Doctor Shopping in Medicare Part D: The Data Behind Multi-Provider Prescribing

"Doctor shopping" — visiting multiple prescribers to obtain controlled substances — is a well-documented pattern in prescription drug abuse. While Medicare Part D data doesn't directly show patient-level shopping, it reveals the provider side of the equation: prescribers whose patterns are statistically consistent with being sources in a doctor-shopping network.

What Does Doctor Shopping Look Like in Data?

While we can't see individual patients visiting multiple doctors, we can identify providers whose prescribing patterns create the supply side of doctor shopping. Key indicators:

• Extreme opioid prescribing rates — more than half of all claims are opioids
• High cost per beneficiary — patients receive unusually expensive regimens
• Opioid+benzodiazepine co-prescribing — a dangerous combination linked to overdose deaths
• High long-acting opioid rates — Schedule II drugs with higher abuse potential
• OIG exclusion status — providers already flagged by federal investigators

The Highest Opioid-Rate Prescribers

Among the top 1,000 opioid prescribers in Medicare Part D, these providers have the most extreme opioid prescribing rates:

Cost-Per-Beneficiary Outliers

Another doctor-shopping signal: providers with extremely high cost per beneficiary. When a small number of patients generate outsized drug costs, it can indicate concentrated prescribing to a few individuals seeking large quantities.

Specialties Most Associated With Opioid Prescribing

Certain specialties have structurally higher opioid rates. Understanding these baselines is critical — a pain management specialist at 40% opioid rate may be normal, while a family practice doctor at 40% is a statistical outlier.

The Opioid + Benzodiazepine Red Flag

Co-prescribing opioids and benzodiazepines is one of the strongest indicators of problematic prescribing. The FDA has issued black box warnings about this combination. In our dataset, 0 providers among the top opioid prescribers are flagged for opioid+benzo co-prescribing.

Detecting Doctor Shopping Networks

True doctor-shopping detection requires patient-level data (which CMS restricts for privacy). However, provider-level patterns can identify the supply nodes in these networks:

1. Statistical peer comparison: Compare each provider's opioid rate against their specialty median
2. Cost outlier detection: Flag providers whose per-beneficiary costs exceed 3x the specialty average
3. Drug combination analysis: Identify concurrent opioid + benzo + muscle relaxant prescribers
4. Geographic clustering: Multiple high-rate prescribers in small areas suggest coordinated networks
5. ML pattern matching: Our machine learning model identifies patterns resembling confirmed fraud cases

Policy Implications

States that have implemented Prescription Drug Monitoring Programs (PDMPs) have seen measurable reductions in doctor shopping. Our data shows that:

• States with stronger PDMP enforcement tend to have lower average opioid rates
• The geographic concentration of high-rate prescribers suggests enforcement gaps
• Excluded providers still billing Medicare (0 in our opioid dataset) represent a systems failure