 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
a University of California, Los Angeles/RAND Program on Latino Children with Asthma, Los Angeles, California
b Centers for Disease Control and Prevention, Hyattsville, Maryland
c Center for the Advancement of Underserved Children, Department of Pediatrics, Medical College of Wisconsin and Children's Hospital of Wisconsin, Milwaukee, Wisconsin
d University of Michigan School of Public Health, Ann Arbor, Michigan
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
METHODS. Weighted logistic regression analyses of merged 1997 to 2001 National Health Interview Survey data were used to estimate the prevalence of asthma diagnosis and asthma attacks in a sample of 46511 children (age: 2–17 years) living in the 50 states and the District of Columbia.
RESULTS. Puerto Rican children had the highest prevalence of lifetime asthma (26%) and recent asthma attacks (12%), compared with non-Hispanic black children (16% and 7%, respectively), non-Hispanic white children (13% and 6%, respectively), and Mexican children (10% and 4%, respectively). Adjustment for asthma risk factors did not change these comparisons appreciably. Compared with non-Hispanic white children, the adjusted odds ratios (ORs) for a lifetime asthma diagnosis were 2.33 (95% confidence interval [CI]: 1.90–2.84) for Puerto Rican children, 1.16 (95% CI: 1.04–1.29) for non-Hispanic black children, and 0.90 (95% CI: 0.79–1.03) for Mexican children. Birthplace influenced the association between ethnicity and lifetime asthma diagnosis differently for Puerto Rican and Mexican children. Compared with United States-born non-Hispanic white children with United States-born parents, the adjusted ORs were 1.95 (95% CI: 1.48–2.57) for Puerto Rican children in families with the child and parent(s) born in the 50 states/District of Columbia and 2.50 (95% CI: 1.51–4.13) for island-born Puerto Rican children with island-born parents. The corresponding adjusted ORs for Mexican children were 1.05 (95% CI: 0.90–1.22) for families born in the 50 states/District of Columbia and 0.43 (95% CI: 0.29–0.64) for those born in Mexico. The results were similar for recent asthma attacks.
CONCLUSIONS. The appreciably higher asthma morbidity rates experienced by Puerto Rican children cannot be explained by sociodemographic and other risk factors measured in the National Health Interview Survey. The heterogeneity of asthma among Hispanic subgroups should be considered in developing effective public health prevention and intervention strategies.
Key Words: asthma • Hispanic • Latino • Puerto Rican • Mexican
Abbreviations: NHIS—National Health Interview Survey • OR—odds ratio • CI—confidence interval
Poor and minority children in the United States suffer a disproportionately high burden of asthma.1–7 Multiple studies document increased asthma prevalence, health care utilization, and mortality rates among non-Hispanic black children,1,5,6,8–10 but less is known about Hispanic children with asthma, particularly regarding differences among children of Puerto Rican, Mexican, Cuban, or other Hispanic heritage.11–14
Mounting evidence demonstrates that Hispanic/Latino individuals (hereafter referred to as Hispanic, to be consistent with US Census nomenclature) are a heterogeneous group, with varying health status and associated risk factors.15–17 Among Hispanic children, Puerto Rican children have higher rates of chronic illness and low birth weight.18,19 Asthma studies to date in Hispanic populations confirm this pattern, with Puerto Rican individuals living in the 50 states and the District of Columbia having higher asthma morbidity and mortality rates than other Hispanic subgroups.11–14,20,21 Less comparable information is available about persons with asthma living on the island of Puerto Rico.22,23 In contrast, Mexican children living in the 50 states and the District of Columbia have been shown to have lower morbidity rates, despite having risk factors for poor health.11,18,24
Since the dramatic increase in asthma prevalence over the past 2 decades, prevalence differences among Hispanic subgroups of children in the United States have not been analyzed in detail.6,25 Although a few studies have explored some possible explanatory factors for these differences (eg, socioeconomic, access to care, or environmental factors), no study to date, to the best of our knowledge, has evaluated a wide range of possible risk factors simultaneously.11,12 In addition, the observation of better-than-expected health indicators for certain Hispanic groups, compared with other populations that experience poverty, low education, and problems accessing health care,15 has not been explored fully for asthma. The National Health Interview Survey (NHIS) provides an opportunity to analyze multiple factors that may be associated with differences in asthma burdens among different Hispanic subgroups.
We used the 1997 to 2001 NHIS to obtain adequate sample sizes for analysis of asthma prevalence among Puerto Rican, Mexican, Cuban, and Dominican children. In particular, we hypothesized that Puerto Rican children would have an ethnic-specific predisposition for high asthma burden after adjustment for possible confounding factors. We also analyzed the association between asthma prevalence and place of birth of the child and parents, as a crude proxy of acculturation, to determine whether the association varied among Hispanic subgroups.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Lifetime asthma diagnosis rates were determined on the basis of responses to the question, "Has a doctor or other health professional ever told you that your child had asthma?" The 12-month period prevalence of asthma attack was estimated on the basis of positive responses for lifetime asthma diagnosis and positive responses to the question, "During the past 12 months, has your child had an episode of asthma or an asthma attack?" The precision of crude estimates of lifetime asthma diagnosis rates and asthma attack prevalence across racial/ethnic groups was evaluated with SEs.
We analyzed the association between these outcomes and available covariates, ie, known or possible risk factors for asthma selected on the basis of previous studies.6,8,14,28–34 Demographic variables included gender, young age (2–5 years versus 
5 years), and race/ethnicity (non-Hispanic white, non-Hispanic black, Puerto Rican, Mexican, Cuban, Dominican, other Hispanic, or other non-Hispanic [all other races/ethnicities]). Hereafter we refer to children of Mexican heritage (whether born in the United States or in Mexico) as Mexican and children of Puerto Rican heritage (whether born on the island of Puerto Rico or in the 50 states and the District of Columbia) as Puerto Rican, and we refer similarly to Cuban and Dominican children. We refer to non-Hispanic white children and non-Hispanic black children as white and black, respectively. Sociodemographic variables included family income, household structure (single-parent household versus 2 adults in the household), and highest level of education of all parents in the household (no high school diploma versus high school diploma or higher). We excluded 2534 children with missing information regarding parental education. Nearly 20% of families did not report approximate family income. However, <5% of observations were missing a response to the question, "Is your family income above or below $20000?" Our initial analyses did not differ appreciably with the use of approximate family income to categorize families as poor versus nonpoor or the use of family income above versus below $20000. Therefore, we used family income above/below $20000 as a proxy for poverty, to ensure adequate sample sizes for Hispanic subgroups, and excluded children with missing information on this variable (n = 2439).
Behavioral/environmental factors included measures of overweight and exposure to tobacco smoke. The NHIS includes proxy-reported data on height and weight, which were used to calculate BMI. To remove extreme outliers, we used Epi Info 2002 software35 to calculate BMI z scores according to age and gender. According to the World Health Organization Fixed Exclusion Range,36 observations with BMI z scores of less than –4.0 or >5.0 were excluded, as were those missing either height or weight data (n = 7967). BMI was categorized as >95th percentile, >85th to 95th percentile, 5th to 85th percentile, or <5th percentile.37 Height was also included in the model, because BMI alone may not characterize adequately the joint relationships between health outcomes and body composition and body size.38 The NHIS does not include a question about smoking for all persons in the household. However, the sample adult is asked whether he or she smokes cigarettes currently. Children living with a smoking sample adult were classified as "exposed" to tobacco smoke. Children for whom the sample adult did not smoke were classified as having "unknown exposure," because there might have been other adult smokers in the household. To evaluate this classification, we analyzed data for children living with only 1 adult (ie, the sample adult). This sensitivity analysis yielded similar results and increased our confidence that the group with unknown exposure represented predominantly children not exposed to cigarette smoke.
We included 2 crude indicators of acculturation. The first indicator was whether the interview was conducted completely in Spanish. Unfortunately, the NHIS contains no question about the primary language spoken in the household or about the parent's English proficiency. The second indicator was place of birth of the parent(s) and the sample child, with 3 categories, ie, child and parent(s) born in the 50 states/District of Columbia, child and parent(s) born in the 50 states/District of Columbia and outside the 50 states/District of Columbia (any combination), and both child and parent(s) born outside the 50 states/District of Columbia. For each Hispanic subgroup, <1% of children were born neither in the 50 states/District of Columbia nor in the country of their ethnic origin. That is, for Hispanic children, being born outside the 50 states/District of Columbia represented predominantly being born in Puerto Rico for Puerto Rican children, Mexico for Mexican children, Cuba for Cuban children, and the Dominican Republic for Dominican children.
Geographic variables included region of residence in the mainland United States (Northeast, South, Midwest, or West) and central city versus non-central city residence. Having health insurance was used as a crude proxy for access to health care and was dichotomized. Respondents were also asked whether their child had a usual source of care when sick; these responses were also dichotomized. Children who used the emergency department as their usual source of sick care were considered not to have a usual source.
After all exclusions, the final sample size was 46511 (an overall decrease of 22%). No imputation methods were applied to missing data. Observations excluded for missing or implausible information for BMI, parental education, and family income were more likely to be for children of racial/ethnic minority groups. The white sample was decreased by 15%, whereas the sample reduction among minority groups was 28% for black, 26% for Puerto Rican, 35% for Mexican, 23% for Cuban, and 29% for Dominican. Because of the small sample sizes for Cuban and Dominican children, there is a higher level of uncertainly associated with the estimates for these groups and estimates are less reliable.
All variables were entered simultaneously into a logistic regression model by using PROC RLOGIST in SUDAAN software.27 We checked for colinearity by comparing SEs between models that included or omitted selected variables. Interactions between race/ethnicity and other covariates were investigated with product terms. The only interaction that changed the interpretation for race/ethnicity and the outcomes was place of birth of the child and parent(s). Odds ratios (ORs) and 95% CIs were estimated for race/ethnicity and place of birth strata with logistic regression.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Overall, 13.1% of children had received a diagnosis of asthma in their lifetimes and 5.9% had experienced an asthma attack in the previous 12 months. Table 2 shows the prevalence of a lifetime asthma diagnosis and recent asthma attack according to race/ethnicity. The familiar pattern of higher asthma prevalence among black children, compared with white children, was seen, but the most striking finding was that Puerto Rican children had dramatically higher rates of a lifetime asthma diagnosis and asthma attacks compared with all other children. Cuban and Dominican children had a lifetime asthma diagnosis prevalence similar to that of black children and asthma attack prevalence comparable to that of white children. Mexican children had the lowest lifetime asthma diagnosis rates and asthma attack prevalence compared with all other children.
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Second, parental reports of childhood asthma prevalence were related to the place of birth of the parent and child. Stratified analyses showed that, for black, white, Mexican, and perhaps Dominican children, asthma prevalence was lower among children born outside the 50 states/District of Columbia than among children born in the United States. In contrast, for Puerto Rican and perhaps Cuban children, asthma prevalence was higher among those born outside the 50 states/District of Columbia. However, interpretation of the results for Cuban and Dominican children is limited by small sample sizes. Because Puerto Rican and Mexican children are the largest Hispanic subgroups and because they define the extremes of asthma prevalence, focusing on the association between place of birth and asthma outcomes for these groups may lend insight into the impact of acculturation on asthma outcomes. Klinnert et al42 described a possible protective effect of lower acculturation, defined as a lower prevalence of several biological and psychosocial asthma risk factors, among a group of predominantly Mexican American children with asthma. Our findings support this pattern among Mexican children but show the opposite effect for Puerto Rican children. Therefore, the finding of a decreased risk of certain health conditions among some Hispanic children born outside the 50 states/District of Columbia15,18,42 is not generalizable to Puerto Rican children with asthma.
Although place of birth is only a crude proxy for acculturation, our findings suggest that acculturation is an important factor to consider in understanding asthma diagnosis, risk factors, and the effectiveness of clinical and public health interventions. One would expect that acculturation, the process of cultural transition in which an individual is exposed to and adopts characteristic ways of a foreign culture,15 would be somewhat different for Puerto Rican children. For example, in contrast to other Hispanic children born outside the 50 states/District of Columbia, island-born Puerto Rican children are US citizens and have fewer potential barriers to the health care system, such as fewer problems with language and obtaining insurance, than do foreign-born Mexican children.43 The possible beneficial effect of acculturation among Puerto Rican children was also described by Pachter and Weller,44 who found that acculturation among Puerto Rican children could be associated with greater compliance with medical regimens.
Why is the burden of asthma greatest among Puerto Rican children? We interpret most of the differences in reported prevalence as actual differences in the true incidence of disease. Our overarching hypothesis for future research is an ethnic group-specific genetic predisposition among Puerto Rican children that interacts with early life physical and social environmental exposures (eg, molds and/or family psychosocial stressors).45,46 Exposures that may contribute to the higher prevalence of asthma among Puerto Rican children include their lower rates of breastfeeding and higher rates of low birth weight, compared with other groups.47–49 Because our results also suggest that island-born Puerto Rican children have a higher prevalence of asthma than do Puerto Rican children born in the 50 states/District of Columbia, birth and residence in the island of Puerto Rico may expose children with a strong genetic predisposition toward asthma to certain environmental risk factors.
Factors other than true differences in disease expression could also account for part of the large observed differences between the Puerto Rican group and other groups. An ascertainment bias may be present. Perhaps children of Puerto Rican origin have a greater severity of asthma and, given the recurrent nature of asthma among children, it is possible that Mexican or other Hispanic subgroups have less opportunity to have an asthma diagnosis made by a provider. There may also be other reasons for differences between ethnic groups in acquiring an asthma diagnosis, such as differential recognition of disease symptoms by health care professionals. Prevalence measures depend on parental reports, and there may also be differences in the accuracy and recall of parental reporting of a child asthma diagnosis, related to racial/ethnic group-specific or acculturation-related differences in knowledge or awareness of asthma.
Our results showed that being insured and having a source of sick care were associated with higher reported asthma prevalence, which suggests that contact with the health care system is related to a higher likelihood of being diagnosed as having asthma. Acquiring an asthma diagnosis is likely a multifactorial issue, with acculturation, health care delivery, and behavioral factors influencing decisions to seek care and physician labeling of conditions. In general, speculation about what unmeasured factors (eg, health care, disease severity, environmental, genetic/biological, or behavioral factors) could explain the excess asthma prevalence in the Puerto Rican group is difficult, because it is challenging to establish a priori which asthma risk factors might differ systematically between the Puerto Rican group and all other Hispanic and non-Hispanic groups.
Our study has several limitations. Parental reporting of asthma diagnoses and symptoms might differ from a clinical assessment. Although previous studies showed that agreement between parental reporting and medical records is among the highest levels for asthma, the overall agreement is only moderate.50 Because the NHIS is not administered outside the 50 states and the District of Columbia, we were not able to compare the effect of birth and residence in Puerto Rico or Mexico. The NHIS lacks data about other important asthma risk factors, such as maternal history of asthma, indoor and outdoor environmental exposures, and medical examination findings (eg, child atopy measured with skin tests),51,52 which might have explained the marked differences among the groups. A national data set that includes a broader set of asthma risk factors, the National Examination and Nutrition Survey, does not have large enough sample sizes for evaluation of differences among Hispanic subgroups, specifically Puerto Rican children. The NHIS also does not contain information about hospitalizations for treatment of asthma and, although information about emergency department use was available, the sample sizes were too small for analysis of data for smaller ethnic groups. In addition, some bias might have been introduced by the differential rates of certain missing data in the NHIS data set across ethnic groups. The direction and degree of this possible bias, however, are difficult to estimate.
Furthermore, throughout the study period (1997–2001), the NHIS implemented changes to improve the translation and cultural adaptation methods used in the field, such as following a documented process for translation and validation of the survey, instead of depending on interviewers to perform translations in the field as needed (D. Rose, PhD, verbal communication, 2002). To the best of our knowledge, possible differences in how Hispanic subgroups interpret the NHIS asthma questions have not been evaluated. In contrast to other studies in which language was found to be a marker for acculturation or other important health characteristics and processes,42,53–55 we found that lifetime prevalence and recent attack prevalence were not related to interview language. It may be that there is indeed no relationship between language and asthma outcomes, that the measure of language in the NHIS assesses English proficiency and language acculturation inadequately, and/or that possible measurement error was introduced from translation problems in the earlier years of the study period. Given the stated limitations of the language variable, we were careful not to overinterpret associations with other variables.
Our findings have implications for clinicians and health care delivery organizations. Increased awareness of the higher risk of asthma among Puerto Rican children could facilitate better targeting of intervention resources. Health care providers could be more vigilant and aggressive in making sure that Puerto Rican children receive appropriate asthma treatment, follow-up care, and education about asthma triggers and prevention strategies. Underuse of preventive antiinflammatory medications was reported in studies of mostly Puerto Rican populations.56,57 Asthma studies also showed that continuity of care, family beliefs and compliance with therapy, and the use of ethnomedical therapies can vary among Hispanic subgroups.44,58–60 Moreover, health care organization policies promoting cultural competence were shown to be associated with better quality of care among children with asthma.61 Therefore, cultural paradigms need to be considered in the design and implementation of successful clinical and educational interventions. Specifically, in geographic areas with a high density of Puerto Rican individuals, health care provider organizations and public health agencies could find that tailoring programs to the specific needs of the Puerto Rican population is an effective population-based strategy. Similarly, the difference in prevalence between foreign-born and United States-born Mexican children could prompt vigilance among public health and health care professionals regarding a possible lack of awareness of asthma and less access to the health care system among groups with lower levels of acculturation.
Although the Centers for Disease Control and Prevention and other organizations are starting to address the need for improved asthma surveillance among Hispanic subgroups, more comparable data for evaluation of differences in risk and prognostic factors among geographically diverse Hispanic subgroups would be desirable. Approaches to promoting comparability could include, for example, additional standardization of methods for appropriate translation and cultural adaptation of survey instruments and/or expansion of asthma-related surveys in Puerto Rico. There is a current lack of information regarding Hispanic subgroups in national and state data sets, such as data on hospitalization and emergency department use. Efforts need to be made to improve the recording of Hispanic ethnicity in medical records. If such data were available, it would facilitate inclusion of information about Hispanic subgroups in asthma-related national reports, such as the National Healthcare Disparities Report.62
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
We thank Dr Neal Halfon for his assistance in identifying sources of funding for the initial phase of this work and Marian Branch, Linda Escalante, and Louis Ramírez for their editorial and administrative assistance. M.L. thanks her husband, Richard Greenberg, and their daughter, Serena Michelle Lara-Greenberg, for their unwavering support of this and other projects. This article is dedicated to all Hispanic children with asthma and their families.
|
FOOTNOTES |
|---|
Address correspondence to Marielena Lara, MD, MPH, RAND Health, 1776 Main St, Santa Monica, CA 90407. E-mail: lara{at}rand.org
The authors have indicated they have no financial relationships relevant to this article to disclose.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  S. N. Brim, R. A. Rudd, R. H. Funk, and D. B. Callahan Asthma Prevalence Among US Children in Underrepresented Minority Populations: American Indian/Alaska Native, Chinese, Filipino, and Asian Indian Pediatrics, July 1, 2008; 122(1): e217 - e222. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Gollenberg, P. Pekow, G. Markenson, K. L Tucker, and L. Chasan-Taber Dietary behaviors, physical activity, and cigarette smoking among pregnant Puerto Rican women Am. J. Clinical Nutrition, June 1, 2008; 87(6): 1844 - 1851. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. S. Sawicki, L. Smith, B. Bokhour, C. Gay, K. H. Hohman, A. A. Galbraith, and T. A. Lieu Periodic Use of Inhaled Steroids in Children With Mild Persistent Asthma: What Are Pediatricians Recommending? Clinical Pediatrics, June 1, 2008; 47(5): 446 - 451. [Abstract] [PDF]
|
|
|
|
 |
|
 N. C Smeeton, R. J Rona, J. Gregory, P. White, and M. Morgan Parental attitudes towards the management of asthma in ethnic minorities Arch. Dis. Child., December 1, 2007; 92(12): 1082 - 1087. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. D. Cabana, M. Lara, and J. Shannon Racial and Ethnic Disparities in the Quality of Asthma Care Chest, November 1, 2007; 132(5_suppl): 810S - 817S. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Choudhry, M. A. Seibold, L. N. Borrell, H. Tang, D. Serebrisky, R. Chapela, J. R. Rodriguez-Santana, P. C. Avila, E. Ziv, W. Rodriguez-Cintron, et al. Dissecting Complex Diseases in Complex Populations: Asthma in Latino Americans Proceedings of the ATS, July 1, 2007; 4(3): 226 - 233. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. Martin, M. U. Shalowitz, T. Mijanovich, E. Clark-Kauffman, E. Perez, and C. A. Berry The Effects of Acculturation on Asthma Burden in a Community Sample of Mexican American Schoolchildren Am J Public Health, July 1, 2007; 97(7): 1290 - 1296. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. T. Cohen, G. J. Canino, H. R. Bird, S. Shen, B. A. Rosner, and J. C. Celedon Area of Residence, Birthplace, and Asthma in Puerto Rican Children Chest, May 1, 2007; 131(5): 1331 - 1338. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Choudhry, E. G. Burchard, L. N. Borrell, H. Tang, I. Gomez, M. Naqvi, S. Nazario, A. Torres, J. Casal, J. C. Martinez-Cruzado, et al. Ancestry-Environment Interactions and Asthma Risk among Puerto Ricans Am. J. Respir. Crit. Care Med., November 15, 2006; 174(10): 1088 - 1093. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. M. Davis, R. Kreutzer, M. Lipsett, G. King, and N. Shaikh Asthma Prevalence in Hispanic and Asian American Ethnic Subgroups: Results From the California Healthy Kids Survey Pediatrics, August 1, 2006; 118(2): e363 - e370. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||
