Psychotropic Medication Prescribing for Youth at a Regional Autism Center

Introduction

Autism spectrum disorder (ASD) is a neurodevelopmental disorder characterized by persistent deficits in social communication, interaction, and restricted, repetitive interests and behaviors (APA 2013). ASD is identified in about 1 in 54 children, and it is over four times more common among males than females (Maenner et al. 2020). There are an estimated 5.4 million adults (2.2%) living with ASD in the United States, and within the United States, and there are geographic differences in reported prevalence rates (Dietz et al. 2020; Maenner et al. 2020).

While ASD is reported to occur in all ethnic, racial, and socioeconomic groups (Maenner et al. 2020), disparities in diagnosis have been reported (Mandell et al. 2002; Zuckerman et al. 2017). In an analysis of 2016 data from the Autism and Developmental Disabilities Monitoring (ADDM) Network, the Centers for Disease Control and Prevention (CDC) reported approximately the same ASD prevalence for Black, Asian/Pacific Islander, and White 8-year-old children, and a lower prevalence for Hispanic children (Maenner et al. 2020). Forty-five percent of White, 43% of Hispanic, and 40% of Black children with ASD were evaluated by 36 months of age (Maenner et al. 2020). Differences in identification matter because identifying ASD sooner leads to earlier connections to health care services that can improve outcomes (Maenner et al. 2020).

There appears to be a high variability in prescribing practices for youth with ASD (Jackel et al. 2017). There are no FDA-approved medications for the treatment of core ASD symptoms. Nevertheless, medications are commonly prescribed for irritability and/or comorbid psychiatric conditions such as attention-deficit/hyperactivity disorder (ADHD), depression, and anxiety. Not surprisingly, prevalence estimates of these comorbid psychiatric conditions are higher in clinic sample-based studies than in population-based and registry-based studies (Lai et al. 2019). Risperidone and aripiprazole, two second-generation antipsychotics (SGAs), are FDA approved for the treatment of irritability in youth with ASD (Jackel et al. 2017), although they come with significant side effect burden, notably metabolic effects.

Several recent studies have examined the prescribing patterns as well as factors associated with psychotropic medication use. We identified four relevant studies that examined psychiatric medication prescribing among youth with ASD. They vary in their consideration of patient factors, including comorbid psychiatric symptoms, age, sex, ethnicity, race, and insurance.

A retrospective chart review of youth with ASD (n = 54) at specialized outpatient psychiatric program at a major university center (Shekunov et al. 2017) found a high rate of impairing comorbid psychiatric disorders (mean number of psychiatric diagnosis 2.3). Eighty percent received polypharmacy, with patients receiving an average of three psychotropic medications. The most prevalent comorbidities were ADHD (83%), anxiety disorders (67%), bipolar spectrum (43%) and mood disorder not otherwise specified (44%). The most common classes of medications prescribed were SGAs (74%), selective serotonin reuptake inhibitors (SSRIs) (26%), stimulants (24%), and nonstimulant ADHD medications (22%). The three most prescribed psychotropic medications following SGAs were citalopram (19%), methylphenidate (19%), and atomoxetine (15%). In this study, the relationships between prescribing rates and age, sex, ethnicity, race, and type of insurance were not examined.

Jackel et al. (2017) examined the electronic health record (EHR) data from three DBPNet sites (https://dbpnet.research.chop.edu/) to identify factors associated with psychotropic medication prescribing for children with ASD (n = 5095) who were seen at three academic developmental-behavioral pediatrics clinical programs from January 2010 to December 2012. They identified older age and the presence of comorbidities as predictors of prescribing of psychotropic medications. They also found that 37% of the patients with ASD were prescribed a psychotropic medication over the 2-year period, and 14% received more than one class of psychotropic medication at some point. The most prescribed classes were stimulants (16%), SSRIs (13.5%), SGAs (10.9%), and alpha-2 adrenergic agonists (9.4%). No sex differences in prescribing were found. For children ages 3–5, those on Medicaid were more likely to be prescribed psychotropic medications (particularly alpha-2-adrenergic agonists and SGAs) than those with private insurance. This study did not examine the relationship between prescribing rates and ethnicity or race.

Lake et al. (2017) examined SGA use in youth with ASD (n = 5150) from the Autism Treatment Network in two age groups (2–11, and 12–17 years). They also found no sex differences. In youth, ages 2–11 (n = 4749), SGA use was associated with older age, non-White Hispanic race/ethnicity, oppositional behaviors, prior ASD diagnosis, and use of multiple psychotropic medications. Insurance was not included in the association analysis of this group. In youth, ages 12–17 (n = 401), Non-White Hispanic and White non-Hispanic were the only two categories. Ethnicity and race were not analyzed separately due to small sample size. In this group, older children with any public insurance were more likely to be prescribed SGAs in the bivariate analyses, however, only older age was associated with SGA use in the multivariable regression analysis.

Kamimura-Nishimura et al. (2017) examined data from the National (Hospital) Ambulatory Medical Care Surveys (1994–2009) for youth ages 2–18 years (n = 158,488) and found a positive correlation between older age and psychotropic medication use in ASD visits. Teenagers (12–18 years) and school age (6–11 years) children were more likely to be prescribed medication compared with preschoolers (2–5 years). Older individuals with ASD were more likely to have a comorbid behavioral diagnosis and be prescribed psychotropic medication. There were no sex differences in psychotropic prescribing. No overall differences were found between the rates of prescribing for White and non-White patients. These were the only two categories examined; there were no further details on prescribing for specific ethnic groups or racial minorities. Psychotropic prescriptions were not differentially associated with insurance status (Private vs. Medicaid/State Children's Health Insurance Program, Self-Pay, No Charge/Charity or Other).

While comorbid psychiatric symptoms and age appear to be two major factors associated with psychotropic medication use across studies, other factors, such as sex, ethnicity, race, and insurance were not consistently included in prior studies.

The Seattle Children's Autism Center (SCAC) serves youth throughout Washington state (WA). In this study, we described the demographics of patients seen at SCAC to test a hypothesis that ethnic and racial minorities may be under-represented in the SCAC Psychiatric Medication Evaluation and Management Clinic, based on prior studies of disparities in ASD diagnosis. In addition, we examined the prescriptions provided to the patients to determine whether specific types of psychotropic prescriptions were associated with patient factors, including age, sex, ethnicity, race, insurance, visit number, and diagnoses.

Methods

Results

Table 1 includes the demographic data for our sample, including age, sex, self-identified ethnicity and race, insurance, as well as the percent of commonly observed comorbid diagnoses, including ADHD, anxiety, disruptive behavior, and depressive disorders.

This study identified high prevalence rates of ADHD (53.6%), anxiety disorders (52.8%), disruptive behavior disorders (41.4%), and depressive disorders (12.7%). Among the four classes of medications examined in this study, ADHD medications were most frequently prescribed (55.1%), followed by antidepressants (43.8%), antipsychotics (26.9%), and mood stabilizers (8.5%).

Compared with the WA Census, the SCAC saw a higher percentage of those who identified as Black (5.8% vs. 4.4%) or two or more races (6.4% vs. 4.9%), after dropping patients for whom we did not have data for ethnicity and race (missing N = 87 for ethnicity and N = 188 for race) (Table 2).

The likelihood of receiving a medication prescription by demographic, insurance characteristics, number of visits, and the diagnoses was examined using binary logistic regression. Results are shown in Table 3.

The model for all psychiatric medications explained 70.3% of the variance in prescriptions. Older age, Public (i.e., Medicaid/Medicare) insurance, higher visit number, presence of comorbid diagnoses, including ADHD, anxiety, and disruptive behavior disorder, were associated with a greater likelihood of receiving a psychotropic medication.

The logistic regression model for ADHD medications explained 43.2% of the variance. Public insurance, higher visit number, having diagnoses of ADHD, or disruptive behavior disorder were associated with a greater likelihood of receiving ADHD medication prescriptions, whereas Hispanic ethnicity and a diagnosis of depressive disorders was negatively associated with ADHD medication prescriptions.

The logistic regression model for antipsychotics explained 33.9% of the variance. Older age, Black and N-P races, and a diagnosis of disruptive behavior disorder were associated with a greater likelihood of receiving prescriptions for antipsychotic medications, whereas Hispanic ethnicity and a diagnosis of anxiety or ADHD were negatively associated with antipsychotic medication prescriptions.

The logistic regression model for antidepressants explained 44.5% of the variance. Older age, higher visit number, and a diagnosis of anxiety and/or depressive disorders were associated with a greater likelihood of receiving prescriptions for antidepressants.

Lastly, logistic regression model for mood stabilizers explained 17.9% of the variance. Older age and a diagnosis of disruptive behavior disorder were associated with greater likelihood of receiving prescriptions for mood stabilizers, whereas those with Hispanic ethnicity and Asian race were less likely to receive mood stabilizers.

Discussion

Our study sample demographics were representative of WA, aligning well with the state census. This may be because our study site is a major academic institute-affiliated program that serves patients from around the state.

This study identified high prevalence rates of ADHD (53.6%), anxiety disorders (52.8%), disruptive behavior disorders (41.4%), and depressive disorders (12.7%). This finding is not surprising as our sample is clinic based, specifically referred for psychiatric medication evaluations, compared with community-based or registry-based samples (Lai et al. 2019).

Among four classes of medications examined in this study, ADHD medications were most frequently prescribed (55.1%), followed by antidepressants (43.8%), antipsychotics (26.9%), and mood stabilizers (8.5%). The rates of antipsychotic use in ASD have been higher in other studies (Jobski et al. 2017; Shekunov et al. 2017). In a meta-analysis of 47 studies (Jobski et al. 2017) antipsychotic medication was the most common drug class in 15 of 47 studies ranging from 7.3% to 57.4%, whereas ADHD medications/stimulants were the most frequently prescribed in 12 of 47 studies. Notably, they also found a slightly lower prevalence of antipsychotic use from more recent data (19.5% before 2007 vs. 16.7% for 2007–2012). Antipsychotics are typically utilized to address disruptive/challenging behaviors in individuals with ASD, such as aggression, or self-injurious behavior. The lower rates of antipsychotic prescriptions in our sample may be because our sample included more recent prescription data compared with previous studies. This may better reflect current practice style, where there is an increased effort to identify underlying psychiatric disorders (e.g., anxiety) and the function for the disruptive/challenging behavior, and to pursue alternative treatment options, such as utilizing functional behavior assessment and applied behavior analysis.

In this study, age was consistently and positively associated with prescription of a psychiatric medication, including antidepressants, antipsychotics, and mood stabilizers, consistent with the finding from Jacket et al. (2017), with exception of ADHD medications that were most commonly prescribed for the 6–12-year age group, followed by the 13–17-year age group, which is also consistent with previous reports of ADHD treatment in youth overall (Danielson et al. 2018).

In our review of studies of youth with ASD, there were a lack of sex differences in prescribing (Jackel et al. 2017; Kamimura-Nishimura et al. 2017; Lake et al. 2017). While males had higher rates of ADHD and disruptive behavior diagnoses and females had higher rates of anxiety and depressive disorder diagnoses in our sample, when diagnoses were included in the regression model, we did not find sex differences in prescriptions for any of the examined psychiatric medication classes.

This study made several observations in relation to ethnicity and race. We found that Hispanic youth were less likely to receive ADHD medications, antipsychotics, and mood stabilizers, and Black and N-P youth were more likely to receive antipsychotics. In studies of youth overall, White youth are found to be prescribed more psychotropic medications than Hispanic and/or Black youth (Zito et al. 1998; Cook et al. 2017), including antipsychotics (Smith 2010). It is not clear whether this difference is related to an ASD and/or related developmental disorder diagnosis, or the use of a different statistical model. Regardless, it is noteworthy that Black (N = 53) and N-P (N = 18) youth received antipsychotics more frequently compared with other race categories. While we cannot confirm if this finding is related to relatively small sample size and related selection bias, this may be an example of health inequity.

Specifically, antipsychotics are often prescribed for children with ASD and challenging behaviors such as aggression. Black youth have been shown to be diagnosed with ASD at older ages (Maenner et al. 2020), which may reflect biases and/or less access to expert evaluation. In addition, a review of studies of perceptional biases have identified that Black young men tend to be perceived as taller and more physically threatening than White young men (Wilson et al. 2017). We must consider whether this bias could contribute to Black youth with ASD being perceived as aggressive and prescribed antipsychotic medications more readily than White youth. Disentangling reasons for prescribing, including objective differences in symptoms, is needed.

We found a significant association between Public insurance and any psychiatric medication and ADHD medications, as well as a trend toward a positive association between Public insurance and antipsychotic medications. In prior studies of psychotropic medication prescribing rates in youth with ASD who have public (e.g., Medicaid) versus private insurance, there have been conflicting results (Jackel et al. 2017). Children from low-income families are more likely to have Medicaid as their insurance source (https://www.medicaid.gov/medicaid/national-medicaid-chip-program-information/medicaid-childrens-health-insurance-program-basic-health-program-eligibility-levels/index.html). In the United States, 34.3% of Non-Hispanic Black, 22.5% of Hispanic, 17.0% of Non-Hispanic White, and 15.8% of Non-Hispanic Asian adults (ages 18–64) had public health insurance, according to 2019 estimates (Cohen et al. 2020). This suggests a possible confound between low-income and psychotropic medication prescribing. Low-income families may have less access to behavioral interventions, and therefore may be more likely to receive psychotropic prescriptions to manage symptoms (Jackel et al. 2017).

This study has several limitations. Patients or their caregivers were asked to self-identify the patient's race and ethnicity upon their new referral to the SCAC; these options were similar to the WA census. In contrast to the WA census, SCAC provided an option to select “Other” category for race. Some chose not to identify their race or ethnicity, leading to missing data. The WA census includes adults, in contrast to our sample of youth only, which may have created a discrepancy in the racial demographics. For those seen at SCAC, a larger proportion may have chosen to identify, based on the options provided, as two or more races. The SCAC is in King County, Washington, which has different demographics than WA overall.

Some patients were seen only once, which would make it less likely that their medications would be prescribed in the clinic, and this likely contributes to the association between higher number of visits and prescribing of medication. These data did not capture medications for those patients who were seen at SCAC and receiving prescriptions elsewhere. Therefore, we are likely underestimating the prevalence of those prescribed psychotropic medications. On the contrary, it is also possible that prescriptions provided during the visit may not have been used by the patients, given the nature of EHR review.

Medications classified as mood stabilizers, such as valproic acid and lamotrigine, may have been used for seizures rather than for mood symptoms, although this is less likely as prescriptions were linked to psychiatric clinic visits. Alpha-2 agonists were included in the ADHD medication group but may have been prescribed for sleep rather than ADHD.

We have not included data on the severity of ASD, which would have been helpful to contextualize our sample. ASD severity and comorbidities influence prescribing practices (Cook et al. 2017; Jackel et al. 2017; Shekunov et al. 2017). For instance, individuals with ASD and severe irritability are commonly prescribed SGAs (Fung et al. 2016). In addition, the diagnoses were captured by ICD-10 codes associated with visits, but not all relevant diagnoses may have been captured. This is a limitation of the review of extracted data from EHR, as compared with a more in-depth clinical encounter note review.

Conclusion

This examination of factors associated with psychiatric medication prescriptions for youth with ASD at one large Autism Center provides more specificity than prior studies about sex, ethnic, racial, and insurance-related differences that are relevant for this population. Older age and higher visit number were among the most significant factors associated with psychotropic prescriptions. We found that older youth were prescribed more medications, except for ADHD medications, which were most frequently prescribed to 6- to 12-year-old children. We found that Hispanic youth received fewer prescriptions than non-Hispanic youth for ADHD medications, antipsychotics, and mood stabilizers. We found higher rates of antipsychotic prescriptions for Black and N-P youth than youth of other races. There were also increased prescription rates among those with Medicaid insurance. These findings can serve as an impetus to examine the reasons for variance.

Clinical Significance

ASD is reported to occur in all ethnic, racial, and socioeconomic groups (Maenner et al. 2020), however, studies have identified ethnic and racial disparities in ASD care (Zuckerman et al. 2017). The SCAC serves youth throughout WA. The authors examined whether the ethnicity and race of patients (N = 1112) seen at the SCAC aligned with the demographics reported in the WA census, and whether patient demographics were associated with specific psychotropic prescriptions. These demographic differences in prescribing for youth with ASD provide more specificity than prior studies about sex, ethnic, racial, and insurance-related differences, and can serve as an impetus to examine the reasons for variance.