89-92. ' - Park RM, Maizlish NA, Punnet L, et id. (1991): A comparison of PMRs and SMRs as estimators of ' occupational mortality. Epidemiology 2:49-59. ' - Passey RD (1922* Experimental soot cancer. Br Med j 1112-1113. Pettrscn GR, Milham S(1974):.Hodgkin'idisease mortality and occupational exposure to wood. J Natl Cancer Inst 53:957-958. ' ' . Reagan 17, Freiman IS (1973): Multiple, cerebral gliomas in multiple sclerosis. J Neurol Neurosurg Psychiatry 36:523-528. . Redmond CK, Gooco A, Lloyd JW, et al. (1972): Long-term mortality study of steelworkers. VI. ~ Mortality from malignant neoplasms among coke oven workers. J Occup Med. 14:621-629. Registrar General's Decennial Supplement for England and Wales,-1970-72 Occupational Mortality - (1978): Office of Population Censuses and Surveys, Her Majesty's Stationary Office, London. Registrar General's Decennial Supplement for England and Wales, 1979-80, 1982-83 Occupational Mortality (1986): Office of Population Censuses and Surveys, Her Majesty's Stationary Office, London. 70 Robinson et aL ' / . * - * . ** .. * * .' Rothman KJ (1986): "Motfcrn Epidemiology.'' Boston: Little, Brown and Company. Savitz DA, Calif BE'(1987): Leukemia andoccupational exposure to electromagnetic Gelds: review of ' . epidemiological surveys. )OM 29:47--51. ., . Sehade'Wj, Swanson' CM (1988): Coenpaiisoaof death certificate occupation and industry, data with - lifetime Occupational h&oriesdbtiibed by interview: variation] in the accuracy ofdeath certificate . t eutijes. AmerI fed 14:121-136:! .. ` . . . . . . Schneider S, SusieF (1993): AnTpvestyatiouof Health Hazards on a New Construction Project," Washington.DC: Center to.ptotectWuiten' Rights. . . Schortehfeld D, Banmem D (1982): ``Cincer. Epidemiology aid Prevention." Philadelphia: Saunders. Selikoff Uvttammond EC, ScidmanH(1979): Mortality experience of insulation workers in the United' States ahd Canada.1943-1976. Amj NYAcadSci 330S1-U6. - Singleton.']A' (1989k COMS D-rCaHftmiut' occupational mortality 1979-1981 adjusted for stndkmg, alcohol, arid sodoecopbraic status (Dam), . -* Sfeenbud K, Beaumont ] (1984): the accuracy of occupation and industry dab on death certificates. ] . . becup hied 26(4)288--29$,. ' ... . ,. / . .1 Steealand R and Brown D (1995):,Maftality-istudy of gold miners exposed to silica and nasasbestifonn ';. amphibole minerals:.An updaic.vrithJl'i, idoth years of fotfow-up.Am J Ind Med 27:217-229. Stout-WeSgand N (1988): Fatal (xeuptiitrnal;injjBries in the United States la 1980-1984: Results of the . first national census of truumaticpccugdtional fatalities. Scand I Wort Environ and Health 14: ; suppi 1^0--92. . ' . Thomas TL, Warvreita RJ(l986):Scand JWorkEnvnon and Health 12:1-15.' '. . U.S. Bureau of the Census (1980):' 1980 Census ofPopulation. Alphabetical index of Industries and Occupations. PH C80-R3, US GPO. U.S. Bureau of the Census (1982): -i980 Census of Population and Housing, Supplementary Report: Provisional Estimates of Social, Economic and Housing Characteristics. U-S. Department of Commerce. PHC80-S1-1. .. . U.S. Bureau of the Census (199Z): Statistical Afctraciofthe United-States: 1992112th ed., Washington, DC. - - ' ' . ' ' U.S. Health 1990 (1991): National Center for Hedtb Statistics. DHHS Pnblicatiou No. (PHS) 9I-PHS, Washington, DC, GPO, Table 37. ... - Wegman DH, Peten JM (1978):'Oat cell lungcaaoer in selected occupations. I Occup Med 20:793--795. World Health Organization (WHO) (1978): International Classification of Diseases, adapted, 9th rev.