The following are pre-publication drafts of articles from the Morbidity and Mortality Weekly Report 44(01) dated January 13, 1995. Late-breaking articles, and final editorial revisions are not included. Therefore, these articles should not be quoted without consulting the official printed copy that is released to the public, as received from the CDC.--CDC.Also, please refer to the printed copy for any charts or graphically represented data. ********************************************************** CONTENTS -- Injuries Associated With Use of Snowmobiles -- New Hampshire, 1989-1992 -- Physician Vaccination Referral Practices and Vaccines for Children -- New York, 1994 -- Differences in Maternal Mortality Among Black and White Women -- United States, 1990 -- Proportionate Mortality from Pulmonary Tuberculosis Associated With Occupations -- 28 States, 1979-1990 -- Erratum: Vol. 43, No. 40 *********************************************************** Injuries Associated With Use of Snowmobiles -- New Hampshire, 1989-1992 Recreational use of snowmobiles is popular in New Hampshire during the winter months; from 1982 to 1992, the annual number of registered snowmobiles ranged from approximately 21,200 to 42,500. During this period, 26 deaths associated with use of snowmobiles in New Hampshire accounted for 822 years of potential life lost before age 65 years. To assist in the development and evaluation of injury-prevention programs for users of off-highway recreational vehicles (OHRVs) (e.g., all-terrain vehicles, trail bikes, and snowmobiles), the State of New Hampshire Department of Fish and Game (DFG) and the New Hampshire Department of Health and Human Services examined reports of injuries resulting from OHRV use in New Hampshire from January 1989 through February 1992*. This report summarizes information about snowmobile-associated fatal and nonfatal injuries during this period. Since 1981, New Hampshire has required reporting of OHRV incidents resulting in injury. A standard report form must be completed by a person involved in the event or by a law enforcement agent and filed with DFG within 5 days of the incident. Information collected on the form includes demographic characteristics of the operator, type of vehicle, environmental conditions, date and time of the incident, whether the operator reported having taken an OHRV safety course, type of injury, excessive speed, and use of alcohol and helmets. During January 1989-February 1992, DFG received reports of 164 snowmobile incidents resulting in injury. Of the 164 incidents, 155 involved 188 vehicles and resulted in 163 nonfatal injuries, and nine involved 13 vehicles and resulted in 12 fatalities and two nonfatal injuries (Table 1). All fatal incidents were reported by law enforcement agents. Of the 155 reports of nonfatal incidents, 103 (66%) were completed by a law enforcement agent. All operators involved in fatal (13) and most involved in nonfatal (161 [86%]) incidents were male. Seven (54%) operators involved in fatal incidents and 75 (40%) operators involved in nonfatal incidents were aged 20-29 years; no operators involved in fatal incidents and 40 (21%) involved in nonfatal incidents were aged less than 20 years. No operator involved in a fatal incident and 14 (7%) of those involved in a nonfatal incident were reported to have taken an OHRV safety course. Of nine fatal events and 155 nonfatal events, seven (78%) and 64 (46%), respectively, occurred during darker periods (i.e., 4 p.m.-8 a.m., November-March). No fatal and 25 (16%) nonfatal events occurred during periods of precipitation or other inclement weather (i.e., fog or active snow, sleet, or rain). Operating on a frozen body of water was reported for five of nine fatal and 36 (23%) of 155 nonfatal events. Overall, 67% of fatal incidents were associated with alcohol use and 67% with excessive speed. Of the 103 police-reported nonfatal incidents, 16 (16%) involved alcohol use, and 36 (35%) involved excessive speed; in comparison, of 52 incidents reported only by persons involved in the incident, one (2%) and three (6%), respectively, reported use of alcohol or excessive speed. Of eight deaths resulting from incidents occurring on a frozen body of water, three resulted from hypothermia and five from either head and neck injuries (three) or multiple trauma (two). Three other deaths were attributed to head and neck trauma and one to multiple trauma. Of 165 persons nonfatally injured, 104 (63%) were reported to have been wearing helmets. Helmets were reported to have been worn by 31 (57%) of 54 persons with nonfatal head injuries, compared with four of six persons with fatal head injuries. Reported by: A Hewitt, State of New Hampshire Dept of Fish and Game; Bureau of Vital Records and Health Statistics; D Solet, M Kiely, Office of Chronic Disease and Health Data, New Hampshire Dept of Health and Human Svcs. Div of Field Epidemiology, Epidemiology Program Office, CDC. Editorial Note: In New Hampshire, most fatal snowmobile incidents involved male operators in their 20s, use of alcohol, or excessive speed; half of persons killed sustained head injuries. In addition, fatalities occurring as a result of operating on frozen bodies of water were associated with either severe trauma or events related to falling through the ice (i.e., hypothermia). These findings are consistent with previous studies of fatalities associated with the use of OHRVs (1,2). For example, contributing factors for nondrowning deaths following incidents on frozen water surfaces have included high speeds attained on such open surfaces and unexpected uneven terrain (e.g., ice ridges) (1). The findings in this report also indicate that some snowmobile drivers and passengers did not wear helmets. Although this investigation could not assess the effectiveness of helmet use, a previous study estimated that helmet use can reduce the risk for death among all-terrain vehicle operators by approximately 42% and can reduce the likelihood of head injury in a nonfatal incident by approximately 64% (3). The findings in New Hampshire are subject to at least three limitations. First, rates of injury and death could not be determined because of the lack of an accurate denominator. Although previous studies have used registered OHRVs as a denominator, this number may vary in relation to season and other environmental factors (e.g., inclement weather). Second, because approximately one third of nonfatal injury reports were completed only by persons involved in the incident, some information reported may not be valid (e.g., helmet use, speed, and alcohol use). Finally, these findings probably underestimate the true incidence of snowmobile-associated injuries because of underreporting. Review of hospital emergency and discharge records could assist in evaluating the extent of underreporting. Information from the injury reporting system in New Hampshire may be useful for public health surveillance and assessment of snowmobile and other OHRV injuries (4). In addition, this data source can be used by the New Hampshire Snowmobile Association and other organizations to target high-risk groups for intervention programs. Since 1975, DFG has operated a safety training course for OHRV users. State law requires that any OHRV operator driving off their private property either possess a valid driver's license (minimum age: 16 years) or have taken this course. Operators aged less than 30 years should especially be targeted by any intervention strategy; in particular, young operators with a valid driver's license are encouraged to take the DFG safety course. References 1. Rowe B, Milner R, Johnson C, Bota G. Snowmobile-related deaths in Ontario: a 5-year review. Can Med Assoc J 1992;146:147-52. 2. Hargarten SW. All-terrain vehicle mortality in Wisconsin: a case study in injury control. Am J Emerg Med 1991;9:149-52. 3. Rodgers GB. The effectiveness of helmets in reducing all-terrain vehicle injuries and deaths. Accid Anal Prev 1990;22:47-58. 4. Smith SM, Middaugh JP. An assessment of potential injury surveillance data sources in Alaska using an emerging problem: all-terrain vehicle-associated injuries. Public Health Rep 1989; 104:493-8. * Because the standard reporting form was changed in 1992, comparison with later years was not possible. Physician Vaccination Referral Practices and Vaccines for Children -- New York, 1994 Although vaccinations are among the most effective preventive public health measures available, many children are not vaccinated on time (1). One identified barrier to timely vaccination is referral of children by primary-care physicians to other medical settings for vaccination (2). This report summarizes a survey by the New York State Department of Health of vaccination referral practices among New York physicians and describes the implementation in New York of Vaccines for Children (VFC), a national program making federally purchased vaccines available at no cost to health-care providers for administration to eligible children (3). During April 1993, a random sample of 1137 licensed pediatricians and family-practice physicians (from a total n=5392) in New York were surveyed by mail about vaccination practices. Of 752 (66%) responses, 502 (67%) were from actively practicing primary-care physicians. Of these, 250 (50%) referred all or some of their patients elsewhere for vaccinations. Of referring physicians, 228 (91%) referred patients to a local health department clinic; 109 (44%) had increased the number of patient referrals during 1983-1993, while seven (3%) had decreased referrals. In addition, 63 (25%) reported that the number of well-child-care visits had decreased during 1983-1993, while five (2%) reported increases during that time. Of the 250 referring physicians, 246 provided reasons for referral and rated those reasons as "very important," "somewhat important," or "not important" (Table 1). Financial hardship was a "very important" reason for referral for 217 (88%) of those surveyed; the lack of vaccination coverage by private insurance was "very important" for 132 (54%). Physicians also were asked whether the government should underwrite the cost of mandatory vaccinations. Overall, 409 (54%) respondents indicated that some or all of the costs of childhood vaccination should be underwritten. Since October 1, 1994, free vaccine has been available to VFC participating providers in New York. Categories of federally eligible children aged less than 19 years include those on Medicaid, those who are uninsured, those who are underinsured who visit federally qualified health centers, and American Indians/Alaskan Natives. New York also provided free vaccine to underinsured children who receive care in any medical setting and any child served at a local health department. Medical-care providers were recruited for VFC through articles published in professional organization newsletters and by mailing of registration packets to licensed pediatricians, family physicians, osteopathic physicians, and medical facilities. As of December 27, 1994, a total of 1378 physician practices in New York (including at least 1972 individual physicians and at least 362 health-care facilities) were participating in VFC. To determine the extent of enrollment by Medicaid providers, the list of VFC enrollees was compared to a list of providers who billed Medicaid for childhood vaccines during federal fiscal year 1993. Of 2169 physicians who billed Medicaid for childhood vaccines in 1993, a total of 1213 (56%) had enrolled. Among the 166 physicians who submitted a minimum of 1000 claims for individual vaccines, 143 (86%) had enrolled, while 653 (68%) of 956 physicians not yet enrolled had submitted fewer than 50 claims. In September 1994, the New York State Department of Health conducted a telephone survey of health-care providers who had returned registration forms and declined participation in the program to determine reasons for nonparticipation and to guide future recruitment efforts. Of the 41 physicians who had declined, 29 (71%) were contacted. Of these, five (17%) were retired, five (17%) did not accept patients aged less than 19 years, six (21%) were subspecialists or in academic medicine and did not provide vaccinations, six (21%) indicated that most of their patients would not be eligible, one (2%) had multiple reasons for not registering, and the six (21%) with patients who could benefit from VFC agreed to register as a result of the phone call. From September 15, 1994 (the first date vaccines could be ordered), through December 27, 1994, a minimum of 2,496,000 doses of vaccine had been ordered and approximately 2,456,000 doses were shipped to VFC participants. The average time between placement of orders and receipt of vaccine by providers was 1 week. Reported by: HG Cicirello, MD, GA Bunn, DR Lynch, SC Meldrum, MS, GS Birkhead, MD, D Morse, MD, State Epidemiologist, New York State Dept of Health. S Freidman, MD, N Jenkusky, MPH, New York City Dept of Health. EE Schulte, MD, Albany Medical Center, Albany, New York. National Immunization Program, CDC. Editorial Note: In the United States, approximately 2 million children need one or more doses of recommended vaccines (4). In a 1993 retrospective survey of children entering kindergarten in New York, only 53% had been appropriately vaccinated with the recommended vaccines by age 2 years (5). Important barriers to timely vaccination include missed opportunities to vaccinate at each health-care visit, inconvenient clinic hours, inadequate parental awareness of the need for timely vaccination, inadequate vaccination tracking, the costs of vaccines, and the referral of children from the private sector to the public sector. This report demonstrates that vaccination referrals, in part attributable to vaccine costs, are common among New York primary-care providers. Implementation of VFC is expected to reduce these financial barriers. The actual number of VFC participants in New York probably is underestimated by VFC enrollment figures because many physicians work in facilities or in group practices where only the physician-in-chief is registered in the program. In addition, the impact of the program on vaccination coverage and overall occurrence of vaccine-preventable diseases cannot be determined yet. However, VFC has allowed New York to increase provision of vaccine to more children in primary-care settings where they first seek care. In states where vaccines have been made available for all children, vaccination rates of preschoolers are approximately 10% higher than the national average (AG Holtmann, University of Miami, unpublished data, 1993). The Childhood Immunization Initiative (CII) has designated vaccination of preschool-aged children a national priority and has established 1996 and year 2000 goals of vaccinating at least 90% of children by age 2 years with the recommended number of doses of diphtheria and tetanus toxoids and pertussis, polio, Haemophilus influenzae type b, hepatitis B, measles, mumps, and rubella vaccines. The five strategies of CII, which address both financial and nonfinancial barriers to vaccination, are to 1) improve the delivery of vaccines; 2) reduce the cost of vaccines for parents (VFC); 3) enhance awareness, partnerships, and community participation; 4) monitor vaccination coverage and disease; and 5) improve vaccines and their use. VFC, as an integral part of this initiative, will assist physicians and other health-care providers in reaching these national goals. References 1. NCHS. Health, United States, 1993. Hyattsville, Maryland: US Department of Health and Human Services, Public Health Service, CDC, 1994; DHHS publication no. (PHS)94-1232. 2. Bordley WC, Freed GL, Garrett JM, Byrd CA, Meriwether R. Factors responsible for immunization referrals to health departments in North Carolina. Pediatrics 1994;94:376-80. 3. CDC. Vaccines for Children program, 1994. MMWR 1994;43:705. 4. CDC. Vaccination coverage of 2-year-old children--United States, 1993. MMWR 1994;43:705-9. 5. Hamburg BG, Dowling TP, Kramer JI, Randles RH, Rodewald LE. Immunization of pre-school children. New York: New York State Public Health Council, Ad Hoc Immunization Committee, March 1993. Differences in Maternal Mortality Among Black and White Women -- United States, 1990 The risk for maternal mortality has consistently been higher among black women than white women. The 1990 national health objective of reducing maternal mortality to no more than five deaths per 100,000 live births for any racial/ethnic group was nearly achieved for white women, for whom the maternal mortality ratio* was 5.7 in 1990 (1); for black women, however, the ratio was 18.6. The year 2000 national health objectives include reducing the overall maternal mortality ratio to no more than 3.3 deaths per 100,000 live births and to no more than five for blacks (objective 14.3) (2). This report summarizes race-specific differences in maternal mortality among black and white women for 1990 and compares these with trends in mortality from 1940-1990. Maternal mortality ratios were calculated at 10-year intervals from 1940 to 1990 using data contained on death certificates filed in state vital statistics offices and compiled by CDC in a national database (3,4). Maternal deaths were defined as those for which a maternal condition was designated as the underlying cause of death, as recorded on the death certificate by the attending physician, medical examiner, or coroner.** This report compares maternal mortality only for black and white women because data for other racial/ethnic groups were not available for all years; data for Hispanic women are included in the totals for both blacks and whites. In 1990, the overall maternal mortality ratio was 8.0 deaths per 100,000 live births, a 98% decline from 363.9 in 1940. From 1940 to 1990, race-specific ratios declined substantially, from 319.8 to 5.7 for white women and from 781.7 to 18.6 for black women. Although the percentage decline was similar for black women and white women (97.6% and 98.2%, respectively), the ratios for black women were consistently two to four times higher than those for white women. For example, compared with that for white women, the maternal mortality ratio for black women was 2.4 times greater in 1940, 3.6 times greater in 1950, 4.1 times greater in 1960, 3.9 times greater in 1970, 3.4 times greater in 1980, and 3.3 times greater in 1990 (Figure 1, page 13). From 1960 through 1990 (years for which more detailed data were available), the maternal mortality ratio was higher for black women in all age groups and for each of the major causes of death. The black-white differential was greatest for pregnancies that did not end in a live birth, such as ectopic pregnancy, spontaneous abortion, induced abortion, and gestational trophoblastic disease.*** Reported by: Div of Reproductive Health, National Center for Chronic Disease Prevention and Health Promotion; Div of Vital Statistics, National Center for Health Statistics, CDC. Editorial Note: Despite overall improved maternal survival during 1940-1990, black women were more than three times more likely than white women to die from complications of pregnancy, childbirth, and the puerperium. Although the reasons for this disparity are unclear, possible explanations include differences in pregnancy-related morbidity, access to and use of health-care services, and content and quality of care. Maternal hospitalization, except when associated with delivery, can serve as a marker for severe maternal morbidity. For example, during 1987-1988, a study of pregnancy-related hospitalizations indicated the ratio for black women was 1.4 times that for white women (6); during the same period, the black-white maternal mortality ratio was 3.1. However, in a study of women in the military--who have unrestricted access to prenatal care--there was virtually no difference between black and white women in the overall prevalence of antenatal hospitalization and in the indications for hospitalization (7). Early entry into prenatal care (i.e., during the first trimester)--one indicator of access to and use of pregnancy-related health care--has been assessed for women whose pregnancies ended in a live birth. During 1980-1990, although 76% of all mothers received early prenatal care, the percentage of black women who did not receive early prenatal care was nearly twice that for white women (8). In 1990, 39.4% of black mothers did not receive early prenatal care, compared with 20.8% of white mothers. Once women enter prenatal care, studies indicate differences between black and white women in the advice given to them and use of technology (9,10). Data describing access to pregnancy-related health care other than prenatal care (e.g., gynecologic services) or the content and quality of health care once women obtain these services are limited. Narrowing discrepancies in maternal mortality between black and white women will require evaluating and addressing race-specific differences in morbidity and in access to and use and content of pregnancy-related care. Addressing discrepancies in maternal mortality also may improve maternal morbidity and infant survival. References 1. Public Health Service. Promoting health/preventing disease: objectives for the nation. Washington, DC: US Department of Health and Human Services, Public Health Service, 1980. 2. Public Health Service. Healthy people 2000: national health promotion and disease prevention objectives. Washington, DC: US Department of Health and Human Services, Public Health Service, 1991; DHHS publication no. (PHS)91-50213. 3. NCHS. Vital statistics of the United States, for years 1939- 1991. Vol I-natality. Hyattsville, Maryland: US Department of Health and Human Services, Public Health Service, CDC. 4. NCHS. Vital statistics of the United States, for years 1939- 1991. Vol II-mortality, part A. Hyattsville, Maryland: US Department of Health and Human Services, Public Health Service, CDC. 5. NCHS. Estimates of selected comparability ratios based on dual coding of 1976 death certificates by the eighth and ninth revisions of the International Classification of Diseases. Hyattsville, Maryland: US Department of Health and Human Services, Public Health Service, CDC, 1980. (Monthly vital statistics report; vol 28, no. 11, suppl). 6. Franks AL, Kendrick JS, Olson DR, Atrash HK, Saftlas AF, Moein M. Hospitalization for pregnancy complications, United States, 1986-1987. Am J Obstet Gynecol 1992;166:1339-44. 7. Adams MM, Harlass FE, Sarno AP, Read JA, Rawlings JS. Antenatal hospitalization among enlisted servicewomen, 1987-1990. Obstet Gynecol 1994;84:35-9. 8. NCHS. Health, United States, 1993. Hyattsville, Maryland: US Department of Health and Human Services, Public Health Service, CDC, 1994; DHHS publication no. (PHS)94-1232. 9. Kogan MD, Kotelchuck M, Alexander GR, Johnson WE. Racial disparities in reported prenatal care advice from health care providers. Am J Public Health 1994;84:82-8. 10. Brett KM, Schoendorf KC, Kiely JK. Differences between black and white women in the use of prenatal care technologies. Am J Obstet Gynecol 1994;170:41-6. * The maternal mortality ratio is the number of maternal deaths per 100,000 live births. CDC's National Center for Health Statistics (NCHS) uses the term maternal mortality rate as required by the World Health Organization. In this report, the term "ratio" is used because the numerator includes some maternal deaths that were not related to live births, and thus were not included in the denominator. For this analysis, 3 years of data were combined to calculate maternal mortality ratios to promote statistical reliability and stability in the estimates. For example, 1990 ratios are based on data from 1989 through 1991. In addition, beginning with the 1989 data year, NCHS began using race of mother instead of race of child to tabulate live birth and fetal death data by race. In this analysis, race for live births is tabulated by the race of the child for maternal mortality to maintain comparability of ratios. ** An underlying cause of death is defined by the International Classification of Diseases, Ninth Revision (ICD-9), as "a) the disease or injury which initiated the train of morbid events leading directly to death, or b) the circumstances of the accident or violence which produced the fatal injury." In 1979, the ICD-9 provided the first formal definition of maternal mortality, defining maternal death as the death of a woman while pregnant or within 42 days of termination of pregnancy. This definition differed from that used previously by NCHS, which included deaths up to 1 year after termination of pregnancy. However, the change from the 1-year limit used in the eighth revision to the 42-day limit used in the ninth revision did not greatly affect the comparability of maternal mortality statistics (4). *** The ICD code is revised approximately every 10 years. In the ninth revision, ectopic pregnancy (ICDA code 631) was transferred from complications of pregnancy (ICDA codes 630--634) to pregnancy with abortive outcomes (ICD codes 630--638) (5). In this report, maternal deaths from ectopic pregnancy are included with abortive outcomes for all time periods. Proportionate Mortality from Pulmonary Tuberculosis Associated With Occupations -- 28 States, 1979-1990 The risk for occupational exposure to tuberculosis (TB) is increased among health-care and other workers exposed to persons with active TB, workers exposed to silica or other agents that increase the risk of progression from latent infection to active TB, and workers in occupations associated with low socioeconomic status (SES). Accurate estimates of and surveillance for occupationally associated TB are limited because reports of incident TB cases lack comprehensive occupational data (1). Although occupation is routinely recorded on death certificates, this information is not routinely coded and entered into vital statistics data files. To identify occupations associated with increased risk for TB mortality, CDC's National Institute for Occupational Safety and Health (NIOSH) used data from the National Occupational Mortality Surveillance (NOMS) database* to conduct a proportionate mortality study of persons with pulmonary TB by occupation for 1979-1990 (the most recent year for which data were available). This report presents the findings of the study. Data collected in the NOMS database include each decedent's usual industry and occupation, coded using 1980 U.S. census codes (2). During 1979-1990, approximately 3.4 million mortality records that included a usual occupation (excluding "housewife") were reported to NOMS. For this study, data for blacks and whites were analyzed separately because of substantial differences in race-specific rates of pulmonary TB (1,3); numbers of deaths for other racial groups were too small to calculate stable estimates. Ethnicity was not routinely coded on death certificates until 1989 and could not be analyzed. Indirectly age-standardized, race- and sex-specific proportionate mortality ratios (PMRs) were calculated by comparing the proportion of pulmonary TB deaths (International Classification of Diseases, Ninth Revision, code 011) among decedents in each occupational group to the proportion of pulmonary TB deaths among all decedents with a coded occupation (4). Confidence intervals (CIs) were calculated using the Mantel-Haenszel chi-square test or, when less than 1000 deaths were observed, the variance from a Poisson distribution. This analysis examined a total of 458 separate or combined occupational groups, comprising all 503 census occupational codes; of these 458 groups, 329 (71.8%) had one or more deaths attributed to pulmonary TB. Data are presented for the 21 occupational groups in which at least one of the race- and sex-specific categories had both 1) four or more pulmonary TB deaths and 2) either a PMR greater than 200 or a PMR with a 95% CI that did not include 100. From 1979 through 1990, a total of 2206 deaths was attributed to pulmonary TB. Of these deaths, 1024 (46.4%) occurred among workers in the 21 occupational groups that met the selection criteria. The 21 occupational groups were categorized into four risk groups (Table 1): 1) high potential for exposure to persons with TB (based on published reports and NIOSH health hazard evaluations [HHEs]**); 2) potential for substantial exposure to silica (determined using unpublished data from the National Occupational Exposure Survey [5] and the National Occupational Health Survey of Mining [6]); 3) low SES occupation (defined as a Nam-Powers socioeconomic index score less than 30 [where 1 signifies the lowest possible SES occupation and 100 the highest]) (7) without other recognized risk factors; and 4) unknown risk factors. Of the 21 occupational groups that met the selection criteria, two race- and sex- specific groups were associated with potential workplace exposure to persons with TB: white male health services workers (health and nursing aides, orderlies, and attendants) (PMR=350; seven deaths)*** and white male funeral directors (PMR=299; four deaths) (Table 1). Six of the 21 occupational groups were associated with potential for high silica exposure. For white males, these groups comprised mining machine operators; operators of machinery used to grind, abrade, buff, or polish; nonconstruction laborers; and construction workers (particularly brick and stone masons, construction laborers, carpenters, and roofers); PMRs ranged from 134 (169 deaths) to 290 (six deaths). For black males, these groups comprised construction workers (particularly construction laborers); mixing and blending machine operators; and furnace, kiln, and oven operators, except food; PMRs ranged from 128 (105 deaths) to 376 (five deaths). For two of the occupations associated with low SES that met the selection criteria (food preparation and service workers [particularly bartenders and cooks] and farm workers), previous reports have documented increased risk for TB (8). The other low SES occupations that met the selection criteria (e.g., housekeepers and butlers and nonfarm animal caretakers) have not previously been associated with increased risk for TB. Reported by: Surveillance Br, Div of Surveillance, Hazard Evaluations and Field Studies, and Epidemiological Investigations Br, Div of Respiratory Disease Studies, National Institute for Occupational Safety and Health, CDC. Editorial Note: In comparison with surveillance for TB based on incident symptomatic cases or TB skin-test conversions, mortality-based surveillance may be a relatively insensitive indicator of current occupational risks. However, except for studies of selected occupations, the findings in this report comprise the only available information about possible associations between occupation and TB infection. Furthermore, these findings identify several additional occupations associated with increased risk for TB, for which there are biologically plausible explanations. For example, the increased risk for funeral directors may reflect increased likelihood of exposure to infection from cadavers, and for mining machine operators, may reflect exposure to silica, which increases the risk of progression from latent infection to active TB. For occupational groups categorized "unknown risk," reasons for their elevated PMRs cannot be explained. These groups may be at increased risk through factors other than the three recognized in this analysis or may show elevated PMRs by chance alone. The findings in this report are subject to at least four limitations. First, mortality-based surveillance data are not sensitive indicators of risk for disease because mortality is affected by a combination of several interacting factors. In particular, mortality from TB reflects exposure, infection, SES, access to and adequacy of medical care, and underlying medical conditions. Overall, such factors probably contributed to the approximately threefold greater proportion of pulmonary TB-related mortality among blacks than among whites in this study (0.18% and 0.05%, respectively). Second, the method of death certificate-based PMR analysis described in this report is subject to possible misclassification of usual occupation and cause of death and to potential biases inherent in the use of the PMR statistic as a risk estimator (4) and fails to compensate statistically for multiple comparisons. Third, death certificates lack information about lifestyle and other risk factors; therefore, no adjustment was made for possible confounding. Fourth, the timeliness of death certificates as a source of data is constrained by processing delays. Despite these limitations, these findings are important because of the strength and biological plausibility of many of the reported associations. The recent increase in TB incidence and the occurrence of multidrug-resistant TB have focused attention on particular populations at high risk for disease and on the potential for transmission of infection to health-care workers. Occupation-based surveillance for TB can assist in identifying groups at high risk and indicate trends in the occurrence of infection in these workers. These surveillance findings also can assist in evaluating the effectiveness of prevention measures for groups previously established to be at risk, such as miners and agricultural workers. To improve the detection and control of TB among occupational groups, CDC has proposed two surveillance activities (9): 1) collection and analysis of occupational information from TB case reports and 2) serial cross-sectional surveys of known high-risk populations (including selected occupations) to determine the prevalence of TB skin-test positivity. These data, in combination with ongoing collection and analysis of occupational information from death certificates, will enable comprehensive surveillance for monitoring recognized high-risk populations and identifying new at-risk groups. References 1. CDC. Tuberculosis morbidity--United States, 1992. MMWR 1993;42:696-7,703-4. 2. Bureau of the Census. 1980 Census of population: alphabetical index of industries and occupations--final edition. Washington, DC: US Department of Commerce, Bureau of the Census, 1982; publication no. PHC80-R3. 3. CDC. Prevention and control of tuberculosis in U.S. communities with at-risk minority populations: recommendations of the Advisory Council for the Elimination of Tuberculosis. MMWR 1992;41(no. RR-5):1-11. 4. Decoufle P, Thomas T, Pickle L. Comparison of the proportionate mortality ratio and standardized mortality ratio risk measures. Am J Epidemiol 1980;111:263-9. 5. Seta JA, Sundin DS, Pedersen DH. National Occupational Exposure Survey, Vol 1. Cincinnati: US Department of Health and Human Services, Public Health Service, CDC, NIOSH, 1988; DHHS publication no. (NIOSH)88-106. 6. Groce DW, Carr WG, Hearl FJ. The National Occupational Health Survey of Mining. In: CDC surveillance summaries (August). MMWR 1986;35(no. SS-2):17. 7. Nam CB, Terrie EW. Comparing the Nam-Powers and Duncan SEI occupational scores. Tallahassee, Florida: Florida State University, Center for the Study of Population, 1986. 8. Weeks JL, Levy BS, Wagner GN, eds. Tuberculosis. In: Preventing occupational disease and injury. Washington, DC: American Public Health Association, 1991:572-81. 9. CDC. National action plan to combat multidrug-resistant tuberculosis. MMWR 1992;41(no. RR-11):1-48. * Through a collaborative project with CDC (specifically NIOSH and the National Center for Health Statistics) and the National Cancer Institute, the following 28 states contributed occupation-coded death certificate data to NOMS for 1 year or more during 1979-- 1990: Alaska, California, Colorado, Georgia, Idaho, Indiana, Kansas, Kentucky, Maine, Missouri, Nebraska, Nevada, New Hampshire, New Jersey, New Mexico, New York, North Carolina, Ohio, Oklahoma, Pennsylvania, Rhode Island, South Carolina, Tennessee, Utah, Vermont, Washington, West Virginia, and Wisconsin. ** The NIOSH HHE program is a federally mandated program for investigation of suspected hazardous exposures or disease outbreaks in the workplace. As of March 1994, a total of 50 HHE requests related to TB had been received by NIOSH and 25 completed (NIOSH, unpublished data, 1994). *** Based on the criteria used in this analysis, no other health-care workers had elevated PMRs. Erratum: Vol. 43, No. 40 In "Adult Blood Lead Epidemiology and Surveillance--United States, Second Quarter, 1994," on page 741, the first sentence of the second paragraph should read "The cumulative number of BLL reports for the first and second quarters of 1994 increased 48% over the comparable time period for 1993 (Table 1)." On pages 741- 742, the last sentence of the third paragraph should be deleted. The following table contains the corrected numbers and footnotes and replaces Table 1 on page 742. * DISTRIBUTED BY GENA/aegis (714.248.2836 * 8N1/Full Duplex * v.34; your online global HIV/AIDS gateway to people, information, and resources.