Unmet Need and Financial Impact Disparities for US Children with ADHD

Abstract

Objective:

The 5.1 million US children with ADHD have pronounced needs in education, occupational and speech therapy, and medical and behavioral treatments. Given known associations of ADHD diagnosis with race/ethnicity and parent education, this study aimed to assess how measures of socioeconomic status correlate both with adverse family financial impact of ADHD and with disparities in unmet treatment need for ADHD.

Methods:

Secondary analysis of children age 8-17 whose households participated in the 2014 National Survey of the Diagnosis and Treatment of ADHD and Tourette Syndrome. Using bivariate testing, we examined associations among measures of socioeconomic status with unmet ADHD treatment need and family financial impact. Logistic regression models estimated the odds of having unmet treatment need, adjusting for demographic factors and family financial impact.

Results:

Among US school-aged children with a current ADHD diagnosis, 44.3% experienced adverse family financial impact from ADHD and 11.6% had unmet need for ADHD treatment. Children with younger age at first ADHD diagnosis were more likely to experience adverse family financial impact. Children from non-English-speaking households were less likely to report unmet need compared to those from primarily-English speaking households. Adjusted odds of unmet need were twice as great among those who reported adverse family financial impact.

Conclusion:

Deeper understanding of household language’s influence is important in ADHD needs assessments. Considering overall family financial circumstances may also be pertinent, particularly as children age, because earlier diagnosis was associated with adverse financial outcomes. These findings could shape future clinic policies for targeting community resources.

BACKGROUND

Attention-Deficit/Hyperactivity Disorder (ADHD) is characterized by symptoms of inattention and hyperactivity that impair function in multiple settings, typically home and school, for at least six months and before age seven.^1,2 The estimated lifetime prevalence of ADHD in US children aged 4-to-17 years is 11%,^2,3 incurring financial and emotional costs to families and society.⁴ The 5.1 million US children with ADHD have pronounced needs in education and healthcare: challenges in disorder management have led to falling behind academically⁴ and adulthood concerns such as loss of work productivity and decreased likelihood to take preventative health measures.⁴ ADHD impairs quality of life and function^1,5 and is also associated with other impairing conditions such as conduct disorder and oppositional defiant disorder.⁶ Consequently, compared with their typically developing peers, children with ADHD have more challenges with school, the juvenile justice system, and maintaining relationships with peers and family.

Disparities in ADHD diagnosis due to race, ethnicity, health insurance, or parental education are established.^7-9 ADHD tends to be diagnosed less among African- and Hispanic-American patients.^8-10 Non-Hispanic white children and those with higher parental educational attainment are more likely than counterparts to receive comprehensive neurodevelopmental evaluations for ADHD.¹¹ In ethnically diverse families, differences in cultural practice, knowledge, and beliefs shape familiarity with ADHD, how to seek diagnosis, and misunderstandings due to language and stigma surrounding ADHD.¹⁰ Treatment utilization shows similar disparities: African-American and Hispanic patients with ADHD have lower odds of medication use and decreased receipt of school-based services⁹ compared with non-Hispanic white children.⁸

Previous studies have not examined how adverse financial experiences from ADHD (hereinafter “ADHD financial impact”) relate to unmet need for ADHD treatment. Literature indicates higher annual medical costs among children with ADHD compared to controls,¹² and given the known association between financial burden and racial-ethnic and educational disparities, ADHD financial impact may underlie disparities in receiving ADHD treatment. Families with fewer financial resources overall may also have more difficulty accessing ADHD care that their child needs.

This study aimed to show how measures of socioeconomic status correlate both with ADHD financial impact and unmet need for ADHD treatment. Our study’s Conceptual Framework (Figure 1) was informed by Andersen and Aday’s behavioral model of health care utilization.¹³ The Conceptual Framework shows how predisposing, pre-illness factors (e.g., age and sex) and enabling, means-to-service factors (e.g., income and insurance) interrelate with one another, with elements of need or illness level, and with psychosocial factors to impact both ADHD financial impact and unmet need.^13,14 Based on this Conceptual Framework, we predicted that among US children with ADHD, those who were non-white or Hispanic, had a household language other than English, and/or had low parent education would experience significantly more ADHD financial impact and unmet need. We also hypothesized that greater ADHD severity and comorbidity would affect the relationship between ADHD financial impact and unmet need. Finally, we refined our focus on unmet need to assess if ADHD financial impact influences the relationship with predisposing and enabling measures of socioeconomic status.

Figure 1.

Conceptual framework of study objectives

METHODS

Data Source and Study Population

We conducted a secondary analysis of the 2014 National Survey of the Diagnosis and Treatment of ADHD and Tourette Syndrome (NS-DATA), a cross-sectional survey of US children sponsored by the Centers for Disease Control and Prevention (CDC), the National Center on Birth Defects and Developmental Disabilities, and the National Center for Health Statistics (NCHS).¹⁵ NS-DATA followed the 2011–2012 National Survey of Children’s Health (NSCH), both of which were nationally representative for non-institutionalized US children through a cross-sectional random-digit-dial survey of US households with weighting to account for sampling bias in certain groups. If the household’s survey-eligible child had ever been diagnosed with ADHD or Tourette Syndrome (TS) by the time of NSCH per parent report, NS-DATA re-contacted the household 2–3 years after NSCH. NS-DATA shares the NSCH design of clustered children within households and stratification by state and sample by landline or cell phone.¹⁵

NS-DATA completed interviews for 2966 US children with ADHD who had been diagnosed at time of NSCH. Families were eligible for NS-DATA if their child was still under 18 years old, the child continued to live in the household, and respondents confirmed the child had ever been diagnosed with ADHD.¹⁵ The survey was administered in English and Spanish, though administered language was suppressed for confidentiality. The response rate for the 2011–2012 NSCH was 23%. The NS-DATA completion rate was 47.2%, a product of re-contact rate and interview completion among those re-contacted. Respondents were the parent or guardian (hereinafter “parent”) most familiar with the healthcare of the child with ADHD.¹⁵ The authors’ institutional review board determined this study to be exempt as nonhuman subjects research.

Defining the Analytic Sample

This analytic sample (Figure 2) was US non-institutionalized children ever diagnosed with ADHD at the time of NSCH who had a current ADHD diagnosis at NS-DATA. NS-DATA grouped together all children 7-years-old and younger at time of follow-up; we thus limited our analytic sample to children age 8–17, since ADHD diagnosis is not stable in preschool-aged children.⁵ We proceeded with our analytic sample, aware that the limitations of NS-DATA completion rate, the parent respondents, and age may not accurately represent the national distribution of children with ADHD, similar to other national surveys like NS-CSHCN.¹⁶ The final sample size of US children age 8–17 with a current diagnosis of ADHD captured by NS-DATA was 2406.

Figure 2.

Flow chart describing the analytic sample

Measures of Interest

Dependent Variable of Interest: Unmet Need for ADHD Treatment

The primary dependent variable of interest was “unmet need for ADHD treatment.” This was characterized by the NS-DATA survey item: “In the past 12 months, did your child need an ADHD treatment that [he/she] was unable to get?”¹⁵ If yes, this child was considered to have an unmet need in our analysis. NS-DATA follow-up questions to a yes response clarified if the treatment that the child needed was medication, school-based behavioral therapy, behavioral treatment outside of school, and/or some other treatment outside of school. Though not incorporated in the scope of our primary analysis, respondents could also cite reasons why the child could not obtain said treatment, including cost, availability, or lack of information (Appendix A).

Primary Independent Variable of Interest: ADHD Financial Impact

NS-DATA included a survey section entitled Family Impact, in which parents described the child’s ADHD financial impact. Section items asked:

• Has [the child’s] ADHD caused financial problems for your family?

• Have you or other family members stopped working because of [the child’s] ADHD?

• Not including family members who stopped working, have you or other family members cut down on the hours you work because of [the child’s] ADHD?

• Have you or other family members avoided changing jobs because of concerns about maintaining health insurance for [the child]?

Each item included the options to respond with a yes, no, don’t know, or refusal.¹⁵ In our analysis, a family was considered to have reported ADHD financial impact by saying yes to any of these section items.

Model covariates: Predisposing and Enabling Measures

We characterized the children with current ADHD diagnosis and their families by several measures of socioeconomic status and controlled for these in our multivariable analyses. Per our Conceptual Framework (Figure 1), predisposing measures included the child’s sex, race/ethnicity, primary household language, US Census Region, highest level of parent education, age at NS-DATA interview, and age at first ADHD diagnosis. Enabling measures included insurance type (public, private, or uninsured), insurance continuity in the past year, and household income based on Department of Health and Human Services (DHHS) poverty guidelines.¹⁵

We included parent-reported severity of the child’s ADHD in our multivariable analyses, as a confounding variable for the dependent variable of interest. In NS-DATA, ADHD severity was assessed by the question “Would you describe [the child’s] ADHD as mild, moderate, or severe?”¹⁵ Parent-reported severity has been shown to be a valid monitoring measure of ADHD severity, as compared with professionally reported ADHD severity in health plan medical records.¹⁷

The presence of other conditions was included in multivariable analyses due to the high rate of comorbidity in ADHD.^6,11 NS-DATA assessed the presence of co-occurring disorders through a series of items worded as “Does [the child] currently have [co-occurring condition]?”¹⁵ NS-DATA’s list of 13 comorbidities included oppositional defiant disorder, autism spectrum disorder, sleep disorder, learning and language disorders, mood disorders, and a suppressed category of other disorders including intermittent explosive disorder, eating disorder, and substance use disorder. If a parent said yes to any of these co-occurring disorders, the child was categorized as having any comorbidity.

Statistical Analyses

We described the sample in terms of their socioeconomic status, per the predisposing and enabling factors in the Conceptual Framework (Figure 1), ADHD severity, and comorbidities (Tables 1 and 2). We also computed descriptive statistics to assess the overall prevalence of unmet need and adverse financial impact (Table 3). We then computed chi-square tests to determine bivariate associations among unmet need, adverse financial impact, and the covariates of interest (Table 4). A logistic regression model was fit to examine the adjusted odds of unmet need in relationship to family financial impact and the covariates (Table 5). Given our Conceptual Framework and literature supporting race and ethnicity-based disparities in ADHD diagnosis^3,7-11 we included all covariates of interest in the model. To check for bias due to multicollinearity, we computed variance inflation factors, which were all < 2 suggesting limited bias. All analyses were weighted per guidance from NCHS to account for the complex survey sampling design.¹⁵ STATA 14.2 was utilized for this analysis.¹⁸

Table 1.

Predisposing and Enabling Measures of Socioeconomic Status, US Children with ADHD, 8-17 years (total n=2406)

Characteristic (Subgroup n)	Frequency, Unweighted	% of subgroup, Weighted
Predisposing Measures of Socioeconomic Status
Sex (n=2406)
Male	1692	70.1
Female	714	29.9
Race (n=2392)
White, Non-Hispanic	1767	63.4
Black, Non-Hispanic	201	14.4
Hispanic	198	15.4
Other	226	6.9
Primary Language in Household (n=2403)
English	2213	91.3
Language Other than English	190	8.7
Language Other than English, Hispanic*	15	1.2
Language Other than English, Non-Hispanic and Other Race/Ethnicity*	173	7.4
US Census Regions (n=2406)
South	934	39.2
West	624	24.9
Midwest	459	19.3
Northeast	389	16.7
Highest Level Education Attainment, by either parent or guardian (n=2262)
Less than High School	288	13.4
High School Graduate	771	34.3
More than High School	1203	52.3
Age at NS-DATA Interview (n=2406)
8–9yo	227	10.5
10–13yo	1009	44.1
14–17yo	1170	45.4
Age at First ADHD Diagnosis (n=2373) - continuous variable in meaningful categories
1–4yo	277	14.3
5–9yo	1703	70.5
10–14yo+	393	15.3
Enabling Measures of Socioeconomic Status
Insurance, by Type and Continuity over Last Year (n=2366)
Continuous Private	1397	47.9
Continuous Public	822	44.1
Noncontinuous Public	69	4.2
Noncontinuous Private	42	2.2
Currently Uninsured	36	1.7
Household Income based on DHHS poverty guidelines, Imputed (n=2406)
0–99% FPL	359	15.2
100–199% FPL	400	18.1
200–399% FPL	769	30.5
400% FPL or greater	878	36.2

ADHD indicates Attention Deficit/Hyperactivity Disorder; NS-DATA, 2014 National Survey of the Diagnosis and Treatment of Attention-Deficit/Hyperactivity Disorder and Tourette Syndrome; FPL, federal poverty level.

Subgroup n describes the number of participants who answered and did not skip the survey question from the total 2406.

^*Though 190 reported speaking a household language other than English, there were two missing responses when further categorized by race/ethnicity.

Source: 2014 National Survey of the Diagnosis and Treatment of ADHD and Tourette Syndrome

Table 2.

ADHD severity characteristics (total n=2406)

Characteristic (Subgroup n)	Freq, Unweighted	% of n Responses, Weighted
Would you describe the child's ADHD as mild, moderate, or severe? (n=2388)
Mild	800	31.0
Moderate	1201	49.4
Severe	387	19.6
Of children with a current diagnosis of ADHD:
The number of children who reported any of the 13 listed comorbidities (n=1635)	1400	88.0

ADHD indicates Attention Deficit/Hyperactivity Disorder. Subgroup n describes the number of participants who answered and did not skip the question from the total 2406.

Source: 2014 National Survey of the Diagnosis and Treatment of ADHD and Tourette Syndrome

Table 3.

Unmet Need and ADHD Financial Impact (total n=2406)

Variables of Interest (Subgroup n)	Freq, Unweighted	% of n Responses, Weighted
In the past 12 months, did your child need an ADHD treatment that s/he could not get? (n=2391)
Yes	218	11.6
Of above, the child needed MEDICATION for treatment but could not get it	84	53.2
Of above, the child needed SCHOOL-BASED BEHAVIORAL THERAPY but could not get it	115	50.8
Of above, the child needed OTHER TREATMENT but could not get it	76	38.2
Of above, the child needed OUTSIDE SCHOOL THERAPY but could not get it	81	36.1
Family Impact (n=2390)
Responded YES to any of the family impact questions below	971	44.3
Child's ADHD has caused financial problems for family (n=2382)	454	22.4
You or other family members cut down on hours you work because of child's ADHD (excluding family members who have stopped working) (n=2387)	459	21.3
You or other family members avoided changing jobs because of concerns about maintaining health insurance for child (n=2388)	390	15.8
You or other family members have stopped working because of child's ADHD (n=2387)	267	14.4

ADHD indicates Attention Deficit/Hyperactivity Disorder. Subgroup n describes the number of participants who answered and did not skip the question from the total 2406.

Source: 2014 National Survey of the Diagnosis and Treatment of ADHD and Tourette Syndrome

Table 4.

Associations of Measures of Socioeconomic Status with Unmet Need and ADHD Family Impact among US Children age 8-17 with a Current Diagnosis of ADHD (n=2406)

Any Unmet Need (n=2391)				Any Family-Reported Impact (n=2390)
		Yes (n=218)			Yes (n=971)
% of Yes		11.6%			44.3%
Predisposing measures of socioeconomic status
Sex (n=2391)				(n=2390)
Female		8.8%	[5.8, 13.1%]		39.6%	[33.5, 46.0%]
Male		12.8%	[10.0, 16.2%]		46.3%	[42.0, 50.6%]
P-Value	.12			.09
Race and Ethnicity (n=2383)				(n=2377)
White, Non-Hispanic		9.0%	[6.8, 11.8%]		44.2%	[40.0, 48.5%]
Black, Non-Hispanic		16.4%	[9.7, 26.3%]		45.5%	[35.4, 56.1%]
Hispanic		16.5%	[9.9, 26.3%]		43.2%	[33.0, 54.0%]
Other		11.8%	[5.5, 23.8%]		46.3%	[34.7, 58.4%]
P-Value	.07			.98
Primary Language in Household (n=2388)				(n=2387)
English		12.3%	[9.9, 15.1%]		44.6%	[40.9, 48.4%]
Other Language		4.4%	[2.1, 8.9%]		40.1%	[28.8, 52.6%]
P-Value	.004			.49
US Census Region (n=2391)				(n=2390)
South		10.6%	[7.5, 14.7%]		46.9%	[41.3, 52.6%]
West		10.8%	[6.9, 16.4%]		44.2%	[37.2, 51.4%]
Midwest		9.7%	[5.9, 15.5%]		40.4%	[32.6, 48.7%]
Northeast		17.2%	[10.8, 26.2%]		42.4%	[33.9, 51.3%]
P-Value	.23			.60
Highest Level Education Attained by Parent or Guardian (n=2248)				(n=2246)
Less than High School		10.8%	[5.8, 19.4%]		43.5%	[33.8, 53.6%]
High School Graduate		13.4%	[9.4, 18.8%]		43.2%	[37.2, 49.4%]
More than High School		10.7%	[7.8, 14.4%]		43.8%	[38.7, 49.0%]
P-Value	.60			.99
Age at First NS-DATA Interview (n=2391)				(n=2390)
8–9yo		14.5%	[8.2, 24.4%]		50.3%	[39.3, 61.3%]
10–13yo		11.2%	[8.1, 15.3%]		44.0%	[38.6, 49.6%]
14–17yo		11.3%	[8.2, 15.4%]		43.1%	[38.0, 48.3%]
P-Value	.71			.50
Age at First ADHD Diagnosis (n=2358)				(n=2357)
1–4yo		13.6%	[8.1, 21.8%]		58.1%	[48.2, 67.5%]
5–9yo		11.6%	[9.0, 14.9%]		41.9%	[37.7, 46.2%]
10yo+		9.9%	[5.5, 17.3%]		43.0%	[34.6, 51.8%]
P-Value	.72			.008
Enabling measures of socioeconomic status
Insurance, by Type and Continuity over Last Year (n=2353)				(n=2365)
Continuous Private		10.1%	[7.4, 13.6%]		39.7%	[35.3, 44.3%]
Continuous Public		12.1%	[8.6, 16.7%]		47.7%	[41.8, 53.6%]
Noncontinuous Public		19.2%	[8.1, 38.8%]		57.9%	[38.6, 75.1%]
Noncontinuous Private		17.5%	[4.3, 50.2%]		53.5%	[27.1, 78.1%]
Currently Uninsured		13.2%	[5.1, 30.0%]		30.4%	[13.6, 54.9%]
P-Value	.54			.10
Household Poverty Level (n=2391)				(n=2390)
0–99% FPL		15.9%	[10.0, 24.2%]		47.6%	[38.6, 56.8%]
100–199%		7.8%	[3.9, 15.0%]		40.7%	[32.6, 49.4%]
200–399%		13.1%	[9.2, 18.2%]		45.1%	[38.9, 51.5%]
400% or greater		10.5%	[7.2, 15.0%]		43.9%	[38.0, 50.0%]
P-Value	.26			.73
ADHD Severity Measures
Reported ADHD Severity (n=2373)				(n=2373)
Mild		7.6%	[4.6, 12.3%]		31.7%	[26.3, 37.8%]
Moderate		11.6%	[8.6, 15.5%]		47.6%	[42.5, 52.7%]
Severe		18.3%	[12.5, 25.9%]		54.8%	[46.0, 63.3%]
P-Value	.01			<.001
Any Current Comorbidities (n=1624)				(n=1626)
No		5.2%	[1.9, 13.6%]		33.4%	[23.7, 44.8%]
Yes		12.6%	[9.8, 16.2%]		52.9%	[48.2, 57.5%]
P-Value	.07			.002
Any Family Reported Impact (n=2376)				Any Unmet Need (n=2376)
No		7.2%	[5.0, 10.3%]		41.5%	[37.8, 45.4%]
Yes		16.9%	[13.0, 21.6%]		65.1%	[53.7, 75.0%]
P-Value	<.001			<.001

Bold denotes the value is significant at p-value <0.05, computed by Chi-square association tests

ADHD indicates Attention Deficit/Hyperactivity Disorder; NS-DATA, 2014 National Survey of the Diagnosis and Treatment of Attention-Deficit/Hyperactivity Disorder and Tourette Syndrome; FPL, federal poverty level.

Subgroup n describes the number of participants who answered and did not skip the survey question from the total 2406.

Source: 2014 National Survey of the Diagnosis and Treatment of ADHD and Tourette Syndrome

Table 5.

Adjusted OR (95% CI) for Unmet Need

	Model (n=1454)
Constant (Intercept)	0.06 (0.003, 1.18)
ADHD Severity Measures
Reported Severity
Mild	Referent
Moderate	1.56 (0.64, 3.74)
Severe	3.02 (1.19, 7.65)
Any Current Comorbidities
No	Referent
Yes	2.63 (0.89, 7.74)
Predisposing measures of socioeconomic status
Sex
Male	Referent
Female	0.57 (0.29, 1.14)
Race and Ethnicity
White, Non-Hispanic	Referent
Black, Non-Hispanic	1.43 (0.49, 4.16)
Hispanic	2.09 (0.95, 4.58)
Other	1.19 (0.44, 3.22)
Primary Language in Household
English	Referent
Other Language	0.28 (0.09, 0.88)
US Census Regions
Northeast	Referent
South	0.73 (0.30, 1.79)
Midwest	0.62 (0.21, 1.84)
West	0.52 (0.19, 1.46)
Highest Level Education Attainment, by Either Parent or Guardian
Less than High School	Referent
High School Graduate	0.84 (0.29, 2.45)
More than High School	0.82 (0.28, 2.44)
Age at First NS-DATA Interview
8–9yo	Referent
10–13yo	1.05 (0.33, 3.23)
14–17yo	1.50 (0.43, 5.23)
Age at First ADHD Diagnosis, Continuous Variable in Years
1–4yo	Referent
5–9yo	1.14 (0.46, 2.81)
10yo+	0.64 (0.22, 1.85)
Enabling measures of socioeconomic status
Current Insurance and Continuous Insurance over the past year
Noncontinuous Public	Referent
Noncontinuous Private	1.35 (0.17, 10.86)
Continuous Public	0.34 (0.12, 0.998)
Continuous Private	0.50 (0.17, 1.46)
Currently Uninsured	0.81 (0.13, 5.05)
Household Poverty Level
0–99% FPL	Referent
100–199%	0.68 (0.22, 2.07)
200–399%	0.93 (0.35, 2.47)
400% or greater	0.67 (0.24, 1.86)
Any Family Impact
No	Referent
Yes	*2.81 (1.41, 5.60)

Bold denotes the value is significant at P<0.05.

^*P<0.01

^**P<0.001

Logistic regression model to examine the adjusted odds of association between the measures of socioeconomic status covariates of interest, family financial impact, and unmet need for ADHD treatment. Model began with all covariates present to examine the odds of unmet need, as in our Conceptual Framework.

ADHD indicates Attention Deficit/Hyperactivity Disorder; NS-DATA, 2014 National Survey of the Diagnosis and Treatment of Attention-Deficit/Hyperactivity Disorder and Tourette Syndrome.

Subgroup n describes the number of participants who answered and did not skip the survey questions from the total 2406.

Source: 2014 National Survey of the Diagnosis and Treatment of ADHD and Tourette Syndrome

RESULTS

Sample Characteristics

Table 1 displays the characteristics of the analytic sample, representing non-institutionalized US children age 8–17 ever-diagnosed with ADHD and with a current diagnosis.³ Most of the children had English as their primary household language (91.3%), were male (70.1%), identified as non-Hispanic white (63.4%), and had a parent with more than high school education (52.3%). Only 15 of the 190 who had a primary household language other than English also identified as Hispanic (1.2%). A majority of the children had at least one comorbidity (88.0%) and were first diagnosed with ADHD between five to nine years old (70.5%) (Tables 1 and 2). A plurality had moderate ADHD severity (49.4%), had continuous private insurance in the past year (47.9%), lived in the South (39.2%), and had a household income at 400% or greater of the federal poverty line (36.2%).

Unmet Need and ADHD Financial Impact

Table 3 shows the prevalence of any unmet need for ADHD treatment, as well as any ADHD financial impact among US children with ADHD. In the year prior to NS-DATA, 11.6% of children had unmet need. The most common unmet need was medication, which 53.2% of those with unmet need reported; 50.8% had unmet need for school-based behavioral therapy, 38.2% had unmet need for other types of treatment, and 36.1% had unmet need for therapy outside of school. Provider issues, cost, and availability tended to be the most common reasons for unmet ADHD treatment need (Appendix A).

Among US children with ADHD, approximately 44.3% experienced one or more ADHD financial impacts. The most common impact experienced was 22.4% reporting ADHD had caused financial problems for the family. Additionally, for 21.3%, family members cut down work hours; for 15.8%, family members avoided changing jobs to maintain insurance for the child with ADHD; and for 14.4%, family members stopped work due to the child’s ADHD.

Measures of Socioeconomic Status Associated with Unmet Need and ADHD Financial Impact

Results from Chi-square tests of association between the sociodemographic determinants of health with unmet need and ADHD financial impact are presented in Table 4.

Associations with unmet need.

Among our predisposing measures of socioeconomic status, bivariate results showed that among children with ADHD, a significantly higher percentage of those with English primary household language experienced one or more unmet needs compared to those with other primary household languages (p=0.004). Additionally, increased parent-reported ADHD severity was significantly associated with any unmet need (p=0.01). Having any co-morbidities was positively correlated with unmet need, though this association was not statistically significant (p= 0.07). Associations of all other measures with unmet need were non-significant (Table 4).

Associations with ADHD financial impact.

Among our predisposing measures of socioeconomic status, bivariate testing showed that among children with ADHD, a significantly higher percentage of those with a younger age at first ADHD diagnosis were more likely to have ADHD financial impact than those diagnosed at older ages (p=0.008). Further, higher parent-reported ADHD severity (p<0.001) and having any reported co-morbidities (p=0.002) were significantly associated with ADHD financial impact. Associations of all other measures with ADHD financial impact were non-significant (Table 4).

Relationship between unmet need and ADHD financial impact.

On bivariate testing, there was a statistically significant positive association between family financial impact and unmet need.

Logistic regression to examine the adjusted odds of unmet need.

Table 5 summarizes logistic regression model results for unmet need. The association of primary language in the household with any unmet need remained statistically significant. Specifically, compared to children with ADHD whose primary household language was English, the adjusted odds of reporting any unmet need were approximately 72% less for children with ADHD whose primary household language was not English, adjusting for all other measures of socioeconomic status. Adjusting for all other measures of socioeconomic status, odds of unmet need were more than twice as great in families who reported ADHD financial impact of ADHD compared to families who did not (p<0.001).

DISCUSSION

Through results from a national study of US children with ADHD, we examined associations between enabling and predisposing measures of socioeconomic status, ADHD severity and comorbidities, and ADHD financial impact on our dependent variable of interest - unmet need for ADHD treatment.

Among US children with unmet need for ADHD treatment, primary language stood out as a strongly statistically significant association, even after adjusting for ADHD severity, race, and ethnicity. The adjusted odds of reporting any unmet need were significantly increased in English primary-language households compared to households where another primary language was spoken, countering our hypothesis.

A likely explanation of the findings regarding language differences in unmet need may relate to cultural differences in how parents perceive service needs. Aligned with research on service under-utilization,^8,9 children from households with non-English primary language may have less unmet need because their families do not perceive as high a level of service need as English primary-language counterparts. If parents do not identify treatment needs, they will not identify those needs as being unmet. An important element of any chronic condition is the recognition and communication of needs. A growing body of research encourages consideration of cultural diversity when managing ADHD and other developmental conditions,¹⁹ examining families’ perceptions of the myriad therapies that a child would need.¹⁰ A lack of available resources and access for non-English speaking families, as when therapies are not offered at all in an area, further feeds a lack of need recognition. As Andersen and Aday discuss, disparities in patient access to treatment may be grounded in the conflict between how patients and professionals define need.^13,14 This conflict is reflected in the tendency of provider issues to contribute to unmet need (Appendix A). Throughout the child’s lifetime, the provider profoundly impacts the continued recognition of therapy modality combinations that best support function at home and at school. Thus, unmet treatment need may be less apparent to parents due to lack of familiarity with ADHD or with services that could be available to their child.⁸ In this study, a majority of those who reported unmet need cited an inability to obtain school-based behavioral therapy. This echoes aforementioned concerns about the multiple settings in which children with ADHD identify their needs. Since our findings suggest that non-English speaking households may be more likely to state that the child’s needs are met even when using fewer services than other families with similar symptoms,¹⁹ a detailed assessment of service need and receipt in the clinical setting may be necessary, including understanding specific services used in the school and home setting, medication use, perceived progress, and barriers to progress. This echoes previous data that patient-centered medical care can decrease risk of financial problems.¹⁶

Another interpretation is that there was less unmet need for ADHD treatment among non-English primary language households because children in non-English language households have milder ADHD, and thus need fewer services. Some literature shows decreased ADHD diagnosis rates among Spanish-speaking patients compared to English-speaking counterparts,¹⁰ supporting this hypothesis. Despite literature on the “immigrant paradox,” that children from immigrant families have better than expected health, particularly pertaining to mental health,²⁰ differences in illness severity do not account for differences in treatment need. Other research in minority health shows that ADHD symptomatology is similar or higher among immigrant and minority groups compared to white children.^8,21 A milder ADHD phenotype may not completely explain language differences in unmet need for ADHD treatment. It is also possible that non-English households are better able to access ADHD services. However, overall service access for mental health among immigrants is poor,^8,22 particularly for child mental health, and we controlled for parent-reported ADHD severity in our models, both of which make this hypothesis less likely.

Younger age of ADHD diagnosis was significantly associated with ADHD financial impact. We may interpret this to mean that the impact of ADHD is additive; costs associated with the condition compound over years.^4,12 Similarly, a younger age at ADHD diagnosis reflects a longer period of time during which the family shares the condition’s burden and is exposed to adverse outcomes.¹²

Finally, this study demonstrated a strong association between ADHD financial impact and unmet need for ADHD treatment. Statistical analysis saw over twice as great the adjusted odds of unmet need among those reporting any ADHD financial impact. Due to the cross-sectional design, we cannot say that ADHD financial impact predicts unmet need. Families with ADHD financial impact may have more unmet need because they are less able to take advantage of ADHD services. Families who do not use services may have children with ADHD symptoms that are more difficult to manage, thus increasing financial adversity. Findings suggest that better elucidating how a family’s overall financial and employment circumstances around ADHD might ameliorate long-term management of the condition. Clinic-based efforts such as social determinants of health screenings²³ or medical-legal partnerships²⁴ may particularly help reduce unmet need for ADHD treatment by improving ADHD financial impact. Likewise, study findings suggest that improved coverage of ADHD treatment, through such actions as mental health parity legislation²⁵ or Medicaid health and community-based service waivers,²⁶ may improve the overall financial stability of families of children with ADHD.

There were several limitations in the study design. We were limited by the cross-sectional design of NS-DATA; we cannot make any kind of causal inference in terms of temporal precedence. The survey did not ask specifically how high families perceived their need for services (ie, if they do or do not feel that as many treatment services are necessary) in contrast to which needs were unmet; we thus could not distinguish whether differences in our dependent variable were due to disparate baseline levels of need. Further, the measures of socioeconomic status in our analytic sample do not match the US population of children with a current diagnosis of ADHD. A majority of our analytic sample were male, non-Hispanic and white, speaking primarily English in the household with continuous private insurance. Another limitation was that health determinants from physical and environmental contexts were not included in the survey and would be an area for future study. Though families could take the survey in English or Spanish, data on which households received which script were suppressed by MCHB to protect participant confidentiality. Survey items lacked detail of other languages spoken in the household or other races and ethnicities with which the children identified. This factor became an increasing challenge during analyses when a significant difference in unmet need was found between English- and non-English primary-language households. Due to the low 11.6% prevalence of our dependent variable in the analytic sample, we could not calculate a continuous variable of number of unmet needs nor analyze reasons for unmet need without dramatically losing power.

ADHD prevalence is 11% among US school-aged children, with repercussions in education, healthcare, and the adult workforce. Disparities by race, ethnicity, and household language are already known in ADHD diagnosis and service use. This study suggests that considering the financial impact of ADHD may be important in assessing unmet need for ADHD treatment. This may be particularly important as children with ADHD age, as earlier diagnosis was associated with more family financial adversity. Likewise, consideration of factors such as household language may be important in ADHD needs assessments. These study findings could shape future clinic policies for targeting resources toward communities, such as immigrant and multi-generational homes, prone to barriers of care. Further research may explore other interacting factors for unmet need besides financial impact, how reasons for unmet need correlate with non-English-speaking households, and disparities in treatment need among adults with ADHD.

What’s New.

Among US children with ADHD, those with adverse family financial impact of ADHD have greater odds of unmet need for ADHD treatment. Individual need requires more understanding of associations between financial impact with age at first ADHD diagnosis and unmet need with household language.

APPENDIX A

Appendix A.

Reasons for Unmet Need for ADHD Treatment (total n=2406)

	Freq, Unweighted	% of n Responses, Weighted
In the past 12 months, did your child need an ADHD treatment that s/he could not get? (n=2391)
Yes	218	11.6
Of above, the child needed MEDICATION for treatment but could not get it	84	53.2
Due to cost: issues related to cost or insurance	32	42.5
Due to provider issues: doctor/school refused to provide, child does not have provider	17	31.0
Due to availability: treatment/service was not available in child's area/school	11	10.3
Due to delays: waiting lists, backlogs, drug shortages	16	9.1
Due to information: Parent/doctor/school did not know or have information about treatment	3	5.4
Due to family issues: child or other family member did not want treatment	4	5.3
Due to eligibility: child was not eligible for treatment/service	1	4.4
Other reasons	7	3.9
Of above, the child needed SCHOOL-BASED BEHAVIORAL THERAPY but could not get it	115	50.8
*Due to provider issues: doctor/school refused to provide, child does not have provider	33	32.3
*Due to availability: treatment/service was not available in child's area/school	30	29.2
*Due to eligibility: child was not eligible for treatment/service	23	25.7
*Due to delays: waiting lists, backlogs, drug shortages	12	10.7
*Due to cost: issues related to cost or insurance	18	8.8
*Other reasons	8	4.6
*Due to information: Parent/doctor/school did not know or have information about treatment	5	3.1
*Due to family issues: child or other family member did not want treatment	0	0.0
*Due to child not yet in school	0	0.0
Of above, the child needed OTHER TREATMENT but could not get it	76	38.2
Of above, the child needed OUTSIDE SCHOOL THERAPY but could not get it	81	36.1
Due to cost: issues related to cost or insurance	33	39.7
Due to availability: treatment/service was not available in child's area/school	29	31.4
Due to provider issues: doctor/school refused to provide, child does not have provider	11	20.6
Due to delays: waiting lists, backlogs, drug shortages	11	18.0
Due to family issues: child or other family member did not want treatment	6	10.4
Due to information: Parent/doctor/school did not know or have information about treatment	6	5.9
Due to eligibility: child was not eligible for treatment/service	3	3.9
Other	5	1.5

^*Calculated from n=112 due to legitimate "Don't know" responses

ADHD indicates Attention Deficit/Hyperactivity Disorder

Subgroup n describes the number of participants who answered and did not skip the survey question from the total 2406.

Source: 2014 National Survey of the Diagnosis and Treatment of ADHD and Tourette Syndrome