Volume 54 | Number 3 | June 2019

Abstract List

Kimberly Narain MD, PhD, MPH, Haiyong Xu Ph.D., Francisca Azocar Ph.D., Susan L. Ettner Ph.D.


To document differences among racial/ethnic/gender groups in specialty behavioral health care () utilization/expenditures; examine whether these differences are driven by probability vs intensity of treatment; and identify whether differences are explained by socioeconomic status ().

Data Source

The cohort consists of adults continuously enrolled in Optum plans with benefits during 2013.

Study Design

We modeled each outcome using linear regressions among the entire sample stratified by race/ethnicity, language and gender. Then, we estimated logistic regressions of the probability that an enrollee had any spending/use in a given service category (service penetration) and linear regressions of spending/use among the user subpopulation (treatment intensity). Lastly, all analyses were rerun with controls.

Data Collection

This study links administrative data from a managed organization to a commercial marketing database.

Principal Findings

We found that in many cases, racial/ethnic minorities had lower specialty expenditures/utilization, relative to whites, primarily driven by differences in service penetration. Among women, relative to whites, Asian non‐English speakers, Asian English speakers, Hispanic non‐English speakers, Hispanic English speakers, and blacks had $106, $95, $90, $48, and $61 less in total expenditures. explained racial/ethnic differences in treatment intensity but not service penetration.


In this population, was not a major driver of racial/ethnic differences in specialty utilization. Future studies should explore the role of other factors not studied here, such as stigma, cultural competence, and geography.