The following are pre-publication drafts of articles from the Morbidity and Mortality Weekly Report dated July 7, 1995. Late-breaking articles, and final editorial revisions are not included; therefore, these articles should be considered preliminary, and not to be released to the public. --CDC -------------------------------------------------------------- Farm-Tractor-Related Fatalities -- Kentucky, 1994 Fatalities associated with farm tractors are the most common cause of work-related death in the U.S. agricultural industry (1). To characterize farm-tractor-related fatalities in Kentucky, the Kentucky Fatality Assessment and Control Evaluation (KY FACE) Project studied all fatal farm injuries occurring among persons in that state during 1994, the initial year of operation for FACE in Kentucky. This report summarizes the results of that study. KY FACE is part of a 14-state surveillance and investigation program coordinated by CDC's National Institute for Occupational Safety and Health (NIOSH) and is designed both to evaluate the circumstances of fatal occupational injuries and to develop prevention strategies. KY FACE employs multiple reporting sources* to identify occupational fatalities throughout the state and conducts follow-up investigations. A farm-tractor-related fatality was defined as a death caused by operating or working on or near a farm tractor. A farm tractor was defined as a two- or four-wheel-drive vehicle or track vehicle with a greater than 20-horsepower engine designed to furnish the power to pull, carry, propel, or drive implements designed for agricultural activities (2). During 1994, the KY FACE surveillance system identified 28 tractor-related fatalities in Kentucky; 14 (50%) of these incidents occurred during June-August. Tractor-related fatalities accounted for 16% of the 176 occupational fatalities recorded in Kentucky during 1994. The most common cause of tractor-related fatalities was rollover (23 [82%]), followed by runover (five [18%]). The most common activity at the time of injury was mowing with a rotary mower trailing a tractor (i.e., bush-hogging) on private farms (32%). Other activities included transporting equipment or farm products (21%); checking livestock or property (14%); pulling logs (11%); and planting, plowing, or cutting hay (11%). Of the 28 deaths, 23 (82%) occurred on farms, and five (18%) occurred on public roadways. Four of those occurring on roadways were attributed to loss of control; one tractor was struck by a truck in a rear-end collision. All decedents were males who ranged in age from 15 to 86 years (median: 46 years); one was aged less than 18 years, and 15 (54%), greater than or equal to 60 years. One death occurred in a 15-year-old student who was killed in a tractor rollover incident while working a summer job plowing tobacco. Farming was listed as the usual occupation on 11 (39%) of the 28 death certificates. Ten (36%) of those fatally injured also held jobs off the farm, and 12 (43%) were retired from nonfarming occupations. Most (53%) fatalities occurred from 12:01 to 6 p.m.; 32% occurred from 7 a.m. and noon, and 14% after 6 p.m. An industrial hygienist conducted on-site investigations of 16 of the incidents. Tractors involved in these 16 incidents ranged in age from 2 to 41 years (median: 23 years). In three of the cases, the operators were driving directly up or down steep slopes (of 8, 14, and 30 degrees); in two of these incidents, the operator lost control while descending, and in the third, the operator rolled over backward while ascending a hill. In eight of the 16 incidents, one or both wheels on one side of the tractor slid down an embankment, causing a rollover. In one case, the operator backed the tractor over an embankment, causing the tractor to roll over backwards. In eight of the incidents, tires were air-filled rather than fluid-filled; fluid-filled tires lower the center of gravity, improve traction, and can prevent skidding, loss of control, and rollover. Only two of the tractors were equipped with front-end counterweights, which improve traction and stability. In eight cases, poor equipment condition (e.g., minimally operable brakes), was a contributing factor. Only one of the tractors involved in a rollover fatality was fitted with a rollover-protective structure (ROPS); in this incident, a tractor manufactured in 1962 had been retrofitted with a ROPS but not equipped with seatbelts. Reported by: TW Struttmann, MSPH, C Spurlock, PhD, SH Pollack, MD, E Moon-Hampton, Kentucky Injury Prevention and Research Center; SR Browning, PhD, R McKnight, ScD, Southeast Center for Agricultural Health and Injury Prevention, Univ of Kentucky, Lexington. R Finger, MD, State Epidemiologist, Dept for Health Svcs, Kentucky Cabinet for Human Resources. Div of Safety Research, National Institute for Occupational Safety and Health, CDC. Editorial Note: During 1994, the fatality rate for civilian workers in the agriculture/forestry/fishing industry in Kentucky was 85 per 100,000 workers, a rate more than three times greater than that for the industry in the United States (26 per 100,000 workers in 1993) (3). Operating tractors is a particularly hazardous activity for older workers and adolescents. The proportion of Kentucky tractor-related fatalities among workers aged greater than 60 years (54%) was greater than that reported in the NIOSH National Traumatic Occupational Fatalities surveillance system (44%) (4). Operating tractors with a greater than 20-horsepower engine is extremely hazardous to youth, and federal Child Labor Laws prohibit this activity for employees aged less than 16 years; however, children working on their family farm are exempt from Child Labor Laws. In 1994, tractor rollovers and runovers accounted for 62% of agricultural fatalities in Kentucky. The findings of the KY FACE investigations indicated that in most of the incidents rollover fatalities could have been prevented if the tractors had been equipped with ROPS (Figure 1) and the operators secured with seatbelts, which ensure that the operator remains within the ROPS-protected zone during a rollover. ROPS first became available as optional equipment on farm tractors in 1971 (tractors manufactured before 1971 were not designed to accommodate ROPS devices). However, ROPS were not required for new tractors until 1976, when a standard promulgated by the Occupational Safety and Health Administration (OSHA) required employers to provide ROPS and seatbelts for all employee-operated tractors** manufactured after October 25, 1976 (2). Although virtually all tractors sold after 1985 have been equipped with ROPS, farms with less than 11 employees are not subject to OSHA inspection or enforcement, and farms managed by family members with no other employees are not required to comply with OSHA standards; in Kentucky, 94% of the farms are family-owned businesses with less than 11 employees (5). The median age of tractors investigated in this report was 23 years. One fatal tractor rollover in this study involved a 1979 tractor manufactured without ROPS. Because it was purchased for use on a family farm without employees, it was not subject to the ROPS standard. The cost to retrofit tractors manufactured before 1975 ranges from $400 to $1800, and economic constraints associated with farms in Kentucky limit the feasibility of appropriately modifying all tractors. The findings of KY FACE suggest that installation of ROPS and seatbelts on farm tractors could have prevented the 23 tractor rollover deaths. These findings and previous reports (1) underscore the need for economically feasible ROPS retrofit programs. In Kentucky, the FACE program disseminates reports containing investigative findings and recommends intervention strategies to county extension agents, the Kentucky Labor Cabinet Division of Education and Training, the Kentucky Farm Bureau, and the National Safety Council. News media releases assist in disseminating this information further to the agriculture community and the general public. References 1. CDC. Public health focus: effectiveness of rollover protective structures for preventing injuries associated with agricultural tractors. MMWR 1993;42:57-9. 2. Office of the Federal Register. Code of federal regulations: occupational safety and health standards. Subpart C: roll-over protective structures (ROPS) for tractors in agricultural operations. Washington DC: Office of the Federal Register, National Archives and Records Administration, 1994 (29 CFR *** 1928.51). 3. Toscano G, Windau J. The changing character of fatal work injuries. Monthly Labor Review 1994;117(10):17-28. 4. Etherton JR, Myers JR, Jensen RC, Russell JC, Braddee RW. Agricultural machine-related deaths. Am J Public Health 1991;81:766-8. 5. Bureau of the Census. 1992 Census of agriculture: Vol 1, Geographic Area Series, Part 17, Kentucky State and County Data. Washington, DC: US Department of Commerce, Economics and Statistics Administration, 1992 (AC92-A-17). * Notification sources include newspapers, county coroners, emergency medical personnel, Kentucky Labor Cabinet, U.S. Bureau of Labor Statistics Census of Fatal Occupational Injuries, Kentucky Department of Motor Vehicles' Fatal Accident Reporting System, Southeast Center for Agricultural Health and Injury Prevention, Occupational Health Nurses in Agricultural Communities, and Kentucky Vital Statistics. ** The standard provides exemptions for tractors used in special circumstances where vertical clearances may be limited (e.g., in orchards or inside buildings). Mass Treatment of Humans Exposed to Rabies -- New Hampshire, 1994 On October 22, 1994, the laboratory of the New Hampshire Division of Public Health Services (NHDPHS) diagnosed rabies in a kitten that had been purchased from a pet store in Concord, New Hampshire. On October 19, the animal had developed seizures, then died of unknown causes during the night of October 20-21. Approximately 665 persons received rabies postexposure prophylaxis because of exposure to this kitten and other cats from the same pet store. This report summarizes the epidemiologic investigation of the source of the infection and follow-up care of humans and animals potentially exposed to rabies. Because the pet store did not keep records for kittens acquired for sale, the kitten's origin and date of arrival were unknown. However, on September 26, a group of kittens reported to have included the rabid kitten was examined by a veterinarian and given health certificates, in accordance with state law, before being offered for sale by the pet store. The kitten was sold on October 5 and kept by its owners until its death. On October 22, rabies was diagnosed in the kitten by fluorescent antibody testing at the NHDPHS laboratory. At CDC, genetic typing of the rabies virus isolated from the kitten indicated that it was a variant associated with raccoons. The investigation could not determine whether the kitten was infected with rabies before, during, or after its stay in the pet store; two other kittens sold by the pet store during the same period as the infected kitten died of unknown causes at their new homes but were unavailable for testing for rabies. On October 12, a raccoon captured in Henniker (a suburb of Concord), where the kitten was suspected to have originated, tested positive for rabies. Subsequent investigation indicated that the raccoon may have had direct contact with three feral kittens acquired by the pet store on September 20. All three feral kittens developed signs of respiratory illness and died during approximately October 4-October 6--a period overlapping that during which the rabid kitten was in the store. None of these three kittens were available for testing for rabies and all were younger than the minimum age (3 months) recommended for rabies vaccination. From September 19 through October 23 (the last date any potentially exposed kittens were in the pet store), a minimum of 34 kittens had been offered for sale by the store. In addition to the infected kitten, 33 other kittens were included in the investigation: 27 were located and tested negative for rabies, and five died of unknown causes but were unavailable for testing (including the three feral kittens); one kitten was quarantined at the owner's request, and its status is unknown. Because of limitations in the store's records regarding the origins and sale destinations of the kittens, local news media assisted in alerting community residents about the potential exposures to rabies at the store. The kittens had been allowed to roam freely throughout the store, which was frequented by children from child-care centers and a nearby school. As a result, NHDPHS and two major health-care facilities screened approximately 1000 persons who responded to media alerts and referred to private sector health-care providers for definitive evaluation of those persons who might need rabies postexposure treatment. NHDPHS gave medical providers an algorithm to determine the necessity for recommending rabies postexposure treatment. Rabies postexposure treatment, consisting of one dose of rabies immune globulin and five doses of rabies vaccine, was initiated for approximately 665 persons (1). Reported by: M Klauber, Dartmouth-Hitchcock Medical Center, Concord; C McGinnis, DVM, New Hampshire Dept of Agriculture; RT DiPentima, MPH, AE Burns, MS, VC Malmberg, MS, JS Finnigan, MJ Walsh, JE Whitcomb, CE Danielson, MD, MG Smith, MD, State Epidemiologist, Div of Public Health Svcs, New Hampshire Dept of Health and Human Svcs. Viral and Rickettsial Zoonoses Br, Div of Viral and Rickettsial Diseases, National Center for Infectious Diseases, CDC. Editorial Note: This incident of rabies associated with a pet store resulted in the largest number of persons ever reported to have received rabies postexposure treatment as a result of potential contact with a point source in the United States.* At least three factors accounted for the large number of persons requiring treatment. First, the absence of pet store records regarding the source and destination of animals precluded an accurate estimation of the exposure period. Second, the store's popularity and its practice of allowing kittens to roam freely throughout the establishment increased contacts between humans and kittens. Finally, because many children were potentially exposed, accurate exposure histories could not be elicited; as a consequence, many rabies postexposure treatments were administered on the basis of incomplete information or unknown likelihood of exposure. The costs associated with the public health response to exposures to the rabid kitten in New Hampshire are unprecedented in the United States. The overall estimated cost was $1.5 million, including expenditures for rabies immune globulin and vaccine ($1.1 million), laboratory testing of animals ($4200), and investigation by NHDPHS and CDC personnel ($15,000). This cost is nearly 15-fold higher than that ($105,790) associated with rabies postexposure treatment of 70 persons after a single case of rabies occurred in a domestic dog in California in 1981 (2). CDC recommends implementation of four measures to minimize the number of exposed persons and the costs associated with exposures to persons. First, to facilitate efforts to investigate such exposures, pet stores should keep adequate records (e.g., health certificates, animal source identification, and complete sales receipts). Second, to prevent the exposure to, or the transmission of, rabies and other zoonotic diseases--as well as injuries such as bites and scratches--animals should be kept and displayed separate from customers or at least confined to a discrete area within the store. Third, because feral animals are less likely to have been vaccinated and more likely to have been in contact with wildlife disease reservoirs, acquisition and sale of these animals should be monitored closely. Finally, prompt and standardized assessment of exposure by public health officials should help minimize the number of persons who unnecessarily receive rabies postexposure treatment. The rabies virus is transmitted only when introduced into open wounds or mucous membranes through a bite or direct saliva contact. Other forms of contact (e.g., petting a rabid animal or contact with blood, urine, or feces of a rabid animal) do not constitute an exposure and are not indications for prophylaxis (1). Skillful interviewing is essential to assess individual exposures, especially when the potential exposure occurred some time ago or in another family member (e.g., a young child). The rabid kitten involved in this incident had been infected with a rabies virus variant usually associated with raccoons. Since 1977, a raccoon rabies epizootic has spread from a focus in West Virginia to involve all eastern region states (3). During 1993, nearly 6000 raccoons were confirmed with rabies in this region. Although no human rabies cases have been associated with this epizootic, the economic burden related to postexposure prophylaxis has been high. This epizootic and case described in this report underscore the need for intensification of rabies-control measures, including vaccination of all household pets. References 1. ACIP. Rabies prevention--United States, 1991: recommendations of the Immunization Practices Advisory Committee (ACIP). MMWR 1991;40(no. RR-3). 2. CDC. The cost of one rabid dog--California. MMWR 1981;30:527. 3. Krebs JW, Strine TW, Smith JS, Rupprecht CE, Childs JE. Rabies surveillance in the United States during 1993. J Am Vet Med Assoc 1994;205:1695-1709. * On June 22, a second episode of rabies in a kitten associated with a pet store was reported by the Delaware Department of Health and Social Services. Through June 30, approximately 18 persons had received postexposure prophylaxis for potential rabies exposure. An investigation of this incident is ongoing. Measles -- United States, 1994 As of June 13, 1995, local and state health departments in 39 states had reported 958 measles cases to CDC for 1994. This represents the second lowest number of cases ever reported, after the historic low of 312 cases in 1993 (1). In addition, 303 cases were reported for the U.S. territory of Guam (228) and the commonwealths of the Northern Mariana Islands (29) and Puerto Rico (46). This report summarizes the epidemiologic characteristics of measles cases and outbreaks reported in the United States during 1994. Age distribution, complications, and hospitalizations. Of the 954 measles patients for whom age was known, 247 (26%) were aged less than 5 years, including 73 (8%) who were aged less than 12 months and 69 (7%) who were aged 12-15 months. Nearly one half (475) of all measles patients were aged 5-19 years, and 232 (24%) were aged greater than or equal to 20 years. Among the 537 measles patients for whom information was available, 45 (8%) were reported to have been hospitalized; the median duration of hospitalization was 4 days (range: 1-22 days). Among 338 (35%) measles cases for which information on laboratory testing was provided, 229 (68%) were serologically confirmed. Vaccination status. Vaccination status was reported for 848 (89%) measles patients. Among 762 vaccine-eligible persons,* 171 (22%) were reported to have documented receipt of at least one dose of measles-containing vaccine, and 539 (71%) were unvaccinated. Fifty-two (7%) persons with reported unknown vaccination status were considered to be unvaccinated. Four cases occurred among persons with documentation of two appropriately spaced doses of measles vaccine greater than 14 days before onset of symptoms. Among 301 unvaccinated measles patients who were eligible for vaccination and for whom a reason for nonvaccination was reported, 294 (98%) cited a religious (154 [51%]) or philosophic (140 [47%]) exemption to vaccination. Almost all (92% [277]) of these cases occurred in outbreaks in Illinois, Missouri (2), Nevada, and Utah. Cases among persons claiming religious or philosophic exemption to vaccination accounted for 36% of all reported cases in 1994. Case classification. Among 949 reported cases for which the epidemiologic classification is known, 874 (92%) were indigenous to the United States, including 719 (76%) acquired in the state reporting the case and 155 (16%) resulting from spread from known importation from another state. International importations and cases occurring within two generations of these importations accounted for 75 (8%) measles cases in 1994. These cases were reported from 24 states and, for those for whom the country of origin was reported, occurred most frequently among persons arriving from Europe (26 cases) and East Asia (18). Cases resulted from importations from the Americas (eight), the Middle East (six), and Africa (two). Among the 75 persons with internationally imported measles, 23 (31%) were aged less than 5 years; 32 (43%), 5-19 years; and 20 (27%), greater than or equal to 20 years. Outbreaks. Twenty-two outbreaks (clusters of five or more epidemiologically linked cases) were reported by 15 states during 1994 and accounted for 74% (705) of all reported cases. Two of these outbreaks began in 1994 and continued into 1995 (only cases that occurred during 1994 are reported here). Eight outbreaks, which included 12-156 cases, occurred in schools (six outbreaks) or colleges (two), five outbreaks (range: five-22 cases) involved predominantly preschool-aged children, and nine (range: six-134 cases) occurred in other settings and primarily involved young adults. The largest college outbreak (94 cases) resulted from spread from an importation, and two other outbreaks followed known importations. A total of 176 cases (18% of all reported cases) were related to international importations in 1994. A single chain of transmission that was first recognized in a Colorado ski resort (3) extended into nine additional states and resulted in the largest outbreak of 1994 (247 cases); this outbreak involved students who were unvaccinated because of religious exemptions and who attended a college in Illinois or a school in Missouri (2). Two other outbreaks involving persons with philosophic exemption to vaccination occurred in Salt Lake City, Utah (134 cases), and White Pine County, Nevada (12 cases). In outbreaks among persons with religious or philosophic exemption to vaccination, school-aged children accounted for 73% of all cases, and represented 56% of all measles cases among 5-19-year-olds in 1994. Intensive surveillance and case investigation resulted in identification of three large multistate outbreaks during 1994. Epidemiologic linkages were established among 247 cases in 10 states from the outbreak that began in Colorado, among 57 cases in six states resulting from exposures in Las Vegas, and among 146 cases from an outbreak that began in Utah and spread to Nevada. The genomic sequences of viruses isolated from the outbreak in Illinois and Missouri was similar to that of a virus isolated from an earlier outbreak in Memphis, Tennessee. These viruses probably were recently imported into the United States because they were closely related to measles virus strains that had previously circulated in Europe. Four distinct genotypes were identified by genomic sequencing among 10 isolates from four outbreaks and three single measles cases in the United States in 1994. None of these was related to the genotype circulating during the resurgence of 1989-1991, suggesting that all of these viruses were introduced into the United States as a result of importation. Reported by: State and local health depts. Measles Virus Section, Respiratory and Enteric Viruses Br, Div of Viral and Rickettsial Diseases, National Center for Infectious Diseases; National Immunization Program, CDC. Editorial Note: Although measles incidence has increased since the historic low reported in 1993, the number of cases reported during 1994 is the second fewest in the United States since measles reporting began in 1912. Important characteristics of current epidemiologic trends are the shift in age distribution of cases to older persons, the large proportion of cases in groups whose members do not routinely accept vaccination, and the increasing numbers of cases linked to international importations. Since the measles resurgence of 1989-1991, increasing proportions of cases have occurred among school-aged children and adults, and proportionately fewer in preschool-aged children--a substantial change from 1989-1991, when incidence was highest among preschool-aged children, of whom as many as 80% were unvaccinated (4,5). The shift in age distribution probably resulted from record-high measles vaccination coverage levels among preschool-aged children, which reached 90% in the first quarter of 1994 (6). More than half of the cases in persons aged 5-19 years were associated with outbreaks among persons with a religious or philosophic exemption to vaccination. Additional efforts will be necessary to reduce transmission among persons with objections to vaccination. Laboratory and epidemiologic data suggest that measles transmission was interrupted in the United States during late 1993 (7). Because of the effective implementation of a strategy of mass vaccination of children in all countries in Central and South America, importations from the Americas have decreased substantially since 1991 and now represent a small percentage of all importations. However, the continued risk for international importations and spread from importations from other locations represent a challenge to the goal of measles elimination in the United States; known international importations or spread from international importations accounted for almost one fifth of reported measles cases in 1994. The strategy for achieving the Childhood Immunization Initiative goal of eliminating indigenous measles transmission in the United States (8) is based on four components: 1) maintaining high coverage with a single dose of measles-mumps- rubella vaccine (MMR) among preschool-aged children, 2) achieving coverage with two doses of MMR for all school and college attendees, 3) enhancing surveillance and outbreak response, and 4) increasing efforts to develop and implement strategies for global measles elimination. CDC will continue to work with state and local health departments to implement recommendations to achieve high levels of population immunity, rapidly report and investigate all suspected measles cases, and enhance surveillance to facilitate rapid identification and confirmation of cases and implementation of appropriate control measures. References 1. CDC. Table II. Cases of selected notifiable diseases, United States, weeks ending December 31, 1994 and January 1, 1994 (52nd week). MMWR 1995;43:969. 2. CDC. Outbreak of measles among Christian Science students--Missouri and Illinois, 1994. MMWR 1994;43:463-5. 3. CDC. Interstate measles transmission from a ski resort--Colorado, 1994. MMWR 1994;43:627-9. 4. Gindler JS, Atkinson WL, Markowitz LE, Hutchins SS. Epidemiology of measles in the United states in 1989 and 1990. Pediatr Infect Dis J 1992;11:841-6. 5. CDC. Measles surveillance--United States, 1991. MMWR 1992;41(no. SS-6):1-12. 6. CDC. Vaccination coverage of 2-year-old children--United States, January-March 1994. MMWR 1995;44:142-3,149-50. 7. CDC. Absence of reported measles--United States, November 1993. MMWR 1993;42:925-6. 8. CDC. Reported vaccine-preventable diseases--United States, 1993, and the Childhood Immunization Initiative. MMWR 1994;43:57-60. * Persons aged greater than or equal to 12 months who were born after 1957. Persons born in or before 1957 are considered to be immune based on the likelihood of their having had measles before licensure of measles vaccine in 1963. Prevalence of Smoking by Area of Residence -- Missouri, 1989-1991 Variation in smoking prevalence by area of residence may be an important consideration in the development, implementation, and management of programs that promote nonsmoking. In general, the prevalence of cigarette smoking is highest among persons at economic, educational, and social disadvantage (1,2), and the proportion of persons who are disadvantaged is greater in urban and nonmetropolitan areas. Because smoking prevalence varies by area of residence and characterization of these differences can assist in directing efforts to promote nonsmoking, the Missouri Department of Health compared urban, suburban, and nonmetropolitan areas using data from two sources: the Behavioral Risk Factor Surveillance System (BRFSS) for Missouri from 1989 through 1991 (suburban and nonmetropolitan areas) and a survey specially commissioned in 1990 (Smoking Cessation in Black Americans [SCBA]) of persons living in low-income census tracts in north St. Louis and central Kansas City (urban areas). This report summarizes the results of this analysis. BRFSS is a population-based, random-digit-dialed telephone survey of the civilian, noninstitutionalized population aged greater than or equal to 18 years (3). For this analysis, respondents' suburban or nonmetropolitan residence was determined by county of residence: respondents not living in counties composing a metropolitan statistical area (MSA) were categorized as residing in nonmetropolitan areas; respondents living in counties composing MSAs were categorized as residing in suburban areas. Persons living in the urban areas of St. Louis or Kansas City (Jackson County) were excluded from the BRFSS data. However, the SCBA survey was conducted in 60 low-income census tracts to determine smoking prevalence and attitudes among residents of these areas (4). To estimate prevalences, BRFSS data were weighted to reflect the total population in each area (based on the 1990 census) and for respondent probability of selection. Based on the 1990 census, 46% of persons resided in suburban areas, 34% in nonmetropolitan areas, and 20% in St. Louis and Kansas City. BRFSS data were aggregated for 3 survey years to increase the number of respondents in the demographic categories* for the suburban and nonmetropolitan areas, and SUDAAN was used to calculate the variance (5). For both the BRFSS and SCBA, current smokers were defined as persons who had smoked greater than or equal to 100 cigarettes and who reported being a smoker at the time of the interview. The prevalence of cessation was obtained by dividing the number of former smokers by the number of ever smokers (respondents who have ever smoked greater than or equal to 100 cigarettes during their lifetime) and multiplying by 100. Differences in group-specific prevalence rates in this report reflect nonoverlapping confidence intervals. Overall, the prevalence of current smoking was higher among persons residing in the urban areas (32.4%) than in the suburban (24.8%) and nonmetropolitan areas (26.5%) (Table 1). This pattern was consistent across all sex and education subgroups. The prevalence of current smoking also was higher in the urban areas for adults aged 35-54 years and greater than or equal to 55 years. For the 18-34-year age group, the prevalence of current smoking in the urban areas (31.3%) was comparable to that in the suburban (27.8%) and nonmetropolitan (33.5%) areas. For whites, the prevalence of current smoking was higher for those living in the urban areas (34.8%) than in suburban (24.9%) or nonmetropolitan (26.0%) areas. For blacks, the prevalence of current smoking was similar in urban areas (32.0%) and nonmetropolitan areas (32.1%) but higher than in suburban areas (24.0%). Among current smokers, the mean number of cigarettes smoked per day was highest in the nonmetropolitan areas (22.8), lowest in the urban areas (15.0), and intermediate in suburban areas (19.9). The prevalence of cessation was lower in the urban areas (37.4%) than in the suburban (50.0%) or nonmetropolitan areas (47.6%). Reported by: CL Arfken, PhD, W Auslander, PhD, EB Fisher, Jr, PhD, Center for Health Behavior Research, Washington Univ School of Medicine, St. Louis; RC Brownson, PhD, School of Public Health, St. Louis Univ; J Jackson-Thompson, PhD, B Malone, MPA, Div of Chronic Disease Prevention and Health Promotion, Missouri Dept of Health. Office on Smoking and Health, National Center for Chronic Disease Prevention and Health Promotion, CDC. Editorial Note: In Missouri during 1989-1991, the prevalence of smoking generally was highest in urban areas regardless of sex, education, age, and race. These findings are consistent with those of previous reports describing the relation between urban area of residence and smoking status (6,7). The persistence of the association between urban residence and smoking status, despite controlling for demographic characteristics, suggests that other factors contribute to the higher prevalence of smoking in urban areas. Such factors may include cultural norms, the burden and management of stress (8), relative effectiveness of risk-reduction messages (9), and exposure to tobacco advertisement and promotions. Differences in prevalences among racial/ethnic groups may be influenced by differences in educational levels, socioeconomic status, and social and cultural phenomena that require further explanation. The findings in this report are subject to at least three limitations. First, because these estimates are based on self-reported data, prevalences may be underestimated (10). Second, a stratified analysis was conducted to control for each demographic variable individually because combining data from separate surveys with differing sampling designs precluded use of multivariate techniques to control for each variable simultaneously. Third, grouping areas at the urban, suburban, and nonmetropolitan levels may mask important community differences within each of these areas. The findings in Missouri suggest that urban areas are an important target for nonsmoking promotion efforts. In general, local survey data can provide useful information to assist state and local health departments in identifying populations for risk-reduction programs. In Missouri, state and local health departments and community organizations are using these findings to develop programs and activities to reduce the prevalence of smoking among urban residents. For example, in Kansas City, intensive education efforts have been initiated to change social and community norms about smoking through activities such as rallies and town hall meetings and the promulgation of nonsmoking regulations. In St. Louis, activities have included counter-advertising, public service announcements, tobacco education in schools, and training of health-care providers about tobacco-use prevention. References 1. Fisher E Jr, Lichenstein E, Haire-Joshu D. Multiple determinants of tobacco use and cessation. In: Orleans C, Slade JD, eds. Nicotine addiction: principles and management. New York: Oxford, 1993. 2. Novotny TE, Warner KE, Kendrick JS, Remington PL. Smoking by blacks and whites: socioeconomic and demographic differences. Am J Public Health 1988;78:1187-9. 3. Remington PL, Smith MY, Williamson DF, Anda RF, Gentry EM, Hogelin GC. Design, characteristics, and usefulness of state-based behavioral risk factor surveillance: 1981-1987. Public Health Rep 1988;103:366-75. 4. Brownson RC, Jackson-Thompson J, Wilkerson JC, Davis JR, Owens NW, Fisher EB Jr. Demographic and socioeconomic differences in beliefs about the health effects of smoking. Am J Public Health 1992;82:99-103. 5. Shah BV. Software for Survey Data Analysis (SUDAAN) version 5.5 [Software documentation]. Research Triangle Park, North Carolina: Research Triangle Institute, 1991. 6. Wechsler H, Gottlieb NH, Demone HW. Lifestyle, conditions of life, and health care in urban and suburban areas. Public Health Rep 1979;94:477-82. 7. Ingram DD, Gillum RF. Regional and urbanization differentials in coronary heart disease mortality in the United States, 1968-85. J Clin Epidemiol 1989;42:857-8. 8. Sclar ED. Community economic structure and individual well-being: a look behind the statistics. Int J Health Serv 1980;10:563-79. 9. Wing S, Casper M, Riggan W, Hayes C, Tyroler HA. Socioenvironmental characteristics associated with the onset of decline of ischemic heart disease mortality in the United States. Am J Public Health 1988;78:923-6. 10. Klesges L, Klesges R, Cigrang J. Discrepancies between self-reported smoking and carboxyhemoglobin: an analysis of the second National Health and Nutrition Survey. Am J Public Health 1992;82:1026-9.