Developing and Testing Innovative New Methodologies for Analyzing the Impact of Mental Healthcare Benefit Design on Outcomes.”

University Of California Los Angeles, Los Angeles CA

Investigators

Linked publications, trials & patents

Abstract

Project Summary/Abstract Mental health insurance benefits are an important determinant of service utilization and therefore critical to include in studies examining access, affordability and quality of mental health care. Yet reliable mental health insurance benefit design measures (e.g., copayments, coinsurance, and deductibles) are rarely available to researchers. The increasing availability of claims databases offers researchers the opportunity to construct proxy measures. Yet due to the paucity of actual benefit design information, the most accurate algorithms for constructing proxies for benefit design using claims are unknown, as is the validity of these measures when used in policy analyses. The proposed research responds to PAR-17-264 (NIMH Innovative Mental Health Services Research Not Involving Clinical Trials) by exploiting a unique linked database already available to our research team, containing both behavioral health specialty claims and information on actual insurance plan design, to develop and test innovative algorithms for deriving benefit information from expenditure fields in inpatient and outpatient claims data. The study has three aims: (1) Develop algorithms for constructing benefit measures (copayments, coinsurance and deductibles) from specialty mental health claims; (2) Test the validity of constructed benefit measures by comparing the constructed values with ?gold standard? values from actual mental health benefit design data to determine the most accurate algorithm; and (3) Validate the use of constructed benefit measures for policy analyses by testing whether the constructed measures have the same estimated effects on utilization as the ?gold standard? data. Data available for use in this study reflect the team's longstanding collaboration with researchers from the behavioral health division of Optum®, one of the largest managed behavioral health organizations in the country. Optum has provided detailed benefit and claims data for a large sample of commercially-insured enrollees representing all 50 states. For a prior study, our research team has already used these data to create copayment, coinsurance, and deductible variables for inpatient and outpatient mental health care by network status and plan type. These variables will serve as the ?gold standard? measures against which the constructed variables will be compared. The methods for creating constructed values from the claims data will vary by benefit type and sensitivity analyses will be conducted to assess the extent to which researchers using claims data could still produce accurate estimates if they did not have access to certain key variables, such as a subscriber unit identifier. By developing a method to generate benefit measures from a commonly available secondary data source (i.e. claims data) and validating them using data less commonly available to researchers (i.e. actual benefit design from a claims processing database), the proposed study will help to facilitate future, highly policy-relevant studies by other researchers who do not have access to benefit design data, such as examining the implications of state and federal legislation affecting mental health insurance benefits (e.g., benefit mandates or parity laws) for access to care.