By Ellen Feldman, MD
Altru Health System, Grand Forks, ND
- This cross-sectional, electronic health record review of 185,565 patients in Washington (where recreational cannabis has been legal for adults since 2012 and medicinal cannabis has been legal since 1998) drew data from 25 Kaiser Permanente medical clinics, where all primary care patients are screened annually for cannabis use.
- The goal of the study was to find the prevalence of documentation of medicinal cannabis use and associated health conditions in this population.
- Results included medicinal cannabis use in 3,551 patients (2%), nonmedicinal cannabis use in 36,599 patients (20%), and no cannabis use documented in 145,415 patients (78%).
- Among the population with documented medicinal cannabis use, there was a higher prevalence of health conditions for which cannabis could both be helpful and convey risk when compared to the patients with documented nonmedicinal cannabis use or no cannabis use.
SYNOPSIS: This cross-sectional review of 185,565 patients’ medical records found documentation of medicinal cannabis use in 2% of the records. Within this subgroup, 44.5% had documentation of one or more health conditions potentially benefitting from treatment with cannabis, 54.4% had documentation of one or more health conditions potentially worsening with cannabis use, and 36.6% had both types of health conditions.
SOURCE: Matson TE, Carrell DS, Bobb JF, et al. Prevalence of medical cannabis use and associated health conditions documented in electronic health records among primary care patients in Washington state. JAMA Netw Open 2021;4:e219375.
Cannabis now is legal for medicinal use in more than half of U.S. states. Although multiple factors, including federal restrictions on cannabis research, have hindered the development of practical, evidence-based guidelines regarding cannabis, this agent is gaining traction for nonmedicinal purposes in the United States as well, with 18 states currently allowing such use among adults.1,2
Despite state legislation, cannabis is illegal on a federal level. According to the Centers for Disease Control and Prevention, cannabis held the dubious honor of being the most commonly used federally illegal substance in the United States in 2019, with an estimated 48.2 million individuals reporting use during that year.3
Recognizing that “cannabis use is highly relevant to a patient’s care,” Matson et al conducted a descriptive, cross-sectional study of electronic health record (EHRs) in a large health system (Kaiser Permanente) to determine the prevalence of documented medicinal cannabis and diagnoses associated with these patients from Nov. 1, 2017, until Oct. 31, 2018. This investigation took place across 25 primary care clinics in Washington, where medicinal cannabis has been legal since 1998 and recreational cannabis was approved in 2012. All primary care patients are screened annually for cannabis use as part of a routine seven-item behavioral health screen by responding to the question “How often in the past year have you used marijuana?”
However, this screening question does not differentiate medicinal from nonmedicinal cannabis use. To do so, Matson et al employed a natural language processing (NLP) program (a type of artificial intelligence) to find references to medicinal cannabis use in the EHRs within a specified time before and after the appointment.4 To help eliminate bias in reporting, employees of the health system were excluded from the study.
Matson et al used a 2017 comprehensive review of the health effects of cannabis by The National Academies of Science, Engineering, and Medicine (NASEM) to categorize health conditions into one of three sectors: potential benefit from cannabis, potential risk from cannabis, and inconclusive evidence.5
Out of the 185,565 patient charts screened during a primary care visit, 3,551 (2%) had documentation of medicinal cannabis use, 36,599 (20%) documented cannabis use without mention of medicinal use, and 145,415 (78%) did not document cannabis use during the past year. Out of the 40,150 patients documenting cannabis use of any sort, 9% of the charts reflected medicinal cannabis use. Based largely on observational studies, NASEM identified at least five conditions for which cannabis has potential benefits: chronic noncancer pain, multiple sclerosis, muscle spasms or spasticity, severe nausea, and sleep disorder.5
After analyzing the study cohort, including adjustment for demographic and socioeconomic factors, Matson et al found patients with EHR-recorded medicinal cannabis use had a higher prevalence of one or more of these conditions when compared with members of the other two groups (nonmedicinal or nonusers of cannabis). Overall, 49.8% of the medicinal use group had at least one of these diagnoses documented compared with 39.9% of the patients with documented nonmedicinal cannabis use and 40% of the patients with no documented cannabis use. (See Table 1.)
Table 1. Adjusted Prevalence of Health Conditions Potentially Benefitting from Cannabis Use |
|
Documentation
of Cannabis
for Medicinal Use
(n = 3,551) |
Documentation
of Nonmedicinal Cannabis Use
(n = 36,599) |
Documentation of No Cannabis Use
(n = 145,415) |
Chronic noncancer pain |
35.4 (95% confidence interval [CI], 34.1-36.7) |
28.3 (95% CI, 27.8-28.7) |
28.3 (95% CI, 27.8-28.7) |
Multiple sclerosis |
0.6 (95% CI, 0.4-0.8) |
0.4 (95% CI, 0.3-0.5) |
0.3 (95% CI, 0.3-0.4) |
Sleep disorder |
21.8 (95% CI, 20.6-22.9) |
18.1 (95% CI, 17.7-18.5) |
18.5 (95% CI, 18.3-18.6) |
Muscle spasms/spasticity |
5.1 (95% CI, 4.5-5.7) |
3.5 (95% CI, 3.3-3.7) |
3.5 (95% CI, 3.4-3.6) |
Severe nausea |
7.6 (95% CI, 6.9-8.2) |
4.8 (95% CI, 4.6-5.1) |
4.3 (95% CI, 4.2-4.4) |
Any condition potentially benefitting from cannabis use |
49.8 (95% CI, 48.3-51.3) |
39.9 (95% CI, 39.4-40.3) |
40.0 (95% CI, 39.8-40.2) |
However, patients with documented medicinal cannabis use also had a higher prevalence of NASEM-identified health conditions potentially exacerbated or worsened by cannabis use, again adjusted for the demographic and socioeconomic covariates, when compared with patients in the other two groups. These conditions included serious mental illness (e.g., schizophrenia and bipolar), depression, substance use disorder, and respiratory conditions.5 (See Table 2.)
Table 2. Adjusted Prevalence of Health Conditions Potentially at Risk from Cannabis Use |
|
Documentation
of Cannabis
for Medicinal Use
(n = 3,551) |
Documentation
of Nonmedicinal Cannabis Use
(n = 36,599) |
Documentation of No Cannabis Use
(n = 145,415) |
Serious mental illness |
2.8 (95% confidence interval [CI], 2.0-3.2) |
2.0 (95% CI, 1.9-2.1) |
1.3 (95% CI, 1.2-1.3) |
Chronic obstructive pulmonary disease |
15.6 (95% CI, 14.6-16.6) |
15.3 (95% CI, 14.9-15.7) |
14.7 (95% CI, 14.5-14.8) |
Substance use disorder |
21.9 (95% CI, 20.6-23.1) |
14.1 (95% CI, 13.7-14.5) |
7.1 (95% CI, 7.0-7.3) |
Opioid overdose |
0.2 (95% CI, 0.1-0.3) |
0.2 (95% CI, 0.1-0.2) |
0.1 (95% CI, 0.1-0.1) |
Any condition for which cannabis has potential risk |
60.7 (95% CI, 59.0-62.3) |
50.5 (95% CI, 50.0-51.0) |
42.7 (95% CI, 42.4-42.9) |
The adjusted prevalence of health conditions for which there is inconclusive evidence for cannabis use varied across the subgroups. For example, the prevalence of eating disorders, hypertension, cancer, and heart disease did not significantly vary with documentation of cannabis use. The prevalence of anxiety was higher in the medicinal cannabis group (28.7%) compared to 18.5% in the nonmedicinal cannabis group and 13.2% in noncannabis users, while heart disease (13.4%) was slightly higher in the noncannabis users than either of the cannabis-user groups (11% to 11.9%).
When compared with the two other groups, the medicinal cannabis group had the highest prevalence of patients with health conditions that potentially can benefit from cannabis as well as conditions for which cannabis use poses risk. Specifically, the medical cannabis group prevalence for these conditions was 33.6% compared with 25.2% in the nonmedicinal cannabis group and 22.3% in the noncannabis users.
COMMENTARY
Matson et al present a unique, large-scale record review quantifying a 2% provider documentation of medical cannabis use among 185,565 patients and an additional 20% patient-documented cannabis use (without documentation of medicinal use.) Although it certainly is possible that use in any category is underreported, this study is meant to reflect provider and patient documentation and responses. One caveat noted by Matson et al is that, since the medicinal documentation was extracted via a NLP program, it is possible that error could have contributed more to the medicinal cannabis numbers as opposed to the nonmedicinal cannabis group, whose quantification was based on direct responses to a questionnaire. It is important to reiterate that this study is purely observational — there is no evidence or attempt to understand cause and effect, therapeutic benefits, or risks from use of cannabis, or motivation for cannabis use. These remain important areas for further investigation to describe the relative benefits and risks of cannabis use more fully. Further results from this record review reveal that patients with documented medicinal cannabis use in this investigation have a higher prevalence of health conditions potentially benefiting and potentially harmed by cannabis when compared with patients in the other groups. This sparks a valuable takeaway message for primary care providers — be cognizant of asking patients about medicinal and other cannabis use, and review potential risks and benefits.
Unfortunately, there currently is limited (but growing) evidence of the potential risks and benefits for cannabis use with any of the specified conditions. There is evidence of cannabis’ efficacy in treating chronic pain in adults, in symptoms of spasticity in multiple sclerosis, and short-term improvement for insomnia. There is established efficacy for using cannabinoids in treating chemotherapy-induced nausea and vomiting in adults. As more studies are completed, knowledge about dose, interactions, targeted demographic groups and contraindications likely will increase, allowing for the development of treatment guidelines and more widespread use. There is evidence that repeated cannabis use is associated with developing cannabis use disorders and that this risk is higher when cannabis use begins in the early teen years. There is moderate evidence suggesting an association between cannabis use and substance use disorder (alcohol and illicit drugs). There is evidence that cannabis use increases the risk of developing schizophrenia, other psychotic disorders, and social anxiety disorder, and limited evidence of an association with depression. While there is evidence that smoked forms of this agent may worsen pulmonary conditions, this appears reversible when cannabis use is stopped. However, the specifics of these and other risks remain uncharacterized.6,7
This study can serve as a reminder for the integrative provider to screen patients for cannabis use and use this as an opportunity to share what is known and unknown about potential risks and benefits. It is often tempting to ignore what we do not understand — this clearly is not a productive path to take in medicine and can hinder efforts to develop full knowledge of a field. While there is a great deal of ambiguity in discussing cannabis use risks and benefits, offering patients information about what is known and identifying and documenting medicinal and nonmedicinal use are valuable first steps on the way to gain a more comprehensive overview of the field.
REFERENCES
- National Conference of State Legislatures. State medical marijuana laws. Published Aug. 23, 2021. https://www.ncsl.org/research/health/state-medical-marijuana-laws.aspx
- Yu B, Chen X, Chen X, Yan H. Marijuana legalization and historical trends in marijuana use among U.S. residents aged 12-25: Results from the 1979–2016 National Survey on drug use and health. BMC Public Health 2020;20:156.
- Centers for Disease Control and Prevention. Marijuana and public health: Data and statistics. Updated June 8, 2021. https://www.cdc.gov/marijuana/data-statistics.htm
- Yse DL. Your guide to natural language processing (NLP). Towards Data Science. Published Jan. 15, 2019. https://towardsdatascience.com/your-guide-to-natural-language-processing-nlp-48ea2511f6e1
- Committee of the Health Effects of Marijuana. Health effects of marijuana and cannabis-derived products presented in new report. The National Academies of Sciences, Engineering, and Medicine. Published Jan. 12, 2017. https://www.nationalacademies.org/news/2017/01/health-effects-of-marijuana-and-cannabis-derived-products-presented-in-new-report
- Hall W. A summary of reviews of evidence on the efficacy and safety of medical use of cannabis and cannabinoids. European Monitoring Centre for Drugs and Drug Addiction. Published December 2018. https://www.emcdda.europa.eu/system/files/publications/12852/MedicalCannabis-BackgroundPaper.pdf
- National Institute on Drug Abuse. Marijuana research report: Is marijuana safe and effective as medicine? Published July 2020. https://www.drugabuse.gov/publications/research-reports/marijuana/marijuana-safe-effective-medicine
