Racial Health Disparities Among Special Health Care Needs Children With Mental Disorders

Abstract

Background: A health care reform has been taking place to provide cost-effective and coordinated care. One method of achieving these goals is a patient-centered medical home (PCMH) model, which is associated with provision of quality care among children belonging to racial/ethnic minorities. Despite the potential of the PCMH for children of minority backgrounds, little is known about the extent to which minorities with mental disorders have the PCMH. Objective: The study examined racial/ethnic disparities among children with mental disorders in accessing care from the PCMH. Methods: The 2009-2010 National Survey of Children with Special Health Care Needs (CSHCN) was used for this analysis. Multivariate logistic regressions were applied to capture the racial/ethnic disparities and to analyze a composite outcome of the PCMH. Results: An estimated population size of 4 677 904 CSHCN with mental disorders was included. Among them, 59.94% of children reported to have received medical homes. Compared with white children, the odds of receiving any medical home services decreased among Hispanic children (odds ratio [OR] = 0.69; P < .05) and black children (OR = 0.70; P < .05). The likelihood of having a medical home was lower for Hispanic children than white children, when they had attention deficit hyperactivity disorder (ADHD; OR = 0.57; P < .05) and development delay (OR = 0.73; P < .05). Compared with white children with ADHD or depression having a medical home, the odds of black children with ADHD (OR = 0.63; P < .05) and depression (OR = 0.68; P < .05) having a medical home were lower. Conclusions: There were significant racial/ethnic disparities among CSHCN with mental disorders, indicating several sizeable effects of each of the 5 components on Hispanic, black, and other children compared with white children. These differences could be a potential to improve racial/ethnic disparities.

Keywords

access to care children health outcomes patient-centeredness program evaluation

Introduction

In recent years, mental disorders have surfaced as a product of much attention. Advancements in both medicine and technology have shed light on many controversial aspects that surround mental disorders; more specifically the diagnosis and treatment of mental disorders. As capabilities in the fields of science and medicine have expanded, our understanding of the disorders has increased, giving us many possible options for their treatment and management. Along with the additional treatment possibilities, there have been considerable amount of controversies, regarding the high infiltration of stimulants into the nation’s youth population.¹

One child/adolescent in every 10 children/adolescents has a mental illness that can lead to impairment in the future. According to the US Department of Health and Human Services, almost 75% to 80% children needing mental health care are unable to receive it.² About 80% children and 40% youth residing in foster homes, respectively, have a serious behavioral or mental health condition that needs treatment.³ Unfortunately, the existent health disparities among children with mental disorders are because of factors such as geographic location, socioeconomic status, race, and gender. Health disparities among children with mental health illnesses are widespread across the nation. Data from the 2001-2004 National Health and Nutrition Examination Survey indicates that approximately only half the children affected by mental disorders receive professional treatment.⁴ In a study looking at access to mental care among children belonging to racial minorities, it was found that white children residing in areas of high poverty were more likely to receive treatment compared with black and Hispanic children and vice versa.⁵

In the past couple of years in the United States, a health care reform has been taking place to provide cost-effective and coordinated patient care. One method of achieving these goals is by reorganizing primary care and moving toward a patient-centered medical home (PCMH) model. The PCMH model provides a comprehensive, team-based care that can address the different needs of a patient ranging from physical to mental health, can improve access to care by improving accessibility for a patient to receive care, can serve as a source of regular primary care and can ensure better care coordination.^6-8

Children with special health care needs are defined as the children who “require health and related services of a type or amount beyond that required by children generally.”⁹ Medical homes are known to be associated with provision of lower quality care among racial and ethnic minority children with autism and developmental disabilities.¹⁰ A study analyzing the 2005-2006 National Survey of Children with Special Health Care Needs showed that black children with a medical home had a higher probability of visiting the emergency room compared with white children with a medical home.¹¹ The above studies indicate that, in general, racial health disparities exist in children with medical homes. However, there is very little research specifically focusing on the access to medical homes among children with mental diseases, and especially using the criteria of medical homes defined by the American Academy of Pediatrics. Therefore, the aim of this study was to examine racial/ethnic disparities among children with mental disorders in accessing care from a medical home.

Methods

Conceptual Framework

This study employs the Andersen model of health care utilization as a conceptual framework to examine the effects of race/ethnicity on having a medical home (Figure 1). The Andersen model is commonly used to assess factors that are associated with patient utilization of health care services.^12-14 This model assumes that individual determinants of health care utilization are predisposing, enabling, and need factors. Predisposing factors are individual characteristics that preexist to the onset of illness and describe the propensity of individuals to use health care services.¹³ Second, enabling factors are factors that pertain to individuals who have the means that allow them to use health services.¹³ Third, the need for care describes the level of experienced illness. The actual use of health services are triggered by the need factors because the predisposing and enabling factors alone are not sufficient to do the same but are still necessary for the use of health services.¹³

Figure 1.

The Anderson model of health care utilization to explain the effects of race/ethnicity on having a medical home in children with mental disorders.

Data Source

Data from the 2009-2010 National Survey of Children with Special Health Care Needs (NS-CSHCN), conducted by the Centers for Disease Control and Prevention’s National Center for Health Statistics State and Local Area Integrated Telephone Survey program, were used for these analyses. The NS-CSHCN is a nationally representative cross-sectional telephone survey of CSHCN and was conducted in 50 states and the District of Columbia between July 2009 and March 2011. Data were collected using a list-assisted random-digit-dial sample of US households with at least 1 resident child aged 0 to 17 years. The sample design was stratified by state and sample type (landline or cell-phone) and clustered for children within households. A parent or guardian with knowledge of the health and health care of the children in the household participated as a respondent. The majority of the respondents were mothers (75%). A total of 40 242 interviews were completed. The response rate was 43.7% for the landline sample, 15.2% for the cell-phone sample, and 25.5% for the combined sample.

Study Population

The target population in this study comprised all noninstitutionalized CSHCN with mental disorders in the United States. Children were included in this study if they (1) had one or more of the following conditions—attention deficit hyperactivity disorder (ADHD), depression, anxiety problems, behavior or conduct problems, autism or other autism spectrum disorder, developmental delay, and intellectual disability or mental retardation and (2) were aged 6 to 17 years.

Measures

A composite dependent variable was the receipt of a medical home. As defined by the framework of the American Academy of Pediatrics, this composite variable was generated based on the following 5 components of a medical home: (1) the child had a personal doctor or nurse, (2) the child had a usual source of care, (3) the child had family-centered care, (4) the child had problem-free referrals (when applicable), and (5) the child had coordinated care (when applicable).^15,16 Two components (the child had problem-free referrals and the child had coordinated care) were used “when applicable” since some children may not need referrals or care transition and coordination, and thus these criteria were not applicable in these cases. To be considered as the receipt of a medical home, all children should meet the criteria of having a personal doctor or nurse, having a usual source of care, and having family-centered care, while they have to meet the criteria of having problem-free referrals and/or having coordinated care when applicable. The independent variable was race/ethnicity, which was categorized as whites (non-Hispanic), blacks (non-Hispanic), Hispanics, and other races.

The covariates were determined based on the Andersen model. In our study, predisposing factors included age, gender, family structure, and parental education. Age was categorized as a continuous variable and gender was dichotomized as males and females. Family structure was classified as 2-parent, single parent, and others. The parental education level was categorized as less than high school, high school, and more than high school. Enabling factors included geographic residence, family income, and insurance status. Geographic residences were categorized as Northeast, Midwest, South, and West. Family income were measured as a percentage of poverty and categorized as <100% of federal poverty level (FPL), 100% to 199% FPL, 200% to 399% FPL, and ≥400% FPL. Health insurance was categorized as public insurance, private insurance, both private and public insurance, and other forms of insurance and uninsured. Need factors included the perceived severity of mental disorders and comorbidity. A parent- or guardian-reported mental health status were categorized into the 2 groups based on the severity of mental disorders (ADHD, depression, anxiety problems, behavior or conduct problems, autism or other autism spectrum disorder, developmental delay, and intellectual disability or mental retardation): (1) if a child had one or more severe mental disorders, it was defined as severe mental illness and (2) if a child had only mild to moderate mental disorders, it was defined as nonsevere mental illness. Regarding the comorbidity variable, children were classified into the 2 groups based on the presence of physical disorders (asthma, diabetes, seizure, migraines or frequent headache, head injury or concussion, heart problem, blood problem, cystic fibrosis, cerebral palsy, muscular dystrophy, or arthritis or other joint problem): (1) a child who had only mental disorders without physical disorders and (2) a child who had mental disorders with physical disorders.

Statistical Analyses

The sample design effects were incorporated into our secondary analyses of the survey data for the precision of estimators.¹⁷ We used variance estimation based on Taylor linearization for a complex (stratified and weighted) sample.¹⁸ The household identification as the Primary Sampling Unit was not used because there was only 1 child for each household selected in NS-CSHCN data.

Generalized linear regression techniques using multivariate logistic regression were employed to determine the racial and ethnic disparities in having a medical home and each of the five components of a medical home. For each multivariate logistic regression model, we conducted the following 5 steps: (1) We conducted preliminary bivariate association analyses to find potential predictors for multivariate logistic regression model, (2) we built up the initial multivariate logistic regression model using potential significant predictors and variables of interest, (3) we conducted the design-adjusted Wald test to evaluate the categorical predictors in the model, (4) we rebuilt the final multivariate logistic regression model using all predictors with P value less than .05 and the variables of interest, and (5) we used Archer and Lemeshow’s design adjusted test to assess the goodness of fit of the model. The α level for statistical significance was used at .05. All analyses were performed using Stata/IC 12.1 (Stata Corp, College Station, TX).

Results

Population Characteristics

Table 1 shows the characteristics of the study population, that is, CSHCN with mental disorders aged 6 to 17 years. The estimated population size was 4 677 904 children met the study inclusion criteria. Among them, 59.94% children reported to have received medical home care. The majority of the study population were males (65.01%), non-Hispanic whites (63.72%), living in the south (39.37%), had private insurance (44.34%), and lived in 2-parent households (60.77%). More children had mild-to-moderate mental disorders (75.36%) compared with severe mental disorders and had only mental disorders without physical diseases (60.46%).

Table 1.

Descriptive Statistics for the Study Population (Weighted N = 4 677 904 and Unweighted N = 17 059).

Variable	Category	Percentage (95% CI)
Receipt of care from medical home	Yes	59.06 (57.25, 60.84)
Receipt of care from medical home	No	40.94 (39.16, 42.75)
Race	Hispanic	14.81 (10.68, 20.17)
	White, non-Hispanic	63.72 (59.25, 67.96)
	Black, non-Hispanic	13.86 (11.61, 16.46)
	Other	7.62 (6.26, 9.23)
Region	Northeast	18.58 (8.34, 36.25)
	Midwest	23.73 (12.57, 40.22)
	South	39.37 (24.56, 56.42)
	West	18.37 (8.35, 35.74)
Gender	Female	34.99 (33.86, 36.13)
	Male	65.01 (63.87, 66.14)
Insurance type	Private	44.34 (41.56, 47.15)
	Public	39.26 (36.75, 41.84)
	Both private and public	9.85 (8.04, 12.03)
	Other	3.06 (2.46, 3.79)
	Uninsured	3.49 (2.91, 4.18)
Poverty level of the household	Poor (0% to 100%)	24.45 (22.51, 26.50)
Poverty level of the household	Low (>100% to 200%)	23.19 (21.29, 25.20)
	Middle (>200% to 400%)	27.63 (26.24, 29.07)
	High (>400%)	24.73 (21.93, 27.76)
Family structure	Two parents	60.77 (58.31, 63.18)
	Single parent	28.73 (27.25, 30.27)
	Other	10.50 (9.29, 11.84)
Highest parental education level	Less than high school	12.12 (10.64, 13.76)
Highest parental education level	High school	21.29 (19.48, 23.22)
	More than high school	66.59 (64.08, 69.01)
Severity of mental illness	Yes	24.64 (23.32, 26.01)
Severity of mental illness	No	75.36 (73.99, 76.68)
Comorbidity	Mental illness only	60.46 (58.75, 62.15)
	Mental illness plus physical disease	39.54 (37.85, 41.25)
Age in years, mean (95% CI)		11.85 (11.76, 11.93)

Receipt of a Medical Home Among Children With Mental Disorders

Figure 2 shows the receipt of a medical home among different races/ethnicities. Overall, 63.74% white children, 49.87% black children, and 49.25% Hispanic children received receiving medical home services. The results for the 6 logistic models on the 5 components of a medical home and the composite of these components in children with any mental disorders are reported in Table 2. In the composite model, the race/ethnicity, age, gender, family structure, parental education, region, insurance status, family income, severity of mental disorders, and comorbidity variables were used as the potential significant predictors in the initial multivariate logistic regression model; and then the family structure, parental education variables were excluded by the design-adjusted Wald test to evaluate the categorical predictors. Finally, the race/ethnicity, age, gender, region, insurance status, family income, severity of mental disorders, and comorbidity variables were used in the final multivariate logistic regression model. All design-adjusted Wald tests showed that all included predictors in the final multivariate logistic regression model were statistically significant.

Figure 2.

The receipt of a medical home among different races/ethnicities.

Table 2.

The Likelihood of Having a Medical Home and 5 Components of a Medical Home Among All Children With Any Mental Disorder.^a

Variable	Odds Ratio (95% CI)
Variable	Composite Model	Model 1: Personal Doctor/Nurse	Model 2: Usual Source of Care	Model 3: Family-Centered Care	Model 4: Problem-Free Referrals	Model 5: Coordinated Care
Predisposing factors
Race
Hispanic	0.69 (0.59, 0.81)***	1.00 (0.72, 1.40)	0.65 (0.51, 0.83)**	0.66 (0.53, 0.84)**	0.53 (0.41, 0.70)***	1.78 (1.27, 2.49)**
Black, non-Hispanic	0.70 (0.56, 0.88)**	0.72 (0.56, 0.94)*	0.87 (0.66, 1.16)	0.58 (0.40, 0.84)**	1.05 (0.76, 1.47)	0.79 (0.47, 1.31)
Other	0.84 (0.69, 1.02)	0.72 (0.49, 1.08)	0.84 (0.65, 1.10)	0.98 (0.72, 1.33)	0.97 (0.67, 1.41)	1.22 (0.87, 1.72)
Age	0.99 (0.97, 1.01)	0.97 (0.93, 1.01)	0.99 (0.96, 1.02)	0.98 (0.96, 1.01)	0.99 (0.95, 1.02)	0.98 (0.95, 1.02)
Gender
Male	0.86 (0.78, 0.96)**	0.97 (0.75, 1.25)	0.81 (0.66, 0.99)*	0.85 (0.69, 1.04)	1.02 (0.89, 1.18)	0.90 (0.72, 1.14)**
Enabling factors
Region
Midwest	1.21 (1.02, 1.44)*	0.91 (0.71, 1.17)	0.73 (0.58, 0.92)**	1.40 (1.12, 1.75)**	1.07 (0.79, 1.45)	1.20 (0.85, 1.68)
South	1.06 (0.91, 1.25)	0.87 (0.68, 1.10)	0.76 (0.62, 0.95)*	1.05 (0.81, 1.36)	1.09 (0.79, 1.49)	1.10 (0.78, 1.58)
West	0.88 (0.75, 1.02)	0.80 (0.63, 1.01)	0.76 (0.62, 0.93)**	1.09 (0.87, 1.37)	0.71 (0.52, 0.97)*	0.81 (0.55, 1.20)
Insurance type
Private	2.17 (1.67, 2.84)***	3.05 (1.83, 5.08)***	2.64 (1.82, 3.84)***	2.16 (1.52, 3.07)***	2.83 (1.72, 4.65)***	0.74 (0.32, 1.70)
Public	1.85 (1.37, 2.49)***	2.65 (1.73, 4.05)***	3.39 (2.47, 4.65)***	1.56 (0.93, 2.61)	2.60 (1.62, 4.15)***	0.96 (0.52, 1.78)
Both private and public	2.18 (1.66, 2.86)***	3.18 (1.93, 5.24)***	2.78 (1.90, 4.06)***	2.06 (1.24, 3.44)**	3.25 (1.91, 5.54)***	1.65 (0.73, 3.73)
Other	2.81 (1.88, 4.18)***	3.52 (1.70, 7.31)**	3.87 (1.78, 8.40)**	1.88 (1.05, 3.36)*	2.61 (1.34, 5.09)**	1.56 (0.77, 3.17)
Poverty level of the household
≤100%	0.63 (0.52, 0.78)***	0.32 (0.21, 0.50)***	0.44 (0.31, 0.63)***	0.55 (0.37, 0.82)**	0.88 (0.63, 1.24)	1.30 (0.77, 2.21)
>100% to 200%	0.72 (0.62, 0.83)***	0.47 (0.32, 0.68)***	0.61 (0.43, 0.85)**	0.56 (0.42, 0.74)***	0.75 (0.53, 1.05)	1.13 (0.81, 1.58)
>200% to 400%	0.85 (0.75, 0.98)*	0.67 (0.49, 0.93)*	0.73 (0.58, 0.93)*	0.61 (0.51, 0.74)***	1.04 (0.77, 1.43)	1.04 (0.74, 1.47)
The need for care
Severity of mental illness
Severe	0.66 (0.56, 0.78)***	1.05 (0.83, 1.31)	1.24 (0.91, 1.68)	0.83 (0.64, 1.09)	0.60 (0.46, 0.79)***	0.93 (0.58, 1.50)
Comorbidity
Mental and physical disease	0.89 (0.81, 0.98)***	1.24 (0.95, 1.62)	1.49 (1.12, 1.97)**	0.94 (0.80, 1.11)	0.73 (0.60, 0.91)**	1.28 (0.99, 1.65)

Reference categories: female, white, northeast, uninsured, poverty level of the household >400% federal poverty level, not having severe mental illness, mental illness only. The race/ethnicity, age, gender, family structure, parental education, region, insurance status, family income, severity of mental disorders, comorbidity variables were used in the initial multivariate logistic regression model; and after the design-adjusted Wald test, the race/ethnicity, age, gender, region, insurance status, family income, severity of mental disorders, comorbidity variables were used in the final multivariate logistic regression model.

P < .05. **P < .01. ***P < .001.

Among the predisposing factors, race/ethnicity, age and gender were retained in the final models. Overall, the odds of receiving medical home services decreased among Hispanic children (odds ratio [OR] = 0.69; 95% confidence interval [CI] = 0.59-0.81) and black children (OR = 0.70; 95% CI = 0.56-0.88) compared with white children. Regarding each of the specific 5 components, when compared with white children, Hispanic children were less likely to have a usual source of care (OR = 0.65; 95% CI = 0.51-0.83), family-centered medical home (OR = 0.66; 95% CI = 0.53-0.84), and problem-free referrals (OR = 0.53; 95% CI = 0.41-0.70) whereas black children were less likely to have a personal doctor or nurse (OR = 0.72; 95% CI = 0.56-0.95) and family-centered medical home (OR = 0.58; 95% CI = 0.40-0.84) respectively. Hispanic children were more likely to have coordinated care than white children (OR = 1.78; 95% CI = 1.27-2.49). With regard to other predisposing factors, male children were associated with a lower likelihood of having a medical home compared with female children (OR = 0.86; 95% CI = 0.78-0.96), indicating significantly lower probability of having usual source of care and coordinated care.

All the enabling factors were retained in the final models. Regarding insurance status, children with insurance had a higher likelihood of having a medical home compared with those without insurance depending on the type of insurance; specifically, the likelihoods of having a medical home were 2.17 times higher in children with private insurance, 1.85 times higher in children with public insurance, and 2.18 times higher in children with both private and public insurance than that in children without insurance, controlling for the other predictor. Additionally, the odds of having a personal doctor or nurse, usual source of care, and problem-free referrals increased across all types of insurance. Regarding FPL, compared with households with FPL higher than 400%, the odds of having a medical home were significantly lower in households with (1) <100% FPL (OR = 0.63; 95% CI = 0.52-0.78), (2) 100% to 199% FPL (OR = 0.72; 95% CI = 0.62-0.83), and (3) 200% to 399% FPL (OR = 0.85; 95% CI = 0.75-0.98). This means that the odds of having a medical home decreased by a factor of 0.63 in households with <100% FPL, 0.72 in households with 100% to 199% FPL, and 0.85 in households with 200% to 399% FPL when compared with the odds in households with >400% FPL, controlling for the other predictor. Similar results were seen when households with FPL higher than 400% were compared with households with <100% FPL, 100% to 199%, and 200% and 399% FPL for individual components: having personal doctor or nurse, usual source of care, and family-centered care respectively.

Both the need for care factors were retained in the final models. First, for children with severe mental disorders, the odds of the receipt of a medical home were 0.66 times the odd of children with nonsevere mental disorders, holding constant the other predictors (OR = 0.66; 95% CI = 0.56-0.78). Specifically, children with severe mental disorders were less likely to having problem-free referrals. Second, for children with both mental and physical disorders, the odds of the receipt of a medical home were 0.89 times the odd of children with only mental disorders, controlling the other predictors (OR = 0.89; 95% CI = 0.81-0.98); among the 5 components, having a usual source of care and problem-free referrals were significantly related to physical comorbidities.

Receipt of a Medical Home Among Children With Each Mental Disorder by Subgroup Analyses

Table 3 shows the results for subgroup analyses on the receipt of a medical home among different races/ethnicities, controlling for covariates, according to the mental disorder type. In the composite model, compared with white children, the likelihood of having a medical home in Hispanic children was 0.57 times lower when they had ADHD (OR = 0.57; 95% CI = 0.44-0.73) and 0.73 times development delay (OR = 0.73; 95% CI = 0.60-0.94). The odds of black children having a medical home were 0.63 times and 0.68 times lower than those in white children with ADHD (OR = 0.63; 95% CI = 0.50-0.80) and depression (OR = 0.68; 95% CI = 0.48-0.97), respectively.

Table 3.

The Likelihood of Having a Medical Home and 5 Components of a Medical Home by Race/Ethnicity by Subgroup Analyses in Each Mental Disorder After Controlling for Covariates (Age, gender, Family Structure, Parental Education, Region, Insurance Status, Family Income, Severity of Mental Disorders, Comorbidity).^a

Variable	Odds Ratio (95% CI)
Variable	Composite Model	Model 1: Personal Doctor or Nurse	Model 2: Usual Source of Care	Model 3: Family-Centered Care	Model 4: Problem-Free Referrals	Model 5: Coordinated Care
Attention deficit hyperactivity disorder (weighted N = 3 052 912)
Hispanic	0.57 (0.44, 0.73)***	0.76 (0.48, 1.21)	0.49 (0.33, 0.74)**	0.68 (0.47, 0.99)*	0.51 (0.33, 0.79)***	1.42 (0.85, 2.38)
Black, non-Hispanic	0.63 (0.50, 0.80)***	0.63 (0.43, 0.92)**	0.71 (0.49, 1.03)	0.63 (0.45, 0.89)**	1.07 (0.67, 1.69)	0.78 (0.47, 1.29)
Other	0.83 (0.62, 1.12)	0.74 (0.45, 1.23)	0.69 (0.44, 1.07)	1.15 (0.75, 1.77)	1.26 (0.78, 2.04)	0.93 (0.46, 1.89)
Depression (weighted N = 1 066 685)
Hispanic	0.91 (0.68, 1.23)	1.97 (0.80, 4.88)	0.63 (0.28, 1.45)	0.72 (0.43, 1.22)	0.77 (0.49, 1.22)	1.10 (0.64, 1.89)
Black, non-Hispanic	0.68 (0.48, 0.97)*	0.82 (0.43, 1.56)	0.75 (0.40, 1.41)	0.78 (0.44, 1.39)	0.85 (0.49, 1.49)	0.58 (0.26, 1.28)
Other	0.85 (0.55, 1.31)	0.51 (0.26, 1.01)	0.79 (0.28, 2.22)	0.84 (0.44, 1.62)	1.42 (0.66, 3.07)	1.19 (0.53, 2.70)
Anxiety problems (weighted N = 1 702 375)
Hispanic	0.86 (0.67, 1.11)	1.03 (0.52, 2.05)	0.71 (0.37, 1.36)	0.78 (0.54, 1.13)	0.58 (0.41, 0.83)**	2.23 (1.30, 3.85)**
Black, non-Hispanic	0.69 (0.47, 1.01)	0.75 (0.41, 1.39)	1.04 (0.45, 2.40)	0.76 (0.44, 1.32)	0.66 (0.45, 0.96)*	0.78 (0.33, 1.85)
Other	0.92 (0.70, 1.20)	0.70 (0.36, 1.37)	0.92 (0.42, 2.02)	1.00 (0.61, 1.63)	1.20 (0.73, 1.98)	1.22 (0.67, 2.24)
Behavior or conduct problems (weighted N = 1 300 016)
Hispanic	0.99 (0.71, 1.40)	0.97 (0.57, 1.63)	0.89 (0.49, 1.63)	0.72 (0.45, 1.14)	0.94 (0.62, 1.43)	2.70 (1.05, 6.95)*
Black, non-Hispanic	0.94 (0.72, 1.24)	0.8475 (0.60, 1.19)	0.93 (0.49, 1.77)	0.95 (0.64, 1.42)	1.40 (0.89, 2.21)	0.84 (0.50, 1.41)
Other	1.10 (0.82, 1.50)	0.63 (0.30, 1.34)	0.91 (0.45, 1.82)	0.68 (0.42, 1.09)	1.59 (0.78, 3.25)	1.25 (0.71, 2.19)
Autism or other autism spectrum disorders (weighted N = 687 200)
Hispanic	0.81 (0.51, 1.28)	0.57 (0.27, 1.20)	0.98 (0.57, 1.70)	0.72 (0.46, 1.13)	0.38 (0.16, 0.88)*	1.82 (0.97, 3.45)
Black, non-Hispanic	1.29 (0.86, 1.94)	1.23 (0.57, 2.70)	4.21 (1.73, 10.29)**	0.82 (0.49, 1.71)	1.49 (0.49, 4.56)	1.42 (0.55, 3.69)
Other	1.09 (0.77, 1.55)	2.59 (0.94, 7.09)	1.22 (0.57, 2.61)	1.17 (0.70, 1.98)	1.04 (0.55, 1.97)	0.77 (0.27, 2.15)
Developmental delay (weighted N = 1 454 883)
Hispanic	0.73 (0.60, 0.94)*	1.12 (0.51, 2.45)	0.86 (0.49, 1.51)	0.77 (0.51, 1.17)	0.56 (0.32, 0.97)*	1.96 (1.28, 2.99)**
Black, non-Hispanic	0.74 (0.54, 1.01)	0.65 (0.35, 1.20)	1.48 (0.87, 2.52)	0.61 (0.34, 1.10)	1.52 (0.82, 2.83)	0.64 (0.29, 1.38)
Other	0.88 (0.66, 1.18)	1.17 (0.68, 2.01)	1.08 (0.68, 1.73)	1.16 (0.73, 1.83)	1.10 (0.69, 1.75)	0.87 (0.48, 1.59)
Intellectual disability or mental retardation (weighted N = 533 540)
Hispanic	0.77 (0.53, 1.10)	0.78 (0.33, 1.82)	2.35 (0.98, 5.65)	0.57 (0.32, 1.01)	0.46 (0.24, 0.86)*	1.36 (0.73, 2.52)
Black, non-Hispanic	0.74 (0.48, 1.13)	0.55 (0.18, 1.74)	2.03 (0.90, 4.62)	0.29 (0.12, 0.71)**	1.40 (0.67, 2.91)	0.36 (0.11, 1.23)
Other	0.71 (0.48, 1.03)	1.51 (0.71, 3.21)	1.38 (0.40, 4.77)	1.43 (0.49, 4.15)	0.54 (0.28, 1.05)	0.90 (0.39, 2.09)

Reference category: white. The race/ethnicity, age, gender, family structure, parental education, region, insurance status, family income, severity of mental disorders, comorbidity variables were used in the initial multivariate logistic regression model; and after the design-adjusted Wald test, the race/ethnicity, age, gender, region, insurance status, family income, severity of mental disorders, comorbidity variables were used in the final multivariate logistic regression model.

P < .05. **P < .01. ***P < .001.

Several favorable effects in each of the 5 components in Hispanic, black, and other children were significantly different from those in white children based on metal disease type. The odds of having a personal doctor or nurse were 0.63 times lower in black children with ADHD compared with white children with ADHD (OR = 0.63; 95% CI = 0.43-0.92). Hispanic children with ADHD were 0.49 times less likely to have a usual source of care compared with white children with ADHD (OR = 0.49; 95% CI = 0.33-0.74), while black children with autism were 4.21 times more likely to have a usual source of care than white children with autism (OR = 4.21; 95% CI = 1.73-10.29). Among the ADHD population, Hispanic children were 0.68 times (95% CI = 0.47-0.99) and black children were 0.63 times (95% CI = 0.45-0.89) less likely to have family-centered care compared with white children. Also, among children with intellectual disability or mental retardation, the odds of having a family centered home among black children were 0.29 times lower than in white children (OR = 0.29; 95% CI = 0.12-0.71). The likelihood of having problem-free referrals in Hispanic children was 0.51 times, 0.38 times, 0.56 times, and 0.46 times lower compared with white children with ADHD (OR = 0.51; 95% CI = 0.33-0.79), autism (OR = 0.38; 95% CI = 0.16-0.88), developmental delay (OR = 0.56; 95% CI = 0.32-0.97), and intellectual disability or mental retardation (OR = 0.46; 95% CI = 0.24-0.86), respectively. Among children with anxiety problems, the likelihoods of having problem-free referrals in Hispanic and black children were 0.58 times and 0.66 times lower compared with white children (OR = 0.58; 95% CI = 0.41-0.83 and OR = 0.66; 95% CI = 0.45-0.96, respectively). Only Hispanic children were approximately twice more likely to have coordinated care compared with white children, when they have anxiety (OR = 2.23; 95% CI = 1.30-3.85), behavior or conduct problems (OR = 2.70; 95% CI = 1.05-6.95), and developmental delay (OR = 1.96; 95% CI = 1.28-2.99). However, the likelihood of coordinated care in black children did not increase when they had anxiety, behavior or conduct problems or developmental delay, when compared with white children.

Discussion

The United States has higher health care spending per capita than any other developed country.¹⁹ Ironically, the Unites States has not only relatively lower utilization rates—such as the number of hospital days and physician visits—but also poor health care outcomes—such as lower life expectancy and higher infant mortality rate.¹⁹ Thus, the provided care is expensive while outcomes from that care are less than efficient and effective. Additionally, there is dissatisfaction among health care providers and patients with respect to the care provided within the US health care system.²⁰ Hence health care policy experts seek a better model to improve quality of care in the United States. The PCMHs were developed to remodel the health care system more efficiently. Although PCMH is considered a promising model to improve health care quality, previous studies have demonstrated that the racial/ethnic disparities exist in the utilization of medical homes among CSHCN with ADHD and other mental disorders.²¹

However, many unanswered questions still remain about the racial/ethnic disparities existent in having a medical home among children with mental disorders. To answer the unanswered questions, we tried to examine the specific components of a medical home where racial/ethnic disparities among CSHCN exist. Moreover, we tried to examine the type of mental disorders where racial/ethnic disparities among CSHCN exist to have a medical home. This study adds more evidence by examining the existence of racial/ethnic disparities in the receipt of a medical home among CSHCN with mental disorders using the most updated nationally representative data of children in the United States. The Andersen model was used as a conceptual framework to provide the more concrete analyses. Mainly, we found that the racial/ethnic disparities were shown, but the trends were different depending on the components of a medical home, the type of mental disorders, and/or racial/ethnic groups.

Overall, nearly two thirds of CSHCN had a medical home in our study. We found that approximately 59.94% (95% CI = 57.25-60.84) of CSHCN with mental disorders aged 6 to 17 years had access to a medical home. A previous study using the 2007 National Survey of Children’s Health reported that 52.2 % children with 1 behavioral health condition, including ADHD, depression, anxiety, conduct disorder, autism spectrum disorder, and developmental delay have a PCMH.²² Another study using 2005-2006 NS-CSHCN reported that nearly half of CSHCN aged 3 to 17 years with emotional or behavioral problems (46.8%) had a medical home.²³

Generally, previous studies have found racial/ethnic disparities in mental health care utilization, including higher unmet need for services and lower access to services.²⁴ We also found substantial racial/ethnic disparities in having a medical home in children in mental disorders. These were existent in both the medical home composite and its 5 components. Although both Hispanic and black children were less likely to have a medical home compered with white children, the trends seen with each 5 of the components were different between Hispanic and black children. Specifically, Hispanic children were less like to have a usual source of care, family-centered care, problem-free referrals and coordinated care compared with white children while black children were less like to have a personal doctor or nurse and family-centered care compared with white children. However, the likelihood of having family-centered care were lower in both Hispanic and black children compared with white children. Lower likelihood of having family-centered care component implied communication issues with child’s doctors and other health care providers. Previously, linguistic issues have been identified as influencing factors in worsening mental health care outcomes.²⁵ Accordingly, there is need for studies about developing interventions on family-doctor communication and translators to help non–English-speaking patients. The pattern of receipt of a medical home among CSHCN with mental disorders varied with the types of mental disorders. Both Hispanic and black children were associated with a lower likelihood of having a medical home or one of components in ADHD, anxiety, autism or other autism spectrum disorders, and intellectual disability or mental retardation subgroups. Among children with behavior or conduct problems and developmental delays, only Hispanic children were less likely to having a medical home, while only black children were less likely to having a medical home among children with depression. In addition to racial/ethnic disparities, enabling factors, including insurance type and family income, were significantly associated with the likelihood of having a medical home in CSHCN with mental disorders in both the medical home composite and most of its 5 components.

These disparities with respect to race/ethnicity and socioeconomic status can lead to differences in health status because PCMH programs in medical homes improve the possibility of better health outcomes. For example, when CSHCN’s parents can easily get the same nurse to talk to or communicate with their child’s doctor, and get referrals, children’s hospitalizations and emergency department visits get reduced.²⁶ Accordingly, reducing the racial/ethnic disparities in having a medical home could be related to preventing more serious health problems. Fortunately, federal government has started to support and promote the PCMH. Many states have been seeking an appropriate PCMH model to improve the medical home in their Medicaid or Children’s Health Insurance Program.²⁷ In the future, effective policies for addressing disparities should move beyond health insurance policies and extensions of programs, including early childhood intervention programs, the Supplemental Nutrition Program for Women, Infants, and Children, and nurse home-visiting programs, which might help alleviate the racial/ethnic disparities.^27,28

This study has several limitations. First, there is the possibility of recall bias since the survey data are self-reported. Second, another limitation related to using to survey database is that the interpretation of the result is limited to the correlation between race/ethnicity and the receipt of a medical home. Since the cross-sectional study design cannot explain causation, our results fail to address the causation of racial/ethnic disparities on the receipt of a medical home. Third, a parent or guardian participated as a respondent, but parent- or guardian-reported data for children’s mental disorders has been argued.²⁹ Thus, the accuracy of the data, including the diagnosis of disease, is not well established.

Conclusion

Overall, there were significant racial/ethnic disparities among CSHCN with mental disorders, including ADHD, depression, anxiety problems, behavior or conduct problems, autism and other autism spectrum disorders, developmental delays, intellectual disabilities or mental retardation. The components of the medical homes, which showed significant racial/ethnic disparities varied with the types of mental disorders. These results can help shape future interventions to improve PCMH model, which can lessen the gap among the different racial/ethnic minorities.

Footnotes

Acknowledgements

The authors appreciate CAHMI housed at Oregon Health & Science University for the dataset support (2009-2010 National Survey of Children with Special Health Care Needs)

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Chanhyun Park, RPh, MS, is a PhD candidate at the University of Texas, College of Pharmacy. Her research focuses on the patient-centered medical home, health outcomes and pharmaceutical health services research.

Xi Tan, PharmD, is a PhD cadidate at the University of Michigan, College of Pharmacy, Her research areas focus on the health policy evaluation and mental health.

Isha Patel, PhD, is a Visiting assistant research Scientist at the University of Michigan, College of Pharmacy. Her research focuses on the mental health and developmental disorder.

Amanda Reiff is a MPH candidate at the Penn State University College of Medicine.

Rajesh Balkrishnan, PhD, is an Associate Professor at the University of Michigan, College of Pharmacy and School of Public Health. He is an internationally recognized authority on clinical study design as well as measurement and estimation issues in observational studies. He has published more than 250 research articles, and has advised 20 PhD students towards successful degress completion in the area of pharmaceutical health services research.

Jongwha Chang, PhD, is an Assistant Professor at the Samford University, McWhorter School of Pharmacy. He has a strong interest in quantitative methodology in health services research focsed on health ooutcomes, healthcare management and physician-prescribing behaviors. He has more than 30 research articles.

References

Linares

Martinez-Martin

Castellanos

. Stim ulant and atypical antipsychotic medications for children placed in foster homes. PLoS One. 2013;8:e54152.

The Child Wellfare League of America. Child mental health—Facts and figures. http://www.cwla.org/programs/bhd/mhfacts.htm#FACTSHEETS. Accessed May 10, 2013.

Pecora

Jensen

Romanelli

Jackson

Ortiz

. Mental health services for children placed in foster care: an overview of current challenges. Child Welfare. 2009;88:5-26.

Merikangas

Brody

Fisher

Bourdon

Koretz

. Prevalence and treatment of mental disorders among US children in the 2001-2004 NHANES. Pediatrics. 2010;125:75-81.

Le Cook

Manning

Alegria

. Measuring disparities across the distribution of mental health care expenditures. J Ment Health Policy Econ. 2013;16:3-12.

Berry

Mijanovich

Albert

. Patient-centered medical home among small urban practices serving low-income and disadvantaged patients. Ann Fam Med. 2013;11(suppl 1):S82-S89.

Lotstein

Ghandour

Cash

McGuire

Strickland

Newacheck

. Planning for health care transitions: results from the 2005-2006 National Survey of Children with Special Health Care Needs. Pediatrics. 2009;123:e145-e152.

Lotstein

McPherson

Strickland

Newacheck

. Transition planning for youth with special health care needs: results from the National Survey of Children with Special Health Care Needs. Pediatrics. 2005;115:1562-1568.

McPherson

Arango

Fox

. A new definition of children with special health care needs. Pediatrics. 1998;102:137-140.

10.

Magana

Parish

Rose

Timberlake

Swaine

. Racial and ethnic disparities in quality of health care among children with autism and other developmental disabilities. Intellect Dev Disabil. 2012;50:287-299.

11.

Raphael

Zhang

Liu

Tapia

Giardino

. Association of medical home care and disparities in emergency care utilization among children with special health care needs. Acad Pediatr. 2009;9:242-248.

12.

Aday

Andersen

. A framework for the study of access to medical care. Health Serv Res. 1974;9:208-220.

13.

Andersen

McCutcheon

Aday

Chiu

Bell

. Exploring dimensions of access to medical care. Health Serv Res. 1983;18:49-74.

14.

Andersen

. Revisiting the behavioral model and access to medical care: does it matter? J Health Soc Behav. 1995;36:1-10.

15.

Strickland

McPherson

Weissman

van Dyck

Huang

Newacheck

. Access to the medical home: results of the National Survey of Children with Special Health Care Needs. Pediatrics. 2004;113(5 suppl):1485-1492.

16.

Strickland

Singh

Kogan

Mann

van Dyck

Newacheck

. Access to the medical home: new findings from the 2005-2006 National Survey of Children with Special Health Care Needs. Pediatrics. 2009;123:e996-e1004.

17.

Burt

Cohen

. A comparison of methods to approximate standard errors for complex survey data. Rev Public Data Use. 1984;12:159-168.

18.

Vermaa

Bettia

. Taylor linearization sampling errors and design effects for poverty measures and other complex statistics. J Appl Stat. 2011;38:1549-1576.

19.

Anderson

Frogner

. Health spending in OECD countries: obtaining value per dollar. Health Aff (Millwood). 2008;27:1718-1727.

20.

Helfgott

. The patient-centered medical home and accountable care organizations: an overview. Curr Opin Obstet Gynecol. 2012;24:458-464.

21.

Toomey

Homer

Finkelstein

. Comparing medical homes for children with ADHD and asthma. Acad Pediatr. 2010;10:56-63.

22.

Knapp

Woodworth

Fernandez-Baca

Baron-Lee

Thompson

Hinojosa

. Factors associated with a patient-centered medical home among children with behavioral health conditions. Matern Child Health J. 2013;17:1658-1664.

23.

Ghandour

Perry

Kogan

Strickland

. The medical home as a mediator of the relation between mental health symptoms and family burden among children with special health care needs. Acad Pediatr. 2011;11:161-169.

24.

Lopez

Barrio

Kopelowicz

Vega

. From documenting to eliminating disparities in mental health care for Latinos. Am Psychol. 2012;67:511-523.

25.

Flores

. Devising, implementing, and evaluating interventions to eliminate health care disparities in minority children. Pediatrics. 2009;124(suppl 3):S214-S223.

26.

Palfrey

Sofis

Davidson

Liu

Freeman

Ganz

. The Pediatric Alliance for Coordinated Care: evaluation of a medical home model. Pediatrics. 2004;113(5 suppl):1507-1516.

27.

Kaye

Buxbaum

Takach

. Building medical homes: lessons from eight states with emerging programs, 2011. http://www.commonwealthfund.org/~/media/Files/Publications/Fund%20Report/2011/Dec/1569_Kaye_building_medical_homes_v2.pdf. Accessed May 10, 2013.

28.

Currie

. Policy interventions to address child health disparities: moving beyond health insurance Pediatrics. 2009;124(suppl 3):S246-S254.

29.

Bhatia

. Childhood and adolescent depression. Am Fam Physician. 2007;75:73-80.