Association of Mandatory-Access Prescription Drug Monitoring Programs With Opioid Prescriptions Among Medicare Patients Treated by a Medical or Hematologic Oncologist

Abstract

This cohort study examines the association between prescription drug monitoring program mandates with and without a cancer exemption and the percent of oncologists’ patients with any opioid prescription covered by Medicare Part D.

More than 30 states have enacted laws mandating use of prescription drug monitoring programs (PDMP) to reduce inappropriate opioid prescribing. Clinicians in those states must check the PDMP database before writing opioid prescriptions. These mandates vary substantially across states in their timing and scope; for example, some exempt patients with cancer.^1,2 We examine the association between state mandatory-access PDMPs and changes in the percent of oncologists’ Medicare patients with any opioid prescription fills.

Methods

We used the physican-level Medicare Part D Prescriber files for 2013 to 2017 and restricted the sample to physicians specialized in medical or hematologic oncology. For each year, we classified states as having: no mandatory-access PDMP or a mandatory-access PDMP with or without a cancer exemption.

We used linear regression with physician and year fixed effects to assess the association between PDMP mandates with and without a cancer exemption and the percent of oncologists’ patients with any opioid prescription covered by Medicare Part D. This approach measured within-physician changes in opioid prescribing after the implementation of mandated PDMPs compared with physicians in states without mandates, adjusting for secular time trends.

The institutional review board of Emory University determined that study approval was not required because all data analyzed were publicly available and deidentified.

Results

By 2017, 21 states had implemented mandatory-access PDMPs, including 5 states that explicitly exempted the reviewing requirement for patients with cancer (Figure 1). Compared with oncologists in states with no mandated PDMP, the proportion of oncologists’ patients who filled an opioid prescription declined by 1.15 percentage points (95% CI, −1.57% to −0.73%; a 4.8% decline) and by 0.67 percentage points (95% CI, −0.94% to −0.41%; a 2.8% decline) in states that implemented PDMP mandates with and without cancer exemptions, respectively (Figure 2). To protect the privacy of Medicare beneficiaries, this data set suppresses observations when physicians had 1 to 10 Part D claims in a year with an opioid prescription (12 297 of 53 036 [23.1%]). Results from models where we imputed the missing number of patients with any opioid prescription as either 1, 5, or 10 were similar in magnitude and significance to the presented results with missing values omitted.

Figure 1. Timeline of Number of States That Implemented Mandatory-Access Prescription Drug Monitoring Programs (PDMP).

Figure 2. Adjusted Percent of 40 739 Patients Treated by a Medical or Hematologic Oncologist With Any Opioid Prescription.

Discussion

The share of oncology patients who filled an opioid prescription declined by 4.8% and 2.8% in states that enacted mandatory-access PDMPs—with and without exemptions for patients with cancer, respectively. Although recent studies^3,4 have shown that mandated PDMPs are associated with an 8% to 12% reduction in opioid prescribing, this is the first to show that mandated PDMPs—with or without an explicit exemption for patients with a cancer diagnosis—are associated with decreases in opioid prescribing by medical and hematologic oncologists.

Implementation of PDMPs was intended to curb inappropriate opioid prescribing, not legitimate use among patients undergoing oncology treatment, who are often undertreated for pain.⁵ From our early results, we find that exemptions for patients with a cancer diagnosis did not shield Medicare patients treated by a medical or hematologic oncologist from the unintended spillovers of mandated PDMP requirements. This analysis was limited by lack of patient-level data, including cancer type and stage, and short follow-up period. In addition we did not examine changes in per-patient opioid dose. Future studies should examine the effect of PDMP policies with more years of follow-up data and adjusting patient-level characteristics. It is possible that with more time to learn about the nuances of the PDMP mandate, prescriber practices will adjust.

Conclusions

Although policymakers are motivated to prevent opioid misuse, there is growing concern that some physicians—burdened by the task of consulting a PDMP and added scrutiny over their prescribing—have reduced their opioid prescribing even for patients with legitimate pain management needs.⁶ These results show that with or without an exemption for patients with cancer, the percent of patients treated by a medical or hematologic oncologist receiving opioids declined after mandatory-access PDMPs were implemented. As more states contemplate policies to alleviate the opioid crisis, it is critical to understand how they affect both problematic and legitimate opioid use.