Evaluating Legal Interventions Designed to Improve Behavioral Health Outcomes

Relatively recent recognition of the significant burden of behavioral health disorders has elevated their status as public health issues in need of population-level interventions. This dissertation focuses on evaluating two types of legal interventions that have enjoyed widespread but heterogeneous adoption and that are aimed at improving behavioral health care and outcomes: mental health parity laws and prescription drug monitoring programs. I further probe the appropriate means of regulatory intervention given the nature of behavioral health challenges targeted by these laws, existing evidence bases, and legal considerations. Chapter 1, entitled “Association of Federal Mental Health Parity Legislation with Health Care Use and Spending Among Persons with Mental Illness”, undertakes a novel examination of the Mental Health Parity and Equity Addiction Act (MHPAEA) of 2008. MHPAEA is a robust federal law that requires equity in insurance coverage between mental health and medical/surgical benefits. I use claims data from a large national commercial health insurer to assess the effects of MHPAEA on health care utilization and out-of-pocket spending among enrollees diagnosed with a mental health disorder. I review state laws that preceded MHPAEA to identify 24 states that exempt small-employer plans from parity requirements. Within these states, I develop a “control” group (of small-employer plan enrollees not subject to parity throughout the study period) and an “exposure” group (of self-insured plan enrollees newly subject to MHPAEA) for assessing the effects of MHPAEA. Employing a difference-in-differences design and propensity score matching techniques, I show that, among the exposure group, the out-of-pocket cost per outpatient mental health visit slightly declined while covered outpatient mental health visits modestly increased (4-7%) after MHPAEA. I also find that total out-of-pocket spending on outpatient mental health services as well as inpatient and emergency department use are unchanged by the law. I conclude that further investigation is needed to assess whether parity is associated with changes in clinical outcomes. Chapter 2, a normative piece entitled “Preventing Opioid Misuse with Prescription Drug Monitoring Programs: A Framework for Evaluating the Success of State Public Health Laws” argues that successful policymaking to address prescription opioid misuse requires evaluation against a defined set of criteria. I articulate and synthesize three key criteria with which to dynamically evaluate and justify state prescription drug monitoring programs designed to curb high opioid prescribing and misuse: (1) legal powers to regulate, (2) effectiveness of regulation, and (3) ethical considerations. PDMPs are widespread and promising, but heterogeneous and largely uninformed by robust evidence or a systematic assessment of best practices. I demonstrate my framework’s utility in guiding public health lawmaking given the complexities and persistent magnitude of prescription opioid misuse and the rich arsenal of policy options available to address it. I conclude by recommending that PDMPs be implemented with the following features: timely and complete data; strong incentives for prescriber participation; user guidelines and education; integration into clinical work flow; and robust confidentiality and privacy protections. Ongoing evaluation of programs to identify features appropriate for retention and replication also is crucial if PDMPs are to fulfill their potential. Chapter 3 seeks to assess the impacts of more recent PDMP features and is entitled “Effects of Robust State Prescription Drug Monitoring Programs on Opioid Prescribing and Use.” Based on a review of PDMP laws, I identify five “intervention states” with robust PDMPs, as principally identified by the presence of a requirement that prescribers query the data before prescribing opioids. I compare commercial pharmacy claims for enrollees in these intervention states to those in respective controls (with weak or no PDMPs) using an interrupted time series design. I find that the percent of enrollees filling opioid prescriptions and the mean morphine equivalent dosage dispensed per enrollee each decline following robust PDMP implementation in most intervention states relative to controls, with Kentucky exhibiting the most dramatic relative reductions. I recommend that other states seeking to curb high opioid prescribing consider implementing Kentucky’s particularly robust PDMP features, which include a use mandate paired with a registration mandate and careful implementation supervision.Health Polic