Cannabis Use Rose 900% Among Rheumatology Patients With Major Racial Disparities | Research

The takeaway

NLP analysis of 2.6 million medical records found cannabis use documentation rose 900% among autoimmune patients, with Black and Hispanic patients using cannabis more often for pain and having higher ER utilization.

Rheumatology patients, rheumatologists, health equity researchers, and chronic pain specialists.

What the researchers found

Cannabis use documentation rose from 0.1% to 1.1% (900% increase) over 2004-2024. 24.5% of patients had at least one note of current use. Pain was the leading motive (37.9%), higher among Black (54.5%) and Hispanic (43.2%) patients. Cannabis users had higher comorbidity, more ER visits (2.1 vs. 1.3/year), more hospitalizations (1.4 vs. 0.9), and more opioid prescriptions (65% vs. 32.7%).

Why it matters

Autoimmune patients often use cannabis for pain that persists despite controlled inflammation. The significant racial disparities — with Black and Hispanic patients more likely to use for pain — suggest unequal access to other pain management options.

The numbers in context

5,051 patients, 2.6 million notes. 24.5% with current use documented. 900% increase 2004-2024. Black patients: 36.2% use rate, 54.5% for pain. Hispanic: 30.1%, 43.2% for pain. White: 26.5%. Cannabis users: 2.1 ER visits/year vs. 1.3. Opioid scripts: 65% vs. 32.7%.

How the study worked

NLP applied to 2.6 million EHR notes from 5,051 adults with autoimmune rheumatic diseases at a tertiary center (2004-2024). Classifiers achieved F1=0.85 for use status and F1=0.83 for reasons. Analyzed use patterns, reasons, sociodemographic disparities, and healthcare utilization.

What this study cannot tell us

Single tertiary center. NLP may miss undocumented use. Association between cannabis use and higher healthcare utilization may reflect sicker patients, not cannabis effects. Documentation ≠ actual use rates. Cannabis users may be more willing to discuss use at this center.

How to read the evidence

Novel NLP methodology with large dataset and validated classifiers, though single-center and observational.

When this study was published

Published in 2026, spanning two decades of EHR data capturing the evolution of cannabis use in rheumatology.

The bigger picture

The fact that cannabis users also have more opioid prescriptions suggests these patients have more severe or refractory pain, not that cannabis is replacing opioids. The racial disparities in pain-motivated use may reflect unequal access to pain management rather than differing preferences.

Questions still open

Would better pain management reduce cannabis use in these patients? Are racial disparities in cannabis-for-pain driven by undertreated pain? Could medical cannabis be formally integrated into rheumatology care?

Common questions

Are rheumatology patients using cannabis for pain?

Yes — pain was the top reason (37.9% of users), and this was even higher among Black (54.5%) and Hispanic (43.2%) patients. Cannabis users also had more opioid prescriptions, suggesting particularly challenging pain conditions.

Why are there racial differences in cannabis use for pain?

The study can't determine the cause, but possibilities include unequal access to other pain treatments, differences in pain severity or management, and cultural factors. The disparity highlights a potential health equity concern.

Read the original research

Pharmacoepidemiologic characterization of cannabis use and symptomatology in rheumatology using natural language processing of electronic health record clinic notes.

The journal of pain, 39, 105633

View on PubMed ↗Publisher’s paper ↗

Citation

Falasinnu, Titilola; Le, Nathan; Wang, Yiyu; Alagappan, Aishwarya; Walker, Andrew; Park, Tricia; Leung, Jerik; Chaichian, Yashaar; Weisman, Michael; Kenney, Martha; Irani, Anushka; Bozkurt, Selen. (2026). Pharmacoepidemiologic characterization of cannabis use and symptomatology in rheumatology using natural language processing of electronic health record clinic notes.. The journal of pain, 39, 105633. https://doi.org/10.1016/j.jpain.2025.105633