Racial and Ethnic Differences in Psychotropic Prescription Receipt Among Pediatric Patients Enrolled in North Carolina Medicaid

We analyzed North Carolina Medicaid 2017–2018 institutional, professional, and prescription claims; enrollment data; and provider data obtained from the North Carolina Department of Health and Human Services (DHHS). Briefly, the enrollment file contains beneficiary information, including date of birth, sex, self- or caregiver-reported race and ethnicity, county of residence, Medicaid eligibility category, and enrollment dates. The prescription claims file contains fill date and prescriber and pharmacy data for filled prescription drugs that were prescribed by providers in outpatient settings. The institutional and professional files contain claims for inpatient and outpatient services. We used the provider file to identify the taxonomy associated with each provider. The Duke University Health System Institutional Review Board and the North Carolina DHHS approved this study.

We used North Carolina Medicaid claims and enrollment data to identify pediatric patients ages 0–21 who were enrolled in Medicaid on October 1, 2017. We included patients ages 0–21 because of the North Carolina foster care program, which includes individuals up to age 21. To be included in the analyses, patients had to have at least one professional or institutional claim with a provider type of interest (child or adult psychiatric provider, pediatric provider, family medicine provider, primary care physician assistant, or primary care nurse practitioner; a table showing provider taxonomies is available in the online supplement to this report) between October 1, 2017, and September 30, 2018.

We then identified patients who filled at least one prescription in the most common psychotropic medication classes (i.e., antidepressant, stimulant, alpha agonist, or antipsychotic) during the study period by using the generic name on the prescription claim (see the online supplement). Notably, Medicaid pharmacy claims include only filled prescriptions.

We categorized age in years as follows: 0–5, 6–11, 12–17, and 18–21. Counties were classified as either urban or rural in accordance with North Carolina Office of State Budget and Management guidelines. We used the race (White, Black, Asian, Native American, Hawaiian or Pacific Islander, or unreported) and ethnicity (non-Hispanic, Hispanic [Cuban, Mexican, Puerto Rican, or Hispanic other], or unreported) listed in the member enrollment file. The Medicaid eligibility categories were grouped into children from low-income families (N=905,995); foster care or adoption assistance (N=20,334); blind, disabled, or medically needy (N=49,499); pregnant women (N=712); and other programs (e.g., documented and undocumented immigrants, family planning waiver, and refugee medical assistance) (N=7,346).

We fit unadjusted and adjusted multivariable modified Poisson regression models (10) to examine factors associated with the likelihood of filling any psychotropic prescription and filling a prescription in each of the four included drug classes among the entire pediatric Medicaid population. The modified Poisson model was used because the traditional logistic model does not reliably estimate risk ratios for dependent-variable prevalence rates of the outcome (e.g., prescription rates) >10%. In the multivariable model, we included covariates for age, sex, race, ethnicity, county of residence (rural vs. urban), and Medicaid eligibility category. We conducted a sensitivity analysis restricted to those who were continuously enrolled during the 1-year period (93%, N=916,986) to determine whether there were differences from the primary models. We used a two-tailed alpha of 0.05 to establish statistical significance and reported 95% confidence intervals (CIs).

Of the 983,886 pediatric patients enrolled in North Carolina Medicaid who met eligibility criteria, 10.6% (N=103,817) received at least one filled psychotropic prescription. When examining the proportions of patients who received a filled prescription in each of the four included drug classes by race and ethnicity, a higher percentage of White patients (4.2%, N=24,742 of 590,536) received an antidepressant than Black patients (2.1%, N=7,346 of 345,946) (see the online supplement). However, a similar proportion of White and Black patients received a filled antipsychotic prescription (2.0% [N=11,709 of 590,536] and 1.9% [N=6,407 of 345,946], respectively). Compared with Hispanic patients, a higher percentage of non-Hispanic patients received a filled antidepressant (3.9% [N=29,345 of 751,843] vs. 1.3% [N=2,793 of 215,348]) or antipsychotic (2.2% [N=16,731 of 751,843] vs. 0.5% [N=1,144 of 215,348]) prescription.

After multivariable adjustment, we found racial and ethnic differences in the receipt of any filled prescription (Table 1). Compared with White patients, patients of all other races had a significantly lower likelihood of receiving any filled psychotropic prescription, with Asian patients being the least likely to receive a filled prescription (adjusted risk ratio [ARR]=0.20, 95% CI=0.18–0.22). When examining ethnicity in the adjusted model, we found that Hispanic patients and those with unreported ethnicity had a lower likelihood of receiving any filled psychotropic prescription than non-Hispanic patients.

When examining the multivariable adjusted likelihood of patients receiving a filled prescription in each drug class, we found similar results. Compared with White patients, Black patients were significantly less likely to receive a filled antidepressant (ARR=0.40, 95% CI=0.39–0.41), alpha agonist (ARR=0.61, 95% CI=0.60–0.63), stimulant (ARR=0.67, 95% CI=0.65–0.68), or antipsychotic (ARR=0.63, 95% CI=0.61–0.66) prescription. Hispanic patients were 70%–78% (ARRs=0.22–0.30) less likely to receive a filled prescription in each of the given drug classes compared with non-Hispanic patients. Comparable estimates were found with the unadjusted model as well as with the sensitivity analysis that examined only those with continuous enrollment. The full multivariable model, adjusted for other covariates (e.g., age, sex, county of residence, Medicaid eligibility category), can be found in the online supplement.

In this study, we examined filled psychotropic prescription claims for pediatric patients enrolled in North Carolina Medicaid. We found that 10.6% of all enrolled patients received at least one filled psychotropic prescription and that prescription rates varied by race and ethnicity among the four included drug classes. Similar to prior research that used national samples (4–8), these findings demonstrate that, compared with White and non-Hispanic patients, Black and Hispanic pediatric patients in North Carolina were significantly less likely to receive any filled psychotropic prescription or a filled psychotropic prescription in any of the four major classes of psychotropic medications reported in our study, after adjusting for relevant demographic characteristics. Understanding these patterns within a specific state and determining avenues to address these differences require further research on provider-, patient-, and system-level factors.

Our findings raise questions about whether racial and ethnic differences are due to the lack of access to pharmacological treatment for Black and Hispanic patients, the overuse of psychotropic medications among White and non-Hispanic patients, or the combination of both. Previous research has shown that Black and Hispanic patients have significantly lower rates of outpatient behavioral health service utilization compared with White and non-Hispanic patients, even though no racial or ethnic differences exist in prevalence rates of behavioral health conditions after controlling for provider behavior (11–13). Furthermore, there are widespread sociopolitical and systemic inequalities, such as racism and discrimination, that disproportionately affect Black and Hispanic populations; as a result, these inequalities may affect their access to behavioral health treatment (14). Gaining insight into patients’ perspectives of lived experience with seeking and receiving behavioral health care may also help to address racial and ethnic differences in treatment.

Although we found significant racial and ethnic differences in receipt of filled psychotropic prescriptions, we did not have access to sufficient data to ascertain whether these differences constitute health disparities. Specifically, the Institute of Medicine has defined “disparities” as differences not due to clinical needs, appropriateness, or patient preferences (15). Unfortunately, Medicaid claims data do not include information on clinical need for, appropriateness of, or patient preference for psychotropic prescriptions; thus, this limitation precludes characterizing the differences we found as disparities. Other provider- and system-level factors, such as provider bias, stereotypes and clinical uncertainty, institutional and structural racism, and cultural barriers, may be potential explanations for the differences in receipt of filled psychotropic prescriptions found in our study; therefore, these factors should be investigated in future studies.

This study had several limitations. The time frame was opportunistic and limited to a 1-year period in order to inform and support a statewide initiative being implemented in North Carolina. Although we did not restrict our sample to patients with continuous Medicaid enrollment, our sensitivity analysis indicated that this covariate did not affect our results. Our sample included only pediatric patients enrolled in North Carolina Medicaid; thus, results may not be generalizable to other states or patient populations.

We did not include patients’ clinical diagnoses or symptoms; therefore, we are unable to determine whether patients’ prescriptions were indicated. We also recognize that there may be racial and ethnic biases associated with a behavioral health diagnosis, which is typically required when prescribing a psychotropic medication. Additionally, our analyses do not account for off-label prescribing, polypharmacy, the use of psychotherapy and other behavioral health treatments, patient or parent-guardian preference, or treatment response, all of which may influence receipt of a filled psychotropic prescription and are important factors to investigate in future research.

In this study, we examined receipt of filled psychotropic prescriptions among Medicaid-insured pediatric patients in North Carolina. Our results are consistent with those from prior studies, demonstrating that Black and Hispanic patients are significantly less likely to receive filled psychotropic prescriptions compared with White and non-Hispanic patients. Our results can be used by policy makers to inform community outreach and investment, to better understand racial and ethnic differences in prescription receipt, and to promote access to care for underserved populations. In future studies, researchers should seek to understand and address the reasons for these racial and ethnic differences in treatment and investigate patient-, provider-, and system-level factors.