Does Connection to Primary Care Matter for Children With Attention-Deficit/Hyperactivity Disorder?

METHODS

We conducted a cross-sectional analysis of the National Survey of Children's Health, 2003 (NSCH), which is a national survey conducted by National Center for Health Statistics of the Centers for Disease Control and Prevention.^25,26 Survey domains included (1) child and family demographics, (2) children's physical and mental health status, (3) health insurance status and type of coverage, (4) access and use of health care services, and (5) characteristics of the medical home. The NSCH is a telephone survey conducted in English and Spanish between 2003 and 2004 on a random sample in all 50 states and the District of Columbia. Participants for the NSCH were drawn from a subset of telephone numbers selected for the National Immunization Survey. Households were contacted by randomly selecting telephone numbers from this subset. For households with children, 1 child was randomly selected from each household to be the subject of the interview. Postsampling adjustments have been made to account for participation rates among minority groups. The parent or guardian of 102 353 children (69% of those contacted) between the ages of 0 and 17 years completed this 25-minute telephone interview. Additional information about the NSCH is available through the Data Resource Center for Child and Adolescent Health (www.nschdata.org).

We included children who were 6 to 17 years of age and whose parents reported that they had ever received a diagnosis from a doctor or health professional of ADHD or ADD. For our study, to qualify as having asthma, a child (6–17 years of age) must have received a diagnosis of asthma and have had symptoms within the previous 12 months. Children who had both asthma and ADHD were included in the descriptive data but were excluded from all other analyses.

RESULTS

Children in our sample with ADHD (n = 6030) differed from those with asthma (n = 6133) in several important ways (Table 1). Those with ADHD were more likely to be male (72.0% vs 56.8%), were older, and were more likely to be white (71.7% vs 58.6%). Insurance status, household income, region, and highest level of household education were comparable. Fewer children with ADHD and asthma were from non–English-speaking homes compared with all children. The percentage with the presence of comorbidities (63.5% ADHD, 30.6% asthma, and 9.9% all children) and the severity of illness were greater in children with ADHD and asthma in comparison with all children. A total of 841 children had both ADHD and asthma and were included only in the descriptive analyses.

The proportion of children with a medical home was comparable among the 3 groups: 41.1% of children with ADHD, 45.5% of children with asthma, and 41.4% of all children. However, children with ADHD and asthma differed on individual criteria within the definition (Table 2). Children with ADHD were slightly less likely to report having a personal provider or having a provider who communicates well, although similar proportions reported having consistent care. Fewer children with ADHD reported not having problems receiving specialty care (76.5% of children with ADHD, 81.7% of children with asthma; P = .01) and having provider follow-up after being referred for specialty care (51.7% of children with ADHD, 57.5% of children with asthma; P = .02).

The percentage of children with at least 1 unmet need was higher for those with ADHD (16.8%) than with asthma (6.7%; P < .001; Table 3). As would be expected, children with ADHD had more unmet needs for mental health care or counseling than children with asthma (13.2% vs 2.7%; P < .001). Children with ADHD and asthma had similar proportions of unmet needs for medical services (1.2% vs 1.8%) and prescription drugs (1.0% vs 1.2%).

The following covariates were associated with unmet need, met the P-value limit for model inclusion, and therefore were included in the multivariate analysis: insurance status, gender, race, household income, region, household education, language spoken at home, presence of comorbidities, and severity of illness. Age was not associated with unmet need; however, it was included in the multivariate analysis. The relationship of covariates and unmet need was in the expected direction.

Children with ADHD were more likely to have a self-reported unmet need than were children with asthma (unadjusted odds ratio [OR]: 3.5 [95% confidence interval (CI): 3.04–4.03]; P < .001). Our fundamental question was whether having a medical home had the same impact for children with ADHD and asthma. In the unadjusted model, for children with ADHD, having a medical home was associated with a somewhat lower probability of unmet needs (OR: 0.733 [95% CI: 0.567–0.948]; P = .02), whereas for children with asthma, unmet needs for those who have a medical home were dramatically lower (OR: 0.338 [95% CI: 0.203–0.562]; P < .001). Because these effects differed (P = .008), the interaction term was retained for subsequent analyses.

In the multivariate analysis (Table 4), children who had asthma and had a medical home had less than half the likelihood of reporting an unmet need in comparison with those without a medical home (OR: 0.434 [95% CI: 0.259–0.727]; P = .002); however, among children with ADHD, having a medical home was not associated with decreased likelihood of reporting an unmet need (OR: 0.942 [95% CI: 0.706–1.26]; P = .68). These effects were significantly different (P = .01). This multivariate logistic regression model controlled for the following variables: gender, age, insurance status, region, race, highest level of household education, primary language spoken at home, household income, presence of comorbidities, and severity of illness. Parameters for insurance status, age, region, and presence of comorbidities were significantly different from unity. The results from stratified analyses for children with ADHD and children with asthma were consistent with the model that used the interaction term.

DISCUSSION

In this analysis of a nationally representative sample, 16.8% of parents of children with ADHD reported unmet needs. These data suggest that substantially more children with ADHD have unmet needs than children with asthma, a similarly prevalent physical condition also frequently treated in primary care settings. Although the proportions of children who had ADHD and asthma and had a medical home were similar, the impact of having a medical home seems to be dramatically different. Specifically, for children with asthma, having a medial home is associated with approximately half the likelihood of having unmet needs. In contrast, for children with ADHD, no benefit was associated with having a medial home on unmet needs after adjustment for confounding factors.

A number of previous studies showed that CSHCN have more unmet needs than healthy children.²⁹ In regard to mental health care needs, a recent study demonstrated that 25.7% of CSHCN reported a need for mental health care services.¹⁶ The concept of the medical home was developed with the intention to address the needs of CSHCN. Unmet needs for mental health have been associated with lack of a usual source of medical care.¹⁶ We hypothesized that having a medical home, a marker for high-quality primary care, would decrease the unmet needs of children with ADHD. Having a personal provider who communicates well and provides consistent care might facilitate identification of mental health needs and aid in meeting them, especially given the extensive follow-up necessary for making the diagnosis and titrating medication. Being able to receive specialty care and having a provider who follows up on that care might contribute to meeting mental health needs. We would have expected that these components of the medical home would matter for children with ADHD, but they do not. It seems that within the group of CSHCN, differences exist according to disease in the impact of having a medical home on unmet needs.

There are several potential barriers to obtaining all needed care for children with ADHD, and these may not be influenced by having a medical home. For instance, many communities do not have access to enough mental health providers for consultation with the PCP and for referrals and counseling.^7,8 Thus, even if the PCP provides excellent care within a medical home, parents may still experience barriers to obtaining needed services. In addition, studies have shown that although PCPs could be diagnosing and treating uncomplicated ADHD, many are not.^3,10 A variety of barriers to use of the ADHD guidelines have been noted.¹¹ The structure of primary care with relatively short visits and difficulty reimbursing for treatment of mental health conditions may limit the ability of any provider, including those who work in medical homes, to meet all of a family's needs. Co-localization of mental health professionals within the primary care setting might be an approach to address these unmet needs. If we are to expect PCPs to treat ADHD, then perhaps providers require both additional training and changes in reimbursement for mental health services.

Decision support tools might also enhance PCP's abilities to diagnose and treat ADHD. Additional studies that demonstrate that changes such as these actually diminish the unmet needs of children with ADHD are needed.

Another potential explanation for these findings is that the definition of the medical home may not capture attributes of primary care that would satisfy needs of children with mental health conditions. The NSCH based its questions regarding the medical home on the medical home domains outlined by the AAP. Perhaps to address mental health needs, the concept of the medical home needs to be expanded, for instance, to include having a written management plan and/or a comanagement plan with specialists such as mental health providers. It also is possible that the means of measuring the medical home are not adequate to capture the characteristics of the medical home that are particularly important to providing care for children with mental health needs.

Consistent with other studies,^16,30–32 the vast majority of parents reported that their children had a usual source of care. The items that compose the medical home variable can be viewed as measures of satisfaction with primary care and, plausibly, as markers for high-quality primary care. It would not be surprising, then, if parents who respond positively on those items have children with lower unmet needs. This was exactly what our data show for children with asthma. What is striking is that among parents of children with ADHD, even those who respond positively to these items (and are likely reasonably pleased with their primary care) do not report fewer unmet needs. This seeming paradox may relate to parental expectations of their PCPs and the health care system generally. Parents may have the same expectations of their children's physician to care for their behavioral and mental health needs as for physical health conditions. If this is so, then parents may not be taking into consideration the care that their children receive for their behavioral and mental health needs when answering the questions that pertain to the medical home. Additional research is needed to explore parental expectations regarding PCP involvement with behavioral and mental health conditions.

Response to some unmet needs may not be under a PCP's control in the current medical care system; for instance, not being in counseling could be attributable to a shortage of mental health providers rather than resulting from not having a medical home. Moreover, reported unmet mental health care needs may be related not to ADHD but to other psychiatric comorbidities for which the PCP might have a smaller role.

As with any cross-sectional analysis, we cannot infer causality from associations detected. Although the NSCH is among the largest nationally representative surveys of children in the United States, it excludes households that do not have a telephone. There is potential for respondent bias because families who participate might be a significantly different group from families who do not. These concerns were taken into consideration when the sampling weights were designed. All diagnoses and unmet needs are based on self-reporting by the parents, although the NSCH asks parents specifically whether a health care provider diagnosed the child's either ADHD or asthma. In addition, for asthma, we included only children who had had symptoms of asthma within the past year.

In contrast to children with asthma, the NSCH survey did not ascertain whether symptoms were present in the previous 12 months for children with ADHD, so we were unable to restrict the definition of ADHD. Although this could create a misclassification bias, this bias would tend to underestimate the unmet needs of children with ADHD by including some children without current symptoms. All children were asked whether they have an emotional, developmental, or behavioral problem that requires treatment or counseling and, when they did, were asked about their mental health care or counseling needs. We assumed that those who did not report a mental health problem could not have an unmet mental health need. It is therefore possible that unmet mental health needs are disproportionately underreported, in particular for patients with asthma.

CONCLUSIONS

This study found that high-quality primary care in the form of a medical home provided great benefit for children with asthma but little benefit for children with ADHD. We noted that these findings may extend to children with other mental health conditions. This study gives us reason to believe that our current primary care system may not be as successful in meeting the needs of children with behavioral health conditions compared with children with physical conditions. Many have argued that PCPs should care for mental health conditions, such as ADHD. If they are to do so, then this study supports calls for rethinking how screening, diagnosis, and treatment for mental health conditions are integrated into primary care.