Comparison of Medical Cannabis Use Reported on a Confidential Survey vs Documented in the Electronic Health Record Among Primary Care Patients
Disclaimer
This comprehensive review was developed by synthesizing and reformatting information directly from the peer-reviewed article “Comparison of Medical Cannabis Use Reported on a Confidential Survey vs Documented in the Electronic Health Record Among Primary Care Patients” by Lapham et al., published in JAMA Network Open (May 23, 2022). The content was derived from the original study’s abstract, methods, results, discussion, and conclusion sections, along with relevant tables and figures, without altering the meaning or intent of the authors’ findings. This review is intended for educational and informational purposes only and does not constitute medical advice, legal guidance, or an endorsement of any policy. All data, statistics, and interpretations remain the intellectual property of the original authors and their respective institutions, as cited in the source publication.
Study Overview
This cross-sectional study aimed to better understand the prevalence of medical cannabis use among primary care patients and how often such use is documented in electronic health records (EHRs). It was conducted within Kaiser Permanente Washington (KPWA), a large integrated health system in Washington State, where both medical and nonmedical cannabis use are legal.
The key motivation was that patients who use cannabis for medical purposes might benefit from clinician discussions about risks, benefits, and alternative treatments. However, little is known about the actual prevalence of medical cannabis use in primary care settings and whether EHRs reflect these patient reports.
The study used both confidential patient surveys and EHR data to:
Estimate the prevalence of explicit and implicit medical cannabis use.
Compare patient self-reports to EHR documentation.
Key Findings
Prevalence:
Any cannabis use (past year): 38.8% of primary care patients (by survey).
Explicit medical use (patient directly reports using for medical purposes): 26.5%.
Implicit medical use (patient reports using cannabis to manage a health condition without labeling it “medical use”): 35.1%.
EHR-documented medical use: Only 4.8%.
EHR Sensitivity:
When compared to patient-reported explicit medical use, EHR documentation had a sensitivity of just 10%(meaning it missed 90% of cases).
Specificity was high at 97%, indicating that when EHRs documented medical use, it was usually accurate.
Most Common Health Reasons for Cannabis Use:
Pain (28.4%)
Sleep problems (19%)
Stress (19%)
Anxiety/worry (14.6%)
Depression/sadness (9.6%)
Methodology
Population: Adults (≥18 years old) receiving primary care at KPWA who completed a routine cannabis screen between January 28 and September 12, 2019.
Sample Size: Out of 108,950 eligible patients, 5,000 were randomly selected, oversampling frequent cannabis users and racial/ethnic minority patients.
Response Rate: 1,688 completed the survey (34% response rate).
Survey Content: 75 questions, designed with cannabis research experts, covering frequency of use, explicit medical use, implicit health reasons for use, and demographics.
EHR Analysis: Used natural language processing (NLP) to detect mentions of medical cannabis in clinician notes, supplemented by manual review.
Results
Survey vs EHR Documentation:
The survey showed cannabis use was far more common than EHR data suggested.
For patients who reported any cannabis use in the past year, 68% said it was for medical purposes.
In contrast, only 7.7% of cannabis-using patients had EHR documentation of medical use.
EHR Performance:
Sensitivity: 10% (explicit) and 8.4% (implicit).
Specificity: ~97% for both measures.
This means EHRs are good at confirming use when recorded, but poor at capturing most actual cases.
Discussion
The study shows a large gap between patient-reported medical cannabis use and EHR documentation.
Possible reasons for this gap include:
Lack of health system policies requiring or encouraging documentation.
Clinician discomfort or reluctance to discuss cannabis use.
Competing priorities during patient visits.
Limited clinician training on cannabis.
The authors note that asking patients about their health reasons for cannabis use (rather than explicitly about “medical use”) may uncover more cases and improve documentation. This is important for:
Identifying possible drug interactions.
Discussing evidence-based alternatives.
Understanding if patients are substituting cannabis for prescribed medications.
Monitoring potential adverse effects.
Limitations
Response rate was only 34% (though comparable to other large surveys).
Possible differences between survey respondents and the broader population, although weighting helped reduce bias.
Study took place in one health system in one state with legal cannabis—results may not generalize elsewhere.
The time gap between EHR screening and survey completion (average 77 days) could allow changes in cannabis use patterns.
Inclusive survey definition of cannabis may explain why some patients reported use despite no EHR documentation.
Conclusions
Medical cannabis use is common among primary care patients in Washington State—about one-third use it for health reasons—but is rarely documented in EHRs. Most of this usage goes unrecognized by the healthcare system, potentially limiting the quality of patient-clinician communication and care.
Implication: Incorporating targeted questions about the use of cannabis for managing pain, sleep, mood, and other health conditions into routine primary care screening could improve recognition, documentation, and clinical management of medical cannabis use.
RIC & KOWALSKI TAG TEAM
Plain Breakdown & Analysis
Core Study Question:
How many primary care patients use cannabis for medical purposes, and how well do their electronic health records (EHRs) reflect that reality?
Approach:
Confidential patient survey (self-reported use) vs. EHR review (via NLP + manual check).
Focused on Kaiser Permanente Washington (state where cannabis is legal).
Key Stats:
Any cannabis use (past year): 38.8%
Explicit medical use: 26.5%
Implicit medical use: 35.1%
EHR-documented medical use: 4.8%
EHR sensitivity for detecting self-reported medical use: ~10%
Top reasons for use: Pain, sleep problems, stress, anxiety, depression.
Main Finding:
Most medical cannabis use goes undocumented in official medical records — meaning doctors may have no idea about a major part of their patients’ treatment behaviors.
Why This Matters:
Missed opportunities for discussing benefits/risks.
Missed detection of drug interactions.
Missed understanding of how patients are replacing or supplementing prescribed medications.
Limitations:
Conducted in one legal-cannabis state.
Response bias (34% response rate).
Time gap between EHR screen and survey.
Long-form Summaries in 3 Modes
Shapiro Mode
So here’s what the study found — and it’s kind of staggering. Thirty-nine percent of primary care patients used cannabis in the past year. Not just recreationally — over a quarter said explicitly they were using it for medical purposes. If you count those who use it for a health reason without calling it “medical,” that number jumps to thirty-five percent. Now here’s the kicker: in the electronic health records — the supposedly comprehensive, modern, data-driven records — only 4.8% had any documentation of medical cannabis use. Which means that over ninety percent of actual medical cannabis use is invisible to the healthcare system. Why? The study cites several factors: no institutional requirement to document it, doctors uncomfortable talking about cannabis, competing priorities in appointments, lack of training, and so on. But the result is the same: doctors are flying blind on something that could impact drug interactions, treatment plans, and patient safety. The takeaway? Ask patients about why they use cannabis — don’t just ask if they use it — and you might get the truth. Otherwise, your “comprehensive” health record is about as accurate as a political poll on Twitter.
Hunter S. Thompson Mode
Jesus, it’s all there in black and white: thirty-nine percent of these prim, buttoned-down Kaiser patients — not the dreadlocked festival kids or the burned-out hitchhikers — have been huffing the Devil’s Lettuce in the past year. And not for fun, mind you. Twenty-six percent come right out and say it: “Medical.” Pain, sleep, stress — the whole American horror show. But then we dive into the sacred EHR, the official gospel of the medical-industrial complex, and it’s like stepping into a dead city. Four point eight percent. That’s all they’ve got on the books. The rest — ninety percent of the truth — lost in the polite coughs and awkward silences of the exam room. The system is designed not to see it. The doctors have been trained to avert their eyes. “Cannabis” is still the dirty word you keep off your insurance forms. But the reality is boiling under the surface — people are medicating themselves while the great computerized beast pretends nothing is happening. If they’d just ask the right questions, maybe they’d see the blood in the water. But they won’t. Not yet.
George Carlin Mode
You wanna hear a good one? A third of primary care patients in Washington are using weed for medical reasons. Pain, sleep, stress — the usual stuff. And here’s the medical system’s official record of that? Four point eight percent. Yeah. The rest? Poof. Vanished. You know why? Because the doctors never write it down. They’ll log your cholesterol, your toenail fungus, your cousin’s hip replacement, but mention cannabis and suddenly it’s like you just farted in church. “Ohhh, that’s not in the policy.” No, what’s in the policy is keeping their heads buried in paperwork while the patients are out there self-medicating with something that’s been legal for years. And the EHR — this giant, multi-billion-dollar “comprehensive” system — is basically a glorified scrapbook that says, “Nah, we didn’t see anything.” But don’t worry — they’ll keep asking if you’ve fallen in the last six months.