An Examination of the Association of Multiple Acculturation Measures with Asthma Status Among Elementary School Students in El Paso, Texas

BACKGROUND

Variations in the prevalence of lifetime asthma by race and/or ethnicity and within Hispanic subgroups are well documented. ¹ The Centers for Disease Control and Prevention estimates a prevalence of current asthma, whereby a person has been diagnosed with asthma and reports still having it, ² of 9.6% among Hispanics compared to 7.8% among non-Hispanic White patients under the age of 18 years. Within the Hispanic population, which includes immigrants from Mexico, Puerto Rico, Cuba, Central and South America and their descendants, regardless of race, rates are as high as 24.8% among those of Puerto Rican descent and as low as 7.8% among those of Mexican descent. ³

The striking variations in the prevalence of asthma by Hispanic subgroup have prompted efforts to broaden understanding of micro- and macro-level factors driving these differences. The Hispanic epidemiologic paradox (HEP), ⁴ a term first coined in the mid-1980s to describe the contradiction whereby individuals of Mexican descent have better health outcomes than their US-counterparts, despite experiencing poorer socioeconomic conditions, is often used to shed light on variations in outcomes. Sometimes used interchangeably with the Latino epidemiologic paradox, Latino mortality paradox, or Immigrant paradox, researchers have applied the HEP to a myriad of outcomes including, but not limited to, mortality rates, ⁵ low birth weight, ⁶ mental health, ⁷ and asthma. ⁸ Investigators have consistently observed better outcomes among immigrants compared to their US-counterparts initially, and to a lesser extent over time, in the US. In line with the HEP, researchers have generated substantial research in the area of acculturation, a process whereby members of a cultural group may adopt the norms of another,⁹ as a means to elucidate the factors that contribute to these better health outcomes and the point at which they no longer confer a protective effect. ^10–12

Prior scholarship examining acculturation and diagnosed asthma demonstrates lower rates of asthma among Mexico-born Mexican-Americans than their US-counterparts. ^13,14 Some researchers posit that the lower prevalence among those of Mexican descent may result from under-diagnosis of asthma symptoms ¹⁵ rather than an absence of illness. In the early 1980’s, Dodge’s examination of respiratory health differences between Mexican-American and non-Mexican-American White children identified a higher rate of physician-diagnosed asthma among non-Mexican-American than Mexican-American school children (6.5% vs 1.9%), despite similar rates of asthma symptoms. ¹⁶ Indeed, the body of research to date suggests that asthma is under-diagnosed and undertreated. ^17–19 In the United States, prevalence estimates for undiagnosed asthma, wherein symptoms are present but a diagnosis has not been established, vary by race/ethnicity as well as geography, notably urban vs. rural residence. ^15,20–25 A 2007 report from the American Thoracic Society approximates “that the prevalence of undiagnosed asthma in children ranges from 3 to 20%.” ²⁶ This raises concerns about an underestimated burden of disease, particularly among high-risk underserved populations, in whom extrinsic factors may impede the assessment of symptoms, subsequent diagnosis, and initiation of treatment. ²⁷

Whereas the influence of factors associated with the prevalence of asthma is well established, ^28,29 less established is the influence of factors associated with undiagnosed asthma, and even less so, cultural aspects that may promote or impede diagnosis. Some authors looking at parental language and the likelihood of receiving an asthma diagnosis point to possible underdiagnosis of asthma among Spanish-speaking respondents compared to English-speaking respondents, ³⁰ while others describe that nativity (foreign-birth) confers a protective effect against asthma, an effect that diminishes with duration in the US. ^31,32 Among Mexican-Americans in particular, some researchers recommend that nativity as well as acculturation be considered risk factors for asthma. ^13,14 However, ambiguity about what acculturation means and how it is measured has led some to question its use in health research, ^33,34 as defining acculturation as a singular construct, one measured by language or nativity alone, is inherently challenging. ³⁵

Traditionally, proxies for acculturation include static, unidimensional measures like nativity, immigrant status, years in the country, and language use. ^9,36 More nuanced, multidimensional measures of acculturation that tap into behaviors, values, and cultural preferences, are less common given their complexity. ³⁷ Capturing these measures is important, as they can provide insight into differences in prevalence estimates that could otherwise be missed. For instance, differences in health outcomes like childhood obesity ³⁸ and health behaviors like dietary fiber intake, have been observed to differ by level of acculturation. ³⁹

At a minimum, undiagnosed asthma can negatively impact school attendance, academic achievement, and social functioning, particularly if symptoms are uncontrolled. ²³ Despite estimates of lower asthma rates among Mexicans and those of Mexican descent overall, evidence of a higher prevalence of asthma among US-born compared to Mexico-born Mexican-Americans is growing. ^13,14,40 There is also evidence of increased asthma hospitalizations among US-Mexico border children compared to non-border children, ^41,42 which is reflective of poorly controlled asthma, and underscores the need to critically appraise factors like acculturation, and its influence on asthma in this area.

The US-Mexico border region allows for a unique opportunity to examine associations of asthma prevalence with measures of acculturation, and decipher whether patterns in asthma status exist. This study aims to determine the frequency and characteristics of asthma status (no wheezing/asthma symptoms, possible undiagnosed asthma, and diagnosed asthma) among a population of elementary school-aged children in the El Paso Independent School District (Texas, USA), as well as identify and compare measures of acculturation associated with asthma status, controlling for sociodemographic factors among the children. These multivariate analyses of possible undiagnosed asthma extend previous research by including both traditional unidimensional proxies for acculturation, and a theory-based multidimensional measure of acculturation. Combined, these may support novel insight into potential patterns of asthma status.

Theoretical and Conceptual Framework

We drew on some elements from Canino et al.’s Conceptual Framework of Pediatric Asthma Disparities to inform our work and the selection of relevant covariates. ⁴³ This conceptual framework delineates four multifaceted and interconnected domains that influence disparities in the prevalence, morbidity, and severity of asthma among Latino children: individual characteristics, the environment, the healthcare system, and provider’s characteristics. In line with this framework, we focus on the domains of the individual (e.g., age, gender, acculturation), and the health care system (e.g., insurance coverage), to better understand disparities in asthma (e.g., diagnosed and undiagnosed asthma burden) among this group of Latino children.

Methods

Participants

Data were collected through the Environment Core of the Hispanic Health Disparities Research Center at the University of Texas, El Paso (UTEP), which carried out a cross-sectional study aimed at examining inequities in the respiratory health of elementary school children in the US-Mexico border region of El Paso County (EPC).⁴³ The expanded examination of acculturation strategies from these data, which are the focus of the current study, have not been previously investigated. EPC is located in the US-Mexico border region, which stretches over 2,000 miles from California to Texas and is home to an underserved population that is expected to reach 9.8 million by 2020. ⁴⁴ In 2012, approximately 82% of EPC’s 827,398 residents identified as Hispanic; approximately 78% of these identified as of Mexican descent in particular. ⁴⁵ The median household income of county residents ($40,157) was lower than that of the State of Texas ($51,900) and the United States ($53,046). In line with these estimates, 23.3% of persons were below poverty level in EPC, compared to 17.6% in the State of Texas, and 15.4% in the United States. ⁴⁶ The population for these analyses consists of Hispanic students of Mexican descent enrolled at El Paso Independent School District (EPISD), which is the largest school district in EPC.

Data Collection

In May of 2012, all households of 4^th and 5^th grade students enrolled at EPISD (n=6,295) were mailed a survey packet in English and Spanish by a team of researchers at UTEP. ^47–53 The survey packet included a close-ended questionnaire, consent form, postage-paid return envelope and $2 token of appreciation. A follow-up bilingual reminder postcard was mailed a week later. All non-respondents received another survey packet the following week, resulting in final response rate of 30.2% (1,904/6,295). Household surveys with similar response rates have been shown to be representative of target populations. ^54–58 Descriptive statistics demonstrated that the demographic characteristics of this group of 4^th and 5^th grade students at EPISD are similar to those of the student population throughout EPISD. ⁵⁹ The protocol for the parent study was approved by the Institutional Review Board (IRB) at the University of Texas, El Paso. The questionnaire made explicit that completing and returning the survey to the study research team established informed consent to participate. All questionnaires were reviewed by UTEP data managers for completeness and to ensure accuracy of participant responses. The protocol for this current study was approved by the Institutional Review Board (IRB) at the University of Texas, School of Public Health (UTSPH IRB). For the purposes of this work, we restricted our sample to Hispanic students of Mexican descent (n=1,305).

Measures

The study instrument included questions about age, race/ethnicity, nativity, language use, income, education, health insurance, acculturation and the child’s respiratory history. ⁵⁰ The latter included caretaker report of physician diagnosed asthma, asthma symptoms, and treatment in line with the International Study of Asthma and Allergies in Childhood (ISAAC) questionnaire, a valid instrument for determining asthma symptoms in the past 12 months. ⁶⁰ The ISAAC questionnaire has been found to have sensitivity and specificity of 0.85 (0.73–0.93) and 0.81 (0.76–0.86), respectively, as well as test-retest reproducibility. ⁶¹

Dependent and Independent Variables

The primary outcome variable was asthma status (0=no wheezing or no asthma symptoms, 1=possible undiagnosed asthma, and 2=diagnosed asthma), ascertained from primary caretaker responses on the ISAAC questionnaire. ⁶¹ Specifically, a negative report of wheezing/asthma symptoms (coughing, wheezing or whistling in the chest, shortness of breath, chest tightness or phlegm when one does not have a cold or respiratory infection) currently or in the last 12-months, and no physician diagnosis, was used to categorize students with no wheezing/no asthma symptoms. Affirmative reporting of the aforementioned symptoms without a physician diagnosis was used to indicate possible undiagnosed asthma; affirmative reporting of “Has the child ever been told by a doctor or other health professional that he or she has asthma?” was used to categorize students diagnosed with asthma.

The primary independent variables were measures of acculturation as reported by the primary caretaker about the child. Unidimensional measures, including aspects of nativity and preferred language use, have all been associated with diagnosed and possible undiagnosed asthma status. ^30,40,62 Birthplace and language are important predictors of asthma among Mexican-American populations and often used as proxies for immigration and acculturation. ^30,63 We included nativity and survey language used as dichotomous variables (1=US-born, 0=foreign-born; 1=English, 0=Spanish). The multidimensional measure of acculturation was operationalized using the shortened version of Mendoza’s Cultural Life Style Inventory (CLSI) with acculturation categories for Mexican oriented (segregation), multicultural oriented (integration), US oriented (assimilation), and non-dominant/eclectic. ^64,65 Evidence of the validity and reliability of the 11-item CLSI has been documented, with a Cronbach’s alpha to assess internal consistency reliability of α =.92. ⁶⁶ This acculturation inventory includes items on the language and social preferences of participants, as well as preferences for the celebration of cultural holidays. For instance, language preferences when reading, writing, thinking, and/or speaking within and outside the home; language preferences related to media, prayer and social interactions; preferences for the celebration of cultural holidays, and the preferred term for identification. The five-point Likert scale for the language items includes: only Spanish, more Spanish than English, both equally, more English than Spanish, or English only. Related to social preferences, items include preferences to engage always with Hispanics, mostly with Hispanics, both with non-Hispanics and Hispanics about equally, mostly with non-Hispanics, always with non-Hispanics. Preferences for the celebration of cultural holidays include: only Mexican holidays, mostly Mexican holidays, both American (US) and Mexican holidays about equally, mostly American (US) holidays, only American (US) holidays; other (specify); do not celebrate any national or cultural holidays. Responses for the CLSI items were scored to correspond with Mendoza’s suggested categories for dominant oriented acculturation, whereby segregated individuals gravitate toward Mexican culture, assimilated individuals gravitate toward those of the mainstream US culture, integrated individuals maneuver both cultures without clear preference, and non-dominant/eclectic individuals exhibit no clear acculturation pattern. ^66,67 All 11-items were coded to correspond to a “0” if not answered or a “1” if answered. All items were summed and divided by the total number of items answered, resulting in proportions for each of three categories between 0 and 1. Respondents were categorized according to their dominant typology in line with the category that included their highest proportion, as recommended by Mendoza. ⁶⁷ For instance, a respondent with a segregation proportion of 0.18, an integration proportion of 0.72, and an assimilation proportion of .09 would be categorized as “integrated” with integrated =1, segregated=0, and assimilated=0. Tied scorers with no clear acculturation pattern (non-dominant/eclectic) were excluded from these analyses (7.7%).

Covariates

The selection of covariates was based on prior literature and the availability of data. Proxies for socioeconomic status included the whether the child had a regular doctor for routine medical care (1=yes, 0=no) and whether the family postponed or did not seek medical treatment due to problems with cost in the last year (1=yes, 0=no). ⁶⁸ Additional variables examined due to their association with an increased risk of asthma included: gender (1=male, 0=female), the child’s history of allergies (1=yes, 0=no), and insurance coverage in the last year (1=yes, 0=no). Age ranged from 9–13 in this group of 4^th and 5^th grade students. Descriptive statistics indicated that age-adjustment was not needed in these analyses.

Analyses

We performed bivariate associations between the child’s asthma status (no wheezing/asthma symptoms, possible undiagnosed asthma, diagnosed asthma) and variables of interest using chi-square and Fisher’s exact tests for categorical variables and t-tests for normally distributed continuous variables. Missing data were excluded from these analyses (see Table 1), and complete case analyses were performed for all analyses.

Table 1.

Characteristics of the Study Population

	N=1305	Mean (± SD) Or %	% Missing
Age (range: 9–13)	1265	10.38 (±0.77)	3.06
Gender (Male)	1263	48%	3.2
Child's History of Allergies	1276	49%	2.2
Unidimensional Measures
Nativity (US-born)	1301	91%	0.31
Survey Language (English)	1305	55%	0
Multidimensional Measure (CLSI)
Child Acculturation CLSI Category	1304		7.7
Segregated	328	27%
Integrated	173	14%
Assimilated	704	58%
Health Care System Domain
Continuous Insurance Coverage	1281	83%	1.8
Child has Doctor for Routine Care	1280	90%	1.9
Family Postpones Care Due to Cost	1294	27%	0.8
Disparities Domain	1251		4.1
No Wheezing/Asthma Symptoms	947	76%
Possible Undiagnosed Asthma	115	9%
Diagnosed Asthma	189	15%
Total Asthma Burden	304	24%

Notes

^aThe mean for dichotomous variables above can be interpreted as proportions. For instance, the mean for total asthma burden (where 0=no asthma symptoms/wheezing, and 1=possible undiagnosed asthma and diagnosed asthma combined) indicates that 24% of students in this sample are experiencing respiratory distress.

Multinomial logistic regression models were estimated to test the association between the child’s asthma status (with no wheezing/asthma symptoms as referent group) and acculturation measures (nativity, language, CLSI category). Our first model (n=1,095) included gender, history of allergies, insurance coverage, whether the child has a regular doctor for routine care, and whether the family postpones or does not seek medical treatment due to cost (model 1). Subsequent models included nativity (model 2), language (model 3), and CLSI category (model 4), entered separately. Multi-collinearity was assessed using the variance inflation factor (VIF); values greater than 10 were removed from the models. Results are presented as relative risk ratios (RRs) and corresponding 95% confidence intervals (CIs). Statistical significance is measured as a p-value of .05 or less. Statistical analyses were conducted using STATA® version 12 (STATA Corp., TX, USA).

Results

Descriptive Statistics

Table 1 displays the characteristics of the study population of Hispanic children of Mexican descent (n=1,305). Possible undiagnosed asthma was reported for 9% of children in this group, whereas lifetime asthma prevalence was reported for approximately 15% of children; 76% of children in this sample did not experience wheezing or asthma symptoms. Combined, the results of this study indicate that total burden of asthma (possible undiagnosed and diagnosed asthma) is about 24%. Over 83% of students had continuous insurance coverage over the 12-month period prior to the administration of this survey and 90% described having a regular doctor for routine care, however, almost a third of the population described having postponed or not sought medical treatment due to problems with cost.

Acculturation Measures

Over 90% of the study population was born in the United States and 55% of primary caretakers responded to the English version of the study instrument. Almost sixty percent of students were classified as assimilated, 27% were classified as segregated and 14% were classified as integrated, per Mendoza’s CLSI categories (Table 1). Bivariate analyses (Table 2) showed significant associations for asthma status and our acculturation measures. A higher proportion of those classified as segregated (gravitate toward Mexican culture) or as integrated (maneuver between Mexican and US cultures) reported not having wheezing or asthma symptoms in the last year, whereas a higher proportion of those classified as assimilated (gravitate toward US culture) did experience wheezing or asthma symptoms in the last year. Specifically, we found that the relative risk of possible undiagnosed asthma was significantly higher among children whose parents completed the survey in English (RR: 1.9, 95% CI: 1.3–2.9) and among assimilated children (RR: 2.2–95% CI: 1.3–3.7), compared to children with no wheezing/asthma symptoms.

Table 2.

Bivariate Associations between Child’s Characteristics and Asthma Status

	No asthma/symptoms in the last year	Possible Undiagnosed Asthma	Diagnosed Asthma
Age	(10.40 ± .78)	(10.26 ± .84)	(10.38 ± .72)
Gender, N (%)
Girls	504 (55%)	51 (45%)	74 (41%)
Boys	417 (45%)	62 (55%)	108 (59%)
x2=14.43 p = 0.001
History of Allergies, N (%)
No	560 (60%)	27 (24%)	34 (18%)
Yes	378 (40%)	86 (76%)	151 (82%)
x2=140.01 p < 0.001
Child's Nativity
Foreign-born	99 (11%)	8 (7%)	7 (4%)
US-born	846 (89%)	107 (93%)	181 (96%)
x2=9.53 p = 0.008
Survey Language
Spanish	444 (47%)	36 (31%)	67 (35%)
English	503 (53%)	79 (69%)	122 (65%)
x2=16.33 p < 0.001
CLSI Acculturation Category
Segregated	260 (30%)	18 (17%)	32 (18%)
Integrated	126 (14%)	15 (14%)	19 (11%)
Assimilated	488 (56%)	74 (69%)	129 (72%)
x2= 22.46 p < 0.001
Continuous Insurance Coverage, N (%)
No	171 (18%)	18 (16)	18 (10)
Yes	762 (82%)	96 (84)	167 (90)
x2= 8.37 p = 0.01
Child has Doctor for Routine Care
No	101 (11%)	12 (10%)	5 (3%)
Yes	832 (89%)	103 (90%)	178 (98%)
Fisher's exact = 0.001
Family Postpones Care Due to Cost
No	696 (74%)	73 (63%)	130 (70%)
Yes	246 (26%)	42 (37%)	55 (30%)
x2=13.33 p = 0.05

The relative risk of possible undiagnosed asthma was not significantly higher among children born in the US (RR: 1.6–95% CI: 0.7–3.3), compared to children with no wheezing/asthma symptoms. The relative risk of diagnosed asthma was significantly higher among US-born children (RR: 3.0–95% CI: 1.4–6.6), children whose parents completed the survey in English (RR: 1.6–95% CI: 1.1–2.2), and assimilated children (RR: 2.1–95% CI: 1.4–3.3), compared to children with no wheezing/asthma symptoms.

Multivariate Analyses

Relative risk ratios for the multinomial logit models of asthma status are shown in Table 3. Model 1 shows that the relative risk of possible undiagnosed asthma was higher for children with a history of allergies (RR: 4.7, p<0.01) and children from families who postpone care due to cost (RR: 2.07, p<0.01), compared to those with no wheezing/asthma symptoms, holding other model variables constant. The relative risk of possible undiagnosed asthma was lower among children who have a doctor for routine care (RR: 0.89, NS). The relative risk of possible undiagnosed asthma was not significantly different between US- and foreign-born children (model 2). We did observe a marginally significant higher relative risk of undiagnosed asthma among children whose parents took the survey in English (RR: 1.58, p <0.1) (model 3). Model 4 shows no significant difference by acculturation category.

Table 3.

Multivariate Regression Models of Child’s Asthma Status by Measures of Acculturation (using no wheezing/asthma symptoms as referent group)

	Model 1	Model 2	Model 3	Model 4
Possible Undiagnosed Asthma
Gender (male)	1.47	1.46	1.47	1.48
History of Allergies	4.68**	4.67**	4.21**	4.37**
Insurance Coverage	1.31	1.30	1.31	1.34
Child has Doctor for Routine Care	0.89	0.88	0.79	0.82
Family Postpones Care Due to Cost	2.07**	2.07**	2.08**	2.07**
Nativity (US-born)		1.04
Language (English)			1.58
Acculturation
Segregated (reference)
Integrated				1.59
Assimilated				1.60
Diagnosed Asthma
Gender (male)	1.79**	1.77**	1.79**	1.79**
History of Allergies	6.67**	6.59**	6.66**	6.17**
Insurance Coverage	1.61	1.57	1.6	1.65
Child has Doctor for Routine Care	3.11*	2.88	3.10*	2.84
Family Postpones Care Due to Cost	1.71*	1.73**	1.70*	1.70*
Nativity (US-born)		1.51
Language (English)			1.01
Acculturation
Segregated (reference)
Integrated				1.22
Assimilated				1.55
Fit Statistics
Log-likelihood	−709.27	−708.84	−707.49	−706.47
df	10	12	12	14
chi2	167.8	168.65	171.37	173.4
Likelihood-ratio test		0.65	0.16	0.26

Notes: complete case analysis n=1,095; statistical significance:

^*p<0.05;

^**p<.01

Similarly, the relative risk of diagnosed asthma was higher for boys (RR: 1.79, p<0.01), children with a history of allergies (RR: 6.7, p<0.01), children who have a doctor for routine care (RR: 3.11, p<0.05), and children from families who postpone care due to cost (RR: 1.71, p<0.05), compared to those with no wheezing/asthma symptoms, holding other model variables constant. We did not observe significant differences by nativity (model 2) or language (model 3). We did observe a marginally significant higher relative risk of diagnosed asthma among assimilated children (RR: 1.55, p<0.1).

Discussion

Incorporating unidimensional and multidimensional measures of acculturation in assessments of asthma status can reveal important relationships that could otherwise be missed. Few studies have examined factors associated with the burden of possible undiagnosed asthma. Fewer still have examined these factors along with well-defined acculturation measures that take into account more than language use, nativity or years in the United States, as traditionally, unidimensional constructs of acculturation are used. ^21,40

Our bivariate analyses showed significant associations between asthma status and all measures of acculturation (p <0.05). In terms of nativity, a higher proportion of children with no wheezing/asthma symptoms were born outside of the US, whereas a higher proportion of students with possible diagnosed asthma and diagnosed asthma, respectively, were US-born. We found a similar pattern with language and CLSI category: a higher proportion of children whose parents took the survey in Spanish did not have wheezing/asthma symptoms, whereas a higher proportion of children whose parents took the survey in English did have wheezing/asthma symptoms; a higher proportion of children with no wheezing/asthma symptoms were classified as segregated (i.e., connected more strongly with Mexican culture), a higher proportion of students with symptoms were classified as integrated/assimilated.

In our multivariate analyses, nativity was not significantly associated with possible undiagnosed asthma or diagnosed asthma compared to no wheezing/asthma symptoms, however, language and CLSI category were marginally significant for possible undiagnosed asthma and diagnosed asthma, respectively (p<0.1). While prior scholarship describes that foreign birth has been associated with undiagnosed asthma as well as lower odds of reporting diagnosed asthma, we did not find statistically significant differences in this sample. ^{62, 69–71} Related to language, children whose parent took the survey in English were more likely to have possible undiagnosed asthma than children whose parent took the survey in Spanish, although not significantly. In their analysis of language and asthma status among Hispanic school children in Chicago, Mosnaim and colleagues report higher rates of asthma among children of parents who completed questionnaires in English, and posit that an under diagnosis of asthma may be related to parental preference for Spanish communication. ³⁰ This is the opposite of what we found here. In our case, the majority of Spanish-speaking children have no symptoms (81% vs. 71% of English-speaking children, data not shown), which may relate to why children from English-speaking families have higher relative risk of undiagnosed asthma.

In terms of the acculturation categories, there was no statistically significant difference in the relative risk of undiagnosed asthma for integrated or assimilated vs. segregated children. This suggests that children who identify strongly with Mexican culture were more likely to be undiagnosed than children who identify with both Mexican and US culture. This finding is somewhat related to Eldeirawi et al., who found a higher likelihood of asthma among highly acculturated Mexican-Americans (OR=2.57, 95%CI: 1.58–4.19) compared to their lower acculturated counterparts. ⁶³ In terms of why the acculturation findings were not significant, there are several possible reasons. One explanation relates to the Hispanic Health Paradox and the very low rates of asthma symptoms and wheezing reported by immigrant children and children from immigrant families.⁵⁰ Another reason may relate to the location of the study on the US-Mexico border. Acculturation, language preferences and nativity may be less important in shaping a child’s relative risk of remaining undiagnosed, when the majority of people in the community speak Spanish and are comfortable in the Mexican culture. A previous study found that acculturation was closely related to having an asthma diagnosis vs not having a diagnosis,⁵¹ suggesting that those factors do matter in some ways. It is clear that culture-related axes are not driving disparities in undiagnosed asthma among this sample of Hispanic children. Instead, access to care seems to be more important since our results reveal a lower relative risk of possible undiagnosed asthma among those with a regular doctor for routine care (RR: 0.26; 95% CI: 0.1–0.9, not shown), compared to diagnosed asthma. Further information is required as to the means by which measures of acculturation can impact respiratory health. While one of the key strengths of this study is the level of detail from measures included in the study instrument, this is also one of its key limitations. The data for this analysis were obtained from a self-administered questionnaire about the child’s language and social preferences. It is possible that the respondent’s perception of the child’s preferences differ from those of the child’s.

The most important finding from this study is not related to acculturation and instead relates to the high burden of asthma uncovered in this analysis: the prevalence of possible undiagnosed asthma is 9% and that of diagnosed asthma is 15%; combined, these estimates show that almost a quarter of this population is experiencing some respiratory distress. Our findings are congruent with those of Shalowitz and colleagues in Chicago, who identified a total potential burden of asthma of about 25%. ⁷² Studies examining racial and ethnic disparities of diagnosed and undiagnosed asthma, for example, Quinn et al., have found more children (14.5%) with respiratory symptoms consistent with possible asthma, than had diagnosed asthma (12.2%), ¹⁵ supporting the notion that asthma may be under-diagnosed. In our sample, we found a prevalence of possible undiagnosed asthma closer to that expected for diagnosed asthma, suggesting that indeed, asthma may be under-diagnosed.

CONCLUSION

Examining contextual factors associated with asthma status among US-Mexico border children presents an opportunity to disentangle the influence of such factors on the burden of illness experienced by this population. There remains a need to better assess the health status and outcomes of a burgeoning population with health problems that may be masked by social or cultural factors. In line with scholarship on the influence of race/ethnicity on health outcomes, where the social and economic effects that the construct of race/ethnicity has on outcomes may mask associations, ⁷³ a similar pattern may occur with acculturation. However, it is also important to consider that capturing acculturation on the border may not be as informative as in other areas. The current study advances previous work with a broadly representative group of Hispanic schoolchildren throughout El Paso County using a validated asthma survey for previously diagnosed asthma and possible undiagnosed asthma as well as more fine-tuned measures for acculturation to evaluate potential disparities. Future studies would benefit from the inclusion of qualitative interviews with representative sample of the study population of interest to assess aspects of acculturation and tease out differences among those grouped into a given acculturation category. Including measures that allow primary caretaker to describe their definition of wheezing and/or asthma may reveal whether self-report is in fact lower or whether the severity is not deemed urgent enough to warrant reporting by the primary caretaker.

The Hispanic population in the U.S. is growing at an astounding rate. With the shift in the population – from foreign-born to a largely U.S.-born – comes the opportunity to continue examining how acculturation impacts health outcomes. In the case of Mexican-Americans in particular, the largest Hispanic subgroup, it remains to be seen whether the lower levels of asthma observed now will continue in ensuing generations -- or whether an under diagnosis may lead to an increased burden of disease over time.