Wonk Data Drop: Race is not the strongest predictor of child psychotropic use
Despite White children taking more psychotropics, no racial differences in the gap between use and mental health diagnosis
By Meredith Dost, PhD and Robin Ghertner, MPP
BLUF
Millions of children take psychotropic medications - many without a documented diagnosis. While there are reasons that can be perfectly appropriate, it’s the easiest signal to study for potentially inappropriate prescribing.
Policy discussions often raise race as predicting prescribing patterns; we tested that with data.
This analysis finds important differences in medication use and relevant diagnosis across children in the largest racial and ethnic groups. We find:
White non-Hispanic children were more likely than other racial groups to use psychotropics in 2021-2022.
11.3% of non-Hispanic white children used these medications in 2021-2022,
That’s more than Hispanic (7.6%), non-Hispanic Black (5.5%), and non-Hispanic Asian children (3.2%).
White children were also more likely to be diagnosed with a behavioral health condition (15.2%) compared with Hispanic (8.0%) and Black (5.3%) children.
So once you account for whether a child has a diagnosis, plus poverty, age, and sex, there are no racial differences in the gap between taking a psychotropic and having a diagnosis.
ICYMI
Many children using psychotropic medications don’t have a mental or behavioral health diagnosis
Our recent Data Drop found that 9% of all children took psychotropic medications in 2021-2022, and 11% of all children had a behavioral health diagnosis.
30% of all children using psychotropics didn’t have a related diagnosis.
Psychotropic use was higher among kids in poverty and kids on Medicaid, than their counterparts.
What’s Missing From the Discussion: Differences by Race and Ethnicity
We know Black and Hispanic children are disproportionately involved in the child welfare system, more likely to be in poverty, and that they face less comprehensive behavioral health treatment.
At the same time, research has established that White children are more likely to use psychotropics (Sepe-Forrest et al., 2025).
What we don’t know is how the gap between diagnosis and medication use differs by race and ethnicity. This question matters in a policy moment with significant discussions on the appropriate role of psychotropic medication in children’s mental health.
Our new analysis re-examines race and ethnic differences in psychotropics, using the most recent national data. In particular, we look at racial differences in the diagnosis gap, defined as use of a psychotropic without a behavioral health diagnosis.
We use data from the Medical Expenditure Panel Survey (MEPS) to compare rates of psychotropic use and behavioral health conditions, and the gap between prescribing and formal diagnosis, across racial groups.
Our results should not be interpreted to mean that a given race “causes” a child to take prescriptions. They simply describe differences in utilization.
White Children Are More Likely to Use Psychotropics and Have a Diagnosis
Figure 1 breaks down differences by race and ethnicity. Key takeaways:
11.3% of White children utilized psychotropic medications in 2021-2022
This is a higher share than Hispanic (7.6%), Black (5.5%), and Asian non-Hispanic (3.2%) children.
White children were also more likely to be diagnosed with a behavioral health condition:
White children with a behavioral health condition: 15.2%;
Hispanic children with a behavioral health condition: 8.0%;
Black children with a behavioral health condition: 5.3%; and
Asian children with a behavioral health condition: 2.8%.
Poverty is a well-established driver of psychotropic use.
In earlier analyses, we found that children in poverty — and children on Medicaid — are more likely to take these medications than their peers.
We also know from Census data that White children are less likely to live in poverty than children of color. The intuitive read is that poverty could be a proxy for race in our earlier study, but it’s not.
This raises a puzzle: if White children are less likely to live in poverty, why do they have the highest rate of psychotropic use? To answer, we go deeper.
No Racial Differences In the Diagnosis Gap
Once we account for behavioral health diagnosis, poverty, age, and sex, the racial differences in psychotropic use disappear.
White children were not at greater risk of taking psychotropics without a behavioral health diagnosis than their peers. Figure 2 shows our estimates.
This means that the higher prescribing rates shown earlier are explained by higher rates of diagnosis– not by disproportionate prescribing without a diagnosis.
A White child with a behavioral health diagnosis had a 57% chance of taking a psychotropic medication, which is statistically indistinguishable from other race and ethnic groups.
A White child without such a diagnosis had a 3 percent chance of taking such a medication, again, not statistically different from other groups.
Note, while the estimate for Asian children looks different from White children, that difference isn’t statistically significant.
Poverty is the one thing that remained steady across this and previous analyses. Though we didn’t show it in the figures, children in poverty are more likely to take psychotropics even when accounting for their behavioral health condition and their race-ethnic group.
These results point to more questions to consider.
The diagnosis gap is real, but it does not differ by race. Poverty is the driver that holds steady.
That raises strategic questions for policymakers and system leaders:
Are children being diagnosed appropriately?
For children using psychotropics without a diagnosis, we need to understand whether they should have a diagnosis.
Prescribing without a diagnosis is not automatically bad, but it does invite discussion.
This can help decision makers determine whether the issue warrants a policy response, and may help shape and guide that response.
Do families have access to treatment beyond medication?
If psychotropics are being used as a stand-in for therapy or combined care , it may signal coverage, billing, or access problems.
Those each raise different policy issues worth exploring.
Are these race and ethnic differences in psychotropic use and behavioral health diagnoses also present for children in foster care?
Black and Hispanic children are more likely to enter foster care. Does foster care involvement narrow or widen the gap?
Our previous analysis showed substantial differences in mental health diagnoses for children in foster care vs those not in care.
What happens to psychotropic use among Black children once they enter care?
We also found in a prior Premium Data Drop - using a different data source - that children in foster care with a mental health condition were more likely to have medication combined with therapy, than children not in foster care.
Do children of different race or ethnic groups have equal access to therapy?
This analysis also doesn’t focus on children in foster care, especially those in congregate settings where overprescribing is a particular concern.
These are questions Wonk will continue to grapple with. We’ll be limited by what data are available nationally. This is an area where states can produce better information to drive decisions. State data systems have comprehensive information on foster care involvement, behavioral health records, and child demographics.
Methodological Details
This section describes the data, coding decisions, and methodological approach used in this analysis.
We use a nationally representative survey commissioned by the U.S. Department of Health and Human Services called the Medical Expenditure Panel Survey (MEPS), for years 2021-2022 (latest available). The MEPS surveys families and individuals along with their medical providers and employers to provide in-depth data on the cost and use of health care in the U.S. One important limitation is that the MEPS only breaks down race and ethnicity into the largest groups (White, Black, Hispanic and Asian); sample size and the survey design don’t permit looking at other groups, such as Native American or Pacific Islander.
For each year, we link the full-year consolidated data files with the medical conditions file and the prescriptions file. Finally, we pool all three years’ linked files together and merge in the pooled linkage file for common variance structure which allows us to compute accurate estimates that take into account the complex sampling structure. We divide the person-level survey weight (PERWT) by two, the total number of pooled years, to compute estimates that represent annual averages. All analyses were conducted in R and primarily used the “survey” package.
We coded children’s medical conditions and prescription medications taken based on the following definitions:
Mental or behavioral health condition: An individual is coded as having a mental or behavioral health condition if they are treated for any mental, behavioral, or neurodevelopmental disorder (includes all MEPS condition codes beginning with “MBD”).
Psychotropic medication use: An individual is coded as taking at least one psychotropic medication if they take a psychotherapeutic agent or a central nervous system agent. These include antidepressants, antipsychotics, anticonvulsants, antimanic, antiparkinsonian, anxiolytics-sedatives, benzodiazepines-barbiturates, central nervous system agents, hypnotics, and stimulants.
The analyses underlying Figure 2 are based on a logistic regression predicting psychotropic use (1=any use, 0=no use), with the following categorical variables: mental/behavioral health condition status, race-ethnic group, family poverty status as % of Federal Poverty Level, age, and sex. All differences that we discuss are statistically significant (p<0.05).
References
Sepe-Forrest, L., Meraz, R., Adams, S. M., Chang, Z., D’Onofrio, B. M., & Quinn, P. D. (2025). Diagnostic patterns and racial/ethnic differences in antipsychotic prescribing among privately insured youth. Translational Issues in Psychological Science.https://doi.org/10.1037/tps0000457
