Risk scores are statistical indicators based on prescribing patterns compared to specialty peers. They are NOT allegations of fraud, misconduct, or improper care. Many legitimate medical reasons can explain outlier prescribing.
Read our methodology →The Controlled Substance Pipeline: Beyond Opioids
Analysis · March 2026
The opioid crisis dominates headlines, but it's only one part of a broader controlled substance prescribing picture. Benzodiazepines kill more people than any opioid except fentanyl. Stimulant prescriptions have surged amid the Adderall shortage. Gabapentin — technically not a controlled substance at the federal level — has become one of the most misused drugs in America. And some providers prescribe across multiple controlled substance categories simultaneously, creating compounded risk for their patients.
450,343
Opioid Prescribers
113,169
High-Rate Opioid Rx
6,680
Opioid-Flagged Providers
6,706
Multi-Flag Providers
The Four Horsemen of Controlled Substances
1. Opioids: The Known Crisis
Opioids remain the most scrutinized controlled substance category in Medicare. Among 1,380,665 prescribers,450,343 (33%) write at least some opioid prescriptions. Of these, 113,169 have opioid rates that exceed their specialty's norm — the providers most likely to be contributing to the crisis.
But the opioid story has evolved. The current wave of overdose deaths is driven primarily by illicit fentanyl, not prescription opioids. Still, prescription opioids remain a gateway: approximately 80% of people who use heroin first misused prescription opioids. Medicare data captures the prescription pipeline — the legal supply that can feed into the illicit market.
2. Benzodiazepines: The Silent Killer
Benzodiazepines — Xanax (alprazolam), Valium (diazepam), Klonopin (clonazepam), Ativan (lorazepam) — are prescribed for anxiety, insomnia, and seizures. They're the most commonly prescribed psychotropic drugs in the US, and they're remarkably dangerous:
- 12,499 overdose deaths involving benzodiazepines in 2021 (CDC)
- Synergistic with opioids — combining benzos with opioids dramatically increases overdose risk (the FDA issued a Black Box Warning)
- Dependence develops quickly — physical dependence can develop in as little as 2-4 weeks
- Elderly are most vulnerable — falls, cognitive impairment, and respiratory depression disproportionately affect Medicare-age patients
The Beers Criteria — the American Geriatrics Society's list of potentially inappropriate medications for older adults — explicitly includes benzodiazepines. Yet millions of Medicare patients receive them.
The Opioid-Benzodiazepine Combination
Concurrent prescribing of opioids and benzodiazepines is one of the most dangerous patterns we track. Both drug classes suppress breathing; together, they dramatically increase the risk of fatal respiratory depression. Despite the FDA's 2016 Black Box Warning, this combination remains alarmingly common in Medicare. Our pill mill analysis identifies providers who co-prescribe these drug classes at high rates.
3. Stimulants: The ADHD Surge
While stimulants (Adderall, Ritalin, Vyvanse, Concerta) are most associated with younger populations, their prescribing has implications for Medicare in several ways:
- Growing ADHD diagnosis in adults 50+ — recognition of adult ADHD has increased, leading to more stimulant prescriptions in the Medicare-eligible population
- The Adderall shortage (2022-present) — DEA manufacturing quotas haven't kept pace with demand, leading to supply disruptions, pharmacy hopping, and desperate patients turning to telehealth mills
- Diversion risk — stimulants are among the most commonly diverted prescription drugs
- Cardiovascular risk — in elderly patients, stimulants carry significant cardiovascular risks including increased blood pressure, tachycardia, and potential for stroke
The Adderall shortage has been particularly revealing. When legitimate supply channels fail, it exposes the fragility of the controlled substance distribution system and creates opportunities for exploitation by providers willing to prescribe more freely.
4. Gabapentin: The Unofficial Controlled Substance
Gabapentin (Neurontin) occupies a unique position in the controlled substance landscape. It's not federally scheduled, but:
- 8 states have classified gabapentin as a Schedule V controlled substance
- Misuse is widespread — an estimated 40-65% of gabapentin prescriptions are misused
- Potentiates opioids — gabapentin enhances the euphoric effects of opioids, making it popular for co-abuse
- Found in 1 in 10 overdose deaths — gabapentin is increasingly detected in fatal overdose toxicology reports
- Massive prescribing volume — gabapentin is one of the most prescribed drugs in the US (over 60 million prescriptions annually)
In Medicare, gabapentin is often prescribed alongside opioids for pain management. While this can be legitimate (gabapentin is FDA-approved for post-herpetic neuralgia), the combination raises risk flags that our scoring system tracks.
Multi-Category Controlled Substance Prescribers
The most concerning pattern isn't any single controlled substance category — it's providers who prescribe heavily across multiple categories. A provider who writes high volumes of opioids AND benzodiazepines AND gabapentin is creating compounded risk for their patients.
Providers with Multiple Risk Flags
Our risk scoring system flags providers based on individual risk factors. Those with flags across multiple controlled substance categories warrant the closest scrutiny:
| Provider | Specialty | Location | Risk Score | Risk Flags |
|---|---|---|---|---|
| Kamyar Cohanshohet | Internal Medicine | Beverly Hills, CA | 77/100 | extreme_opioid_vs_peers95th_pctile_opioidelevated_costhigh_la_opioid_vs_peersopioid_benzo_coprescriberleie_excludedlow_drug_diversity |
| Stephen Kelly | Family Practice | Oklahoma City, OK | 70/100 | extreme_opioid_vs_peers95th_pctile_opioidhigh_la_opioid_vs_peershigh_antipsych_elderlyopioid_benzo_coprescriber |
| Eleanya Ogburu-Ogbonnaya | Neurology | Columbia, SC | 66/100 | extreme_opioid_vs_peers90th_pctile_opioidelevated_costhigh_la_opioid_vs_peerselevated_antipsych_elderlyopioid_benzo_coprescriberhigh_fills_per_patient |
| Rodolfo Herrera | Family Practice | Bedford, TX | 66/100 | extreme_opioid_vs_peers95th_pctile_opioidelevated_la_opioidhigh_antipsych_elderlyopioid_benzo_coprescriber |
| Jennifer Jenkins | Nurse Practitioner | Morristown, TN | 64/100 | extreme_opioid_vs_peers95th_pctile_opioidhigh_la_opioid_vs_peerselevated_antipsych_elderlyopioid_benzo_coprescriber |
| Susan Frady | Nurse Practitioner | Canon City, CO | 64/100 | extreme_opioid_vs_peers99th_pctile_opioidhigh_la_opioid_vs_peersopioid_benzo_coprescriber |
| Karolina Cichocka | Nurse Practitioner | Farmington, CT | 64/100 | extreme_opioid_vs_peers99th_pctile_opioidhigh_la_opioid_vs_peersopioid_benzo_coprescriber |
| Sabera Shabnam | Family Practice | Branson, MO | 64/100 | extreme_opioid_vs_peers95th_pctile_opioidhigh_la_opioid_vs_peerselevated_antipsych_elderlyopioid_benzo_coprescriber |
| Joanne Vogel | Nurse Practitioner | Colorado Springs, CO | 64/100 | extreme_opioid_vs_peers95th_pctile_opioidhigh_la_opioid_vs_peerselevated_antipsych_elderlyopioid_benzo_coprescriber |
| Victoria Francis | Nurse Practitioner | Ann Arbor, MI | 64/100 | extreme_opioid_vs_peers99th_pctile_opioidhigh_la_opioid_vs_peersopioid_benzo_coprescriber |
Specialty Patterns in Controlled Substance Prescribing
Different specialties have vastly different relationships with controlled substances. Pain Management and Anesthesiology top the opioid rate charts by design — but when Family Practice or Internal Medicine providers match those rates, it's a red flag:
| Specialty | Providers | Opioid Rate | Flagged Count |
|---|---|---|---|
| Nurse Practitioner | 258,730 | 0.0% | 3,249 |
| Dentist | 131,349 | 0.0% | 44 |
| Physician Assistant | 129,444 | 0.0% | 945 |
| Internal Medicine | 128,226 | 0.0% | 693 |
| Family Practice | 116,540 | 0.0% | 553 |
| Student in an Organized Health Care Education/Training Program | 66,682 | 0.0% | — |
| Emergency Medicine | 54,622 | 0.0% | 65 |
| Pharmacist | 36,120 | 0.0% | 78 |
| Obstetrics & Gynecology | 34,229 | 0.0% | — |
| Optometry | 33,191 | 0.0% | — |
| Psychiatry | 23,909 | 0.0% | 72 |
| General Surgery | 22,097 | 0.0% | — |
| Orthopedic Surgery | 21,058 | 0.0% | — |
| Ophthalmology | 19,647 | 0.0% | — |
| Hospitalist | 19,524 | 0.0% | 36 |
Geographic Patterns
Controlled substance prescribing follows clear geographic patterns that reflect a combination of cultural factors, regulatory environments, provider density, and patient demographics:
| State | Opioid Rate | Providers | Total Cost |
|---|---|---|---|
| CA | 0.0% | 139,057 | $27.10B |
| NY | 0.0% | 104,092 | $22.46B |
| FL | 0.0% | 93,928 | $20.29B |
| TX | 0.0% | 92,813 | $19.28B |
| PA | 0.0% | 64,171 | $13.21B |
| OH | 0.0% | 53,444 | $10.85B |
| NC | 0.0% | 44,019 | $9.76B |
| MI | 0.0% | 46,434 | $9.40B |
| IL | 0.0% | 53,002 | $9.23B |
| GA | 0.0% | 36,843 | $8.63B |
| NJ | 0.0% | 36,308 | $7.73B |
| MA | 0.0% | 39,945 | $7.22B |
| TN | 0.0% | 29,129 | $6.89B |
| IN | 0.0% | 26,170 | $6.29B |
| MO | 0.0% | 25,377 | $5.75B |
The Regulatory Patchwork
Controlled substance regulation in the US is a complex patchwork of federal and state rules that creates gaps exploitable by bad actors:
Federal (DEA) Scheduling
The Controlled Substances Act classifies drugs into Schedules I-V based on abuse potential and medical use. Schedule II (opioids, stimulants) requires new prescriptions each fill; Schedule IV (benzodiazepines) allows refills. Gabapentin is not federally scheduled.
State Variations
States can impose stricter rules: some require PDMP checks before every controlled substance prescription, some limit initial opioid prescriptions to 3-7 days, and 8 states have scheduled gabapentin. These variations create "arbitrage" opportunities where patients seek prescriptions in less restrictive states.
PDMP (Prescription Drug Monitoring Programs)
All 50 states have PDMPs, but interstate data sharing remains inconsistent. Only about 50% of states participate in the PMP InterConnect system for real-time interstate data. This means a patient getting benzodiazepines in Ohio and opioids in Kentucky may not trigger any alert.
What the Data Can't Tell Us
It's important to acknowledge the limitations of Medicare Part D claims data for controlled substance analysis:
- No diagnosis codes — we can't see why a drug was prescribed
- No benzodiazepine-specific rates — our data tracks opioid rates but not benzo rates at the provider level
- No gabapentin abuse indicators — gabapentin claims look identical whether used for neuropathy or misused for recreation
- No patient-level tracking — we see provider-level aggregates, not individual patient journeys across multiple providers
- No illicit use — Part D data captures only legitimate prescriptions, not the downstream diversion that fuels abuse
Recommendations
Policy Changes Needed
The Bigger Picture
The controlled substance prescribing landscape is more complex than any single drug category can capture. While opioids get the attention, benzodiazepines, stimulants, and gabapentin each present their own risk profiles — and the real danger often comes from their intersection.
Among the 1,380,665 prescribers in our dataset, the vast majority prescribe controlled substances responsibly. But the minority who prescribe across multiple categories at high volumes — especially those flagged by our risk scoring system — represent a concentrated source of patient harm and potential diversion that deserves focused regulatory attention.
The pipeline of controlled substances flowing through Medicare Part D is massive, geographically concentrated, and only partially visible through claims data. Improving that visibility — through better data, better interstate coordination, and better analytical tools — is essential for reducing harm while preserving access for patients who genuinely need these medications.
📚 Data Sources & Methodology
Provider-level data from CMS Medicare Part D Prescriber Public Use File (2023). Opioid prescribing rates are CMS-calculated. Risk flags are from our multi-factor scoring model. Overdose statistics from CDC WONDER. Gabapentin misuse data from the DEA and peer-reviewed literature. PDMP participation data from the National Alliance for Model State Drug Laws (NAMSDL).
Explore Related Analysis
Related Analysis
The Opioid Prescribing Crisis in Medicare
OpioidsGeographic Opioid Hotspots
OpioidsAnatomy of a Pill Mill
CostCost Outlier Providers