Abstract
The rapid rise in opioid overdose deaths in the past 17 years has coincided with a tripling of prescription opioid
(PO) prescriptions dispensed, largely to treat chronic pain. Increased co-prescription of benzodiazepines
(BZDs) and opioids also substantially increased the risk of overdose. Cannabis has been proposed as an
alternative treatment for chronic pain that could ameliorate opioid withdrawal symptoms and assist in recovery
from opioid and BZD dependence. Two major policy shifts are likely to change prescribing practices and abuse
of POs and BZDs. First, in the past decade, most states have enacted policies that regulate PO prescribing
and dispensing. Second, since 1996, 29 states have legalized use of cannabis for medical purposes, and 9
states have legalized cannabis for recreational use. Policies that regulate access to POs may decrease the
number of opioid prescriptions for chronic pain and co-occurring PO/BZD prescribing, while laws that allow
greater access to cannabis may offer a substitute for POs, heroin, and BZDs. The combination of stricter PO
policies and less restrictive cannabis laws may reduce opioid-related harm to a greater extent than either
measure alone. As states make unprecedented changes to PO policies and cannabis laws, we need to
examine the independent and synergistic contributions that both types of measures have on opioid prescribing
practices and opioid overdoses, with and without BZDs. We propose to pursue this aim in two populations: (1)
in the U.S. population, using repeated cross-sectional data of individuals nested in states from the National
Survey on Drug Use and Health; and (2) among Medicaid patients with chronic pain (who have 10 times
greater risk of opioid use disorder relative to privately insured patients), using a 45-state Medicaid Analytic
Extract longitudinal cohort. Our specific aims are: (1) to examine, in NSDUH, whether nonmedical use of POs,
BZDs and heroin, and opioid and BZD use disorders decrease (and cannabis use/disorder increases) following
enactment of more restrictive PO policies and less restrictive cannabis laws in 2004-2019, compared to trends
in states that did not enact these measures; and (2) to test whether Medicaid patients are less likely to have
claims for opioid prescribing (alone and overlapping with BZDs), clinic visits for chronic pain, and opioid
overdoses (with and without co-occurring BZD overdose) following enactment of more restrictive PO policies
and less restrictive cannabis laws in 2001-2019, compared to patients in states that did not enact these
measures. For Aim 2, we will sample 10,000 Medicaid patients with chronic pain per year, follow each cohort
for 4 years (n=190,000), and combine the cohorts to construct an accelerated longitudinal cohort. States will be
classified by how restrictive or lenient their PO policies are (e.g., prescription drug monitoring programs, pain
clinic regulations), and whether they legalized cannabis for medical and/or recreational use. At a time when
opioid overdose deaths are increasing at an unprecedented rate, this study will provide critical, policy-relevant
findings about the types of policies that are most likely to end the opioid epidemic.
