Project Summary
In the early 2000s, an estimated 1% of United States youth received antipsychotic prescriptions. Of these
prescriptions, approximately 65% were given for non-FDA-approved indications. This led to urgent calls to curb
antipsychotic prescribing, based on both the lack of research on drug benefits in youth and the extensive
evidence demonstrating adverse metabolic and neurologic consequences. More than a decade later, the
characteristics of the current pediatric population receiving antipsychotics, and the degree to which these
medications alleviate psychological symptoms, remain unclear. The objective of this proposal is to understand
the pediatric population receiving antipsychotics and examine their effectiveness in reducing psychiatric
hospitalizations compared to alternate medication options. It will use a new-user patient sample (N = 46,425)
that is large enough to adequately study hospitalization risk and examine differential effects based on
demographic factors and conditions that have scarcely been studied (e.g., depression), yet appear frequently
in antipsychotic users. This project will use a large United States health claims dataset to accomplish this
objective through two specific aims: (1) characterize the youth population using antipsychotics by examining
individual factors that predict antipsychotic initiation and (2) contrast medication effectiveness at reducing rates
of psychiatric hospitalizations using active comparator designs. In Aim 1, antipsychotic users (cases) will be
matched 1:1 with non-users (controls). The odds of using antipsychotics will then be evaluated based on pre-
existing psychiatric diagnoses, race, and socioeconomic status. In Aim 2, survival analysis will be used to
compare rates of psychiatric hospitalizations across different comparisons. First, differences in rates of
psychiatric hospitalizations will be compared for aripiprazole and risperidone users both across the sample and
in demographic subgroups. Second, rates of hospitalizations will be contrasted for antipsychotics versus ADHD
stimulants in users with ADHD and antipsychotics versus antidepressants in users with depression. The active
comparator design and propensity score matching will be applied to ameliorate confounding by indication and
related factors that obscure the ability to identify causal effects of the medications. This work is innovative as it
will provide insight into comparative effectiveness of antipsychotics and alternative treatment options for
pediatric conditions (e.g., ADHD) that have rarely been studied. Furthermore, the results will have a significant
public health impact by providing greater insight on disparities in antipsychotic treatment and how
antipsychotics may differentially affect diverse populations. This program of research will dovetail with the
applicant’s prior experience on causal inference methods and their long-term goal of becoming an impactful
pharmacoepidemiologist by training the applicant to independently analyze large-scale health claims datasets,
broadly disseminate research to audiences that include healthcare professionals, and develop strong content
knowledge in pediatric psychotropic use.