Mandatory use of prescription drug monitoring programs.

The United States is in themidst of a prescription opioid overdose and abuse epidemic. The rate of fatal prescription drug overdoses involving opioids almost quadrupled from 1.4 deaths/100 000 people in 1999 to 5.4 deaths/100 000 people in 2011.1 The rate of emergency department visits involving prescription drug misuse—primarily of opioid, antianxiety, and insomnia medications—more than doubled from 214 visits/100 000 people in 2004 to 458 visits/100 000 people in 2011.2 Forty-nine states have responded by developing prescription drug monitoring programs (PDMPs), which digitally store controlled substance dispensing information and make those data accessible to prescribers, pharmacies, and law enforcement officials. Although PDMPs are designed to curb opioid overprescribing, prescriber utilization is low. The median PDMP registration rate among licensed prescribers who issue at least 1 controlled substance prescription is 35%.3 Furthermore, not all enrolled prescribers regularly use PDMPs. Consequently, 22of the49stateswithPDMPsnow legally mandate prescribers to query the system before writing for controlled substances with recognized potential for abuse or dependence.4 These requirements face pushback from prescribers, many of whom consider them to be burdensome incursions into clinical practice.5 For example, physician and dentist group challenges to thebreadthofcircumstancesproposedfor PDMP checks have contributed to a 2-year delay in the final implementation of a legally required mandate in Massachusetts.6 On the other hand, proponents argue that requiredPDMPconsultation isnecessary tochange prescribing behavior, citing early evidence from states that havedeployedmandates to demonstrate their potential to reduce opioid abuse.5 Some studies associate state PDMPs with lower rates of prescription drug abuse and altered prescribing practices, although evidence is mixed and inconclusive.7 Small (if any) demonstrated effect sizes, a dearth of detailed prescribing data prior to PDMP implementation, and a lack of precision in characterizing interventions inexistingstudiesmakeattributingsignificantchanges in totalopioidprescribingorhealthoutcomes to PDMPs a challenge.7 Another reason for inconsistent findings may be low and variable prescriber utilization of PDMPs. Prescribers must actually access PDMP data for the systems to have an appreciableeffect. Inaddition,voluntaryapproacheshaveselfselection bias: already conscientious opioid prescribers are those likely to use PDMPs. Clear benefits can derive from increased prescriber participation in PDMPs. When prescribers query the database for a patient’s prescription history, they have access to information about the dose, supply, and prescriber of scheduled drugs the patient has filled. With knowledge of this information, practitioners can communicate with patients about their histories, avoid polypharmacy, and refrain from supplying opioids to those who “doctor shop” while comfortably prescribing to those who do not. When a critical mass of prescribers use PDMP information, the collective care each patient receives across providers theoretically can be improved and efficiencies are less likely to be compromised by any one uninformed practitioner. Moreover, prescribers may become accustomed to new practice norms, in which improved information and patient outcomes outweigh perceived burdens associated with checking PDMPs. But are mandates an effective way to increase PDMP use and improve prescribing outcomes? Twenty states require licensed prescribers to register with the state PDMP.8 Use mandates go a step further and dictate the circumstances for PDMP queries. Some states require prescribers to access a patient’s prescription history in the database if they suspect drug abuse; others rely on objective criteria (eTable in the Supplement). In Kentucky, Tennessee, New York, and Ohio—early adopters of comprehensive use mandates—there were substantial increases in queries and reductions in opioid prescribing following implementation. In New York, Tennessee, and Ohio, there were declines in doctor shopping.3,5 Although these results must be rigorously validated, for example, by comparing them to outcomes in states without mandates and controlling for co-interventions, they suggest the potential influence of mandates to reduce unsafe opioid prescribing. Mandates face significant prescriber opposition across the country.5,6 Some objections relate to generic problemswithPDMPs thatwouldbeexacerbated under a mandate. Prescribers have difficulty obtaining logins, systems can be “down,” information is not integrated into clinical workflow, and data are often incomplete.9 Moreover, minimal guidance exists to assist users in interpreting query results. These drawbacks burden and create ambiguity for physicians and other prescribers. Other objections are specific to mandating PDMP use. Robust evidence is lacking about how to best target mandates to prescriber types and contexts, which makes defining exemptions a policy challenge. Bluntly framedmandatescouldrequirephysiciansandotherprescribers to search PDMPs when not clinically indicated orwaste time that couldbespentotherwise treatingpatients. Althoughmandates are not meant to deter opiVIEWPOINT