Problems in Defining and Evaluating State Opioid Policies: Insights from RAND's OPTIC Center

The nation is in the midst of an opioid-related public health crisis. The use of opioids has skyrocketed over the last decade, with serious health and social consequences. In response, states have enacted a heterogeneous collection of policies aimed at reducing mortality and morbidity, producing a state policy landscape that is both complex and dynamic. The field faces three central challenges. First, states implementing a given policy may be fundamentally different from states that do not, yielding low quality “control” groups for causal inference. Second, implementation of a given policy will vary by state, and the resulting variation in implementation can have important implications regarding the strength of a signal of any policy effect. Finally, existing evaluations focus primarily on effectiveness of a single policy; however, the actual policy landscape is much more dynamic, as states enact various policies on a rolling basis to address the evolving opioid crisis. Since multiple policies are being used to address the opioid crisis, it is imperative to understand the dynamic impacts of concurrent and sequential opioid policies. This is where RAND’s Opioid Policy Tools and Information Center (OPTIC) of Excellence comes in, with the goal of educating researchers and policy makers of these issues, providing standardized information relevant for conducting policy evaluations, and educating users of these data on issues related to definitions of policies, simultaneous adoption and robust methods for better inference.

In this paper, we document the problem of policy definition, dynamic implementation, and simultaneous policy adoption by demonstrating the evolution of state laws in four frequently evaluated opioid policy spaces: prescription drug monitoring programs (PDMPs), naloxone distribution laws (NALs), Good Samaritan Laws (GSL) and Medical Marijuana laws (MML). We demonstrate that (a) there are numerous ways researchers have operationalized each individual policy in their work, due to some degree because of differences in how policies are defined by data gathering organizations; (b) that most states have adopted at least one of these policies within the period 2013-2017; and (c) many states either simultaneously adopt multiple policies or update previous policies at the same time they adopt new policies, making identification of any one policy’s impact difficult using traditional econometric methods. We demonstrate this clearly by showing impacts of these various state policies on opioid prescribing and overdoses from 2010 forward. Our results demonstrate why the field needs a more complete understanding of the opioid “ecosystem” when trying to consider the impact of a state's opioid policy approach. It further shows the need for more robust methods for identification of causal effects.