INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Although Hispanic populations in the United States have been considered to be at disproportionately high risk for death from asthma in comparison with non-Hispanic whites (1, 2), to date, asthma mortality in Hispanic populations has not been well described at a national level. A previous report examining mortality in U.S. Hispanic populations from 1979 through 1981 was based on information from only 15 states (3). In 1984, only 22 states and the District of Columbia included an item on death certificates regarding Hispanic country of origin (4). Only since 1990 has information on Hispanic origin been available in the mortality data of most states.

The terms "Hispanic" or "Latino" encompass several different national subgroups that share a common language but that may differ in racial composition as well as characteristics particular to the country of origin (5). Given the effects of these characteristics upon health status, it may not be reasonable to present and discuss health outcomes for Hispanic populations as if they were a single monolithic group.

Using U.S. vital statistics data, we examined both national and regional mortality from asthma in U.S. Hispanic populations in comparison with that in non-Hispanic populations. We also compared mortality among Hispanic national groups to determine what differences, if any, exist.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Counts of asthma deaths from 1990 through 1995 were obtained using U.S. vital statistics data, which is compiled and maintained by the National Center for Health Statistics (NCHS). The information in the NCHS system is based upon death certificate information obtained from the 50 U.S. states and the District of Columbia and thus represents all deaths reported in the United States for a particular year (4). Deaths from asthma were those in which the underlying cause of death was coded 493 per the International Classification of Diseases, ninth edition (ICD-9) (6). For this analysis, we coded race of decedents as white, black, and other, and age as 0 to 4 yr, 5 to 14 yr, 15 to 34 yr, 35 to 64 yr, and 65 yr and older. These age strata matched those used in a summary of asthma surveillance data published recently by the Centers for Disease Control and Prevention (CDC) (7). The four U.S. geographic regionsNortheast, Midwest, South, and Westwere coded according to standard definitions developed by the U.S. Bureau of the Census (8) (see Appendix ) and were determined by the state of residence at the time of death.

Decedents were considered Hispanic if their ethnicity was recorded as Central or South American, Cuban, Mexican, Puerto Rican, or other Spanish descent on the death certificates. We defined separate subgroups for analysis based on race and Hispanic ethnicity as coded on the death certificates: non-Hispanic white, non-Hispanic black, non-Hispanic other races, and Hispanic (regardless of race). Within Hispanic decedents, four subgroups were analyzed: Hispanics of Mexican origin (Mexican-Americans), Hispanics of Cuban origin (Cuban-Americans), Hispanics of Puerto Rican origin (Puerto Ricans), and Hispanics of other origin, including Central and South Americans, other Spanish speakers, and other Hispanic-Americans. Because the racial composition of Hispanic subgroups may differ, we also examined the possible effect of race on asthma mortality for all Hispanics and by Hispanic subgroup.

In 1990, 47 U.S. states and the District of Columbia were reporting Hispanic ethnicity on death certificates; Louisiana, Oklahoma, and New Hampshire were the exceptions. By 1993, Louisiana and New Hampshire were reporting these data. For each year of data, we included in the analysis only those states that reported Hispanic country of origin on death certificates. For 1990-1992, we used data derived from 47 states and the District of Columbia, and for 1993 through 1995 we added data from Louisiana and New Hampshire. Thus, the analysis included data for 49 U.S. states and the District of Columbia, and excluded data for Puerto Ricans living on the island of Puerto Rico, as well as residents of other U.S. commonwealths, territories, and protectorates.

Denominators for computing rates were based upon U.S. Census data. Total and strata-specific population estimates for the years 1991 to 1994 were interpolated linearly between 1990 Decennial Census data and 1995 projected U.S. population data (9). Coding for Hispanic ethnicity was available in both the 1990 and 1995 data; however, coding for Hispanic country of origin was available only in the 1990 data. Therefore, to complete the interpolation of population estimates based upon country of origin for Hispanics, we approximated the proportional representation of Hispanic national groups within the entire U.S. Hispanic population for 1995. We assumed that, within each of the four regions, the proportion of Cuban-Americans, Mexican-Americans, Puerto Ricans, and other Hispanic-Americans within the entire regional Hispanic population remained constant by sex, race, and age group from 1990 through 1995. For any year, the population estimates included data only from states that provided Hispanic mortality data as described previously.

Crude and age-adjusted average annual mortality rates per million population were calculated using standard epidemiologic methods (10). To be consistent with the CDC asthma surveillance summary mentioned previously (7), age-adjusted rates were standardized to the 1970 U.S. population. The Statistical Analysis System (SAS) (11) was used for data management and analysis. Rates based on the average population during the 6-yr period were computed according to age group, race, sex, region, Hispanic ethnicity (Hispanic, non-Hispanic), and Hispanic country of origin.

All age groups were included in the analyses. Several previous published reports based on asthma mortality include only decedents 34 yr of age and younger (1, 12), citing the possibility of misdiagnosis in persons 0 to 4 yr of age because of viral bronchiolitis and wheezing illnesses in infancy and in persons 35 yr of age and older because of comorbid conditions such as bronchitis and chronic obstructive pulmonary disease (COPD) (13). Therefore, in sensitivity analyses, we compared results including only deaths occurring in persons 5 to 34 yr of age with those obtained when all age groups were included. Finally, we compared rates in the Northeast region with those in the other three regions combined.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

From 1990 through 1995, there were 30,785 deaths attributed to asthma in the United States. Of these, 21,715 were in non-Hispanic whites, 6,520 were in non-Hispanic blacks, 887 were in non-Hispanics of other races, and 1,663 were in Hispanics. Of the 1,663 Hispanic deaths from asthma, 567 (34%) were in Puerto Ricans, 522 (31%) were in Mexican-Americans, 161 (10%) were in Cuban-Americans, and 414 (25%) were in other Hispanic Americans. The proportional mortality among Hispanics of different national origins was in contrast to the groups' proportional representation in the Hispanic population. Thus, Puerto Ricans, who constituted 34% of reported asthma deaths in Hispanics, made up about 11% of all Hispanics living in the mainland United States in 1997 (14), whereas Hispanics of Mexican origin, who constituted 31% of reported asthma deaths in Hispanics, made up about 63% of the U.S. Hispanic population in 1997 (14).

U.S. asthma mortality by race, Hispanic ethnicity, sex, and geographic region is presented in Table 1. Non-Hispanic blacks had the highest overall and sex-specific asthma mortality rates in both the United States and all regions. In the Northeast, Hispanics had higher mortality rates than did non-Hispanic whites, whereas in other regions, Hispanics had asthma mortality rates that were similar to or lower than those among non-Hispanic whites.

Overall and sex-specific mortality for Hispanic subgroups, comparing rates in the Northeast with rates in the Midwest, South, and West combined is presented in Table 2. Rates for Puerto Ricans, Cuban-Americans, and other Hispanic-Americans are greater in the Northeast than in other regions combined; the number of deaths in Mexican-Americans in the Northeast, however, was too small to provide a stable estimate. Deaths from asthma in Puerto Ricans predominantly occurred in the Northeast; 81% of all asthma deaths reported for Puerto Ricans occurred there. In addition, 44% of all asthma deaths reported for other Hispanic-Americans occurred in the Northeast. Almost all Hispanic asthma deaths were among persons of white race. The total number of asthma deaths from 1990 through 1995 for Hispanic subgroups in the United States were as follows: Mexican-Americans: 514 white, 6 black, 2 other; Puerto Ricans: 553 white, 13 black, 1 other; Cuban-Americans: 152 white, 7 black, 2 other; other Hispanic-Americans: 379 white, 22 black, 13 other.

Age-specific asthma mortality among U.S. Hispanics by region is presented in Figure 1. Rates in the Northeast are higher than rates in other regions for all age strata, but particularly for those in the 5 to 14 yr, 15 to 34 yr, and 35 to 64 yr age strata. Age-specific rates for Hispanics in the Midwest, South, and West show comparable patterns.

View larger version (13K): [in this window] [in a new window]   Figure 1.   Average annual age-specific asthma mortality rates (log scale) in U.S. Hispanics, 1990-1995, by geographic region.

Age-specific mortality among Hispanic subgroups is presented in Figure 2. For all age strata, Puerto Ricans had the highest rates of asthma mortality. Mexican-Americans had the lowest asthma mortality rates in all age strata except the 0 to 4 yr stratum. Cuban-Americans and other Hispanic-Americans had age-specific asthma mortality rates that fell between the rates of the Puerto Ricans and Mexican-Americans.

View larger version (14K): [in this window] [in a new window]   Figure 2.   Average annual age-specific asthma mortality rates (log scale) in U.S. Hispanic subgroups, 1990-1995. * Represents mainland Puerto Ricans,  zero asthma deaths in Cuban-Americans aged 0-4 and 5-14 years for 1990-1995.

Age-specific mortality among Hispanic subgroups is presented in Figure 3, comparing the Northeast region with all other regions combined. Rates for Mexican-Americans are presented in Figure 3A. Rates for Puerto Ricans are presented in Figure 3B. Rates for Cuban-Americans are presented in Figure 3C. Rates for other Hispanic-Americans are presented in Figure 3D. No plot is presented for Mexican-Americans in the Northeast, as only three deaths from asthma were reported for Mexican-Americans in the Northeast from 1990 through 1995. For all age strata, rates in Puerto Ricans, Cuban-Americans, and other Hispanic-Americans were highest in the Northeast. Rates were highest in magnitude mainly in Puerto Ricans. Mexican-Americans had the lowest rates in the combined regions for all age strata.

View larger version (24K): [in this window] [in a new window]   Figure 3.   Average annual age-specific asthma mortality rates (log scale) in U.S. Hispanic subgroups, 1990-1995, Northeast region versus Midwest, South, and West regions combined: A. Mexican-Americans; B. Puerto Ricans; C. Cuban-Americans; D. Other Hispanic Americans. * 3 asthma deaths in Mexican-Americans in Northeast, 1990-1995,  represents mainland Puerto Ricans,  zero asthma deaths in Cuban-Americans aged 0-4 and 5-14 years for 1990-1995, § zero asthma deaths in other Hispanic-Americans aged 0-4 years in Northeast for 1990-1995.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our results indicated that from 1990 through 1995, national and regional asthma mortality rates for Hispanics of Puerto Rican nationality were higher than those seen among other Hispanic groups and non-Hispanic whites. In addition, asthma mortality rates for Puerto Ricans were higher than those of non-Hispanic blacks in the Northeast. Mexican-Americans, on the other hand, had lower national and regional asthma mortality rates than did other Hispanic subgroups, non-Hispanics whites, and non-Hispanic blacks. Such differences between Hispanic subgroups can be obscured when they are combined into one group for analysis.

The results of our regional analysis are consistent with previous studies. We found that age-adjusted asthma mortality rates for all Hispanics tend to resemble those of non-Hispanic whites for the entire United States and the Midwest, South, and West regions, whereas in the Northeast, overall Hispanic asthma mortality rates were higher than those of non-Hispanic whites and approaching those of non-Hispanic blacks. Hispanics zero to 34 yr of age living in New York City were reported to be three times as likely to die of asthma than were non-Hispanic whites within the same age range (1). A study of asthma mortality in New Mexico from 1969 through 1977, however, reported lower rates in Hispanics compared with non-Hispanic whites, although the rates were based on small numbers (15). In addition, a study of asthma mortality in California from 1985 through 1989 among persons zero to 65 yr of age and older reported age-stratified asthma mortality for Hispanics, which was 40 to 80% of that for non-Hispanic whites and 10 to 50% of that for non-Hispanic blacks (16). Given that Hispanics in the Northeast are more likely to be of Puerto Rican or other Hispanic origin, whereas Hispanics in other regions of the country are more likely to be of Mexican origin, our results, along with previous studies, indicate that all Hispanics should not be considered to be at high risk for death from asthma.

The mortality patterns for asthma observed in Hispanic subgroups may reflect differences in prevalence. Unfortunately, asthma prevalence in Hispanic adults is not well known, as most previous studies of asthma prevalence in Hispanics have focused on children. Data from the 1982-1984 Hispanic Health and Nutrition Examination Survey (HHANES) showed a lifetime asthma prevalence of 20.1% in Puerto Rican children, compared with 4.5% for Mexican-Americans, 8.8% for Cuban-Americans, 6.4% for non-Hispanic whites, and 9.1% for non-Hispanic blacks (17). In addition, a study of children in Connecticut reported an asthma lifetime prevalence of 18.4% in children of Hispanic (primarily Puerto Rican) mothers (18). One adult-based study of 475 non-Hispanic white and 371 Hispanic pregnant women in East Boston, Massachusetts, reported a 19% (eight of 43 women) lifetime prevalence of asthma in Puerto Rican women. Hispanic women of Central and South American origin population, who made up the majority of the Hispanic population in East Boston, had a lifetime asthma prevalence of 8% (24 of 296 women) (19). Factors that could explain elevated prevalence and morbidity of asthma in Puerto Rican children, compared with Mexican-American and Cuban-American children, include a greater genetic/biologic predisposition to asthma, poverty-related risk factors such as poor access to care and detrimental home exposures, differences in family structures and associated social risk factors and differential migration patterns, and U.S. geographic sites of residence (5, 20).

In considering a possible genetic predisposition toward asthma and severe presentation of the disease among Puerto Rican Hispanics, the potential interaction of these physical and social environmental risk factors should be taken into account. In addition, differences in acculturation and health beliefs and behaviors among Hispanic subgroups could play a role in explaining increase mortality in Puerto Ricans, as these would impact upon factors such as beliefs about asthma, management of the disease, use of medical services, response to acute attacks, detrimental behaviors such as smoking, and reliance upon folk remedies (21).

Given the high mortality rates for asthma reported in black populations, we considered whether differing racial composition of Hispanic subgroups could be a contributing factor for mortality, especially since in the 1990 Census, 12.0% of Puerto Ricans, 4.7% of Cuban-Americans, and 11.6% of other Hispanic-Americans reported being of black race, compared with 1.8% of Mexican-Americans (8). Our results showing a small number of asthma deaths among black Hispanics from 1990 through 1995 does not indicate elevated asthma mortality in black Hispanics as compared with white Hispanics. Furthermore, the small number of asthma deaths reported for black Puerto Ricans and other Hispanic-Americans suggest that race does not explain asthma mortality rates observed for these groups in the Northeast. Although we acknowledge the potential for miscoding of race for Hispanics on death certificates, we do not believe that such miscoding is seriously influencing our results.

Our finding of higher asthma mortality in Puerto Rican Hispanics is consistent with prior studies demonstrating heterogeneity in health outcomes among persons of different Hispanic subgroups. Although lower than expected mortality rates in Hispanic populations as compared with non-Hispanic populations have been reported when all Hispanic groups are aggregated (25, 26), mortality rates differ according to specific Hispanic heritage. National vital statistics for 1995 demonstrate that among Hispanics, Puerto Ricans have the highest overall age-adjusted mortality (582.9 per 100,000), followed by Cuban-Americans (387.4 per 100,000), and then Mexican-Americans (362.4 per 100,000) (27). Puerto Ricans also have the highest infant mortality rate (8.6 per 1,000 live births) as compared with Mexican-Americans (6.1 per 1,000 live births) and Cuban-Americans (5.1 per 1,000 live births) (27). Some investigators have termed this lower overall mortality among Hispanics, despite poverty and poor access to care (28), as the "Hispanic paradox" (26, 29). Yet this phenomenon has been debated because this "paradox," which has been hypothesized to be related to a "healthy migrant" effect, greater consumption of fruit and vegetables, and protective psychosocial effects (26), appears to apply to Mexican-Americans but not necessarily to Puerto Ricans.

As noted previously, we included asthma deaths for all age groups in our analysis, whereas previous investigators have restricted their analyses to decedents 34 yr of age and younger. Thus, we recognize the potential that overdiagnosis of asthma as the cause of death, especially for older persons, for which obstructive lung disease is often a comorbid factor, may inflate the mortality rates. However, we observed a pattern of asthma mortality among Hispanic subgroups for decedents 5 to 14 yr of age and 15 to 34 yr similar to that for all age groups included.

Although the results presented here show higher asthma mortality rates for Puerto Rican Hispanics than for other Hispanic subgroups, possible bias in both number of deaths and population estimates must be considered. Complete coding of Hispanic ethnicity on death certificates in the past can cut across Hispanic subgroups, especially for Hispanics living in New York City (4). We examined the coding of Hispanic ethnicity in deaths attributed to asthma by state and found that from 1990 through 1993, between 7.4 and 10.3% of asthma deaths in New York state did not include the Hispanic ethnicity of the decedent. This situation had been largely resolved by 1994; the percentage of asthma deaths in New York state with missing data for Hispanic ethnicity was 2.6% in 1994 and 0.6% in 1995. Thus, rates for Puerto Ricans and other Hispanic-Americans may be underestimated. In addition, because Hispanic ethnicity on death certificates typically is coded by a medical examiner or coroner, and usually is based upon report from a family member, there exists a potential for miscoding. The National Mortality Followback Survey, however, found a 98.9% consistency in reporting of Hispanic origin between the death certificate and the questionnaire as completed by the death certificate informant, typically the decedent's next of kin (30). Another study comparing death certificate information with the Current Population Survey (CPS) found 89.7% agreement between the two sources on coding of Hispanic origin (Hispanic versus non-Hispanic). Among the Hispanic subgroups, agreement on subgroup classification between death certificate and the CPS was as follows: Mexican-Americans, 84.9%; Puerto Ricans, 85.9%; Cuban-Americans, 80%, and other Hispanic-Americans, 47.6% (31).

The problems involved in enumerating racial and ethnic minorities, including Hispanic populations, in the United States are well documented (5, 25, 26, 32). In addition, interpolation of population estimates, especially when they are partially based on census projections, may produce inaccurate figures, which will impact the magnitude of the rates produced using them. We recognize that producing accurate population estimates for Hispanic populations is difficult at present and that these difficulties may be reflected in the rates presented here. We are encouraged, however, by the consistency of our results in relation to previous studies of asthma prevalence and mortality in Hispanics discussed previously.

It has been suggested that the number of deaths used to compute mortality rates in Hispanic populations can be underestimated because members of these populations may be more likely to return to their country of origin or birth after retirement or upon knowledge of having a terminal illness (29, 32). Given the usual acute nature of fatal attacks of asthma, we do not believe that this potential bias greatly influences our results.

Additional research is necessary to explain why Puerto Ricans have higher asthma mortality rates than other Hispanic subgroups and non-Hispanic whites. Our results indicate that this higher mortality is not completely accounted for by the higher proportion of Puerto Ricans living in the northeastern United States, where factors such as innercity poverty and environmental exposures have been postulated to be associated with high mortality rates. Puerto Ricans have higher mortality rates than do other Hispanic subgroups in all regions of the United States. This suggests that a genetic or biologic predisposition may exist for asthma among Puerto Ricans. How this predisposition may lead to higher prevalence, mortality, and perhaps severity of asthma among Puerto Ricans is unclear. Longitudinal epidemiologic studies are needed to better describe asthma morbidity and mortality patterns in Hispanic populations and to evaluate the possible interaction of physical and social environmental exposures with a possible genetic predisposition. Migration studies that evaluate possible differences and etiologic factors between mainland Puerto Ricans and island Puerto Ricans with asthma may shed further light on possible environmental and genetic risk factors.

In conclusion, our analysis has shown that asthma mortality rates among Hispanics living in the United States vary dramatically by nationality, with Puerto Ricans having the highest mortality rates. It is not clear whether these differences reflect underlying asthma prevalence or severity or other factors. Furthermore, it is clear that Hispanics do not represent a uniform, discrete group in terms of health outcomes, and that further epidemiologic research and public health research interventions, particularly those concerning asthma, should focus on Hispanic subgroups. Finally, our analysis shows elevated asthma mortality for Hispanic subgroups in the northeastern United States as compared with other regions of the country. The reasons for this difference is possibly related to socioeconomic, health behavioral, or other factors and should be more fully explored.