OBJECTIVES: We examined how Hispanic children, with stratification according to language to approximate acculturation, differed with respect to sociodemographic characteristics and medication use. We also examined how different factors were associated with the use of different classes of prescription medications.
METHODS: We used data from the 2004 Medical Expenditure Panel Survey linked to the National Health Interview Survey. Independent variables were grouped as predisposing characteristics, enabling factors, perceived need, and evaluated need. Multivariate logistic regression was used to assess the impact of independent variables on the outcomes of overall and specific types of medication use.
RESULTS: Hispanic, Spanish-interviewed children were less likely to have a usual source of care than were Hispanic, English-interviewed subjects. Both groups had lower odds of using any prescription medication, compared with white children, which was explained largely by having a usual source of care. The lower use of psychiatric medications in the Spanish-interviewed group was not explained by the independent variables, whereas the difference in the use of antibiotics was.
CONCLUSIONS: There are differences between Hispanic children according to acculturation, and acculturation affects prescription medication use. These findings may be used to address more specifically the needs of Hispanic children, particularly mental health needs.
Hispanic children constitute 20% of children in the United States.1 Research consistently has revealed significant health and health care disparities between Hispanic and white children.2–7 To reduce disparities, it is important to determine accurately where they exist. Most studies treat Hispanic children as a homogeneous group, but the designation of Hispanic ethnicity alone does not capture adequately the underlying diversity of nationality, immigration status, and acculturation.
Prescription medications are essential for the management of many pediatric conditions. Less use of prescription medications in the US Hispanic child population is well documented.3,8,9 One study found the prevalence of prescription medication use to be 70% lower among Hispanic children whose parents responded to an interview in Spanish, compared with those whose parents responded in English.10 Language, level of acculturation, preferences related to seeking and using health care, and barriers to accessing health care all are likely implicated in these disparate patterns of medication use.
We examined the impact of parents' acculturation on children's prescription medication use. Acculturation can be defined as the “acquisition of the cultural elements of the dominant society.”11 Approximately one half of Hispanic children live in less-acculturated households, as measured on the basis of either parents' citizenship or limited English proficiency.10,12 We used language as the measure of acculturation because it is the most robust marker studied.11,13 Language also may be a marker for access to care or direct communication barriers; therefore, we used a multifactorial model to isolate the role of acculturation.14
We hypothesized that Hispanic children from less-acculturated family backgrounds would be less likely to have used any prescription medication, compared with children from more-acculturated backgrounds, primarily because of barriers in accessing care. We also examined the use of specific medication classes, that is, psychiatric, respiratory, antiinfective, and central nervous system (CNS) medications.
We used the full-year consolidated data file and the prescribed medicines file from the 2004 Medical Expenditure Panel Survey (MEPS).15,16 MEPS is a longitudinal survey conducted by the Agency for Healthcare Research and Quality. An overlapping-panel, complex sampling design is used to conduct a series of 5 computer-assisted, in-person interviews with participating households. A new MEPS panel is derived each year from the previous year's National Health Interview Survey (NHIS). We linked the 2004 MEPS data to the 2002 and 2003 NHIS to obtain additional data on mother's citizenship and education. Our study sample was restricted to children 0 to 17 years of age with linked NHIS data and no missing data for the variables of interest (N = 7539). The MEPS person-level sample weights generalize to the US civilian, noninstitutionalized population for each survey year. A designated household respondent (most often the child's mother) provides all demographic and health information for each child in the sample.
Prescribed medication data are collected for every member in the household during each round of the MEPS and are verified with the dispensing pharmacies through a follow-back survey. After data collection, independent coders assigned National Drug Codes to each medical record. The MEPS prescribed medication event public release file also includes therapeutic class variables derived from the Multum Lexicon Plus (Cerner Multum, Denver, CO) classifications of the National Drug Codes.
We used an adaptation of the Behavioral Model of Health Services Use to account for socioeconomic factors, health status, and other factors that may influence the outcome of whether a child used a prescription medication.14 The model has 3 main components, that is, predisposing characteristics, enabling resources, and need, which function together to predict health services use. We divided the concept of need into perceived need (health status) and evaluated need (provider visits).
Use of any prescription medicine was measured with a dichotomous measure corresponding to ≥1 new or refilled prescription obtained for the child during 2004. Additional dichotomous variables measured medication use in each of 4 therapeutic classes, namely, antiinfective, psychiatric, respiratory, and CNS (mostly analgesic) medications. Multiple therapeutic classes were listed for ∼5% of prescription records (eg, amantadine for influenza and for Parkinson disease). For those cases, we used parent-reported data on conditions for which the child was taking the medication to assign a therapeutic class.
Reported race/ethnicity and parental language of interview were collapsed into a single variable assigning children to 1 of 4 categories, that is, Hispanic, English interview; Hispanic, Spanish interview; white (non-Hispanic), English interview; or nonwhite (non-Hispanic), English interview. We excluded children with interviews conducted in both English and Spanish or in any language other than English.
We measured predisposing characteristics by using categorical variables indicating child's age and gender, mother's citizenship, and family size. Enabling resources were evaluated through categorical variables denoting mother's education, family income as a proportion of the federal poverty level, region in the United States, metropolitan statistical area (denoting rural versus urban), type of insurance (private, public, or uninsured for the entire year), having a usual source of care, transportation to the usual source of care, and ethnicity of the provider.
We assessed perceived need for health care in 4 ways, that is, parent-rated overall child health status, parent-reported child mental health status, number of missed school days attributable to illness or injury, and whether a child experienced an illness or injury requiring urgent care in the past year. Parent ratings of child health status were based on scores of 0 to 100 (100 = best health), by using a validated scale with 6 variables from the Child Health Questionnaire.17 We assessed evaluated need through several dichotomous variables that identified children with ≥1 outpatient visit, emergency department (ED) visit, or inpatient hospitalization during the past year. For all measures, responses of “do not know” or “refused” were excluded from the analysis if they represented <1% of the sample.
A secondary analysis examined asthma and attention-deficit/hyperactivity disorder (ADHD). For asthma, respondents were asked whether their child had ever received a diagnosis of asthma. Respondents who said yes were asked whether their child still had asthma. We used the latter question to assess the prevalence of asthma. ADHD data were obtained from the linked NHIS data. Respondents were asked whether their child had any limitation; if they responded yes, then they were given a list of conditions defined as potentially limiting, with ADHD being one.
Analyses were conducted by using SPSS 15.0 with the complex samples module (SPSS, Chicago, IL), to adjust the sample variances for the complex survey design. We used design-based F statistics to compare differences in the distributions of predisposing characteristics, enabling resources, perceived need, and evaluated need. Logistic regression was used to examine the unadjusted associations of these factors with the use of prescription medications. Four multivariate models were estimated in sequential manner, to evaluate the relative contributions to the use of prescription medications (both overall and according to therapeutic class) of the following sets of covariates: predisposing, enabling, perceived need, and evaluated need. Only variables with significant bivariate associations with the use of any prescription medication (P < .05) were included in the multivariate models.
Overall, 10.2% of subjects reported Hispanic ethnicity and were interviewed in English, 7.6% were Hispanic with a Spanish interview, 61.0% were white with an English interview, and 21.2% were nonwhite non-Hispanic with an English interview; 67.8% of the subjects in the latter group were black.
Sociodemographic Differences Between Groups
Children in the Hispanic, Spanish-interviewed group were much less likely to have US citizen mothers, compared with the Hispanic, English-interviewed group (Table 1). The Spanish-interviewed group also was poorer and was twice as likely to have been uninsured the entire year, compared with the Hispanic, English-interviewed group. Only 79.4% of Hispanic, Spanish-interviewed children had a usual source of care, compared with 90.7% of their English-interviewed counterparts. The Hispanic, Spanish-interviewed group also had a smaller proportion with very good or excellent reported mental health status. However, the same group reported missing fewer days of school because of illness or injury and had fewer reported injuries or illnesses. Hispanic, Spanish-interviewed children also had fewer visits, both outpatient and ED, than did other groups.
Ethnicity and Language Associations With Medication Use
In comparison with the Hispanic, English-interviewed group and the white, English-interviewed group, a smaller proportion of the Hispanic, Spanish-interviewed group used any medication (Table 1). We found significant differences among the groups in the use of the different classes of medications. Hispanic, Spanish-interviewed subjects were much less likely to have used psychiatric medications and were somewhat less likely to have used antibiotics. There were no significant differences in the use of respiratory or CNS medications among the groups.
Older children, children from larger families, and children with a noncitizen mother were less likely to have used any medication (Table 2).
Children without a usual source of care had a lower prevalence of use of any medication (24.3%), compared with children with a usual source of care (51.6%) (Table 2). Being uninsured, being poorer, having a mother with less education, and living in the West also were significantly associated with less medication use. Transportation and ethnicity of the provider were not significantly associated with any medication use.
Children with any prescription medication use had a lower mean perceived health status score (score: 77.3) than did those with no medication use (score: 82.9). Missing more school because of illness, having an illness or injury, and having worse perceived mental health status all were significantly associated with more medication use (Table 2).
Children who did not have a provider visit (outpatient, inpatient, or ED) were significantly less likely to have used a medication (Table 2).
Multivariate Logistic Regression Results
Any Prescription Medicine
Hispanic, English-interviewed children were slightly less likely than white children to have used any medication, and both the Hispanic, Spanish-interviewed and the nonwhite, non-Hispanic groups had about half the odds of using any medication as white children (Fig 1A). There was no significant difference between the Spanish-interviewed group and white children after adjustment for enabling factors (Fig 1C). In the final model, after adding all of the explanatory variables, only the non-Hispanic, nonwhite group of children had statistically lower odds of using a medication compared with white children (Fig 1E). Having a usual source of care and having a visit to a provider were strong predictors of prescription medication use (data not shown).
All other groups had lower unadjusted odds of using an antibiotic, compared with white children (Fig 1A). After adjustment for all explanatory variables, the Hispanic, Spanish-interviewed group had the same odds as white children (Fig 1E). Having a usual source of care and having any illness or injury were significant predictors of use, whereas insurance and mother's citizenship were not (data not shown).
None of the groups of children differed significantly in the odds of using a respiratory medication (Fig 1).
In the unadjusted model, Hispanic, Spanish-interviewed children were more likely to have used CNS medications (primarily analgesic medications) than were white children (Fig 1A). The final model strengthened this relationship, with Hispanic, Spanish-interviewed subjects having twice the odds of using CNS medications, compared with white children (Fig 1E). Having a usual source of care was not a significant predictor of any CNS medication use, whereas ED and inpatient visits were (data not shown).
Of Hispanic children, only the Spanish-interviewed group had lower odds for the use of psychiatric medications, compared with white children (Fig 1A). These markedly lower odds remained despite adjustment for all explanatory variables (Fig 1E). Insurance and mother's citizenship were significant predictors in the final model (data not shown). Reported mental health status was a strong predictor, whereas any physical illness or injury was not, which validates the theoretical model.
Condition-Specific Use of Medications
Given the pattern of use for psychiatric medications, we wanted to assess whether this difference could be attributed to differences in ADHD diagnosis or management. We restricted the sample to 6- to 17-year-old children for this analysis. A significantly smaller proportion of Hispanic, Spanish-interviewed children used any psychiatric or CNS stimulant medication (Table 3). The Hispanic, Spanish-interviewed group had a lower prevalence of parent-reported ADHD, although this was not significant (P = .254). Of those with a reported ADHD diagnosis, only 13.0% of Hispanic, Spanish-interviewed subjects used a CNS stimulant medication, compared with 59.6% of Hispanic, English-interviewed subjects and 66.9% of white subjects (P = .062).
We found a significantly lower prevalence of parent-reported current asthma in the Hispanic, Spanish-interviewed group (3.8%) (Table 3). Of those with asthma, there were no significant differences in the use of any respiratory medication or the use of any of the subclasses of respiratory medications.
The Hispanic child population is not homogeneous. We have shown that one important distinguishing characteristic for this population is acculturation, as indicated by the interview language. We found that, in many sociodemographic ways, more-acculturated Hispanic children are quite distinct from less-acculturated Hispanic children, and we saw significant differences in their use of medicines. This study suggests that treating Hispanic children as a single group when conducting research on disparities may be inadequate.
Reasons for Not Using Prescription Medications
We found that a smaller proportion of the less-acculturated Hispanic group used any medication and this difference was largely explained by not having a usual source of care and having fewer provider visits. Although this is not surprising, this does provide additional evidence of the importance of a usual source of care. Our analysis of specific types of medications revealed a much more-complex picture and is the subject of the rest of this discussion.
Children with infections but without a usual source of care or insurance may be less likely to receive antibiotics for their condition. In our study, antiinfective medications were used less by Hispanic children than by white children, which is consistent with literature findings.18 The finding that enabling factors eliminated this difference in the Hispanic, Spanish-interviewed population suggests that, for these types of acute illnesses, the difference in medication use may largely related to access to care.
Another reason for not using medication is that the child may not be sick. There is evidence suggesting different prevalence rates of asthma in the Hispanic population, with Puerto Rican individuals having a higher rate than non-Hispanic white individuals and Mexican Hispanic individuals having a lower rate.19–21 Less-acculturated Hispanic children, particularly Mexican American children, have a lower prevalence of reported asthma than do other children.21 We saw a lower prevalence of use of respiratory medications in our less-acculturated group, which was composed of more Mexican American individuals. However, this group also had a lower prevalence of reported current asthma, and there was no difference in respiratory medication use among those with current asthma. Although this is a potentially encouraging finding and is consistent with the Hispanic epidemiological paradox, our study could not take into account the quality of management. This finding is most relevant in contrast to the results regarding psychiatric medication use.
A third reason why a child may not use medication is the lack of recognition of symptoms as illness or different cultural constructions of illness. There is evidence indicating a lower rate of reporting and diagnosis of ADHD in the Hispanic population,22,23 and Hispanic children were less likely to have used stimulant medications.24,25 Our findings showed that adjustment for access to care and evaluated need variables did not affect the odds of using a psychiatric medication for less-acculturated Hispanic subjects. A study showed markedly lower use of stimulant medications by Hispanic children with ADHD, compared with non-Hispanic children.26 Our study suggests that this observed difference is likely attributable to the subset of less-acculturated Hispanic children, because more-acculturated children were just as likely as white children to have an ADHD diagnosis and to have used stimulant medications. There is some indication that less-acculturated Hispanic individuals are less inclined to interpret similar symptoms as ADHD.27,28 The stark difference in psychiatric medication use according to acculturation in our study and the finding that the mother's citizenship was a significant predictor of use suggest that acculturation may be important in addressing ADHD in the Hispanic population.
Hispanic, Spanish-interviewed children were twice as likely as white children to have used CNS medications (nearly all analgesics). This reversal of the overall trend has no clear explanation in the literature. One clue from this study is that having a usual source of care was not associated with CNS medication use, whereas ED and inpatient visits were. This suggests that these CNS medications might have been given outside the context of continuity of care and might not be the most appropriate for the condition.
We used the interview language as a surrogate measure of acculturation because language is a robust marker for acculturation.11,13 Aside from being a marker for acculturation, language may influence medication use as a communication barrier. Limited English proficiency can restrict parents' ability to communicate effectively with their children's providers,29,30 and we cannot completely separate the effects of acculturation from those of communication.
Medication use is a useful and difficult-to-assess outcome. It is a concrete measure of access and utilization; however, we cannot assess whether the comparison group was using the correct number of medications, too few, or too many. Because our assessment of medication use did not take condition data into account, direct inferences about appropriate medication use are impossible. Finally, Hispanic individuals also differ according to nationality. This variable is less practically meaningful and did not add to the model; therefore, it was not included.
Our study contributes to the field by examining medication use in the Hispanic child population, with stratification according to language acculturation. Because approximately one half of all children used a medication in 2004 and the Hispanic population constituted ∼20% of children, these findings have implications for policy and practice.
The findings regarding psychiatric medications have relevance for practitioners working with less-acculturated Hispanic families, because untreated ADHD can impair significantly children's ability to reach their potential. Using interpreter services and being knowledgeable of cultural preferences may facilitate addressing ADHD in this population. Finally, the importance of a usual source of care is reinforced by these data. In this regard, however, we need not more research but more action.
This publication was made possible with support from the Oregon Clinical and Translational Research Institute, grant UL1 RR024140 from the National Center for Research Resources (a component of the National Institutes of Health), and the National Institutes of Health Roadmap for Medical Research.
We thank Motomi Mori, Michael Lasarev and Cynthia Morris for their input on this manuscript.
- Accepted February 27, 2009.
- Address correspondence to Byron Alexander Foster, MD, MPH, 415 Bainbridge Ave, Bronx, NY 10467. E-mail:
Financial Disclosure: The authors have indicated they have no financial relationships relevant to this article to disclose.
What's Known on This Subject:
There are significant health care disparities among children, particularly Hispanic children, in the United States. Language has been found to be a marker for some of these disparities.
What This Study Adds:
Acculturation may explain some health care disparities among Hispanic children. Hispanic children from less-acculturated family backgrounds are less likely than their more-acculturated peers to use psychiatric medications, but they do not differ in their use of antibiotics or respiratory medications.
- ↵National Center for Health Statistics. Health, United States, 2006, With Chartbook on Trends in the Health of Americans. Hyattsville, MD: National Center for Health Statistics; 2006
- ↵Flores G, Tomany-Korman SC. Racial and ethnic disparities in medical and dental health, access to care, and use of services in US children. Pediatrics.2008;121 (2). Available at: www.pediatrics.org/cgi/content/full/121/2/e286
- Flores G, Tomany-Korman SC. The language spoken at home and disparities in medical and dental health, access to care, and use of services in US children. Pediatrics.2008;121 (6). Available at: www.pediatrics.org/cgi/content/full/121/6/e1703
- Scott G, Ni H. Access to health care among Hispanic/Latino children: United States, 1998–2001. Adv Data.2004;(344):1– 20
- ↵Hahn BA. Children's health: racial and ethnic differences in the use of prescription medications. Pediatrics.1995;95 (5):727– 732
- ↵Cohen J. Design and Methods of the Medical Expenditure Panel Survey Household Component. Rockville, MD: Agency for Healthcare Policy and Research; 1997. MEPS Methodology Report 1, AHCPR Publication 97–0026
- ↵National Center for Health Statistics. Data File Documentation, National Health Interview Survey, 2002 [machine readable data file and documentation]. Hyattsville, MD: National Center for Health Statistics; 2003
- ↵Landgraf JM, Abaetz L, Ware JE. The CHQ User's Manual. Boston, MA: Health Institute; 1996
- ↵Miller GE, Carroll WA. Trends in Children's Antibiotic Use: 1996 to 2001. Rockville, MD: Agency for Healthcare Research and Quality; 2005. MEPS Research Findings 23, AHRQ Publication 05–0020
- ↵Davis AM, Kreutzer R, Lipsett M, King G, Shaikh N. Asthma prevalence in Hispanic and Asian American ethnic subgroups: results from the California Healthy Kids Survey. Pediatrics.2006;118 (2). Available at: www.pediatrics.org/cgi/content/full/118/2/e363
- Lara M, Akinbami L, Flores G, Morgenstern H. Heterogeneity of childhood asthma among Hispanic children: Puerto Rican children bear a disproportionate burden. Pediatrics.2006;117 (1):43– 53
- ↵Schneider H, Eisenberg D. Who receives a diagnosis of attention-deficit/hyperactivity disorder in the United States elementary school population? Pediatrics.2006;117 (4). Available at: www.pediatrics.org/cgi/content/full/117/4/e601
- ↵Visser SN, Lesesne CA, Perou R. National estimates and factors associated with medication treatment for childhood attention-deficit/hyperactivity disorder. Pediatrics.2007;119 (suppl 1):S99– S106
- ↵Perry CE, Hatton D, Kendall J. Latino parents' accounts of attention deficit hyperactivity disorder. J Transcult Nurs.2005;16 (4):312– 321
- ↵Schmitz MF, Velez M. Latino cultural differences in maternal assessments of attention deficit/hyperactivity symptoms in children. Hisp J Behav Sci.2003;25 (1):110– 122
- Copyright © 2009 by the American Academy of Pediatrics