Our study aims to examine disparities in mental health services uptake and variation in mental health care settings across racial-ethnic and rural-urban subpopulations before and after the COVID-19 pandemic. We will employ two nationally representative datasets with different sampling methods: the 2018-2021 Medical Expenditure Panel Survey (MEPS) files and the 2018-2021 National Survey on Drug Use and Health (NSDUH).

First, using the MEPS data, we will examine the trends in three measures of mental health service access in the past year.

• Any mental health care
• Outpatient mental health care
• Psychotropic medication

We will categorize mental health events into four settings:

1. Outpatient or office-based visits to a mental health specialist or specific visits to primary care provider
2. Hospital inpatient stays
3. Emergency room visits for any disorder covered by mental health or behavioral health disorder ICD-10 codes F01-F99 or Conditions with Clinical Classifications Software Refined codes MBD000-MBD034 (including mental, behavioral, and neurodevelopmental disorders as well as opioid-, alcohol-, and substance abuse-related conditions)
4. Any psychotropic medication fills associated with episodes of mental health or substance use disorders

We will identify participants in need of mental health care by extracting data from the Kessler-6 scale, the two-item Patient Health Questionnaire, self-reported poor mental health status, and the mental health component of the SF-12 health status instrument. We will use the most updated Urban Influence Codes (UIC) to define metropolitan counties as urban and micropolitan or non-core counties as rural. Because UIC codes are not available in MEPS public use data files, we have requested access to these restricted data, particularly state and county FIPS Codes and Area Health Resources Files.

Second, using NSDUH, we will assess trends in four measures of past-year mental health care access including:

1. Any mental health services
2. Specialty and non-specialty mental health services
3. Psychotropic medication
4. Virtual mental health services

Using the population weights constructed by MEPS, we will compare national rural-urban and racial differences in participant characteristics, mental illness status, and in mental health access patterns during 2018-2019 and 2020-2021. To estimate the independent differences in mental health care patterns among individuals with mental health diagnosis or self-reported mental health conditions/needs, we will use generalized linear probability models and/or logistic regressions, with complex survey weights controlling for individual sociodemographic characteristics, mental health severity, and residence state fixed effects.