Association for Public Policy Analysis & Management

In response to growing abuse of prescription opioid painkillers, 49 U.S. states have implemented electronic prescription drug monitoring programs (PDMPs) which record patients into a state-wide system when an opioid is received. Prescribers can query patient information within a state’s database to identify possible patterns of abuse, including “doctor shopping,” where a patient receives many overlapping prescriptions from many doctors. Low prescriber utilization of the prescription monitoring databases led to the passage of mandates that strengthened the prescription monitoring programs by legally requiring prescribers to queue information about patients before writing any controlled substance prescription. This paper uses a difference-in-differences regression framework and interactive fixed effects factor models to identify the effect of the prescription drug monitoring programs and their later mandates on the strengths of oxycodone painkiller prescriptions and on drug crime rates, with an emphasis on heroin and opioid seizures and financially-motivated crimes.

While past work has examined the effect of PDMPs on aggregate oxycodone amounts, this paper is the first in the literature examining prescription opioid supply-side interventions to examine heterogeneous effects of a program on oxycodone pills of varying dosage strengths. Analysis of prescription effects are performed on data from the Medicaid State Drug Utilization data which tracks counts of drugs by state, quarter and National Drug Code, an identifier that allows oxycodone to be classified by pill strength.

This paper is part of a growing economics literature linking prescription opioid abuse and heroin abuse, and is the first empirical study at the national level to examine the link between Prescription Drug Monitoring Programs and heroin crime outcomes. This paper is also the first empirical study to investigate a link between opioid abuse and non-drug crime outcomes such as shoplifting. Crime rates are aggregated by county and month using the National Incident-Based Reporting System (NIBRS), a detailed incident-level dataset of crimes from 2000-2015. Heroin substitution behaviors are measured using heroin seizure incidents.

 The implementation of prescription drug monitoring databases causes a decrease in oxycodone pills in the Medicaid population, and this effect is concentrated among high-dose pills. Prescription drug monitoring programs cause a nearly immediate 30% increase in the number of crime incidents where an offender is carrying heroin. The programs do not directly affect the number of likely drug dealers selling heroin, suggesting a demand-side response indicating that abusers of high-dose pills substitute to heroin following a PDMP. Prescription drug monitoring programs have an ambiguous effect on illegally obtained prescription opioids because the databases create opposing market forces through channels of supply and demand. Preliminary results on later mandates of prescriber use of PDMPs suggest that mandates further decrease prescription oxycodone and further increase heroin seizures. Thefts and shoplifting increase in response to the databases, suggesting that addicts steal to support their drug addiction. Results imply that policymakers should anticipate additional substitution behavior in response to future supply-side interventions in the prescription opioid market and future policy should take into account these additional responses.