Parents with language barriers reported less timely illness care and routine care for children in managed care Medicaid.
: To examine the relationship of race/ethnicity, language, communication with providers, and interpreter use with timeliness of illness and routine care for children in managed care Medicaid.
: Cross-sectional study using data from the Consumer Assessment of Health Plans Survey, which was administered to parents of children less than 15 years of age in Washington State managed care Medicaid in 2000.
: Predictors of interest were childâ€™s race/ethnicity, parentâ€™s primary language at home, difficulty communicating with the childâ€™s providers because of language, and need and use of interpreters. The primary outcomes were how often (always, usually, sometimes, or never) the child received illness and routine care as soon as wanted. Analyses used multivariate ordered logistic regression and model-based direct adjustment.
: A total of 5142 children were included. In adjusted analyses, children received illness and routine care significantly less often than wanted if the parentâ€™s primary language at home was not English; the parent ever had difficulty communicating with the provider; and the parent needed but did not always get an interpreter.
: Language barriers were associated with decreased timeliness of care for children in managed care Medicaid.
(Am J Manag Care. 2008;14(7):417-426)
This research suggests that language barriers are associated with delayed access to illness and routine care for children in managed care Medicaid.
Children received timely care less often if their parentsâ€™ primary language was not English or their parent reported language problems in communicating with providers. However, having access to interpreters was associated with getting timely care more often.
Racial/ethnic disparities in timeliness of care did not persist after controlling for language barriers.
Providers, health plans, and policymakers should consider steps to address language barriers in order to reduce disparities in the timeliness of childrenâ€™s healthcare.
Timeliness of care has been cited by the Institute of Medicine as a key component of quality healthcare.1 Delayed access to care is cause for concern because of its association with increased morbidity, longer hospital stays, increased costs, and increased emergency department (ED) use.2-6 Children from racial/ethnic minority families and those covered by Medicaid are at risk for suboptimal access to healthcare7-10 and may be at risk for less timely care as well.8 Patient–provider racial/ethnic and language concordance appears to mitigate disparities in quality of care for adults,11-14 although less clearly for children.15-17
Language may play an important role in the existence of racial/ethnic disparities in access to timely care. In 2000, 18% of the US population over age 5 years spoke a language other than English at home, an increase over the previous 10 years.18 A number of studies have demonstrated the effect of language barriers on access to and use of healthcare. Children whose parents’ primary language is not English are less likely to have received the recommended number of preventive care visits.19 Adults who are limited in English proficiency (LEP) use physician services and preventive care less often, are less likely to have a regular source of care, and use the ED more often as the primary source of care20-22; children with LEP parents are less likely to be brought in for needed care.23 Lack of trained medical interpreters is common and is associated with decreased quality of care for adult populations.24-26 However, the relationship between race/ethnicity, language barriers, and delayed access to care for children has been less well elucidated. Prior studies of racial/ethnic disparities in timeliness of care for children have not adequately accounted for the confounding effects of language-related factors such as interpreter use and communication problems.20,21,27,28
Thus, the objective of this study was to examine the association of race/ ethnicity, parent’s primary language at home, difficulty communicating with providers, and use of interpreters with report of receiving illness care and routine care as soon as wanted for children covered by Medicaid.
The primary outcomes of the study were timeliness of both illness care and routine care. To measure timeliness of illness care, parents were first asked whether their child had an illness or injury that they believed needed immediate care from a doctor’s office, clinic, or ED in the previous 6 months. Those who responded affirmatively were asked “in the last 6 months, when your child needed care right away for an illness or injury, how often did your child get care as soon as you wanted?” Choices for responses were never, sometimes, usually, or always. The validity of this measure of delayed care has been demonstrated in previous work.5 To measure timeliness of routine care, parents were first asked whether they had made any appointments in the last 6 months for their child with a doctor or other health provider for regular or routine healthcare. If they had, they were asked “in the last 6 months, how often did your child get an appointment for regular or routine health care as soon as you wanted?” Choices for responses were never, sometimes, usually, or always.
Bivariate analyses were done using the χ2 test or Fisher exact test to assess the association between the predictor of interest and timeliness of care, categorized as always, usually, sometimes, or never getting care as soon as wanted. Multicollinearity of the key race/ethnicity and language predictor variables was tested using the Pearson correlation coefficient, and variables with r values less than 0.80 were included in the multivariate analyses. Data in the multivariate analyses were weighted to reflect the under-representation of children younger than age 48 months that occurred because of the exclusion of children who had participated in the earlier MAA survey on development. We used ordered logistic regression to determine the relationship between each predictor variable and whether the child always, usually, sometimes, or never received care as soon as the parent wanted. Ordered logistic regression maximizes statistical power and allows data from each distinct category of response to be analyzed as it was reported, rather than lumped together to create a binary outcome; this method has been used in other published literature assessing healthcare quality using CAHPS measures.31 Separate models were run for illness care and routine care. In addition to race/ethnicity and the language variables of interest, other variables in the model included child’s age, sex, health status, presence of a special healthcare need, parental age and education, and having a regular provider.
To make the results of the ordered logistic regression more easily interpretable, we performed simulations to derive the adjusted probabilities of always, usually, sometimes, or never getting care as soon as wanted for race/ethnicity and language predictors. To do this, we used model-based direct adjustment, using the entire study sample as the standard population.32 An overall P value for categorical predictor variables with more than 2 categories was calculated using the multivariate Wald test. All analyses were done using Stata 8.0 software (Stata Corporation, College Station, TX). A P value of <.05 was considered significant for all analyses.
Of the 9450 children sampled, 319 were ineligible because of death, language barriers, undeliverable surveys, not being a member of the intended managed care plan, being in the plan less than 6 months, or not being a Washington state resident. Of the 9131 eligible children, the survey response rate was 56.3%, which yielded a sample size of 5142 children. Characteristics of the study sample are presented in , as well as characteristics of children in the analyses of timeliness of illness care (those with an illness or injury that needed care in the previous 6 months) and timeliness of routine care (those whose parent made an appointment for routine care in the previous 6 months). Fourteen percent of parents reported that they had ever (ie, sometimes, usually, or always) had a hard time speaking with or understanding their child’s provider because they spoke different languages. Six percent of parents reported that they needed an interpreter to speak with their child’s provider in that past 6 months; of these, more than half (52.7%) reported that they were not always able to get an interpreter if needed.
Forty percent of children were reported to have had an illness or injury needing care in the past 6 months. Among these children, there were significant differences by race/ethnicity in parental report of getting care as soon as wanted, with 60.0% of black children and 60.9% of Hispanic children always getting illness care as soon as wanted compared with 71.5% of white children (P = .018) (). Among the 58% of children whose parents made them an appointment for routine care in the past 6 months, significant differences by race/ethnicity were also seen in report of getting care as soon as wanted. Only 50.5% of black children and 48.5% of Hispanic children always got routine care as soon as the parent wanted, whereas 57.8% of white children did (P <.001). Children whose parents’ primary language at home was not English were less likely to report always receiving illness care as soon as wanted compared with those whose parents’ primary language at home was English (42.5% vs 72.3%, respectively, for illness care [P <.001]; 37.8% vs 57.8%, respectively, for routine care [P <.001]). Children of parents who reported ever having problems communicating with the child’s provider because of language were less likely to always receive illness care and routine care as soon as wanted compared with children of parents who never had such communication problems (42.1% vs 73.5%, respectively, for illness care [P <.001]; 38.5% vs 58.5%, respectively, for routine care [P <.001]). The frequency of always getting illness and routine care as soon as wanted was lowest for children whose parents did not always get an interpreter when needed, somewhat higher for those whose parents always got an interpreter when needed, and highest for those who did not need an interpreter (30.2% vs 57.1% vs 70.4%, respectively, for illness care [P <.001]; 27.7% vs 43.3% vs 56.0%, respectively, for routine care [P <.001]).
Of parents who did not always get an interpreter when needed, 72.4% had ever had a hard time communicating with the child’s provider because of language differences, compared with 35.6% of parents who always got an interpreter when needed and 11% of those who did not need an interpreter (P <.001).
In testing for multicollinearity among race/ethnicity and language-related variables, the largest Pearson correlation coefficient was r = 0.63 for the relationship between Hispanic race/ethnicity and having a parent whose primary language at home was not English; given the lack of correlation coefficients greater than or equal to 0.80, all variables were included in the multivariate models.
Results of the ordered logistic regression are presented separately for illness care and for preventive care (). The regressions show the relationship between the predictors and increasing frequency of getting care as soon as wanted for the child; a positive beta coefficient indicates more often getting care as soon as wanted. Compared with white children, there were no significant differences in frequency of receiving illness or routine care as soon as wanted for children of other race/ethnicities in these adjusted analyses.
Children had a significantly decreased frequency of getting illness or routine care as soon as wanted if the parent’s primary language at home was not English or if the parent ever had a problem communicating with the child’s provider because of language. Compared with children whose parents did not need an interpreter, children whose parents always got an interpreter when needed were reported to significantly more often receive illness and routine care as soon as wanted, while there was no significant difference for those whose parents did not always get an interpreter when needed.
Using model-based direct adjustment, the simulations in and provide an estimate of the size of the differences in adjusted probabilities of getting care as soon as wanted for the race/ethnicity and language variables. These adjusted probabilities can be considered to be the probability of each outcome that would have been observed for each predictor if subjects had the same distribution of covariates as did the study population as a whole. The adjusted probability of always getting illness care as soon as wanted was significantly and substantially different according to the other language variables, with an adjusted probability of 0.55 for children whose parents’ primary language at home was not English compared with 0.74 for those whose parents’ primary language at home was English (P = .002); 0.58 for children whose parents ever had problems communicating with the child’s provider because of language compared with 0.73 for those whose parents never had a problem (P = .005); and 0.59 for those whose parents needed but did not always get an interpreter compared with 0.71 for those whose parents did not need an interpreter and 0.86 for those whose parents always got an interpreter when needed (overall P = .006 by Wald test) ().
For getting routine care as soon as wanted, the adjusted probabilities followed patterns similar to those for illness care (). The adjusted probability of always getting routine care as soon as wanted was significantly and substantially different for each of the other language variables: 0.43 for children whose parents’ primary language at home was not English compared with 0.60 for children whose parents’ primary language at home was English (P = .002); 0.48 for children of parents who ever had difficulty communicating with their child’s provider due to language compared with 0.58 for those who never had such a problem (P = .005); and 0.49 for children whose parents needed but did not always get an interpreter compared with 0.57 for those who did not need an interpreter and 0.71 for those who always got an interpreter when needed (overall P = .004 by Wald test).
Our findings must be interpreted in light of several limitations. Like other observational studies, this study was not able to control for bias due to unmeasured confounders such as income, parental employment status, transportation availability, immigration status, natality, or length of residence in the United States. Our population of children receiving Medicaid was likely to be more homogeneous with respect to factors such as income and parental employment than the general population, which may attenuate the confounding effects of these unmeasured factors; only children in families with incomes under 200% of the federal poverty level were eligible for Medicaid managed care in Washington State during the study period. We were able to control for parental education, an important proxy for unmeasured factors such as income and employment. However, the lack of a significant association between education and timeliness in adjusted models was unexpected, although not inconsistent with some studies.47-49 The lack of significant association in our study may be because of the relatively small number of subjects in our sample with less than a high school education. The lack of a large significant association between lower education and less timely care at least suggests that there was no substantial confounding from unmeasured factors for which education may be a proxy (eg, income, employment).
Our study considered the parent’s primary language to be that which he/she spoke at home, but this primary language may not be the language used in the healthcare setting. Parental LEP may be a better measure of language barriers that affect access to care than the primary language spoken at home,23 but our survey was unable to measure this important construct. Because the survey was only available in English and Spanish, it included only a small percentage of parents whose primary language at home was neither English nor Spanish. As a result, our findings may not be generalizable to families who speak other languages and may underestimate the barriers faced by parents who speak less common languages. Also, our results may not be generalizable to other regions of the country with different ethnic, cultural, and linguistic groups.
CAHPS does not include information about the illnesses or routine care to which the parents refer, so it is not possible to know the degree to which this care was truly needed. Although our measure of perceived delay is subjective, the validity of this and other measures of reported delay have been demonstrated.5,50
The survey response rate of 56% is a limitation. However, our response rate is reasonable compared with other published studies in Medicaid populations, where the response rate can range from 40% to 50%.49,51-53 If nonrespondents were more likely to have limited education and literacy, then our results may underestimate the risk for untimely access due to language barriers. The education level of our study sample does appear high, with about 80% of respondents having a high school or greater education. However, this level is comparable to that of the general population of parents of lowincome children in Washington State, of whom 78% have a high school or greater education based on 2006 data from the Annual Social and Economic Supplement of the Current Population Survey.54
2. Oâ€™Toole SJ, Karamanoukian HL, Allen JE, et al. Insurance-related differences in the presentation of pediatric appendicitis. J Pediatr Surg. 1996;31(8):1032-1034.
4. Brousseau D, Bergholte J, Gorelick M. The effect of prior interactions with a primary care provider on nonurgent pediatrics emergency department use. Arch Pediatr Adolesc Med. 2004;158(2):78-82.
6. Smink DS, Fishman SJ, Kleinman K, Finkelstein JA. Effects of race, insurance status, and hospital volume on perforated appendicitis in children. Pediatrics. 2005;115(4):920-925.
8. Agency for Healthcare Research and Quality. Statistical Brief 3: Childrenâ€™s Health Care Quality, Fall 2000. May 8, 2002. http://www.meps.ahrq.gov/mepsweb/data_files/publications/st3/stat03.pdf. Accessed
9. Newacheck PW, Pearl M, Hughes DC, Halfon N. The role of Medicaid in ensuring childrenâ€™s access to care. JAMA. 1998;280(20):1789-1793.
11. Saha S, Komaromy M, Koepsell TD, Bindman AB. Patient-physician racial concordance and the perceived quality and use of health care. Arch Intern Med. 1999;159(9):997-1004.
13. Perez-Stable EJ, Napoles-Springer A, Miramontes JM. The effects of ethnicity and language on medical outcomes of patients with hypertension or diabetes. Med Care. 1997;35(12):1212-1219.
15. Stevens GD, Shi L, Cooper LA. Patient-provider racial and ethnic concordance and parent reports of the primary care experiences of children. Ann Fam Med. 2003;1(2):105-112.
17. Hampers LC, McNulty JE. Professional interpreters and bilingual physicians in a pediatric emergency department: effect on resource utilization. Arch Pediatr Adolesc Med. 2002;156(11):1108-1113.
19. Cohen AL, Christakis DA. Primary language of parent is associated with disparities in pediatric preventive care. J Pediatr. 2006;148(2):254-258.
21. Kirkman-Liff B, Mondragon D. Language of interview: relevance for research of southwest Hispanics. Am J Public Health. 1991;81(11):1399-1404.
23. Flores G, Abreu M, Tomany-Korman SC. Limited English proficiency, primary language at home, and disparities in childrenâ€™s health care: how language barriers are measured matters. Public Health Rep. 2005;120(4):418-430.
25. Flores G. The impact of medical interpreter services on the quality of health care: a systematic review. Med Care Res Rev. 2005;62(3):255-299.
27. Flores G, Abreu M, Olivar MA, Kastner B. Access barriers to health care for Latino children. Arch Pediatr Adolesc Med. 1998;152(11):1119-1125.
29. Homer CJ, Fowler F, Gallagher P, et al. The Consumer Assessment of Health Plan Study (CAHPS) survey of childrenâ€™s health care. Jt Comm J Qual Improv. 1999;25(7):369-377.
31. Christakis DA, Wright JA, Zimmerman FJ, Bassett AL, Connell FA. Continuity of care is associated with high-quality care by parental report. Pediatrics. 2002;109(4):e54.
33. Adam MB, McGuire JK, Walsh M, Basta J, LeCroy C. Acculturation as a predictor of the onset of sexual intercourse among Hispanic and white teens. Arch Pediatr Adolesc Med. 2005;159(3):261-265.
35. Huang ZJ, Yu SM, Ledsky R. Health status and health service access and use among children in U.S. immigrant families. Am J Public Health. 2006;96(4):634-640.
37. Manson A. Language concordance as a determinant of patient compliance and emergency room use in patients with asthma. Med Care. 1988;26(12):1119-1128.
39. Flores G, Olson L, Tomany-Korman SC. Racial and ethnic disparities in early childhood health and health care. Pediatrics. 2005;115(2):e183-e193.
41. Mosen DM, Carlson MJ, Morales LS, Hanes PP. Satisfaction with provider communication among Spanish-speaking Medicaid enrollees. Ambul Pediatr. 2004;4(6):500-504.
http://www.cmwf.org/usr_doc/youdelman_languageinterp_541.pdf. Accessed May 3, 2006.
44. Black G. Medicaid interpreter services facing budget ax. Seattle Post-Intelligencer. February 11, 2002:B1.
46. Chen AH, Youdelman MK, Brooks J. The legal framework for language access in healthcare settings: Title VI and beyond. J Gen Intern Med. 2007;22(suppl 2):362-367.
48. Stevens GD, Seid M, Mistry R, Halfon N. Disparities in primary care for vulnerable children: the influence of multiple risk factors. Health Serv Res. 2006;41(2):507-531.
50. Hackett T, Cassem N, Raker JW. Patient delay in cancer. N Engl J Med. 1973;289(1):14-20.
52. Sisk JE, Gorman SA, Reisinger AL, Glied SA, DuMouchel WH, Hynes MM. Evaluation of Medicaid managed care. Satisfaction, access, and use. JAMA. 1996;276(1):50-55.
54. National Center for Children in Poverty. Washington: Demographics of low-income children. http://www.nccp.org/profiles/WA_profile_6.html. Accessed January 22, 2008.